Australia’s NationalScience Agency Soils and land suitability for theSouthern Gulf catchments A technicalreport from the CSIRO Southern GulfWater ResourceAssessment for theNationalWater Grid Mark Thomas1,Seonaid Philip1, Peter Zund1,Uta Stockmann1,Jason Hill2, Linda Gregory1, IanWatson1and Evan Thomas3 1CSIRO; 2Northern Territory Department of Environment,Parks and Water Security; 3Queensland Department of Environment and Science A logo with black text Description automatically generated ISBN 978-1-4863-2070-7 (online) A logo with black text Description automatically generated ISBN 978-1-4863-2069-1 (print) Citation Thomas M, Philip S, Zund P, Stockmann U, Hill J, Gregory L, Watson I and Thomas E (2024) Soils and land suitability for the Southern Gulf catchments. A technical report from the CSIRO Southern Gulf Water Resource Assessment for the National Water Grid. CSIRO, Australia. Copyright © Commonwealth Scientific and Industrial Research Organisation 2024. To the extent permitted by law, all rights are reserved and no part of this publication covered by copyright may be reproduced or copied in any form or by any means except with the written permission of CSIRO. Important disclaimer CSIRO advises that the information contained in this publication comprises general statements based on scientific research. The reader is advised and needs to be aware that such information may be incomplete or unable to be used in any specific situation. No reliance or actions must therefore be made on that information without seeking prior expert professional, scientific and technical advice. To the extent permitted by law, CSIRO (including its employees and consultants) excludes all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication (in part or in whole) and any information or material contained in it. CSIRO is committed to providing web accessible content wherever possible. If you are having difficulties with accessing this document please CSIRO Southern Gulf Water Resource Assessment acknowledgements This report was funded through the National Water Grid’s Science Program, which sits within the Australian Government’s Department of Climate Change, Energy, the Environment and Water. Aspects of the Assessment have been undertaken in conjunction with the Northern Territory and Queensland governments. The Assessment was guided by two committees: i. The Governance Committee: CRC for Northern Australia/James Cook University; CSIRO; National Water Grid (Department of Climate Change, Energy, the Environment and Water); Northern Land Council; NT Department of Environment, Parks and Water Security; NT Department of Industry, Tourism and Trade; Office of Northern Australia; Queensland Department of Agriculture and Fisheries; Queensland Department of Regional Development, Manufacturing and Water ii. The Southern Gulf catchments Steering Committee: Amateur Fishermen’s Association of the NT; Austral Fisheries; Burketown Shire; Carpentaria Land Council Aboriginal Corporation; Health and Wellbeing Queensland; National Water Grid (Department of Climate Change, Energy, the Environment and Water); Northern Prawn Fisheries; Queensland Department of Agriculture and Fisheries; NT Department of Environment, Parks and Water Security; NT Department of Industry, Tourism and Trade; Office of Northern Australia; Queensland Department of Regional Development, Manufacturing and Water; Southern Gulf NRM Responsibility for the Assessment’s content lies with CSIRO. The Assessment’s committees did not have an opportunity to review the Assessment results or outputs prior to their release. This report was reviewed by Prof Budiman Minasny (School of Life and Environmental Sciences; the University of Sydney, Australia) and Prof George van Zijl (Natural and Agricultural Sciences; North-West University, South Africa). The Digital Soil Mapping and Land Suitability analysis reported in this Technical Report would not have been possible without the help, support, encouragement and advice from a large number of people. Our gratitude extends to local Indigenous organisations and their staff who have advised us on access and permission, in particular the Northern Land Council, Doomadgee Shire Council and the Carpentaria Land Council Aboriginal Corporation. Many CSIRO staff who were not part of the immediate team or report authors, provided a range of inputs to the land suitability activity. This includes technical advice, project design, project management oversight, financial controls, administration support, contract management, database and query advice, communications, help with document files and input from other activities. The CSIRO staff we wish to thank (in no particular order) include: Cuan Petheram, Simon Irvin, Arthur Read, Linda Karssies, Georgia Reed and Caroline Bruce. We are greatly indebted to Peter R Wilson recently retired from CSIRO for sharing his knowledge, guiding and questioning our rationale and providing a few laughs from the first Assessment in 2012. Similarly, a number of people from associated jurisdictions provided valuable input to the project but are not named as part of the team especially the staff from the Qld DESI Chemistry Centre. Our fieldwork was improved by the support of the local grazing industry. They gave us ‘the time of day’, shared experiences and on ground knowledge, and provided the local context that is so important for work of this kind. Our documentation, and its consistency were much improved by Joely Taylor who provided great service and patient application of the Assessment’s style and convention standards and Nathan Dyer took some wonderful photos and videos. Finally, we’d like to thank our family, friends and colleagues who picked up the slack when we were away from home or so busy that we (temporarily) forgot that they existed. Acknowledgement of Country CSIRO acknowledges the Traditional Owners of the lands, seas and waters, of the area that we live and work on across Australia. We acknowledge their continuing connection to their culture and pay our respects to their elders past and present. Photo Soil sampling on the most south-easterly part of Barkly Tablelands in the upper Gregory catchment, Qld. Source: CSIRO Director’s foreword Sustainable development and regional economic prosperity are priorities for the Australian, Queensland and Northern Territory (NT) governments. However, more comprehensive information on land and water resources across northern Australia is required to complement local information held by Indigenous Peoples and other landholders. Knowledge of the scale, nature, location and distribution of likely environmental, social, cultural and economic opportunities and the risks of any proposed developments is critical to sustainable development. Especially where resource use is contested, this knowledge informs the consultation and planning that underpin the resource security required to unlock investment, while at the same time protecting the environment and cultural values. In 2021, the Australian Government commissioned CSIRO to complete the Southern Gulf Water Resource Assessment. In response, CSIRO accessed expertise and collaborations from across Australia to generate data and provide insight to support consideration of the use of land and water resources in the Southern Gulf catchments. The Assessment focuses mainly on the potential for agricultural development, and the opportunities and constraints that development could experience. It also considers climate change impacts and a range of future development pathways without being prescriptive of what they might be. The detailed information provided on land and water resources, their potential uses and the consequences of those uses are carefully designed to be relevant to a wide range of regional-scale planning considerations by Indigenous Peoples, landholders, citizens, investors, local government, and the Australian, Queensland and NT governments. By fostering shared understanding of the opportunities and the risks among this wide array of stakeholders and decision makers, better informed conversations about future options will be possible. Importantly, the Assessment does not recommend one development over another, nor assume any particular development pathway, nor even assume that water resource development will occur. It provides a range of possibilities and the information required to interpret them (including risks that may attend any opportunities), consistent with regional values and aspirations. All data and reports produced by the Assessment will be publicly available. Chris Chilcott Project Director C:\Users\bru119\AppData\Local\Microsoft\Windows\Temporary Internet Files\Content.Word\C_Chilcott_high.jpg The Southern Gulf Water Resource Assessment Team Project Director Chris Chilcott Project Leaders Cuan Petheram, Ian Watson Project Support Caroline Bruce, Seonaid Philip Communications Emily Brown, Chanel Koeleman, Jo Ashley, Nathan Dyer Activities Agriculture and socio- economics Tony Webster, Caroline Bruce, Kaylene Camuti1, Matt Curnock, Jenny Hayward, Simon Irvin, Shokhrukh Jalilov, Diane Jarvis1, Adam Liedloff, Stephen McFallan, Yvette Oliver, Di Prestwidge2, Tiemen Rhebergen, Robert Speed3, Chris Stokes, Thomas Vanderbyl3, John Virtue4 Climate David McJannet, Lynn Seo Ecology Danial Stratford, Rik Buckworth, Pascal Castellazzi, Bayley Costin, Roy Aijun Deng, Ruan Gannon, Steve Gao, Sophie Gilbey, Rob Kenyon, Shelly Lachish, Simon Linke, Heather McGinness, Linda Merrin, Katie Motson5, Rocio Ponce Reyes, Jodie Pritchard, Nathan Waltham5 Groundwater hydrology Andrew R. Taylor, Karen Barry, Russell Crosbie, Margaux Dupuy, Geoff Hodgson, Anthony Knapton6, Stacey Priestley, Matthias Raiber Indigenous water values, rights, interests and development goals Pethie Lyons, Marcus Barber, Peta Braedon, Petina Pert Land suitability Ian Watson, Jenet Austin, Bart Edmeades7, Linda Gregory, Ben Harms10, Jason Hill7, Jeremy Manders10, Gordon McLachlan, Seonaid Philip, Ross Searle, Uta Stockmann, Evan Thomas10, Mark Thomas, Francis Wait7, Peter Zund Surface water hydrology Justin Hughes, Matt Gibbs, Fazlul Karim, Julien Lerat, Steve Marvanek, Cherry Mateo, Catherine Ticehurst, Biao Wang Surface water storage Cuan Petheram, Giulio Altamura8, Fred Baynes9, Jamie Campbell11, Lachlan Cherry11, Kev Devlin4, Nick Hombsch8, Peter Hyde8, Lee Rogers, Ang Yang Note: Assessment team as at September, 2024. All contributors are affiliated with CSIRO unless indicated otherwise. Activity Leaders are underlined. 1James Cook University; 2DBP Consulting; 3Badu Advisory Pty Ltd; 4Independent contractor; 5 Centre for Tropical Water and Aquatic Ecosystem Research. James Cook University; 6CloudGMS; 7NT Department of Environment, Parks and Water Security; 8Rider Levett Bucknall; 9Baynes Geologic; 10QG Department of Environment, Science and Innovation; 11Entura Shortened forms For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Units For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Common and scientific names For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Preface Sustainable development and regional economic prosperity are priorities for the Australian, NT and Queensland governments. In the Queensland Water Strategy, for example, the Queensland Government (2023) looks to enable regional economic prosperity through a vision that states ‘Sustainable and secure water resources are central to Queensland’s economic transformation and the legacy we pass on to future generations.’ Acknowledging the need for continued research, the NT Government (2023) announced a Territory Water Plan priority action to accelerate the existing water science program ‘to support best practice water resource management and sustainable development.’ Governments are actively seeking to diversify regional economies, considering a range of factors, including Australia’s energy transformation. The Queensland Government’s economic diversification strategy for North West Queensland (Department of State Development, Manufacturing, Infrastructure and Planning, 2019) includes mining and mineral processing; beef cattle production, cropping and commercial fishing; tourism with an outback focus; and small business, supply chains and emerging industry sectors. In its 2024–25 Budget, the Australian Government announced large investment in renewable hydrogen, low-carbon liquid fuels, critical minerals processing and clean energy processing (Budget Strategy and Outlook, 2024). This includes investing in regions that have ‘traditionally powered Australia’ – as the North West Minerals Province, situated mostly within the Southern Gulf catchments, has done. For very remote areas like the Southern Gulf catchments (Preface Figure 1-1), the land, water and other environmental resources or assets will be key in determining how sustainable regional development might occur. Primary questions in any consideration of sustainable regional development relate to the nature and the scale of opportunities, and their risks. How people perceive those risks is critical, especially in the context of areas such as the Southern Gulf catchments, where approximately 27% of the population is Indigenous (compared to 3.2% for Australia as a whole) and where many Indigenous Peoples still live on the same lands they have inhabited for tens of thousands of years. About 12% of the Southern Gulf catchments are owned by Indigenous Peoples as inalienable freehold. Access to reliable information about resources enables informed discussion and good decision making. Such information includes the amount and type of a resource or asset, where it is found (including in relation to complementary resources), what commercial uses it might have, how the resource changes within a year and across years, the underlying socio-economic context and the possible impacts of development. Most of northern Australia’s land and water resources have not been mapped in sufficient detail to provide the level of information required for reliable resource allocation, to mitigate investment or environmental risks, or to build policy settings that can support good judgments. The Southern Gulf Water Resource Assessment aims to partly address this gap by providing data to better inform decisions on private investment and government expenditure, to account for intersections between existing and potential resource users, and to ensure that net development benefits are maximised. Preface Figure 1-1 Map of Australia showing Assessment area (Southern Gulf catchments) and other recent CSIRO Assessments FGARA = Flinders and Gilbert Agricultural Resource Assessment; NAWRA = Northern Australia Water Resource Assessment. The Assessment differs somewhat from many resource assessments in that it considers a wide range of resources or assets, rather than being a single mapping exercises of, say, soils. It provides a lot of contextual information about the socio-economic profile of the catchments, and the economic possibilities and environmental impacts of development. Further, it considers many of the different resource and asset types in an integrated way, rather than separately. The Assessment has agricultural developments as its primary focus, but it also considers opportunities for and intersections between other types of water-dependent development. For example, the Assessment explores the nature, scale, location and impacts of developments relating to industrial, urban and aquaculture development, in relevant locations. The outcome of no change in land use or water resource development is also valid. The Assessment was designed to inform consideration of development, not to enable any particular development to occur. As such, the Assessment informs – but does not seek to replace – existing planning, regulatory or approval processes. Importantly, the Assessment does not assume a given policy or regulatory environment. Policy and regulations can change, so this flexibility enables the results to be applied to the widest range of uses for the longest possible time frame. It was not the intention of – and nor was it possible for – the Assessment to generate new information on all topics related to water and irrigation development in northern Australia. Topics For more information on this figure please contact CSIRO on enquiries@csiro.au not directly examined in the Assessment are discussed with reference to and in the context of the existing literature. CSIRO has strong organisational commitments to Indigenous reconciliation and to conducting ethical research with the free, prior and informed consent of human participants. The Assessment allocated significant time to consulting with Indigenous representative organisations and Traditional Owner groups from the catchments to aid their understanding and potential engagement with its requirements. The Assessment did not conduct significant fieldwork without the consent of Traditional Owners. CSIRO met the requirement to create new scientific knowledge about the catchments (e.g. on land suitability) by synthesising new material from existing information, complemented by remotely sensed data and numerical modelling. Functionally, the Assessment adopted an activities-based approach (reflected in the content and structure of the outputs and products), comprising activity groups, each contributing its part to create a cohesive picture of regional development opportunities, costs and benefits, but also risks. Preface Figure 1-2 illustrates the high-level links between the activities and the general flow of information in the Assessment. Preface Figure 1-2 Schematic of the high-level linkages between the eight activity groups and the general flow of information in the Assessment Assessment reporting structure Development opportunities and their impacts are frequently highly interdependent and, consequently, so is the research undertaken through this Assessment. While each report may be read as a stand-alone document, the suite of reports for each Assessment most reliably informs discussion and decisions concerning regional development when read as a whole. For more information on this figure please contact CSIRO on enquiries@csiro.au The Assessment has produced a series of cascading reports and information products: • Technical reports present scientific work with sufficient detail for technical and scientific experts to reproduce the work. Each of the activities (Preface Figure 1-2) has one or more corresponding technical reports. • A catchment report, which synthesises key material from the technical reports, providing well- informed (but not necessarily scientifically trained) users with the information required to inform decisions about the opportunities, costs and benefits, but also risks, associated with irrigated agriculture and other development options. • A summary report provides a shorter summary and narrative for a general public audience in plain English. • A summary fact sheet provides key findings for a general public audience in the shortest possible format. The Assessment has also developed online information products to enable users to better access information that is not readily available in print format. All of these reports, information tools and data products are available online at https://www.csiro.au/southerngulf. The webpages give users access to a communications suite including fact sheets, multimedia content, FAQs, reports and links to related sites, particularly about other research in northern Australia. Executive summary The Southern Gulf Water Resource Assessment (the Assessment) was commissioned by the Australian Government and led by CSIRO in collaboration with the Northern Territory (NT) and the Queensland (Qld) governments. The Assessment area spans the NT and Qld border and comprises the Settlement Creek, Gregory – Nicholson, Leichhardt and Morning Inlet river catchments, and the selected larger islands of the adjacent Wellesley Island group in the Gulf of Carpentaria. Combined, these lands cover 108,200 km2 and is dominated by extensive cattle grazing. The Assessment aims to support decision making for sustainable regional development for the Southern Gulf catchments in the NT and Qld by clarifying the scale and nature of opportunities and limitations for agriculture and other uses of water resources. A fundamental input to any assessment of water resource development for agriculture is an understanding of the soil and land resources that are present, their spatial distribution, and the limitations to their uses. Specifically, knowledge is needed of the potential suitability of soils for a range of crops, planting seasons and irrigation management. This report details the digital soil mapping (DSM) and subsequent land suitability analysis for the catchment. Four major tasks were completed: • New soil data were collected covering some of the gaps in pre-existing surveys. • Soil attributes selected to support land suitability analyses (e.g. surface soil pH, clay content, A horizon depth, soil thickness, available water capacity and soil types (Soil Generic Groups – SGGs)) were modelled with a spatial resolution of approximately 30 m on the ground. • Digital data and maps were combined with environmental digital data (e.g. climate) into a land suitability modelling framework to determine suitability of land for a range of crop groups under various management scenarios (i.e. planting seasons and irrigation types (land uses)). Rainfed agriculture was also included as a land use. • Reliability of mapped outputs were statistically tested and mapped, and accompany mapped outputs. Reliability maps inform users of the level of confidence they can place in mapped outputs according to personal risk/reward preferences. The DSM applied new soil survey data from 97 sampling sites, which were augmented by soil survey data from 712 pre-existing sampling sites. New soil survey sites were located using statistical methods to infill geographic gaps in the pre-existing dataset. Due to access limitations not all statistically generated observation sites were able to be sampled. The DSM used predictor covariates from various national datasets, including the Soil Landscape Grid of Australia. Access-to-land issues restricted the sampling of more sites, especially in the lower parts of the catchment. For example, we have less data for the Doomadgee Plain than we would have liked, which should be noted when its prospects are discussed. The digital land suitability analysis was completed for agriculture and aquaculture. Agricultural land suitability analysis was completed for 21 crop groups, including horticulture and silviculture. Crop groups comprise similar sets of crops with similar growing and management needs. The land suitability activity followed the standard Food and Agriculture Organization of the United Nations (FAO) schema, which estimates suitability of crops under various management scenarios using a 5- class ranking system from ‘land highly suitable with negligible limitations’ (class 1) to ‘land unsuitable with extreme limitations’ (class 5). The overall suitability for each 30 m grid of the resulting suitability data is determined by the most limiting soil and land factor for the grid cell. The land suitability framework does not include flooding as commonly applied in the standard FAO schema; surface hydrology is covered by parallel Assessment activity. The framework and associated system can be updated as new data for soils, crop varieties, land management practices and climate become available. The suitability of 58 land use combinations (i.e. crop group by season by irrigation type) were evaluated and is presented in the report. We report of 58 land use combinations, crop group 14 (perennial grass Rhodes) under spray irrigation is the most suited to the Assessment area conditions with 5,103,095 ha (47.1% of the area) suitable. Other prospective land uses include crop group 3, intensive horticulture (e.g. cucurbits) under dry-season trickle irrigation (4,918,110 ha; 45.4%), crop group 9, small-seeded crops (e.g. chia) under dry-season spray irrigation (4,757,511 ha; 43.9%), and crop group 12, annual forage (e.g. sorghum) under dry-season spray irrigation (4,714,796 ha; 43.5%). The most versatile agricultural lands are on the Doomadgee Plain dominated by brown, yellow and grey loamy soils, and brown, yellow and grey sandy soils, the Armraynald Plain and Barkly Tableland with cracking clay soils, and parts of the Gulf Fall with red sandy soils, brown, yellow and grey sandy soils, and smaller tracts of cracking clay soils. In terms of marine aquaculture, there is 300,206 ha (2.8%) suited to lined aquaculture and 193,600 ha (1.8%) for earthen ponds. For freshwater aquaculture, lined options are suitable for 6,265,400 ha (57.9%) and 2,408,273 ha (22.3%) for earthen ponds. The report presents key land and soil vulnerabilities to degradation through land development or ongoing agricultural practices. These vulnerabilities relate to the inherent attributes of the land and soil properties. It is important to emphasise what is reported is suitable to be used for coarse-scale land appraisal, which is consistent with a reconnaissance-type of land assessment. As such, the reported outputs are not suitable for planning and development needs at scheme, property or paddock level. Satisfying those needs requires additional investigation at a commensurate intensity. Given the system-wide approach to assessing opportunity and constraints to intensification of agriculture in the region, this report should be considered within the context of the other Assessment activity reports. These activities include climate; surface water hydrology; groundwater hydrology; agriculture and socio-economics; water storage; socio-economics; Indigenous Peoples’ water values, rights and development aspirations; and terrestrial, aquatic and marine ecology. These companion studies are also published. All data from the Assessment are available to the public via CSIRO’s Data Access Portal (CSIRO Data Access Portal ). Contents Director’s foreword .......................................................................................................................... i The Southern Gulf Water Resource Assessment Team .................................................................. ii Shortened forms .............................................................................................................................iii Units ............................................................................................................................... iv Common and scientific names ........................................................................................................ v Preface ............................................................................................................................. viii Executive summary ........................................................................................................................ xii 1 Introduction ........................................................................................................................ 1 1.1 Assessment area .................................................................................................... 3 1.2 Summary of previous soil investigations ............................................................... 6 1.3 Land resource activity design, data and inference system ................................... 7 1.4 Report objectives ................................................................................................... 9 2 Methods ............................................................................................................................ 10 2.1 Pre-existing soil data ........................................................................................... 10 2.2 New soil data and field methods ......................................................................... 12 2.3 Laboratory methods ............................................................................................ 17 2.4 Digital soil mapping ............................................................................................. 18 2.5 Land suitability analysis ....................................................................................... 27 3 Results ............................................................................................................................. 43 3.1 Survey data .......................................................................................................... 43 3.2 Digital soil attribute mapping .............................................................................. 44 3.3 Landscapes and Soil Generic Groups ................................................................... 49 3.4 Soil attribute data and maps ............................................................................... 81 3.5 Land suitability ..................................................................................................... 89 4 Synthesis ......................................................................................................................... 109 References ........................................................................................................................... 113 Appendices ........................................................................................................................... 122 Available water capacity estimation method .................................................... 123 Supplementary data in digital soil mapping ...................................................... 127 Land use combinations for crop groups and suitability analyses ..................... 132 Land suitability rules for land uses .................................................................... 134 Land suitability rules for aquaculture ................................................................ 162 Soil data sites used in digital soil mapping ........................................................ 163 Maps of land suitability options ........................................................................ 166 Figures Preface Figure 1-1 Map of Australia showing Assessment area (Southern Gulf catchments) and other recent CSIRO Assessments .................................................................................................... ix Preface Figure 1-2 Schematic of the high-level linkages between the eight activity groups and the general flow of information in the Assessment ........................................................................ x Figure 1-1 Location of the Southern Gulf Water Resource Assessment study area in northern Australia, including settlements and the major rivers overlaid on hillshade relief. Previous assessment areas are also shown ................................................................................................... 2 Figure 1-2 The Southern Gulf Water Resources Assessment study area (108,200 km2) overlaid by physiographic units and hillshade relief .................................................................................... 5 Figure 1-3 Land suitability activity workflow and key inputs and processes .................................. 8 Figure 2-1 Location of pre-existing and new soil sampling sites in and neighbouring the Southern Gulf catchments ............................................................................................................ 11 Figure 2-2 Collecting soil cores. A trailer-mounted push core rig was used to collect samples to a maximum depth of 1.5 m.............................................................................................................. 14 Figure 2-3 Digital soil mapping models built from the spatial intersection of field and laboratory observations (drill arrows) and the covariate stack ..................................................................... 19 Figure 2-4 Selection of covariates used in digital soil mapping with underlying hillshade, including (a) slope %, (b) ternary gamma radiometrics, (c) weathering intensity index, (d) aridity index, (e) mean fractional vegetation cover and (f) bare earth – SWIR1 band ............................ 23 Figure 2-5 Examples of heavily dissected sections of floodplains on (a) the lower Gregory River and (b) Beames Brook ................................................................................................................... 40 Figure 3-1 The Soil Generic Groups of the Southern Gulf catchments produced by digital soil mapping. The inset map shows the data reliability, based on the confusion index as described in Section 2.4.4 .................................................................................................................................. 54 Figure 3-2 Soil profile of the Grey Vertosol (SGG 9) sampled on the northern part of the Armraynald Plains physiographic unit .......................................................................................... 57 Figure 3-3 Grey Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood on the northern part of the Armraynald Plains physiographic unit ......................................................... 58 Figure 3-4 Soil profile of the Grey Vertosol (SGG 9) sampled on the southern part of the Armraynald Plains physiographic unit .......................................................................................... 58 Figure 3-5 Grey Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood and Bauhinia on the southern part of the Armraynald Plains physiographic unit .............................. 59 Figure 3-6 Soil profile of the Brown Vertosol (SGG 9) sampled on Armraynald Plains physiographic unit, east of the Leichhardt River .......................................................................... 59 Figure 3-7 Brown Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood and gutta percha on the Armraynald Plains physiographic unit, east of the Leichhardt River ........... 60 Figure 3-8 Soil profile of Black Vertosol (SGG 9) sampled on the alluvial plains of the Gregory River west of the Armraynald Plains physiographic unit .............................................................. 60 Figure 3-9 Black Vertosol (SGG 9) landscape of Mitchell Grass Downs on the alluvial plains of the Gregory River west of the Armraynald Plains physiographic unit ................................................ 61 Figure 3-10 Soil profile of Brown Dermosol (SGG 2) sampled on Armraynald Plains physiographic unit, middle reach of the Leichhardt River ................................................................................... 62 Figure 3-11 Brown Dermosol (SGG 2) landscape of buffel grass and open woodland with silver leaf box on the Armraynald Plains physiographic unit, middle reach of the Leichhardt River .... 62 Figure 3-12 Soil profile of Red Chromosol (SGG 1.1) sampled on the southern Armraynald Plains physiographic unit, west of the Leichhardt River ......................................................................... 63 Figure 3-13 Red Chromosol (SGG 1.1) landscape of buffel grass / Mitchell grass in open woodland with silver leaf box and whitewood on the southern part of the Armraynald Plains physiographic unit ......................................................................................................................... 63 Figure 3-14 Soil profile of Red Arenosol (SGG 6.1) sampled near Doomadgee on Armraynald Plains physiographic unit north of the Nicholson River ................................................................ 64 Figure 3-15 Red Arenosol (SGG 6.1) landscape of open woodland of Darwin box, Bauhinia and Cooktown ironwood near Doomadgee on the Armraynald Plains physiographic unit north of the Nicholson River ............................................................................................................................. 64 Figure 3-16 Soil profile of Red Kandosol (SGG 4.1) sampled near Doomadgee on the Armraynald Plains physiographic unit south of the Nicholson River ............................................................... 65 Figure 3-17 Red Kandosol (SGG 4.1) landscape of open woodland with silky browntop grass and black speargrass and Bauhinia, rough leaf cabbage gum and ghost gum near Doomadgee on the Armraynald Plains physiographic unit south of the Nicholson River ........................................... 66 Figure 3-18 Soil profile of Brown Kandosol (SGG 4.2) sampled on an the Armraynald Plains physiographic unit south of the Nicholson River .......................................................................... 67 Figure 3-19 Brown Kandosol (SGG 4.2) landscape of disturbed landscape with bare areas (scalding) on the Armraynald Plains physiographic unit south of the Nicholson River ................ 67 Figure 3-20 Soil profile of Grey Vertosol (SGG 9) sampled on the Barkly Tablelands physiographic unit near Morestone Station ................................................................................. 68 Figure 3-21 Grey Vertosol (SGG 9) landscape of grazed Mitchell Grass Downs showing the common dolomite stones on the Barkly Tableland. Many examples of these soils are less stony ....................................................................................................................................................... 69 Figure 3-22 Soil profile of Yellow Chromosol (SGG 1.2) sampled on the Donors Plateau physiographic unit ......................................................................................................................... 71 Figure 3-23 Yellow Chromosol (SGG 1.2) landscape of woodland with spinifex, silver leaf box and broadleaf paperbark on the Donors Plateau physiographic unit .......................................... 71 Figure 3-24 Soil profile of Calcarosol (SGG 10) sampled in the Dissected Barkly Tableland physiographic unit ......................................................................................................................... 74 Figure 3-25 Calcarosol (SGG 10) landscape of spinifex and bull Flinders grass woodland with many limestone cobbles in the Dissected Barkly Tableland physiographic unit .......................... 74 Figure 3-26 Soil profile of Rudosol (SGG 7) sampled in the Dissected Barkly Tableland physiographic unit ......................................................................................................................... 75 Figure 3-27 Rudosol (SGG 7) landscape of spinifex grassland with snappy gum in the Dissected Barkly Tablelands .......................................................................................................................... 75 Figure 3-28 Rudosol (SGG 7) site of very abundant surface coarse fragments in the Isa Highlands physiographic unit ......................................................................................................................... 76 Figure 3-29 Shallow soil profile of Rudosol (SGG 7) profile sampled in the Isa Highlands physiographic unit ......................................................................................................................... 77 Figure 3-30 Rudosol (SGG 7) landscape of open woodland in the Isa Highlands physiographic unit ................................................................................................................................................ 77 Figure 3-31 Soil Generic Group map showing areas (A–F) referenced in Table 3-7. These locations identify the more extensive areas of potential agricultural development. Inset map shows the data reliability, based on the confusion index as described in Section 2.4.4 ............. 79 Figure 3-32 Distribution of (a) soil thickness and (b) the companion reliability mapping in the Southern Gulf catchments ............................................................................................................ 82 Figure 3-33 Distribution of (a) surface texture class and (b) the companion reliability mapping in the Southern Gulf catchments ...................................................................................................... 83 Figure 3-34 Distribution of (a) available water capacity (AWC) in millimetres to 100 cm depth and (b) the companion reliability mapping in the Southern Gulf catchments ............................. 84 Figure 3-35 Distribution of (a) soil permeability and (b) the companion reliability mapping in the Southern Gulf catchments ............................................................................................................ 85 Figure 3-36 Distribution of (a) surface pH and (b) the companion reliability mapping in the Southern Gulf catchments ............................................................................................................ 86 Figure 3-37 Distribution of (a) surface rockiness and (b) the companion reliability mapping in the Southern Gulf catchments ...................................................................................................... 87 Figure 3-38 Distribution of potential acid sulfate soils in the Southern Gulf catchments ........... 88 Figure 3-39 Modelled land suitability (a) for crop group 7, ‘grain and fibre crops’ such as cotton or sorghum, grown using furrow irrigation and (b) grown during wet season, relying on rainfall ....................................................................................................................................................... 90 Figure 3-40 Modelled land suitability (a) for crop group 10, ‘pulse crops (food legumes)’ such as mungbean or soybean, grown using furrow irrigation in the dry season and (b) wet-season rainfed ........................................................................................................................................... 91 Figure 3-41 Modelled land suitability (a) for crop group 14, ‘hay and forage (perennial)’ such as Rhodes grass, grown using spray irrigation and (b) for crop group 12, ‘hay and forage (annual)’ such as sorghum (forage) and maize (silage), grown using spray irrigation in the dry season .... 92 Figure 3-42 Modelled land suitability (a) for crop group 13, ‘hay and forage (annual)’ such as lablab, grown using furrow irrigation in the dry season and (b) crop group 1 ‘tree crops/horticulture (fruit)’ such as mango and lychee, grown using trickle irrigation .................. 93 Figure 3-43 Modelled land suitability (a) for crop group 2, ‘tree crops/horticulture (fruit)’ such as citrus, grown using trickle irrigation and (b) group 19, ‘oilseeds’ such as sunflower or sesame using wet-season spray irrigation ................................................................................................. 94 Figure 3-44 Modelled land suitability (a) for crop group 9, ‘small-seeded crops’ such as chia, grown using dry-season spray irrigation and (b) for crop group 3 ‘intensive horticulture (vegetables, row crops)’ such as cucurbits using dry-season trickle irrigation ............................ 95 Figure 3-45 Modelled land suitability for (a) crop group 6, ‘root crops’ such as sweet potato, grown using dry-season spray irrigation and (b) crop group 15, ‘silviculture/forestry’ such as Indian sandalwood grown using trickle irrigation ........................................................................ 96 Figure 3-46 Area (ha) of the Southern Gulf catchments mapped in each of the land suitability classes for the 14 exemplar land use options ............................................................................... 97 Figure 3-47 Agricultural versatility for (a) spray, (b) trickle and (c) furrow irrigation types, and (d) rainfed ...................................................................................................................................... 99 Figure 3-48 Agricultural versatility index map combining 14 unique land use options ............. 101 Figure 3-49 Land suitability for marine aquaculture in (a) lined ponds and (b) earthen ponds proximal to coastal areas ............................................................................................................ 102 Figure 3-50 Land suitability for freshwater aquaculture in (a) lined ponds and (b) earthen ponds ..................................................................................................................................................... 103 Figure 3-51 Contiguous suitable areas workflow outputs from the Doomadgee Plains ............ 105 Figure 3-52 Satellite image of land south of the Gregory and Nicholson river confluence showing an extensive network of streams on the Gregory River floodplain. Streams deeper than one metre below land surface are highlighted in red ....................................................................... 106 Figure 3-53 Example propagation of apparent artefacts............................................................ 108 Tables Table 1-1 The land areas of the catchments and islands in the Assessment area ......................... 3 Table 1-2 Physiographic units of the Southern Gulf catchments ................................................... 6 Table 2-1 Field-based soil attributes and methods of analysis. NCST method equates to National Committee on Soil and Terrain (2009) description systems ......................................................... 15 Table 2-2 Soil analyses .................................................................................................................. 17 Table 2-3 List of covariates used in new soil sampling design for new site selection and in digital soil mapping (soil attributes and Soil Generic Groups).  symbol indicates stage used ............. 20 Table 2-4 Land suitability classes based on the Food and Agriculture Organization of the United Nations guidelines (FAO, 1976, 1985) ........................................................................................... 28 Table 2-5 Crop groups (1 to 21) and individual land uses evaluated for irrigation potential as used in the Victoria River Water Resource Assessment (Thomas et al., 2024) ............................ 29 Table 2-6 Land suitability limitations and source data ................................................................. 35 Table 2-7 Rules to satisfy () and or not satisfy () for minimum contiguous area and width for each crop group (Table 2-5) .......................................................................................................... 39 Table 2-8 List of crop groups (Table 2-5) for each minimum contiguous area rule from Table 2-7 ....................................................................................................................................................... 40 Table 3-1 Summary of soil data sites collated for the digital soil mapping component of the Assessment including new and pre-existing data within the Assessment area (Southern Gulf catchments) and pre-existing data outside the Assessment area but within the modelling extent ....................................................................................................................................................... 43 Table 3-2 New site data collected during the 2022 field seasons by site type (see Figure 2-1) .. 44 Table 3-3 Random forest model performance and details for the continuous soil attribute map products ........................................................................................................................................ 46 Table 3-4 Random forest model performances and details for the categorical soil attribute map products ........................................................................................................................................ 47 Table 3-5 Soil Generic Group description, management considerations and correlations to the Australian Soil Classification ......................................................................................................... 50 Table 3-6 The area that each Soil Generic Group area compositions comprises of the Assessment area. Areas without parentheses show mainland hectarages whereas areas in parentheses are for the Wellesley Islands .................................................................................... 55 Table 3-7 Qualitative land evaluation observations of large tracts of land with agricultural potential in the Assessment area ................................................................................................. 80 1 Introduction Knowledge of soils and the landscapes they occupy is critical for determining the opportunities for land intensification, especially for irrigated agriculture. Much of the soil in northern Australia is ancient and highly weathered (Reimann et al., 2012). This means that these soils frequently have a low fertility status (i.e. available phosphorus, total nitrogen, organic carbon and exchangeable cations). Soils may also be saline (Webb et al., 1974) or have poor structure. The often meagre fertility status of these soils results in naturally sparse vegetation, leaving them prone to erosion (Brooks et al., 2009; Pillans, 1997). However, areas do exist where soils are richer in nutrients and are well structured to make them potentially suitable for irrigation; often these are younger soils formed from Quaternary alluvium. There are limited extensive tracts of these and other suitable soils in northern Australia but given the vastness of northern river catchments, areas of good soils may be extensive enough to make irrigated agriculture a viable proposition. In locating these potentially useful pockets of land it is necessary to firstly, understand the location and characteristics of the soils, then secondly, to assess their suitability in the context of broader water, landscape, environmental and economic factors. This report describes the approaches used in the land suitability activity of the Southern Gulf Water Resource Assessment (the Assessment) shown in Figure 1-1. The Assessment encompasses four river catchments flowing into the southern Gulf of Carpentaria on or near the Northern Territory (NT) and Queensland (Qld) border, as well as selected larger islands of the Wellesley Islands group in the Gulf of Carpentaria off the coast of the catchments. The Southern Gulf Water Resource Assessment is the next phase in a sequence of catchment assessments in northern Australia that has included: 1. The Flinders and Gilbert Agricultural Resource Assessment combining the river catchments of the Flinders and Gilbert in northern Queensland and completed in 2014 (Bartley et al., 2013). 2. The Northern Australia Water Resource Assessment, combining WA’s Fitzroy River catchment, the NT’s Darwin river catchments, and Queensland’s Mitchell River catchment, completed in 2018 (Thomas et al., 2018a; Thomas et al., 2018b). 3. The Roper River catchment (Thomas et al., 2022) in the NT. 4. The Victoria River catchment (Thomas et al., 2024) in the NT. Figure 1-1 Location of the Southern Gulf Water Resource Assessment study area in northern Australia, including settlements and the major rivers overlaid on hillshade relief. Previous assessment areas are also shown Northern Australia overview map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-500_NorthAust_Landscape_v4RivTxt_10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au In many respects this activity, as with the previous assessments, captures the meaning of earlier studies on soil and land resources of northern Australia (e.g. Wilson et al., 2009) that were based on limited pre-existing soil data. This data constraint has substantially limited the usefulness and applicability of those studies, so hence the need for new data to deliver new knowledge and insights to the land and utility of northern Australia. The report details approaches applied to assess the suitability of 58 agricultural land intensification (land use) options in the Assessment area that combines plausible land use and management scenarios (land uses) for the agricultural setting. This includes 21 crop groups, three growing seasons (wet season, dry season, perennial) and three irrigation types (furrow flood, spray and trickle irrigation). Rainfed farming as a ‘fourth’ growing option is included because it may be viable under some circumstances in the Assessment area. The suitability of land for aquaculture using freshwater and marine species is also assessed. Before determining the suitability for these land use options was possible, various processes to digitally map land and soil attributes were undertaken, and the maps then incorporated into a digital land suitability decision-support framework to identify areas of potential agriculture and aquaculture. 1.1 Assessment area The Southern Gulf catchments Assessment area shown in Figure 1-1 spans the NT and Queensland border and comprises the Settlement Creek, Gregory – Nicholson, Leichhardt and Morning Inlet river catchments, and the selected larger islands of the adjacent Wellesley Island group in the Gulf of Carpentaria. The combined lands cover 108,200 km2 (Table 1-1). The rivers in the Assessment area flow into the southern Gulf of Carpentaria. Table 1-1 The land areas of the catchments and islands in the Assessment area For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au According to the National Native Title Tribunal (2022), the Assessment area is home to eleven Indigenous language groups including the Indjalandji – Dhidhanu, Kalladoon, Kaiadilt, Kukatj, Kurtijar, Garawa, Gangalidda, Gkuthaarn, Lardil, Waanyi and Yangkaal peoples who have occupied the Assessment area for tens of thousands of years. The voyage of the HMS Investigator in 1802, led by Matthew Flinders, was the first authentically recorded European exploration of the region, which followed the Gulf of Carpentaria coastline in 1802. After this John Stokes and his crew on the HMS Beagle expedition in 1841 recorded sailing up the Albert River for 80 km, and he wrote of the hinterlands as ‘the Plains of Promise’. In 1845, Ludwig Leichhardt led the first European land exploration of the area and traversed the lower reaches of all the rivers and creeks from the Mitchell River in Queensland to the NT–Queensland border. Augustus Gregory, in 1856, was the first European on record to explore the NT hinterlands of the Assessment area. The first colonial settlement in the area was Burketown in 1864 on the Albert River. However, in 1866, fever swept the area with most of Burketown’s population dying and the survivors evacuating to Sweers Island in the Wellesley Islands. The town was re-established in 1875 and the town’s artesian bore was drilled in 1897 (Perry et al., 1964). Mining began around Mount Isa in the 1930s, and now Mount Isa city is the largest settlement in the region. Century Mine in the Lawn Hill area began production in the 1990s. Other towns in the Assessment area are Camooweal, Doomadgee, and Mornington on Mornington Island (Wellesley Islands). There is rail access at Mount Isa and shipping access at the Port of Dalrymple on the Albert River, north of Burketown. The Southern Gulf catchments have a wet–dry tropical climate with a highly seasonal rainfall with marked inter-year variability. Large variability in annual runoff occurs, and the strong seasonality in rainfall results in large wet-season flows and small dry-season flows (Petheram et al., 2009; Petheram et al., 2013). The climate of the area varies from arid tropical in the south-west to humid tropical in the north-east. Most of the rainfall occurs in the four summer months (December to March) and comes from a north-western influence mainly due to tropical cyclones. The south-west is the driest, becoming gradually wetter in a northerly direction. Contemporary long-term records (Bureau of Meteorology, 2023) confirm the north-western influence for rainfall with higher mean annual rainfall in the west (Wollogorang lat 17.21°S, long 137.95°E is 948 mm) compared to the east (Burketown lat 17.74°S, long 139.55°E is 780 mm) at a similar latitude. The north–south rainfall trend also shows up in the rainfall records; for example, Century Mine (lat 18.76°S, long 138.71°E) in the central part of the Assessment area has a mean annual rainfall of 568 mm. In Camooweal (lat 19.92°S, long 138.12°E), the mean annual rainfall is 423 mm. In winter, mean annual rainfall is practically nil, but slightly more in the south when compared to the north. Mean annual pan evaporation according to contemporary long-term records increases from the coast toward the south-west. For example, in the north-east (Burketown; 2429 mm), central part (Century Mine; 2902 mm) and south-west (Camooweal; 3061 mm). In all cases this exceeds annual rainfall significantly. The rainfall deficit is largest in the south-west and lowest on the coast. Mean annual maximum temperatures are similar on the coast (Burketown; 32 °C) and in the far- south (Camooweal 33 °C), and hottest in the central part (Century Mine; 35 °C). Mean minimum temperatures are similar on the coast (Burketown; 20 °C) and the central part (Century Mine; 20 °C) and coldest in the south (Camooweal; 18 °C). The physiography of the Assessment area is shown in Figure 1-2, which has been adapted from the work of others (Grimes, 1974; Stewart, 1954; Twidale, 1956). Ten units occur in the area; these are listed in Table 1-2. Physiographic units (PU) serve as a useful framework to understand the potential agricultural lands and soils in terms of qualities and limitations because units are formed from distinct groups of lithologies with different geological histories to give rise to a particular set of soil types and geomorphic patterns. Links between PUs and soils are expanded further in this report. Figure 1-2 The Southern Gulf Water Resources Assessment study area (108,200 km2) overlaid by physiographic units and hillshade relief The area comprises the Settlement Creek, Gregory – Nicholson, Leichhardt and Morning Inlet river catchments, and the selected larger islands of the adjacent Wellesley Island group in the Gulf of Carpentaria. Southern Gulf catchments overview map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-501_location_v2_v11_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Table 1-2 Physiographic units of the Southern Gulf catchments For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 1.2 Summary of previous soil investigations Soils were first systematically investigated in the Southern Gulf catchments in the 1940s by Whitehouse et al. (1947), who were investigating the Armraynald Plain as a potential site for an agricultural research station. Their investigations were limited to surface samples from black soils. In 1951 Hubble and Beckmann (1957) briefly visited six properties of the Australian Estate Pastoral Company of which Kamilaroi Station is in the Assessment area. The first general systematic land resource survey in the Assessment area was undertaken by the CSIRO Northern Australia Regional Survey unit in 1947 and 1948 under leadership of GA Stewart (Christian and Stewart, 1953). This survey covered the area west of the Gregory River to the western edge of the Assessment area. The lands east of the Gregory River were systematically surveyed 5 years later (1953 and 1954) by the CSIRO Division of Land Research and Regional Survey (Perry et al., 1964). In 1987, a 1:250,000 scale land system mapping project was carried out by Aldrick and Wilson (1990) in the NT part of the Assessment area. This broad-scale land system information supported the development of a regional land use strategy. From the mid-1980s to the 1990s, 1:100,000 scale land resource surveys were undertaken on pastoral leases across parts of the Barkly Tableland. The survey on Mittiebah Station (Edgoose, 1987) is partly within the Assessment area. This information aimed to support sustainable pastoral land management across one of the NT’s most productive pastoral districts. Since these systematic surveys, only sporadic soil investigations have been undertaken in the Assessment area by the Queensland Government for state-wide and regional projects to assess land resources (QDNRME, 2004), acid sulfate soils (ASS) and wetlands (Bryant et al., 2008). Since the pastoral station land resource surveys in the NT, only soil investigations supporting pastoral carrying capacity research have been carried out. This research was restricted to a small number of sites; however, the collected information included detailed soil landscape descriptions of soil pits, soil chemistry, soil physical properties such as particle size analysis and soil bulk density data. None of the historical soil information was sufficiently comprehensive or of a scale to provide a full regional assessment. A key recommendation common to all these assessments was that further soils data were needed before detailed agricultural suitability assessments could be conducted. 1.3 Land resource activity design, data and inference system The land resource assessment approach described by this report builds on research legacies of the series of past assessments (Bartley et al., 2013; Bui et al., 2020; Clifford et al., 2014; Harms et al., 2015; Thomas et al., 2018a; Thomas et al., 2015; Thomas et al., 2018b; Thomas et al., 2022; Thomas et al., 2024). Those assessments demonstrated the value of modern digital approaches in land suitability analysis, and the benefits accrued through increased operational efficiencies, speed of analyses, utility of digital outputs, and an objective understanding of the quality of outputs. Figure 1-3 shows the broad workflow that has been adopted. This highlights phases of soil sampling design, digital soil mapping (DSM) and land suitability analysis, while also showing the dependencies feeding into these, including soil mapping covariates, soil attribute data, map quality assessment, and the land suitability framework that drives the land suitability analysis. As with previous assessments this workflow includes quantitatively mapped estimations of uncertainty. These phases are described more fully below. Figure 1-3 Land suitability activity workflow and key inputs and processes 1.3.1 Sampling design Reliable DSM requires a sampling approach (McKenzie et al., 2008) that minimises bias in the selection of soil sampling sites and maximises the spread of sites so that the full range of soil variability in the study area is sampled (at least to the extent suggested by the covariate space). The sampling design used conditioned Latin hypercube sampling (cLHS), a form of digital stratified random sampling described in Minasny and McBratney (2006). cLHS ensures sampling points capture the distribution of the environmental covariates1 chosen to represent the drivers of soil variability in the area of interest. Thus, given a gridded study area of N total sites with ancillary variables (X), select a subsample of size n (n ⪡ N) in order that X forms a Latin hypercube, or the multivariate distribution of subsample X is maximally stratified. The number of sampling points, n, is determined a priori as from operational considerations like project resourcing and budget. 1 Environmental covariates – or simply covariates – are spatial geographic information system (GIS) format datasets that, based on the principles of soil formation, are expected to have functional relationships to on-ground soil attributes, and so can contribute to prediction of soil attributes. For example, slope may support prediction of soil depth, relief patterns for soil water accumulation, or remote sensing for soil colour. 1.3.2 Digital soil mapping DSM has a successful track record in delivering land and soil information to large-area assessments in Australia (e.g., Bui et al., 2007; Grundy et al., 2020; Kidd et al., 2015; Kidd et al., 2020; Kidd et al., 2014; Searle et al., 2021; Thomas et al., 2015; Viscarra-Rossel et al., 2015) and elsewhere (e.g. Behrens and Scholten, 2007; Hartemink et al., 2010; Hartemink et al., 2013). The success of the approach lies in the fact that DSM has co-evolved with gains in computing power and adaptation of statistical methods. Australia is well positioned to undertake DSM for catchment-wide studies because of a legacy of reliable covariates including climate, remote sensing, terrain derived from For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au digital elevation models (DEMs), and gamma radiometrics (mineralogy, landscape evolution) data layers (Bui, 2007). DSM outputs include maps of soil attributes and soil types created in geographic information system (GIS) grid data format that follow natural patterns of soil change across landscapes. These qualities make DSM outputs suitable for direct incorporation with land suitability analysis frameworks discussed next. DSM also enables production of companion mapping reliability maps showing where the DSM mapping is more or less reliable. These companion maps are useful to users who make objective decisions on fitness-for-purpose of the mapping. Comprehensive texts on DSM are presented elsewhere for readers to follow (Grundy et al., 2020; McBratney et al., 2003; Searle et al., 2021). 1.3.3 Land suitability analysis and framework Land suitability analysis (land suitability) is the process of determining the potential of land for specific land uses on the basis of the local range of environmental attributes and agricultural (and fisheries) qualities (Rossiter, 1996). Land suitability assessment considers the attributes and qualities that limit the maximum potential yield of a land use, terming these factors ‘limitations’. This Assessment defines limitations and builds the analytical framework following the Food and Agriculture Organization of the United Nations (FAO) Framework for Land Evaluation (FAO, 1976; 1985). The framework involves a broad assessment of land suitability that integrates multiple limitations including biophysical (edaphic and climate), social and economic themes (FAO, 2007). This Assessment, however, deviates from FAO’s integrated framework as it constrains analysis only to edaphic and biophysical themes; other aspects are covered in companion Assessment activities (e.g. surface water hydrology; socio-economics; Indigenous water values, rights and development aspirations; and aquatic and marine ecology). The land suitability approach applied in the Assessment follows the lineage of approaches first developed in river catchment assessments for the Flinders and Gilbert (Bartley et al., 2013; Harms et al., 2015), the northern Australia (Thomas et al., 2018b), the Roper (Thomas et al., 2022), and more recently, for the Victoria catchments (Thomas et al., 2024). The edaphic components of the land suitability mostly relate to soil attributes – or limitations – that have a key bearing on the growth and productivity of irrigated land uses, or the amount of land preparation and maintenance of farming infrastructure needed that may affect the financial viability of the irrigation enterprise. For example, soil permeability affects the rate of water application, and rockiness relates to the intensity of rock picking required in land preparation. 1.4 Report objectives The objective of this report is to describe the methods and outputs used to: • address questions around the scale of opportunity for agricultural and aquaculture land use intensification of the Southern Gulf catchments Assessment area • support other Assessment activities with land and soil datasets addressing their catchment resource questions, including agronomic and surface hydrology. 2 Methods 2.1 Pre-existing soil data Previously, the catchment areas of northern Australia water resource assessments have been within a single state or territory, making extraction of pre-existing data straightforward. This Assessment differs in that the study area (with some catchments, namely of the Settlement Creek and the Gregory – Nicholson rivers) crosses the boundaries of the NT and Queensland (Figure 1-2). The standards to which the data were collected and entered is consistent across the jurisdictions, with data quality of surveys being variable. A significant amount of soil data are available from previous studies that have been carried out in the NT section of the Assessment and surrounding areas, while pre-existing data in the Queensland Assessment area are less numerous, particularly inside the catchment areas. Following a comprehensive review of the data by the NT Department of Environment, Parks and Water Security, records were drawn from the NT’s corporate Soil and Land Information System (SALInfo; Department of Environment Parks and Water Security, 2000). A similar process occurred with the Queensland Department of Environment and Science’s Soil and Land Information platform, which is the Queensland Government’s corporate repository for the input, management and security of soils data and information. The CSIRO-managed National Soil Database (NATSoil; Karssies et al., 2011) also contributed records. Despite the age of data and possible inconsistencies caused by prior, less precise, analytical methods in the intervening years, much data remains usable and valuable for this work subject to the selection criteria described. Data were included from areas beyond the borders of the Southern Gulf catchments by the following extents: south of –16.17° and north of –21.49° latitude, east of 136.38° and west of 140.95° longitude. The extent is shown in Figure 2-1. These areas are deemed sufficiently close to the catchments for shared soil forming conditions – hence a similar suite of soil types – so inclusion would boost digital soil mapping (DSM) models. After extraction from the various sources, candidate soil site records were collated in a relational database using standard protocols described in (Jacquier et al., 2012). The following criteria were applied to select suitable records. Where duplicate records were discovered for each site, the most recent record was accepted. If a record was captured in geographic coordinates to more than four decimal places, the record was accepted. If the record was tagged with a positional accuracy of 50 m or better, the record was accepted. Only records that had some below-surface soil description or data were accepted. Where numerous clustered records exist (e.g. from an intense land-development survey), one representative record was selected to reduce modelling bias. Quality checks were conducted to only accept sites that demonstrate logical consistency according to the modelling need. For example, if a record classified a site as a Hydrosol according to the Australian Soil Classification (ASC) (ASC, Isbell and National Committee on Soil and Terrain, 2021) but the site was also described as well drained following the National Committee on Soil and Terrain (2009), the record was rejected for use in drainage mapping because of inconsistency in the classification and the observed drainage class. Figure 2-1 Location of pre-existing and new soil sampling sites in and neighbouring the Southern Gulf catchments DSM = digital soil mapping. Location of soil sites map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-502_SoilSites_v2_v11_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au 2.2 New soil data and field methods Field methods followed Australian standards for survey guidelines (McKenzie et al., 2008), soil and site description (National Committee on Soil and Terrain, 2009) and soil classification using ASC (Isbell and National Committee on Soil and Terrain, 2021). All fieldwork was vehicle-based, hence vehicle site access sometimes restricted accessibility to sampling sites. Fieldwork was restricted to the dry season (i.e. outside the wet season) when rivers and creeks were passable, and roads and tracks were dry and traversable. Fieldwork was conducted from April to July 2022. Field teams consisted of CSIRO and Northern Territory Government survey experts, and each field trip comprised at least two remote area-ready vehicles working together. Soil chemical and physical attributes for the newly collected samples were analysed using conventional laboratory methods (McKenzie et al., 2002; Rayment and Lyons, 2011; Thorburn and Shaw, 1987). All field observation data were entered into the Queensland Government Soil and Land Information (SALI, Biggs et al., 2000) database and later transferred to a CSIRO soil database for analysis. Site observations can be viewed on the Queensland Government, Queensland Globe website. Access-to-land issues restricted the sampling of more sites, especially in lower parts of the catchment. For example, we have less data for the Doomadgee Plain than we would have liked, which should be noted when its prospects are discussed. 2.2.1 Design of new soil survey A stratified random sampling design based on conditioned Latin hypercube sampling (cLHS) (Minasny and McBratney, 2006) was used to identify new sampling sites for the DSM that firstly, covered the covariate space, and secondly, infilled geographic gaps in the usable pre-existing soil data discussed in Section 2.1 in Table 2-3. This sampling design technique produces an unbiased selection of sampling sites covering the full range of soil variables represented by the covariates. The design was implemented in the R computing environment (R Core Team, 2018) using the clhs package (Roudier, 2022). The soil–landscape variability was captured by selecting covariates representing soil forming factors (Fitzpatrick, 1980; Jenny, 1941). The seven covariates (Table 2-3) selected in the cLHS sampling design included: • monthly mean annual aridity index (monthly ratio of precipitation to potential evaporation) (Harwood, 2019) • a digital elevation model (DEM) (relief patterns, patterns of through-landscape water movement and residence, etc.) (Gallant and Austin, 2015) • slope percentage (mass wasting) (Gallant and Austin, 2015) • dynamic land cover, mean of 2000–2008 time series (seasonal vegetation dynamics) (Lymburner et al., 2011) • gamma radiometrics comprising elemental images of potassium (K), thorium (Th) and uranium (U) (soil history, age) (Minty et al., 2009). The cLHS sampling was constrained using the cost condition, forcing sampling sites to be within a 300 m buffer each side of mapped roads or tracks. This allowed reduction in field time penalties and made vehicular access more practical. The cLHS sampling design also accounted for pre- existing sampled sites within the study area using the must.include condition of the clhs package. Two hundred new ‘primary’ and 50 ‘secondary’ sampling sites were generated. The primary sites were treated as priority sampling sites. The secondary sites were available to be collected opportunistically after collections from the primary sites had been completed (e.g. if there was time left in the day). Furthermore, primary and secondary sites were accompanied by contingency sites, which were available if these sites could not be accessed because of practical or safety reasons (e.g. fence lines, locked gates, flooding or high flows, etc.). These were located within a radius of 300 m of the site. Ensuring similarity of soils was guaranteed using the similarity index methodology of Brungard and Johanson (2015). A similarity index of 0.8 was required to be an eligible contingency site. This meant that there were typically two to five contingency sites available for each primary or secondary site. Twenty out of the primary 200 sites were selected for subsequent laboratory analysis. These samples were selected to capture the full covariate range of soils already sampled using the R program clhs package (Roudier et al., 2012) and involved using the 200 samples as covariates to constrain selections to the original population. 2.2.2 New soil sampling All sampling site locations were recorded using a Garmin global positioning system (GPS) in WGS84. Samples were taken using either: (i) a 50-mm diameter trailer-mounted push corer (Figure 2-2); or (ii) a 75 mm hand auger to a maximum depth of 1.5 m, or to refusal if bedrock or impenetrable layers (e.g. indurated) were encountered. Cores were extruded and described using standard Australian soil survey notation (National Committee on Soil and Terrain, 2009) to 1.5 m and cores were divided into samples of fixed depth intervals: 0–0.1, 0.2–0.3, 0.5–0.6, 0.8–0.9, 1.1– 1.2 and 1.4–1.5 m. Figure 2-2 Collecting soil cores. A trailer-mounted push core rig was used to collect samples to a maximum depth of 1.5 m Each depth increment was analysed in the field for: • pH 1:5 (Raupach, 1957) • electrical conductivity (EC) 1:5 soil/water • sodic dispersion, using a modified ‘Emerson test’ (McKenzie et al., 2002). Approximately 1 kg of soil from each depth increment was bagged for laboratory analysis (Section 2.3). Site and soil descriptions were recorded, transcribed and uploaded to the Queensland Government Soil and Land Information database. The field observations taken are listed in Table 2-1. Additional field observations were made of surface rockiness and shallow soils (<0.5 m) while driving between sample sites without stopping. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Table 2-1 Field-based soil attributes and methods of analysis. NCST method equates to National Committee on Soil and Terrain (2009) description systems For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 2.2.3 Validation survey Validation surveys have been conducted for previous northern Australia water resource assessments (Bartley et al., 2013; Thomas et al., 2018a; Thomas et al., 2018b; Thomas et al., 2022; Thomas et al., 2024) because their outputs offer another, qualitative perspective to the statistical approach by incorporating valuable field knowledge of soil surveyors familiar with the landscapes. A 2-week field trip was planned for this in May 2023 but was cancelled when permissions from the Carpentaria Land Council and Corporation (CLCAC) were not granted in time. While the DSM model performance was principally checked through the statistical validation approach in Section 2.4, field validation would have added to this expert testing of mapped outputs in situ followed by remodelling in the laboratory for those attributes in need of refinement. To partially resolve this, a workshop was conducted in March 2023 to critically review DSM soil attribute and land suitability outputs, including the project team and stakeholders from the NT and Queensland jurisdictions. This process constructed recommendations to either accept or to remodel mapped attributes where improvements were deemed necessary. 2.2.4 Capturing soil and physiographic knowledge An important complement to the DSM approaches used in the study is application of expert knowledge and experience grounded in traditional soil survey and mapping approaches (Hewitt et al., 2008; McKenzie and Grundy, 2008). This grounding is used to tailor computer methods, assess the outputs and maps, and iteratively refine approaches to improve quality. This was achieved in this Assessment mainly through field-based activities that drew on conventional soil–landscape methodology and paradigms (Fitzpatrick, 1971; 1980; Hudson, 1992) and the application of geomorphic principles to assess the distribution of soils and landscapes. From these, descriptions were generated of major geomorphic units within the various geological settings. The process provided an estimate of the age of deposits (e.g. Quaternary, Tertiary, Mesozoic, etc.) and degree of weathering – knowledge that is important for identifying deposits likely to be more suitable for agriculture (e.g.onQuaternaryalluvium) orforunderstanding soil–landscapeprocesses such aserosion anddeposition, leaching,flooding, waterlogging and salinity. Understandingof the soils and landscapeswere refined into knowledgeframeworksdescribing soiltypes and physiographicunits(PUs).Similar to previousassessments(Bartley et al., 2013; Thomaset al.,2018a;Thomas etal., 2018b; Thomas et al., 2022), the soil units,called Soil GenericGroups(SGGs), were designed tosimultaneouslycovera number of purposes:(i)to be descriptive so as toassist non-expert communication regarding soil and resources; (ii) to be relatable to agricultural potential;and(iii)to align, where practical,to ASC (Isbell and National Committee on Soil andTerrain, 2021).The physiographic understanding of thecatchmentsdraws strongly on prior mapping including lithology(Grimes, 1974; Stewart, 1954; Twidale, 1956), land systems mapping(Christian and Stewart,1953)(Perry et al., 1964)andfield observation, and aggregating thisknowledge into so-calledPUscomprising similar lithology, reliefpatterns andweatheringhistories. PUsserve as a framework tounderstand thepotential agricultural lands and soils in terms ofqualities and limitations,andgiven each unit is derived from adistinct group of lithologiesand landforms,each gives rise to a distinct setof SGGs and geomorphic patterns–although notingsome SGGs can span numerous PUs. 2.3Laboratorymethods Soil physicaland chemicalanalytical techniqueswere used on the newlycollected soil samplesthat werediscussed in Section0. Theanalysesused on these newly collected samples are shown inTable2-2. Table2-2Soil analyses MEASUREMENTELEMENTS AND METHODSREFERENCE Particle size (% sand,Sieve andhydrometer methodThorburn and Shaw, 1987silt, clay) MoistureFor more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Rayment and Lyons, 2011pH, electrical 1:5 soil/waterRayment and Lyons, 2011conductivity (EC), chloride, nitrateExchangeable cationsCation exchange capacity(CEC),exchangeableRayment and Lyons, 2011calcium,magnesium,sodium,potassium, exchangeable sodium percent (ESP) Exchange acidityExchangeable aluminium, H+Rayment and Lyons, 2011Bulk densityRing method using oven dry weightsModified from(Cresswell and Hamilton, 2002) Total elements (totalDry furnaceRayment and Lyons, 2011carbon andnitrogen) Extractable traceIron,manganese,copper,zincRayment andLyons, 2011elementsSurface soil fertilityOrganic carbon (Walkley and Black),total Rayment and Lyons, 2011nitrogen (Kjeldahl),extractablephosphorus Chapter2 Methods|17 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 2.4 Digital soil mapping Many of the modelling approaches applied in modern DSM utilise predictive models that establish relationships between point-based soil observations (i.e. geolocated soil and land attributes) and a set of covariates (McKenzie and Ryan, 1999). These rely on the scorpan approach described in McBratney et al. (2003) that applies covariates for soils (s), climate (c), organisms (o), relief (r), parent material (p), age (a) and neighbourhood (n) (McBratney et al., 2003), and a computer adaptation of the soil formation paradigm (Fitzpatrick, 1980; Jenny, 1941). Some of the best- performing computer models use data mining and machine learning to capture spatial distribution of soil properties without prior assumptions about the form of the complex relations between soils and covariates. Here, covariates with a grid resolution of 30 × 30 m were used for the DSM modelling to produce soil attribute data and maps. This resolution is inherited from national covariate datasets (see Table 2-3), and is consistent in terms of output specifications with the Assessment’s regional scope. DSM models can be expressed as statistically based rules representing the relationship between: (i) the soil attribute at the sampling sites (‘obs’) and their (ii) geographic intersections of the covariates (‘x, y’), as per Figure 2-3. Multiple, co-registered covariates are used in environmental correlation – effectively in a stack of gridded covariates (predictors), as represented in Figure 2-3. Applying the model (and its rules) pixel by pixel across the whole mapping extent predicts the target soil attribute at unsampled locations (i.e. at every pixel covering the study area that does not contain a sampling site). This process of rule-to-covariate matching progresses through the whole study area to compile the complete final soil attribute data. In essence the environmental correlation approach is a digital analogue of the traditional soil surveyor’s mapping approach, which relies on expertise of the soil surveyor to build models (rules) developed from time spent in the field and laboratory data from patterns of relief, drainage or vegetation (i.e. soil covariates) (Hudson, 1992; McKenzie et al., 2008). In the DSM analogue, the ‘expert’ knowledge equates to the statistical model that does the prediction. Figure 2-3 Digital soil mapping models built from the spatial intersection of field and laboratory observations (drill arrows) and the covariate stack A major benefit of DSM compared to the traditional soil surveyor’s approach is that it is possible to statistically quantify and map the reliability – sometimes termed uncertainty – associated with the soil attribute prediction at each pixel. DSM also allows mapping approaches to be applied consistently so that there is no methodological or operator bias contained in the mapped outcome. Therefore, map users can be confident that all areas are systematically comparable, given differing quality and intensity of the data record across the area. Furthermore, this makes updating maps a straightforward process once new soil observations or better covariates become available to re-run the modelling. 2.4.1 Identifying digital soil mapping soil attributes The selection of DSM soil attributes to model and map is governed by the needs of the land suitability analysis (i.e. the limitations required) and SGG mapping. These selections in turn inform the selection of covariates used in the DSM process discussed in the next section. The needs of the land suitability analysis are informed by the candidate crops and their growth needs and thresholds (i.e. limitations). The land suitability method and land suitability limitations are presented in full detail in Section 2.5. Table 2-6 identifies the limitations served directly by DSM outputs (see column ‘Source’), and hence the environmental covariates used. For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 2.4.2Environmental covariates Covariateswere selected as proxies for factors ofsoil formation(Jenny, 1941). The 38covariatesare presented inTable2-3.All thecovariateswerein geographic information system (GIS)rasterfile format(GeoTIFF)and co-registeredto theWGS84 datum and re-sampled to a groundresolution of 1 arcsecond, whichequates to approximately 30×30mgrid resolutionat thelatitudes of the Assessment. Covariates were used in twotasks covered in this report and shown inTable2-1: (i) sevencovariateswere used to select new sampling sites (Section2.2.1), and (ii)in general38covariates were used to model new soilattributes (Section2.4). Table2-3List of covariates used in new soil sampling design for new site selection and indigital soil mapping(soil attributes andSoil GenericGroups). symbol indicates stage used ScorpanSOILFORMATIONFACTORCOVARIATEDESCRIPTIONCUSTODIAN/SOURCESNEW SOILSAMPLING DESIGNDSM SOILATTRIBUTEMAPPING SoilKaolinite (%); Illite(%); Smectite (%) ClimateMonthly mean annual aridityindex Temperature, annual Rainfall, annual Thunder days Rainfallseasonality Land surface Bowen ratio OrganismsNationalDynamic Land Cover dataset Mean FPAR Clay mineral surfaces, 0–0.2 and 0.6–0.8 mdepth intervals The monthly ratio ofprecipitation topotential evaporation (pan, free-water surface) Annualmeandailyminimum temperature Meanrainfall–annual Meanannualthunderstorm days Maximum ofmeanmonthlydifferences between successivemonths The ratio of energyfluxes from one state toanother by sensible heatand latent heatingrespectively Vegetation communitytypes basedon seasonaldynamics in ModerateResolution Imaging Spectroradiometer (MODIS)data, responding to climate and soil type Fraction ofphotosyntheticallyactiveradiation (FPAR)– mean Advanced VeryHigh Resolution Radiometertime series CSIRO: Grundy et al. (2015)and ViscarraRossel (2011)  CSIRO: (Harwood, 2019) CSIRO: (Harwood, 2019)  CSIRO: (Harwood, 2019)   CSIRO: (Harwood, 2019)   Geoscience Australia:Lymburneret al. (2011) https://ecat.ga.gov.au/geonetwork/srv/eng/catalog.search-/metadata/71069  Bureau of Meteorology: http://www.bom.gov.au/jsp/ncc/climate_ averages/thunder-lightning/index.jsp CSIRO:https://portal.tern.org.au/actual- evapotranspiration-australia-cmrset- algorithm/21915 CSIRO: https://portal.tern.org.au/fractional- cover-modis-csiro-algorithm/21786 20|Southern Gulf catchments soils and land suitability ScorpanSOILFORMATIONFACTORCOVARIATEDESCRIPTIONCUSTODIAN/SOURCESNEW SOILSAMPLING DESIGNDSM SOILATTRIBUTEMAPPING Mean bare ground fractionalcover Max non- photosynthetic vegetationfractionalcover Greenvegetationpersistence ReliefElevation Fractional cover BareSoil-Mean MODIStimeseries Fractional cover MODISCSIRO Land and Wateralgorithm Australiacoverage LandsatThematicMapper2000–2010 Persistent Green- Vegetation Fraction 1 arc sec (~30 m) digitalelevation model CSIRO: https://portal.tern.org.au/fractional-  cover-modis-csiro-algorithm/21786  Terrestrial Ecosystems ResearchNetwork: https://portal.tern.org.au/seasonal-  persistent-green-australia-coverage- 23885/23885 CSIRO:(Gallant and Austin, 2015) CSIRO: https://portal.tern.org.au/fractional- cover-modis-csiro-algorithm/21786 Slope (%) Local median300 m Relief aspect Focal range1000 m Focal range300 m MultiResolution Valley Floorindex (MrVBF) MultiResolution Ridge Top Flatness index(MrRTF) Plan curvature Profilecurvature Topographicwetness index Parent Potassium (K); materialThorium (Th); Uranium (U) Gravity Magnetics Slope gradient Median of slope % in 300 mwindow Landform solarexposure Elevation range in 1000 mwindow; longerrange landform patterns Elevation range in 300 mwindow; longer rangelandform patterns Landscape erosional and depositional zones High and flatlandscape zones Landform curvature along the contour Landform curvature directly downslope Landscape zones ofwater accumulation Gamma radiometrics Geologic Bouguergravity anomaly Total magnetic intensity –TMI geologicmagnetism CSIRO:(Gallant and Austin, 2015) CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO:(Gallant and Austin, 2015)  CSIRO: (Gallant and Austin, 2012)  Geoscience Australia:Minty et al. (2009)   Geoscience Australia  Geoscience Australia  Chapter2 Methods|21 ScorpanSOILFORMATIONFACTORCOVARIATEDESCRIPTIONCUSTODIAN/SOURCESNEW SOILSAMPLING DESIGNDSM SOILATTRIBUTEMAPPINGSilicaTotal silicaconcentrationGeoscience Australia Barest Earth–30-year time series ofGeoscience Australia:Roberts etal. BlueLandsatThematic(2019) GreenMapperblue, green, RedSWIR1red, SWIR1 and SWIR2wavelength bands  SWIR2 AgeWeathering Intensity Index(WII) Index of soil-regolithweathering andits effect on soil formationGeoscience Australia: Wilford (2012)  Figure2-4shows a selection of covariatesfromTable2-3, eachrepresenting importantscorpansoil formingfactors. Thefollowingputsthese covariates into ascorpanframeworkwith thediscussions drawing onPUspresentedinFigure1-2andTable1-2. Figure2-4(a) shows slope in the Assessment area as an example of oneofthescorpanrelief soilforming factors. These patternslargely reflectthe parent material and landscape history of thearea.The flatter areascoincide with theCloncurry, Doomadgee, Armraynald and Karumba plainsalso known collectively as the Carpentaria Plain (Grimes, 1974).The flatterareas of theAssessment area are likely tohavethe deepest soils, hence the more agriculturally suitable soils. In contrast, thePUswithsteeper slopesaredominatedby shallow soils, which makeslandsunfavourable foragriculture. The gamma radiometrics (Figure2-4(b)) and weathering intensity index (WII) (Figure2-4(c)) together supply information onthe landscape history of the Assessment area. Gamma radiometrics showsthe geochemistry of the landsurface interms of abundancesfor potassium(K), thorium (Th) anduranium (U)(Minty, 1997), whichrelates to parent materialcompositionandstage of weathering (scorpanparent material and agefactors). The WII(Wilford, 2012; Wilford, 1995)indicatesthe ageof landscape and the history of development (scorpanage factor). Signalsshowing fresher (lessweathered) soils in high relief areas coincidewith sedimentary, volcanic andgraniterocks in the Isa Highlandsthrough the K and U-dominated patterns (Figure2-4(b)) and coincide with theweakerWIIsignals fromFigure2-4(c). More stronglyweathered soilpatterns (i.e. patterns dominatedby strong WII and Th signals) are consistent withtheelevated parts of theCarpentariaPlainand theGulf FallPUs(Figure2-4(b)). The strong U signalof the Assessment areacorrespond todeeper,light textured soils (sands) associated with the sandy sedimentsin the Gulf Fall aroundBuddycurrawaCreek, the sandysoils of DoomadgeePlain,the sandy soils of DonorsPlateau north of FourWays,andthe dolomite of theDissectedBarkly Tableland. Theextensivefresher alluvialplainsunits of theNicholsonand Leichhardtrivers(Figure1-2) reflect thefresher uplandparent materials(i.e. dominance of K andlow WII)from whichthese sediments were sourced. 22|Southern Gulf catchments soils and land suitability Figure 2-4 Selection of covariates used in digital soil mapping with underlying hillshade, including (a) slope %, (b) ternary gamma radiometrics, (c) weathering intensity index, (d) aridity index, (e) mean fractional vegetation cover and (f) bare earth – SWIR1 band Covariates used in modelling, map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-517_Covariates_3x2_v2_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au The scorpan climate factor of aridity index in Figure 2-4 (d) shows the monthly ratio of precipitation to potential evaporation, that is dryness of the climate at a given place (Stephen, 2005). Across the Assessment area there is a strong north-to-south trend of increasing aridity, ranging from humid tropical in the north to arid tropical in the southern part. This gradient reflects the episodic but significant influence of cyclones and continental heating/cooling, together imposing on the Assessment area landscape temperature and humidity trends affecting soil forming factors – especially vegetation trends (scorpan organisms factor). Figure 2-4 (e) (mean fractional vegetation cover) represents one of the scorpan organisms factors through seasonal vegetation dynamics. This shows occurrence of seasonally most-diminished canopy coverage to be most strongly associated with the deeply weathered surfaces of the Doomadgee Plain in the north-west of the Assessment area. Figure 2-4 (f) (barest earth) shows the long-term trends in soil bareness depicting consistent, inter- annual bare earth patterns. This shows the sparsest vegetation in the Assessment area to be in the extensive tidal flats (Karumba Plain), the grasslands (Armraynald and Barkly plains) and bare mainly rocky areas (Dissected Barkly Plain). The scorpan factors in Figure 2-4 are a spatial representation of the interrelationships between actions of soil formation and the resulting soil types and agricultural land attributes – thus agricultural and potential. 2.4.3 Soil attribute mapping: continuous, binary, categorical and Soil Generic Groups The R statistical programming environment (R Core Team, 2018) was used for DSM computing. All soil attributes and SGGs were modelled using a random forest (RF) modelling approach (Breiman, 2001) implemented in the ranger R package (Wright et al., 2019). RF models have a proven track record in environmental attribute prediction and have little tendency to overfit (Breiman, 1996). The approach constructs a multitude of decision trees during the algorithm training phase. Decision trees are ideally suited for the analysis of high-dimensional environmental data; a mix of continuous and categorical covariates that exhibit non-linear relationships, high-order interactions, and missing values can be used to predict continuous soil attributes (regression trees) or categorical ones (classification trees). Each individual decision tree divides a dataset into more and more homogeneous subsets. RFs are an ensemble learning method that employs ‘bagging’ (i.e. bootstrap aggregation), that is growing each tree from a random selection (with replacement) of samples in the training set, made with random selection of predictors (the covariates) in order to construct a collection of decision trees with controlled variance (Breiman, 1996). Bagging allows estimation of the error rate; some input data points are omitted each time a tree is built, and then these ‘out-of-bag’ (OOB) sample points are used to test and report the prediction accuracy of the realisation. Random selection of predictors during RF-building allows the relative importance of individual predictors to be assessed – in other words, if a predictor is left out, how poorly does the model perform? After many trees have been fit, training and test error tend to level off. This means that sub-setting the data into a training and test set is not necessary and all the data can be used to grow a RF model (Breiman, 2001). RFs output the class that is the mode of the predicted classes (classification) or mean prediction (regression) of the individual trees. The ranger package is the fastest and most memory-efficient implementation of RF algorithms available in R (Wright and Ziegler, 2015). The train function in the R caret package (Kuhn, 2015) was used to select the optimal mtry and splitrule arguments in the R ranger package. The mtry argument is the number of variables to possibly split at each node; the default is the rounded down square root of the number of variables, or six in this case (√39). Permeability, drainage, surface condition, surface texture, surface structure and SGGs are categorical attributes comprising multiple classes, whereas microrelief, rockiness and surface salinity are binary class attributes (i.e. present or absent). All were modelled using RF of 500 classification trees. Beyond the reported OOB prediction error (proportion of misclassified observations), the Kappa coefficient (Cohen, 1960) of the output confusion matrix was used to assess these RF model results; Kappa adjusts for chance agreement due to size of classes. While Kappa is not a test of mapping accuracy per se to inform users, in this case the test is used to test the performance of the model and the categorical allocations. Depth of A horizon in metres, clay percentage, surface exchangeable sodium percentage (ESP), soil erodibility (K-factor) using methods in Renard et al. (1991), soil thickness (effective rooting depth) in metres, surface pH, and available water capacity (AWC) in millimetres are continuous attributes, thus they were modelled using RF of 500 regression trees. Model reliability was evaluated two ways: (i) the OOB prediction error (mean square error, MSE), and (ii) R2 (computed on OOB data). In all RF models, relative importance of predictors (covariates) was assessed by permutation. During model development, all available point-based observations that were deemed to have a reliable source were used as model training data. For some soil attributes, extra point-based observation datasets were added to the master training data and used in model building. Further detail about how the extra data points were made is documented in Appendix B. The extra data points include the following: • Rockiness field observations – Observations of locations that fit the ‘rocky’ limitation criteria captured during fieldwork plus others captured while interpreting satellite imagery for rocky areas. • Lancewood sites – Locations dominated by Acacia shirleyi (lancewood) extracted from the vegetation database as it is found on particularly ‘rocky’ areas. • Bare rock data – Locations captured from satellite imagery of bare rock. • Shallow soil sites – Locations captured from vegetation data for species that indicate a soil depth less than 0.45 m. • Extra sites for the soil drainage map from vegetation sites where known species are indicators of wet soil conditions. In addition, landscape features such as swamps, salt scalds and ridge tops were identified with satellite imagery and geology mapping and attributed with either poor drainage or well drained where it was obvious. On the distinctive beach ridges of the Karumba Plain (both mainland and Wellesley Islands) it is expected that the sandy soils are rapidly drained. • Extra sites for the permeability map were based on expert knowledge and satellite imagery. On the distinctive beach ridges of the Karumba Plain (both mainland and Wellesley Islands) it is expected that the sandy soils have a high permeability. • Extra sites for surface condition and surface texture – On the distinctive beach ridges of the Karumba Plain (both mainland and Wellesley Islands) it is expected that the sandy soils have a loose or soft surface condition and a sandy soil texture. Also, on the Karumba Plain, the tidal areas that are bare of vegetation were expected to have a surface crust and a clayey soil texture. • Extra sites for the surface soil salinity map were based on vegetation mapping data for species tolerant of high salinity as well as obvious salt scalds on the Karumba Plain. • Microrelief ‘yes’ or ‘no’ sites sourced from satellite imagery data – On the clay plains of the Armraynald and Barkly plains, soil gilgai is sometimes obvious in satellite imagery. Imagery patterns were compared to field sites with recorded gilgai greater than 0.3 m deep. • Dataset of estimated AWC 60, AWC 100 and AWC 150 – Pedotransfer functions were run for all pre-existing point data where data to do so were available. • Marine plains data2 for AWC 60, AWC 100 and AWC 150 – Estimates of AWC were made for the Karumba Plain PU based on similar soil near Karumba that had soil observations. • Extra sites for the SGG map were based on vegetation sites, landscape understanding and satellite imagery. For the Wellesley Islands where we did not do any fieldwork, we estimated the SGGs using published information about the soils and cross-referencing of satellite imagery with Donors Plateau, which is expected to have similar soils. 2 ‘Marine plains data’ referred to in Table 3-3. AWC of soils was estimated for soil observation points using pedotransfer functions presented in Appendix A. In general, all RF models that produced acceptable model statistics (R2 >0.25, or Kappa >0.353) were applied to map the soil attributes and their uncertainty over the full extent of the mapping area of the Southern Gulf catchments, predicting the soil attributes at unsampled locations. This process was conducted using the CSIRO High-Performance Computing environment, given the large size of the dataset and the computational effort involved. In cases where the strongest comparable model statistics were achieved for more than one modelled iteration of an attribute, the selection of the model to use in the land suitability analysis was made by the project soil surveyors familiar with the Assessment area through a structured expert elicitation approach. 3 These thresholds draw on expert digital soil mapping and soil surveyor experience and were applied in this instance as an acceptance threshold for maps created from the various model permutations. 2.4.4 Digital soil map compilation and quantifying reliability The 500 individual trees of the RF models were used to generate 500 datasets of each soil attribute and then used to estimate model reliability for each attribute. For categorical values, the method for estimating reliability of predictions follows that described in Burrough et al. (1997) following the formula: CI = Pmax–1/Pmax (1) where: CI is the confusion index Pmax is the probability of the most probable soil class Pmax – 1 is the probability of the second-most probable soil class. A CI of zero is low confusion or in other words, very reliable. Conversely, a CI of 1 is high confusion or very low reliability. For continuous soil attributes the estimate of reliability of predicted values is the coefficient of variation (CV), that is, the standard deviation of the 500 predictions divided by the mean, expressed as a percent, at a particular grid location. A CV of 100% is high variability in the model estimates or low reliability, and a CV of zero % is no variability in the model estimates or high reliability. The model prediction for every grid cell covering the study area was mapped and the resulting soil attribute map evaluated as above in Section 2.4.3 through a structured expert elicitation approach. The DSM models that generated strong reliabilities (CI and CV) and cross-checked through the structured expert elicitation approach to be deemed the best mapped result for the attribute were applied in the land suitability modelling (Section 2.5). 2.5 Land suitability analysis Conventional land suitability analysis (land suitability) is a process of determining the potential of land to be used for specific land uses on the basis of the local range of environmental attributes and qualities (Rossiter, 1996), which are collectively termed land use requirements with associated limitations. The output is a 5-class suitability ranking system described in Section 2.5.1. This Assessment defines limitations and builds the analytical framework following the Food and Agriculture Organization of the United Nations (FAO) Framework for Land Evaluation (FAO, 1976; 1985). This involves a comprehensive assessment of land suitability that integrates multiple limitations including biophysical (edaphic and climate), social and economic themes (FAO, 2007). The land suitability analysis applied in this study deviates from FAO’s framework to constrain analysis to only biophysical themes. The edaphic components of the land suitability assessment mostly relate to soil attributes that have a key bearing on the growth and productivity of the irrigated crops, or the amount and cost of land preparation and maintenance of irrigation infrastructure needed that may affect the financial viability of the farming enterprise. For example, soil permeability determines the rate that water can be applied or held, and rockiness relates to the intensity of rock picking required in land preparation and the routine damage to farm machinery that might be expected. The land suitability candidate crops, and application of the framework, are discussed in further detail in the following sections. 2.5.1Crop suitability classes In the land suitability framework,the growth demandsfor each crop foreach attribute/limitationwas scoredaccording to a 1 to 5limitationclass.For example, on a shallow soila shallow-rooted small crop may be assigned a 1-score whereas a horticultural deep-rootedcrop maybe assigned a 5-score.Whenall thelimitations deemed to have a productionimpactare consideredsimultaneously in the analytical framework, a final suitability 1to 5 class rating is computedaccording to themostlimiting(i.e. highest scoring) limitation(s) astheunderlying assumption applied is thatthe most limiting factor determines the overallsuitability rating. In this simplescenario thehorticultural tree crop would not besuitable, simply on the basis of soildepth alone (i.e. theland would be ranked class 5). The derivation of limitation thresholds and their scoreswere either accessedfrom the literature(e.g. FAO, 1976; 1985), or defined by expertswhoarefamiliar withthe selected crops andtheAssessment area. The standard 5-class land suitability rankingused isbased on guidelinesdevelopedby theFAO(FAO, 1976; 1985)and presentedinTable2-4. The rankingapplies a suitabilitytermto each class: suitable(classes 1 to 3) → currentlyunsuitable(class 4) → unsuitable(class 5). The ranking also appliesa limitations termto eachclass:negligible(class 1) →minor (class 2) → moderate(class 3)→severe (class 4) →extreme (class 5). Class 4 (currentlyunsuitable) acknowledges that theremay be future management optionstooneday makethe land currentlydefined asunsuitable to become suitable. Such shiftstohighersuitability may reflect changes to current technology (e.g. new cropvarieties, pesticides, machinesand soil ameliorants) oreconomic (e.g. reducedfertiliser costs,new markets). Table2-4Land suitability classes based onthe Food andAgricultureOrganization of the United Nationsguidelines(FAO, 1976, 1985) CLASSSUITABILITYLIMITATIONSDESCRIPTION 1 SuitableNegligibleHighly productive land requiringonly simple management practices tomaintain economic production2 SuitableMinorLand with limitations that either constrain production or requiremorethan the simple management practices of class 1 land to maintain economic production3 SuitableModerateLand with limitations that either further constrainproduction or requiremore than those management practices of class 2 land to maintaineconomic production4 CurrentlyunsuitableSevereCurrently unsuitable land due to severe limitationsthat precludesuccessful sustained use of theland for the specified land use. In somecircumstances, the limitations may be surmountable with changes toknowledge, economics or technology5 UnsuitableExtremeThe limitations are so severe that the specified land use is precluded. The benefits would not justify the inputs required to maintain production and prevent landdegradation in the long term Eachdropinthesuitabilityrankingimpliesthat more management input (thus increasing cost of production) isrequired to achievethe same levelofcropproduction.The limitationterm is a proxyfor the level of management requiredtoovercome the current level of limitationor the reductionin crop yield/increase inmanagement coststouse the land with the current level of limitation. By 28|Southern Gulf catchments soils and land suitability convention,limitingfactors increase from class 2 throughtoclass 5indicating a higher level ofmanagement interventionis required to elevatethe classtothe nexthigher suitability class. Forexample,if rockiness is identified as the most limitingfactorin a given scenario, rockpickingovercomes therockinesslimitationtopotentiallyelevate the ranking. However, theranking willnot elevate if the new most limiting factorthat emerges, perhaps soil depth,hasthe same rankingthat rockinesshad originally. 2.5.2Candidate crops Individual land suitability data and mapswere prepared for an extensive set of crops by seasonby irrigation type4(generating >120 land use options) for the Flinders and Gilbert AgriculturalResource Assessment(Bartley et al., 2013)and theNorthern Australia Water ResourceAssessment(Thomas etal., 2018b). These land suitabilityframeworkswere developed with research partners and stakeholders and represent the state of agronomicknowledge and anticipated market needs at the time. TheNorthern TerritoryGovernment recently developedaNorthern BarklyRegion IrrigatedAgricultural Land SuitabilityFramework(Burgess et al., 2022)thatis based onthe Katherine region land suitability assessment (Katherine–Daly Waters)(Burgess etal., 2015;McGrath et al., 2019)andthe Roper River region agricultural land suitabilityframework(Andrews and Burgess, 2021). Theframework used in this Assessment aggregates likecrops andcropping systemsinto crop groups; these are listed 1 to 21 inTable2-5. The list is adapted from Andrews and Burgess (2021),while CSIRO has added new crops tothis list, many of which havebeenharmonised intogroups 1 to 16. Newcrops deemed prospectiveanddesirable but notfittinginto NT’s crop groupingshave been added toTable2-5in this Assessment–these arethe groups17 to 21. Table2-5Crop groups (1 to 21) and individual land uses evaluated for irrigation potentialas usedinthe VictoriaRiver WaterResource Assessment(Thomas et al., 2024) MAJORCROP GROUPCROPGROUPINDIVIDUAL CROPS ASSESSED Tree crops/horticulture(fruit) 1 Monsoonal tropical tree crops (0.5 mroot zone)–mango, coconut, dragon fruit, Kakadu plum, bamboo,lychee2 Tropicalcitrus–lime, lemon, mandarin, pomelo, lemonade, grapefruitIntensive horticulture(vegetables, row crops) 3 Cucurbits–watermelon, honeydew melon, rockmelon, pumpkin, cucumber, Asian melons, zucchini,squash4 Fruiting vegetable crops –Solanaceae (capsicum, chilli, eggplant, tomato), okra,snake bean, drumstick tree 5 Leafy vegetablesand herbs–kangkong, amaranth, Chinese cabbage, bok choy, pak choy, choy sum, basil, coriander, dill, mint, spearmint, chives, oregano, lemon grass,asparagus Root crops6 Carrot, onion,sweet potato, shallots, ginger, turmeric, galangal, yambean, taro,peanut,cassava Grain and fibre crops7 Cotton,grains–sorghum(grain), maize,millet(forage) 4Under Food and Agriculture Organization ofthe United Nations (FAO)terms (FAO, 1976) these are ‘land utilisation types’ (i.e. land usepermutations of crop by management). Chapter2 Methods|29 MAJORCROP GROUPCROPGROUPINDIVIDUAL CROPS ASSESSED 8 Rice(lowland and upland) Small-seeded crops9 Hemp,chia,quinoa,medicinalpoppyPulse crops (food10 Mungbean,soybean,chickpea,navybean,lentil, guarlegumes) Industrial11 Sugarcane Hay and forage (annual)12 Annualgrass hay/forages–sorghum (forage), maize (silage) 13 Legume hay/forages–blue pea, burgundybean, cowpea,lablab, Cavalcade, forage soybean Hay and forage14 Perennial grass hay/forage–Rhodes grass, panics(perennial) Silviculture/forestry15 Indian sandalwood(plantation) 16 African mahogany,Eucalyptusspp.,Acaciaspp. 17 Teak Intensive horticulture18 Sweetcorn(vegetables, row crops) Oilseeds19 Sunflower, sesameTree crops/horticulture20 Banana, coffee 21Cashew, macadamia,papaya Each crop group has specific management requirements with respect toplant growth, machineryuse and land degradation management and not all crop groups have been assessed for eachirrigation methodor season (e.g. cucurbitsarenot assessed for the wet season asthey are unlikelytobeplanted dueto high disease risk; African mahogany is not assessedfor furrow irrigation). Overall, wet-season crops are restricted to cropsthat can withstand seasonalwetness and/or can be managed (cultivated/harvested) effectively during thistime of year.Most of the crops can begrown during thedry season under a range of irrigation methods,with manyof the small cropsgrown only duringthis period. Also, mosthorticultural cropsare grown under micro irrigationtechniques (trickle/drip,micro sprays), whereas grain crops, cotton and sugarcaneuse spray orfurrow irrigation. A limitednumber of crops (sugarcane,cotton, some grains andforage)havebeen assessed for potential economic returns under wet-season rainfed conditions.The suitabilityof 17 crops under furrow or flood irrigationwastested, 23 under spray irrigation, 10 under trickleirrigationand 8 under rainfed conditions. In terms of seasons, the suitability of 22 crop groupswere tested under dry-season conditions, 20 under wet and 16 asperennialsystems of agriculture(seeAppendix C).The limitationstomanagement are reflected in the rules of the suitabilityframework and arepresented in Appendix D. 30|Southern Gulf catchments soils and land suitability 2.5.3 Limitations applied The 17 limitations and their sources used in the land suitability analysis are presented in Table 2-6. Limitations and thresholds were based on local and expert knowledge of crop yields and simulation data, and adaptations from other frameworks from similar climate and physiographic settings (Harms et al., 2015). Of these, 4 were from national climate data, 12 were derived from DSM land and soil attribute mapping (Section 2.4), and one (for acid sulfate soils, ASS) derived directly from the DEM. Limitation rule thresholds are presented in Appendix D. Some limitations are prepared from a combination of DSM land and soil attributes. For example, the erodibility limitation is determined by combining soil erodibility (k-factor, Renard et al., 1991) and slope. Similarly, the soil physical limitation accounts for a range of attributes, including soil surface texture, surface condition, soil structural class and sodicity (ESP). The following sections discuss the limitations in further detail. Areas susceptible to coastal ASS are mapped by a spatial analysis of the DEM to locate all land within 8 m of Australian Height Datum (AHD), hence those areas under marine tidal influence. The 5 mAHD threshold accommodates systematic errors noted (Gallant, pers. comm., 2018) in the elevation data (Table 2-3) in tidal river systems across northern Australia. Climate Annual rainfall The total amount of rainfall (precipitation) which falls during the growing season has a significant impact on the suitability for rainfed cropping (i.e. grown without supplementary irrigation). Given the expanse of geographic area assessed, and the variability of annual rainfall and soil conditions across the area, a total of eight rainfall categories were identified, ranging from less than 500 to greater than 1500 mm. For most of the crops assessed, at least 500 mm is required in combination with suitable soil attributes. Heat stress Parts of northern Australia are known for excessive heat over long periods, particularly during the transition periods between the dry and wet seasons. Intensely hot periods, defined as days with the maximum temperature over 40 °C, particularly when combined with wind, may damage seedlings as well as the leaves and fruit of many horticultural crops. Dark soil colours, prominent in the north, can become extremely hot and exacerbate damage. Frost Low temperatures (<2 °C) can damage sensitive crops and reduce crop yields through damage to flowers and fruits. Generally, there are few frost-prone areas in northern Australia, but they are known in some inland areas, some higher elevated locations and may be localised along low-lying creeks and drainage lines. Dry-season and perennial crops are only likely to be affected. Temperature variation Northern Australia generally experiences warm daytime temperatures, but overnight minimums can drop regularly by 15 to 20 °C, particularly during the dry season in inland locations. While some crops (e.g. chickpeas and lychees) require cool temperatures for seed/fruit set, other crops do not prefer such conditions. Land and soil Water erosion Soil erosion by water, if not minimised, reduces the productive capacity of the land. Several factors influence the erodibility of the soil including the intensity of rainfall, the gradient and length of slopes, and management practices that reduce surface cover or disturb the soil surface. Different soil types also have an inherent susceptibility to erosion, quantified as a soil erodibility factor (K-factor), which is related to soil permeability, surface structure, particle size (clay, silt and sand content) and the organic carbon content (Rosewell and Loch, 2002). The inherent stability of soils, estimated by K-factor and slope, are used in this limitation. Wetness Excessive water in the soil profile due to rainfall and local runoff water can reduce crop growth and quality, restrict machinery and irrigation equipment use and may require expensive drainage reclamation works. The wetness limitation considers permeability class (rate of water movement into and through the soil profile) and drainage class (length of time the soil remains saturated). As wetness can be highly seasonal, drainage and permeability may be considered differently for summer (wet-season) and winter (dry-season) crops. Although a soil may show signs of wetness, a crop grown in the dry season will usually not experience adverse wetness conditions. Soil water availability The AWC within specified rooting depths (relevant to different crops) represents the volume of water in a soil profile between field capacity (upper limit) and wilting point (lower limit) and is estimated using soil texture (clay, silt and sand content), the percentage of coarse fragments in the soil (that reduce water storage space) and soil depth. For rainfed cropping, the soil AWC is generally considered to be the maximum amount of moisture stored to grow a crop. For irrigated cropping the AWC relates to the irrigation frequencies required to obtain optimum crop yields. Soil with reduced AWC can be ‘topped up’ by irrigation, as long as the soil is not too free draining, or infiltration rates too slow to allow water into the profile. In this study, suitability subclasses for irrigated land uses are based on the estimated effort and cost required to maintain sufficient moisture in the soil profile for optimum plant growth, which relates directly to the irrigation interval (i.e. days between required irrigations) during the period of maximum water demand. In addition to soil AWC, data used for this estimation are reference crop evapotranspiration (ETo) supplied by SILO (Jeffrey et al., 2001) along with crop-specific factors and equations supplied in the FAO irrigation and drainage paper (Allen et al., 1998). Nutrient balance (pH) In addition to the total amount of nutrients within the soil (which is generally low across northern Australian soils in their natural state), chemical processes within the soil can affect the availability of nutrients for plant uptake. Soil acidity or alkalinity may lead to certain nutrient deficiencies and/or toxicities. Soil pH, within the top 0.1 m of soil, has been used as an indicator of conditions that affect the availability of plant nutrients. Soils with low cation exchange capacity (CEC), hence diminished buffering capacity because of mineral type and/or low organic carbon content, are at risk of acidification through rate of base removal (e.g. produce export) exceeding rate of base supply (e.g. through fertiliser addition or natural weathering product supply). Sandy soils are the most susceptible to nutrient imbalance acidification. Soil thickness Adequate soil thickness is necessary to provide sufficient depth of soil to support plant root development and structural growth. Deeper soils have more water available for plant growth than shallower soils for the same AWC. Shallow soils cause issues with cultivation, seedling establishment and harvesting, particularly for root crops. Uprooting of tree crops by strong seasonal winds may be exacerbated by shallow soils that prevent adequate root penetration. In some high-value, intensive cropping systems (e.g. Asian vegetables), shallow gravelly soils may be modified by mounding to provide adequate depth, although this may be a significant management input and therefore reduce the suitability of such shallow soils compared to deeper ones. Rockiness Surface gravel, stone and rock outcrop can interfere significantly with planting, cultivation and harvesting machinery used for root crops, small crops, annual forage crops and sugarcane. Microrelief (gilgai) Surface microrelief is common in cracking clay soils where wetting and drying cause shrinking and swelling of the soils, resulting in uneven surface features. Microrelief can be substantial, with greater than 0.3 to 0.4 m of vertical displacement in some areas. Gilgai can affect the establishment of irrigation infrastructure and must often be levelled to allow efficient machinery operation and irrigation practices. Levelling may result in inconsistent surface soil characteristics, particularly where sodic and/or saline subsoils close to the surface are exposed. Soil physical conditions Several soil physical attributes have impacts on agricultural practices, crop establishment and growth and harvesting operations. Soil surface condition (firm, hard setting, crusting or with a coarse structure) affects seedbed preparation, seedling emergence or the development of root crops. Silty, hard-setting soils reduce infiltration of rainfall and irrigation water. Clayey soils are adhesive and sticky when wet and may be hard and difficult to manage when dry. Cracking types of clay soils can also shear tree roots and impact on infrastructure, for example soil swelling and shrinking can undermine farm infrastructure. Soils with thin surfaces over sodic and intractable subsoils are generally of low suitability for cropping as the soils are prone to hardsetting and the clay subsoil intractable and hostile to roots. Irrigation efficiency This relates to the capacity of the soil to facilitate the movement of water into and through the soil profile. For surface irrigation (furrow or border-check methods), surface soils are ideally slowly permeable to allow water to move effectively down furrow or across fields. High infiltration results in uneven rates of water being applied close to the source and minimum or no water being delivered to the ends of furrows. In addition, high rates of deep drainage can occur, resulting in water and nutrient loss below the root zone. Efficient irrigation of any sort requires a careful balance between water application rates, soil permeability, and crop growing stage requirements – the latter influenced by climatic conditions. Acid sulfate soils Acid sulfate soils (ASS) is a broad term given to a range of soils containing sulfurous materials (Sullivan et al., 2010). These soils are either strongly acidic (actual ASS; pH <4) or have the potential to become strongly acidic (potential ASS; pH >4) if exposed to atmospheric oxygen, for example when they are exposed or drained (Fanning and Fanning, 1989). If disturbed or improperly managed and acidification occurs (potential ASS → actual ASS), water can leach the sulfuric acid and dissolved heavy metal contaminants from the oxidising sulfate layers posing serious risks to water quality, public health and the health of aquatic environments (Fältmarsch, 2006; Ljung et al., 2009). The ASS soils in the Assessment area are restricted to the coastal fringes where aquaculture is likely to present the only land use potential. In those locations, land development will attract jurisdictional assessment and legislative guidelines from the Northern Territory and Queensland governments. A simple methodology, consistent with the NT’s regulatory guidelines (Dear et al., 2014), was developed to map potential ASS in the Assessment area to guide land use decisions. These soils are restricted to coastal areas under marine influence, hence they are found within 5 mAHD. Table2-6Land suitability limitations and source data LIMITATIONDESCRIPTIONINPUT DATASOURCE Climate–rainAnnual rainfall. Used for rainfed croppingMeanannual rainfall (years1889–2017) scenarios onlyClimate–heat Excessive heatdamages cropsMeannumber of days>35 °C (years1889–2017)http://www.bom.gov.au/research/publications/researchreports/BRR- stress041.pdf http://www.bom.gov.au/research/publications/researchreports/BRR- 041.pdf Climate–frostImpact on crops due to frostMeannumber of days with minimumtemperatures<2°C (years1889–2017) Climate–tempCool seasonal temperatures are requiredMean minimum monthly temperature <15°C (yearshttp://www.bom.gov.au/research/publications/researchreports/BRR- variationfor some crops1889–2017)041.pdf http://www.bom.gov.au/research/publications/researchreports/BRR- 041.pdf Water erosionSoil loss due to water erosionneeds to beK-factor (soil erodibility factor), % slopeDigital soil mapping (DSM)from field observations, laboratoryminimised measurements and calculated data; CSIRO Shuttle Radar TopographyMission (SRTM) WetnessSite and soil conditions that result in poorSite drainage andsoil profilepermeabilityDSM from field observationssoil aeration and impact oncrop growth Soil waterCapacity of a soil to supply waterforAWCestimated using methodsinAppendix ADSM fromestimated AWCavailabilityplant growth;estimated forthesoil(available waterprofile. A critical parameter for rainfedcapacity,AWC)cropping and applied irrigation water efficiency for irrigated land uses Nutrient balanceImpact of soil pH on plant abilityto utiliseSoil pH in top 10 cm of soilDSM from field estimates andlaboratory analysissoil nutrients Soil depthAdequate soil depth for physicalsupportSoil depth(to 1.5 m)DSM from field observationsand plant edaphic requirements RockinessRockiness of soil, includinghard rock and Rock outcrop, surface gravels and coarse fragmentsDSM from field estimatessignificant gravelcontentimpacts on crop growth and farming practices Gilgai (microrelief)Indicates the extent of land levellingVertical interval of microreliefDSM from field estimatesrequired;levelland is required for even drainage and efficient machineryuse Soil physical Physicalsoilconditions that affect Thickness of Ahorizon;surfaceexchangeable DSM from fieldestimations andlaboratory analysis restrictionsworkability, seedling emergence,sodium percentage; SoilGenericGroup; soil surfaceharvesting (especially for root crops) and condition; soil surface texture; soilsurfacestructurewater Chapter2 Methods|35 LIMITATIONDESCRIPTIONINPUT DATASOURCE Irrigation efficiency(furrow andborder-checkMinimise deep drainageSoil infiltration rate implied from whole soil profilepermeabilityDSM from field estimatessurface irrigation) Irrigation efficiency(spray and trickleirrigation) Ease of soil profile recharge (wetting upof soil profile) Soil infiltration rate implied from whole soil profilepermeabilityDSM from field estimatesClay content (aquaculture) Ring tank suitability% ClayDSM fromlaboratory measurementsSalinity (soilsurface) Plant stress due to high levels ofsalt inthe soil profile, salt toxicityPresence/absence of excessive soil surface salinityDSM from field observationsAcidsulfatesoil potentialPotential for soilsulfidesto oxidise tosulfates(forming sulfuric acid) from sitedisturbance and soildryingElevation above mean sealevel,<5mAHDTopographic maps, CSIRO SRTM and land system mapping where available 36|Southern Gulf catchments soils and land suitability 2.5.4 Limitations not applied As with the other assessments (Bartley et al., 2013; Thomas et al., 2018b; Thomas et al., 2022; Thomas et al., 2024) several limitations that may have bearing on enterprise-level land suitability were not assessed as part of this activity. For example, soil temperature may have a limiting effect on crop germination and performance (Abrecht and Bristow, 1996) and was not included. Other limitations that may feature in some land suitability frameworks, although not in scope in the land suitability in this Assessment, include economics and finances (e.g. subsidies and grants, produce market prices, fertilisers and fuel costs, etc.), proximity to produce processing facilities (e.g. cotton gins, abattoirs), flooding risk, land management-induced secondary salinity, conservation area exclusions and proximity to irrigable water. Caution should be employed when using the land suitability outputs from this activity for planning purposes without consideration of these limitations. Policy and land tenure considerations are not imposed in the land suitability modelling in recognition that these socio-economic and political attributes of the landscape may shift as economic and technologies change, and community sentiments and aspirations shift along with community values. Some of these factors are reported with other activities in the Assessment. 2.5.5 Computing land suitability and quantifying reliability The land suitability modelling in this study applied a set of rules (Appendix D) to the DSM and other attribute layer. The land suitability assessment analysis follows the process as defined by the FAO (FAO, 1976; FAO, 1985). Using standard practice, suitability is calculated and mapped spatially by assessing one set of limitation subclass values per pixel to determine the most limiting subclass, which then becomes the overall suitability value for a given pixel. The processing of translating the limitation layers into crop suitability was done in two stages. The first converted the attribute (e.g. pH) into an attribute code (e.g. Nr1 = pH 5.5–7.0, Nr2 = pH 7.0– 8.5, Nr3 = pH <5.5, Nr4 = pH >8.5). The second then applied the crop-specific suitability subclass values to the layers produced in first phase. For example, for rice grown with flood irrigation, raster cells containing values Nr2 become suitability subclass 1, those containing Nr1 or Nr3 become subclass 2, and those containing Nr4 become subclass 3. The different limitation subclasses (e.g. for pH, soil depth and water erosion) are then assessed to determine the most limiting factor and produce a single suitability class map for each crop group by season by irrigation type combination. Given the use of DSM attributes to generate suitability attribute/limitation data, estimates of uncertainty are made possible through a method described by Malone et al. (2015) to propagate uncertainty of the soil attribute values through to the suitability assessment process to give an indication of the overall certainty of land suitability predictions. Each of the DSM attribute data were generated using a RF model comprising 500 trees (Section 2.4.4). Thus, for each pixel on the map there are 500 individual realisations of a given attribute value. On a pixel basis, the calculation of the overall suitability is similar to that of the standard approach described above, except the calculation has been done 500 times per pixel using the individual DSM realisation values. The overall subclass limitation value is the modal subclass value from the assessment of the 500 individual realisations and the final suitability class for the pixel is inherited from the most limiting attribute, as above. An uncertainty index (UI) can be calculated from the distribution of 500 individual subclass values. The UI, for a given pixel, is the degree of confusion between the most probable class and the class immediately less probable in the probability series and is like the concept of the confusion index used by Burrough et al. (1997). The UI was calculated as: UI = Pmax–1/Pmax (2) where: Pmax is the probability of the most probable class Pmax−1 is the probability of the second-most probable class. When UI tends to zero then one class dominates and there is little confusion in the model and when UI tends to 1 then there is less certainty of the modelled suitability value. The modal values for each of the relevant limitations for each pixel for each land use is then used to determine the most limiting subclass, thus determining the overall suitability for each of the land uses. The UI assigned to each pixel is that of the corresponding most limiting subclass value. Where two or more subclasses are the most limiting (e.g. a subclass of 4 for the erosion limitation and a subclass of 4 for the wetness limitation), the cause of uncertainty is assigned to the pixel with the largest UI of the same (and worst) subclasses. The calculation of the suitability and associated UI maps was undertaken using purpose-written R scripts (R Core Team, 2014). Due to the magnitude of calculations required to assess the uncertainties, the calculations were implemented in a high-performance computing environment. The land suitability framework implemented 58 unique rule sets for crop group by season by irrigation type (i.e. 17 furrow/flood, 23 spray and 10 trickle, and 8 rainfed, see Appendix C). The 58 unique grouped options reported here were derived by aggregating individual crops from the 126 unique land use options from the Northern Australia Water Resource Assessment (Thomas et al., 2018b). The land suitability framework is adapted from the Department of Natural Resources and Mines and the Department of Science, Information Technology, Innovation and the Arts (2013). 2.5.6 Landscape complexity Successful cropping means that management and practices are compatible with the physical constraints of the land parcel size, and there is a minimum size of contiguous area of suitable land necessary to achieve production efficiencies at a scale required to be viable. For example, centre pivots require certain dimensions of land to be available for efficiencies. Land parcel size can be impacted by the juxtaposition of suitable and non-suitable soils, or physical limits to the size, extent or shape of individual parcels (e.g. dissected by anabranching (Taylor, 2002) and incised stream channels). The effect is that, at a broad scale as reported here, the penalty of operational inefficiencies of farming the land outweighs the otherwise positive attributes of the soil. In this analysis two components of landscape complexity are considered using methods from Thomas et al. (2022). • The contiguous suitable area component was applied to the Assessment area based on crop- specific minimum areas and length/width of contiguous land. Contiguous suitable areas were produced as standalone data products for all crop groups (Table 2-7 and Table 2-8). Readers will note that, although the analysis was completed, the land suitability maps presented in Section 3.5.1 show the original extent of land suitability data to ensure consistency of approach across the northern Australia assessments to date. • The stream dissection component reflects elaborate patterns of incised (>1 m depth) anabranched channels in alluvial floodplains detected using Light Detection and Ranging (LiDAR). Contiguous suitable areas The 5-class land suitability mapping data produce inherently speckled output with occurrences of minor suitable areas (few and single pixels and odd shapes) potentially making it difficult for users to interpret and apply on ground. To address the component of the landscape complexity limitation that relates to spatial extent, a spatial filtering method was implemented on the land suitability data to filter out parcels of land unlikely to be operationally viable. The result is data layers where each pixel was deemed to satisfy or fail the crop group rule shown in Table 2-7. Table 2-7 Rules to satisfy () and or not satisfy () for minimum contiguous area and width for each crop group (Table 2-5) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au A two-step process was developed and applied across the catchments as a planning aid tactic. First, the five FAO suitability classes presented in Table 2-4 were aggregated to two: ‘suitable’ for suitability classes 1, 2 and 3, or ‘not suitable’ for classes 4 and 5. Second, to further simplify the data, and to reflect the on-ground spatial constraints of farming practices, isolated one or two pixels of ‘not suitable’ contained in larger ‘suitable’ areas were reclassified as ‘suitable’. For each crop group, a minimum area and width were defined based on knowledge of farming practices. Depending on the possible land use, minimum areas were deemed as 2.5, 5, 10 or 25 ha and minimum widths of 80 or 120 m, as presented in Table 2-7. The minimum width was imposed by removing parts of the suitable area that are narrower (in any direction) than the required minimum width. The remaining groups of connected cells were then tested to see if they meet the required minimum area and removed if they did not. Table 2-8 List of crop groups (Table 2-5) for each minimum contiguous area rule from Table 2-7 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Floodplain stream dissection Figure 2-5 shows examples of anabranched (i.e. dissected) sections of a floodplain on the (a) lower Gregory River and (b) Beames Brook. Anabranching intensities effectively reduce potential paddock sizes comprising suitable land and puts management restrictions on the movement of agricultural plant and equipment, limiting the potential for agricultural development. A method was adopted to spatially identify these areas to provide a ‘flag’ on the suitability data outputs. These dissected areas remain classified in the standard class 1 to 5 land suitability system (Table 2-4) because landscape complexity is not included in the standard land suitability rule set. The stream dissection data applies to all crop groups. Figure 2-5 Examples of heavily dissected sections of floodplains on (a) the lower Gregory River and (b) Beames Brook Stream dissection images \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-507_StreamDissect_v1_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au For the purpose of demonstration, the application of this stream dissection component of landscape complexity followed these steps within the Assessment sub-area presented in Figure 2-5(a): • Using LiDAR 1.0 m ground resolution DEM, areas of channel depth greater than 1 m and closer than 100 m to the next greater than 1 m depth channel were identified. The greater than 1 m depth criterion was derived through consultation with producers who reported this depth meant the difference between viable and non-viable irrigation due to the cost of laser levelling required in land preparation. • Focal Statistics (focalmean) over the LiDAR DEM with a 50 m radius circle was applied. • ‘Raster calculator’ in Quantum GIS was used to extract channels by applying a threshold to the difference between the focalmean analysis and the original LiDAR DEM. A difference threshold of 0.9 m was used to identify channels at least 1 m deep because the focal mean represented a slightly lower bank top elevation. • Using the delineated channels raster, a Euclidean distance grid extracting the areas less than 50 m distant from cells delineated as ‘channel’ was derived. • Polygon data were created. Manual editing and some filtering was used to remove unwanted areas either as small, isolated units and eliminated 1 m deep sumps in the landscape occurring in otherwise channel-free areas. 2.5.7 Versatile agricultural lands Versatile agricultural lands were determined using the methods from earlier assessments (Thomas et al., 2018b; Thomas et al., 2022; Thomas et al., 2024) and following approaches in Kidd et al. (2014). Mapped outputs show cumulative scores of suitable classes (i.e. classes 1 to 3) at the geographic intersection of crop suitability maps. This analysis summarises the suitability of the selected 14 exemplar land management options (see Section 2.5.2 and Figure 3-48) chosen for each pixel and highlights where land is potentially more versatile for agricultural development because the pixels suit a larger range of land uses. Analysis results are displayed as an index ranging between 0 and 1, with the value 0 representing the least versatile land, and the value 1 representing the most versatile. In addition to the selected set of land uses, an index of versatile agricultural land was also calculated for each of irrigation type, including rainfed. As such, an index was calculated for furrow (17 instances), spray (22 instances) and trickle (10 instances) irrigation, and rainfed (8 instances) (Appendix C). 2.5.8 Aquaculture land suitability The suitability of soil and land characteristics for aquaculture development was also assessed using rules from Irvin et al. (2018) with adaptations made if necessary for Assessment area conditions and using the available DSM attribute dataset. The limitations considered included clay content, surface soil pH, soil thickness and rockiness; these mainly relate to geotechnical considerations (e.g. construction and stability of impoundments). Other limitations, including slope, and the likely presence of gilgai microrelief and ASS, infer more difficult, expensive and therefore less suitable development environments, and a greater degree of land preparation effort. Suitability was assessed for lined and earthen impounded ponds. With earthen ponds requiring soil properties that prevent pond leakage, soil permeability is an input to the Assessment. Soil acidity (pH) was also considered for earthen ponds as some aquaculture species can be affected by unfavourable pH values exchanging elements into the water column (i.e. biological limitation). In consultation with aquacultural expertise of the agriculture and socio-economics activity, representative and realistic aquaculture species were selected to characterise environmental needs of marine (prawns) and freshwater (red claw crayfish) species. Additionally, barramundi and other euryhaline species, which can tolerate a range of salinity conditions, may be suited to either marine or fresh water, depending on management choices. Except for marine species aquaculture, which for practical purposes is restricted by proximity to sea water, no consideration was given in the analysis to proximity to suitable water for fresh and euryhaline species aquaculture. The aquaculture suitability rules, including the limitation classes and suitability subclasses for each species by pond configuration, are presented in Appendix E. 3 Results 3.1 Survey data Soil records were collated from 712 sites comprising pre-existing and new soil survey data. Table 3-1 summarises these data (see also Figure 2-1 for the geographic distribution of the sites). In terms of the pre-existing data, 451 records were extracted from the NT SALInfo system (Department of Environment Parks and Water Security, 2000), 252 records from the Queensland Soil and Land Information platform (Biggs et al., 2000) and 9 records from the CSIRO NatSoil database (Karssies et al., 2011). The earliest site data is from 1951 and was sourced from the CSIRO NatSoil database. NT sites were sampled between 1986 and 2010 and Queensland data sites between 1998 and 2022. No sites exist in the public domain for the Wellesley Islands. Table 3-1 Summary of soil data sites collated for the digital soil mapping component of the Assessment including new and pre-existing data within the Assessment area (Southern Gulf catchments) and pre-existing data outside the Assessment area but within the modelling extent For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au This activity sampled 97 new sites collected during the 2022 field season (Table 3-2). The planned field program and data collection was for 200 primary digital soil mapping (DSM) sites and subsidiary sites to be collected if time permitted. However, communication and engagement with the Carpentaria Land Council and Corporation (CLCAC) didn’t progress after initial verbal consent for fieldwork and a decision was made to not access the lands in the northern part of the Assessment area, which coincides with the CLCAC area, resulting in only 92 of these 200 primary DSM sites being captured. Table 3-2 describes the type and numbers of new site data collected by the activity. The region of the catchments that was impacted by a disrupted data collection campaign has a dearth of pre-existing site data (Figure 2-1), with only 27 pre-existing sites. Table 3-2 New site data collected during the 2022 field seasons by site type (see Figure 2-1) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Fifty-three percent of the new sites were within 30 m (i.e. within 1 pixel) of their predetermined location as determined by the conditioned Latin hypercube sampling (cLHS) strategy. Another 18% of new sites were between 30 and 90 m of their predetermined location. Steeper hilly areas were more challenging to access and fenced road reserves restricted access to sites. New and pre-existing sites were used in the DSM modelling – although not all pre-existing site records were used for all DSM attribute predictions as some records may have missed one or more of the soil attributes needed or were excluded because of criteria detailed in Section 2.1. Distribution of sites used in the DSM (Figure 2-1) shows variable density between the NT and Queensland and between catchments and physiographic regions, with valuable numbers falling outside the Assessment area. Many of the sites outside the Assessment area were considered useful in the DSM process because they were likely to support modelling of soils inside the boundaries of the Assessment area given the similarities in formation factors and histories – especially on the Barkly Tableland. The soil data was used in several ways to provide the training values of the attributes to be modelled. Training data were extracted from the soil database using structure query language (SQL) queries developed from explicit extraction rules. There were three methods to derive an attribute value: • Actual value – A direct measured value is extracted for the attribute (e.g. permeability, drainage class, pH). • Synthesised value – The final attribute value is a result of interrogating more than one measured attribute (e.g. soil depth derived from depth to R horizon, depth to C horizon, Australian Soil Classification (ASC) family for soil depth). • Calculated value – The final attribute value is a result of a published calculation (i.e. pedotransfer function) that includes values of attributes (e.g. available water capacity (AWC) calculation including values for percentages of clay, fine sand, coarse sand and silt). 3.2 Digital soil attribute mapping This section presents evaluations on the quality of the DSM attribute data. Model evaluation is based on internal model validation, together with external validation through assessing the map products based on expert knowledge and understanding of the soil distribution in the study area as outlined in Section 2.2.3. The results of DSM attribute qualities are presented below. The distribution and source of soil data used to create the DSM attributes is presented in Appendix F. 3.2.1 Model evaluation Overall, 83 models were generated for the activity, and from these 18 digital soil attribute datasets were produced for the Southern Gulf catchments Assessment area. For all soil attributes, models were generated based on a combination of different soil observation datasets (see Section 2.4.3) together with model performance testing (e.g. weighting of soil attribute ranges or classes not well predicted in the model, removing of covariate layers that negatively contributed to the model predictions, or removing or adding an additional covariate to the covariate stack to improve the model performance). No model for any soil attribute stood out based on statistical measures alone. Final decisions for models to use followed a collaborative and iterative process involving assessment of outputs involving the field survey team, digital soil mappers and other experts with knowledge of soils and landscapes of the Assessment area. Creation of some models involved an iterative optimisation process that included testing expertly selected combinations of covariates (guided by soil forming principles, Fitzpatrick, 1980; Hudson, 1992; Jenny, 1941) and data points before the final model for each soil attribute was chosen bearing the best quantitative and qualitative test outcome. This was the case with drainage, permeability and SGG mapping. The consistency of soil attribute maps that are related (e.g. depth of A horizon, soil thickness and rockiness) was also considered in the final model selections. Table 3-3 shows the internal validation statistics from the random forest (RF) continuous soil variables models that were selected to produce the final soil attribute maps. Table 3-4 shows the internal validation statistics for the categorical soil variables. For continuous soil attributes the out-of-bag (OOB) prediction error is a value in the same units as the attribute, and for categorical attributes it is the proportion of misclassified data points. The final soil attribute maps then go forward into the land suitability analysis in Section 2.5. Table 3-3 Random forest model performance and details for the continuous soil attribute map products For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Table 3-4 Random forest model performances and details for the categorical soil attribute map products For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 3.3 Landscapes and Soil Generic Groups An essential part of the Assessment is to understand the land and its resources in a systematic way. Combinations of climate, land and soil attributes are systematically associated with particular soil types (Fitzpatrick, 1971; 1980; Hudson, 1992), and knowledge of soil type in a particular landscape allows agricultural potential to be inferred or expertly applied to check soil and land suitability potential maps. For the broad understanding of Assessment area landscapes, the 10 physiographic units (PUs) shown in Figure 1-2 from Section 1.1 provide a concept of each unit’s geological origins and development history, and provides a broad overview of the soils and inferred land potentials. The Assessment area soils have been mapped according to the SGGs introduced in Section 2.2.4 and shown in Figure 3-1. Table 3-5 presents SGGs, including descriptions, land features in which they tend to occupy, major agricultural management considerations, and correlations to the ASC (Isbell and National Committee on Soil and Terrain, 2021). 3.3.1 Landscape descriptions The mainland Assessment areas can be split into the uplands and the Carpentaria Plains (Grimes, 1974). The upland area in the south and west reaches 620 m above sea level and is the headwaters for the Assessment area catchments. The uplands can be divided into four PUs shown in Figure 1-2. The oldest, most elevated and rugged unit is the Isa Highlands (Twidale, 1956). It consists of Precambrian (>545 Ma) volcanic and sedimentary rocks that have been metamorphosed, weathered and eroded. Soil parent materials within the Isa Highlands from west to east include rhyolite, basalt, dolomitic sediments, siltstone, meta basalt, granite, quartzite and metasediments. Land surface relief is moderate (200–230 m) and generally has a south to north alignment. The next most elevated upland PU is a small part of Barkly Tableland to the west of the Isa Highland PU. The tableland started out as a sedimentary basin in Precambrian times, which was uplifted, folded and eroded. During the Cambrian period, seas transgressed the area and deposited carbonate sediments in the depressions. Later, the Cambrian-period sediments were exposed and eroded. During the Mesozoic Era, isolated lakes and swamps developed (Randal, 1966), subsequently during the Tertiary period the upland areas experienced deep weathering and laterisation. However, areas covered by lakes and swamps did not undergo strong leaching and, as the landscape dried, the current cracking clay soils formed on relatively fresh sediments (Christian et al., 1954). The clay soils overlie dolomitic rocks. Relief is very low (9–30 m) and Mitchell grasslands dominate. Table 3-5 Soil Generic Group description, management considerations and correlations to the Australian Soil Classification For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Since the Cambrian period, the drainage network that flows toward the Gulf of Carpentaria has dissected the tableland leaving remnant upland features defined by deep narrow gorges. This area is mapped as the Dissected Barkly Tableland PU in Figure 1-2. Dissection has been amplified because the underlying rocks formed from dolomitic sediments are relatively soluble compared to surrounding rocks. These gorges have intersected the groundwater systems of the tableland, resulting in spring-fed permanent creeks and rivers such as the O’Shannassy, Gregory and Lawn Hill creeks. The remaining parts of the uplands, comprising mainly of Mesozoic-era sedimentary formations (sandstones), have been eroded into a complex pattern of easterly flowing streams and valleys separated by ranges and outcrops of sedimentary formations (Mullera Formation, Constance Sandstone and Fickling Beds; Smith and Roberts (1972)). The PU is known as the Gulf Fall (Figure 1-2), and the Nicholson River and South Nicholson Creek are the primary systems draining this area. Musselbrook, Lagoon, Settlement, Gold and Running creeks also drain this area. To the east of the uplands are the Carpentaria Plains that consists of a series of plains, pediments and remanent plateaux that can be divided into six PUs. The most elevated sedimentary plain (30 to 150 mASL) immediately east of the uplands is the Cloncurry Plain PU in Figure 1-2. It consists of gently sloping colluvial and fluvial sedimentary plains and pediments with isolated low hills of Precambrian-period rock. Streams are few and incised into the pediments with narrow alluvial plains (Grimes, 1974). The majority of the Cloncurry Plain PU extends from the middle reach of the Leichhardt River to Lawn Hill Creek. In the northern Assessment area, the Doomadgee Plain PU (Figure 1-2) lies below and adjacent to the Cloncurry Plain and is predominantly a sandy, gently undulating plain overlying a deeply weathered Cenozoic-era land surface. Low eucalypt and paperbark scrub cover the lands. Widely spaced creeks drain the plains currently in a radial north-westerly direction toward the coast. This suggests that the underlying old land surface could have been a large sedimentary fan. Prior streams of sandier soils, shallow swampy and water-filled depressions (particularly between Lilly and Moonlight creeks) and small pits due to ferricrete subsidence occur throughout the plains (Grimes, 1974). In the southern half of the Assessment area, the Armraynald Plain PU (Figure 1-2) lies below and adjacent to the Cloncurry Plain PU and consists of argillaceous Cenozoic-era (Quaternary period) sediments (Armraynald Beds) forming black soils covered in grasslands. Stream channels are few, widely spaced and deeply incised due to sea-level changes. The plains extend up the Lawn Hill Creek, Gregory River and Leichhardt River water courses in the lowlands. Lawn Hill Creek and Gregory River are spring-fed permanent running streams. The Gregory River splits into a wide braided system (20 km at its widest) of permanent streams comprising the Gregory River, Beames Brook, Barkly River and Running Creek downstream of the Gregory Crossing. Monsoonal rainforest grows immediately adjacent to these permanent streams that cross the otherwise grassland plains (Grimes, 1974). Downslope of both the Doomadgee Plain and Armraynald Plain PUs lies the coastal Karumba Plain PU (Figure 1-2). This coastal unit extends 10 to 35 km inland from the Gulf of Carpentaria coast, and the plain is widest near the Albert River mouth. This plain consists of Holocene Epoch beach ridges and tidal and extra-tidal flats and plains. Some of the inland plains only flood when the rivers are in spate or when the north-westerly winds cause exceptionally high tides during the monsoon. Because the plain is wide, flat and with a moderate tidal range of about 3.5 m, tidal waters can rapidly inundate the land. Mangroves and tidal flats dominate the coastline, beaches are few and consist of white shelly sand. Small crescent dunes have formed in places from wind action (Grimes, 1974). Strong north-easternly winds across the bare plains – especially in November – may cause a fog-like effect in Burketown from suspended particles. Due to the flatness of the plain, streams meander in complex patterns. To the east of the Armraynald Plain lies Donors Plateau PU (Figure 1-2). This elevated unit (10– 80 mASL) forms a watershed between the Leichhardt and Flinders catchments and forms the eastern boundary of the Assessment area. The plateau consists of siliceous sediments laid down during the Early Cretaceous Epoch from upland sediment sources of the Normanton Formation. The plain, which was once more extensive, has been deeply weathered and lateralised in the higher residual parts during the Tertiary period, and has subsequently been stripped away in part leaving today’s Donors Plateau with minor areas of exposed older Cretaceous sediments. In parts the older sediments have been covered again by sediments laid down during the Pleistocene forming the Armraynald Plains (Ingram, 1972). Much of the Wellesley Islands in the Gulf of Carpentaria represent remnants of a mainland laterised Cretaceous period plain called the Mornington Plateau PU shown in Figure 1-2. This unit is 5 to 30 mASL. Dissection of this PU not as extensive as the Donors Plateau PU. The Mornington Plateau PU is generally fringed by marine plains consisting of coastal sediments or dune fields that support small mangroves in the lower parts of the landscape, and sea cliffs and wavecut platforms in the higher parts of the landscape. The cliff faces exhibit well-developed laterite profiles. Some of the coastal features in the unit are 5 m above current high-tide levels, which is explained by changes in sea level and upwarping of the islands (Grimes, 1974). 3.3.2 Soils and Soil Generic Groups The distribution of SGGs in the Assessment area are shown in Figure 3-1. Table 3-5 presents SGG descriptions, land features they occupy, major agricultural management considerations, and correlations to the ASC (Isbell and National Committee on Soil and Terrain, 2021). Totals of areas occupied by the SGGs are summarised in Table 3-6 showing that four SGGs dominate; referencing Figure 3-1 SGG mapping and Figure 1-2 PU mapping, shallow or rocky soils (SGG 7) occupy over half (55.8%) of the Assessment area and are associated with uplands and plateaux. In contrast, deep, cracking clay soils (SGG 9; 22.5%) dominate the Barkly Tableland and the Armraynald Plains. Wet soils (SGG 3; 5.9%) dominate the Karumba Plains and non-red sandy soils (SGG 6.2; 7.6%) dominate the Doomadgee Plain. Figure 3-1 The Soil Generic Groups of the Southern Gulf catchments produced by digital soil mapping. The inset map shows the data reliability, based on the confusion index as described in Section 2.4.4 Soil Generic Group map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-503_SGG_v2_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Table 3-6 The area that each Soil Generic Group area compositions comprises of the Assessment area. Areas without parentheses show mainland hectarages whereas areas in parentheses are for the Wellesley Islands For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 3.3.3 Summary of Soil Generic Groups of the Assessment area The soils with some of the most agricultural potential in the Assessment area are the cracking clays (SGG 9). These soils have formed from Pleistocene sediments on the Armraynald Plains between the Gregory and Leichhardt rivers, along Lawn Hill Creek and on the Barkly Tablelands. Both the Armraynald Plains and Barkly Tablelands are extensive natural grasslands with few trees, reflecting the cracking nature of the soils. The soils are imperfectly drained on the Armraynald Plains and moderately well drained on the Barkly Tableland due to the underlying limestone/dolomite karst. These clay soils are medium to heavy clays that crack when dry and swell when wet, reducing the rate of deep drainage. Soils have a self-mulching clay surface with gravel common on the Barkly Tableland. Effective rooting depth is deep to very deep (1.2–1.5 m) and the clay texture means the soils have a very high (>220 mm) soil AWC. On the Armraynald Plains soils are suited to a variety of vegetables (except root crops), rice, sugarcane and dry-season grain, forage, pulse crops, sweetcorn and cotton. On the Barkly Tablelands soils are suited to trickle-irrigated mangos and vegetables as well as wet-season cotton, grain and forage crops. In the middle reaches of the Leichhardt River a moderately well-drained, friable non-cracking clay or clay loam soil (SGG 2) has formed. The soils are moderately well to well drained, have a weakly structured, fine sandy to loam surface soil over a structured sandy clay loam or silty clay subsoil. Effective rooting depth is very deep (>1.5 m) and AWC is moderate (>160 mm). Also on the floodplains of the upper reaches of the Leichhardt River are well-drained, sandy loams over structured red clay subsoils (SGG 1.1). This soil also has a very deep (>1.5 m) effective rooting depth and a high (>180 mm) AWC. On the Doomadgee Plain, Donors Plateau, in the Gulf Fall PUs and on the elevated terraces north of the Nicholson River, sandy soils (SGG 6) have formed from sandy sediments overlying ironstone on the Doomadgee Plains and Donors Plateau PUs, and deep sand deposits on the highest terrace on the northern side of the Nicholson River. Near the Doomadgee township, soils are predominantly red sands (SGG 6.1) that are well drained but have a very low AWC (60–100 mm). Away from the river on the Doomadgee Plains, soils are brown sands (SGG 6.2) that are well drained and commonly limited by ferricrete rock within 0.6 to 1 m of the soil surface. The soil has a very low AWC (25–60 mm), depending on soil depth. On Donors Plateau, the yellow sandy soils have a high ironstone gravel content throughout the profile and are deeper. In the Gulf Fall the brown (SGG 6.2) and red (SGG 6.1) deep sands occur in the Buddycurrawa, Breakfast, Running and Sandy creek sub-catchments, and the upper parts of the Nicholson River catchment. Loamy soils (SGG 4) have formed along the Nicholson River and on the Doomadgee and Cloncurry plains, and other isolated areas. The soils along the lower Nicholson have formed from alluvial sediments on the elevated terraces south of the river. They are silty loams over silty clays. They are either red (SGG 4.1) well drained or brown (SGG 4.2) moderately well drained and have a high AWC (>170 mm). On the Doomadgee Plain PU along the Westmoreland Road/Savannah Highway there are brown, yellow and grey sandy clay loams over sandy clays (SGG 4.2) that are poorly to moderately well drained. Soils vary in depth depending on the depth of the underlying rock and have a low to moderate AWC (70–150 mm). The red loamy soils (SGG 4.1) on the Cloncurry Plain PU are shallower, sandier, and commonly have gravel and ironstone throughout the profile, and accordingly, these soils have a smaller AWC. Irrigation potential is limited to spray and trickle- irrigated crops on the moderately deep to deep (>1 m) soils. Seasonally or permanently wet soils (SGG 3) occur on local alluvia along creeks and in swamps, particularly between Lilly and Moonlight creeks on the Doomadgee Plains, and the tidal flats and wetlands of the Karumba Plains. The soils are very poorly drained. Shallow soils (<0.25 m depth) or rocky soils (>50% rock) (SGG 7) occur extensively in more than half of the Assessment area (Table 3-6), particularly in the Isa Highlands, Gulf Fall, Dissected Barkly Tablelands and Donors Plateau PU. Texture contrast soils that have sodic clay subsoils (SGG 8) are only found in two areas. The first is on Nineteen Mile Creek, a tributary of the Leichhardt River where the Pleistocene Armraynald Beds have been stripped away exposing the underlying Cretaceous calcareous shale. These brown sandy clay loams over sandy clays are highly vulnerable to erosion with many examples of extensive gully erosion found. The second area occurs on a meander plain of the Gregory – Nicholson rivers upstream of the junction. Clay soils with sodic subsoils have developed from this alluvium. The soils in the Assessment area have predominantly formed from their parent materials (under lying rocks or sediments) and landscape position and are closely associated with the PUs shown in Figure 1-2. The soils are discussed below within their PUs, starting with the most agriculturally versatile units. Soils of the Armraynald Plains physiographic unit Deep, slowly permeable, grey, brown and black cracking clays (SGG 9), frequently with small (<0.3 m) gilgai, have formed from alluvial sediments (Pleistocene Armraynald Beds) on the Armraynald Plains (Figure 1-2). Grey and brown cracking clays occur on the older relict sediments and black cracking clays on the recent floodplains. Two grey cracking clay variants are present on these plains. In the northern part a grey cracking clay occurs that is imperfectly drained, has a medium to medium heavy clay texture, gypsum crystals and manganese veins, is strongly sodic and saline at depth and has a deep (<1.2 m) effective rooting depth and high AWC (>210 mm). An example of this soil is shown in Figure 3-2 with its landscape setting in Figure 3-3. Figure 3-2 Soil profile of the Grey Vertosol (SGG 9) sampled on the northern part of the Armraynald Plains physiographic unit Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-3 Grey Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood on the northern part of the Armraynald Plains physiographic unit South of Gregory Crossing and Augustus Downs Station the second grey cracking clay variant is moderately well-drained, non-sodic, light medium to medium clay with calcareous nodules. The soil has a very deep (>1.5 m) effective rooting depth and a very high AWC (>250 mm). A soil profile to 1.5 m depth is shown in Figure 3-4, and the typical landscape setting for this soil is shown in Figure 3-5. Figure 3-4 Soil profile of the Grey Vertosol (SGG 9) sampled on the southern part of the Armraynald Plains physiographic unit Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-5 Grey Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood and Bauhinia on the southern part of the Armraynald Plains physiographic unit On the south-eastern part of the Armraynald Plains east of the Leichhardt River a brown cracking clay soil has developed that is imperfectly to moderately well-drained, sodic at depth, light medium to medium clay, with calcareous nodules. It has an effective rooting depth greater than 1.5 m with a very high AWC (>250 mm). An example of this soil is shown in Figure 3-6 with the landscape setting in Figure 3-7. Figure 3-6 Soil profile of the Brown Vertosol (SGG 9) sampled on Armraynald Plains physiographic unit, east of the Leichhardt River Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.5 m deep sample core was taken. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-7 Brown Vertosol (SGG 9) landscape of Mitchell Grass Downs with whitewood and gutta percha on the Armraynald Plains physiographic unit, east of the Leichhardt River On the more recent sediments deposited on the alluvial plains of the Gregory River, a black cracking clay has developed that is moderately well-drained, sodic (at depth) medium to medium heavy clay with calcareous nodules and an effective rooting depth greater than 1.5 m with a very high AWC (>255 mm). The soil is shown in Figure 3-8 and the landscape it occupies in Figure 3-9. Figure 3-8 Soil profile of Black Vertosol (SGG 9) sampled on the alluvial plains of the Gregory River west of the Armraynald Plains physiographic unit Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-9 Black Vertosol (SGG 9) landscape of Mitchell Grass Downs on the alluvial plains of the Gregory River west of the Armraynald Plains physiographic unit Along the Leichhardt River in the southern part of the Armraynald Plains, friable non-cracking clays or clay loams (SGG 2; Dermosol) and sand or loam over relatively friable red clay subsoils (SGG 1.1; Red Chromosol) have developed. The SGG 2 soils occur on recent sediments and the SGG 1.1 soils occur on older sediments. Both soil groups are deep and moderately permeable. The SGG 2 friable non-cracking clays or clay loams (Figure 3-10) are moderately well to well drained, have a weakly structured, fine sandy to loam surface soil over brown, red or grey structured, non-sodic, sandy clay loam or silty light to medium clay subsoil. Soil colour and drainage depends on proximity to stream channel. The soil has a very deep (>1.5 m) effective rooting depth and a moderate AWC (>160 mm). The landscape setting for these soils is shown in Figure 3-11. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-10 Soil profile of Brown Dermosol (SGG 2) sampled on Armraynald Plains physiographic unit, middle reach of the Leichhardt River Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.5 m deep sample core was taken. Figure 3-11 Brown Dermosol (SGG 2) landscape of buffel grass and open woodland with silver leaf box on the Armraynald Plains physiographic unit, middle reach of the Leichhardt River The Red Chromosols (SGG 1.1) are well drained with moderately thick (<0.2 m), sandy loam surface soil over red, structured, non-sodic, light medium to medium clay subsoils. The soil has a very deep (>1.5 m) effective rooting depth and a high AWC (>180 mm). Figure 3-12 shows an example of the soil and Figure 3-13 the landscape setting. The Dermosols and Red Chromosols are Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au suited to vegetables, sugarcane, oilseed, sweetcorn and dry-season grain, forage, pulse crops and cotton. Figure 3-12 Soil profile of Red Chromosol (SGG 1.1) sampled on the southern Armraynald Plains physiographic unit, west of the Leichhardt River Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.5 m deep sample core was taken. Figure 3-13 Red Chromosol (SGG 1.1) landscape of buffel grass / Mitchell grass in open woodland with silver leaf box and whitewood on the southern part of the Armraynald Plains physiographic unit Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Red sandy soils (SGG 6.1; Red Arenosols and Tenosols) and loamy soils (SGG 4; Red and Brown Kandosols) occur near Doomadgee on the northern side of the Nicholson River on a Pleistocene elevated plain of the alluvial plain. The red sandy soils are highly permeable, well drained and very deep (>1.5 m). The soil has a red or brown loamy sand surface soil over a red, massive, sand to sandy loam subsoil with very deep (>1.5 m) effective rooting depth and very low to low AWC (60– 100 mm), depending on soil texture. An example of the soil profile is shown in Figure 3-14, and its setting in Figure 3-15. Figure 3-14 Soil profile of Red Arenosol (SGG 6.1) sampled near Doomadgee on Armraynald Plains physiographic unit north of the Nicholson River Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Figure 3-15 Red Arenosol (SGG 6.1) landscape of open woodland of Darwin box, Bauhinia and Cooktown ironwood near Doomadgee on the Armraynald Plains physiographic unit north of the Nicholson River Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au The loamy soils (SGG 4) occur downstream of Doomadgee on the southern side of the Nicholson River on a Pleistocene elevated plain of the floodplain. Soils are red and well vegetated in the western half of the plain, and brown and subject to scalding and erosion in the eastern half of the plain. The red loamy soils (SGG 4.1; Red Kandosol) as shown in Figure 3-16 are highly permeable, well drained and very deep (>1.5 m). They have brown, silty loam moderately thick (<0.2 m) surface soils over red, massive, silty light to light medium clay subsoil. The soil has a very deep (>1.5 m) effective rooting depth and high AWC (>170 mm). The typical landscape setting for these soils is shown in Figure 3-17. Figure 3-16 Soil profile of Red Kandosol (SGG 4.1) sampled near Doomadgee on the Armraynald Plains physiographic unit south of the Nicholson River Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-17 Red Kandosol (SGG 4.1) landscape of open woodland with silky browntop grass and black speargrass and Bauhinia, rough leaf cabbage gum and ghost gum near Doomadgee on the Armraynald Plains physiographic unit south of the Nicholson River The brown loamy soils (SGG 4.2; Brown Kandosol) are highly permeable, imperfectly to moderately well drained, very deep (>1.5 m) and hardsetting. An example is shown in Figure 3-18. These soils have brown silty light clay moderately thick (<0.2 m) surface soils over brown, mottled at depth, massive, light medium clay subsoils. The soil has a very deep (>1.5 m) effective rooting depth and a high AWC (>190 mm). Figure 3-19 shows a typical landscape setting for these soils with scalded (bare) patches that are common and extensive in these soils and indicate a weak surface structure leaving the soils vulnerable to degradation through sheetwash erosion and hardsetting. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-18 Soil profile of Brown Kandosol (SGG 4.2) sampled on an the Armraynald Plains physiographic unit south of the Nicholson River Topsoil is in top lefthand corner, each profile segment from left to right represents a 0.3 m depth interval to 1.5 m depth. Figure 3-19 Brown Kandosol (SGG 4.2) landscape of disturbed landscape with bare areas (scalding) on the Armraynald Plains physiographic unit south of the Nicholson River Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soils of the Barkly Tableland physiographic unit Cracking clay soils (SGG 9; Vertosols) dominate this PU and have formed in lacustrine-derived sediments. The soils are self-mulching, grey or occasionally brown, cracking clays, moderately to very deep (>1.2 m). Stones, cobbles and outcropping of dolomitic origin may occur on the surface where soils are moderately deep. Chert gravels and pebbles are common as lag deposits on the surface and throughout the soil profile. The soil is slowly permeable, moderately well drained (brown variants are better drained), frequently with shallow (<0.1 m depth) gilgai depressions or sinkholes and have structured light medium to medium heavy clay textures with calcareous nodules and gypsum crystals. An example of a typical soil profile is shown in Figure 3-20, and its landscape setting in Figure 3-21. The soil has a deep to very deep (1.2–1.5 m) effective rooting depth with a very high AWC (>220 mm). Figure 3-20 Soil profile of Grey Vertosol (SGG 9) sampled on the Barkly Tablelands physiographic unit near Morestone Station Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.4 m deep sample core was taken. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-21 Grey Vertosol (SGG 9) landscape of grazed Mitchell Grass Downs showing the common dolomite stones on the Barkly Tableland. Many examples of these soils are less stony Soils of the Cloncurry Plains physiographic unit Loamy soils (SGG 4.1; Red Kandosols) and cracking clay soils (SGG 9; Black Vertosols) have formed on Pleistocene colluvial sediments of the Cloncurry Plain where soil type depends strongly on the landscape position and parent material. The loamy soils (SGG 4.1) are moderately permeable, moderately well drained and are shallow to very deep (0.3–1.5 m). Many of these soils have chert coarse gravels on the hard-setting surface and throughout the profile. Upper profiles are thick (<0.4 m) dark, brown or red, massive, sandy loam to sandy clay loams over red massive sandy clay loam to light clay subsoils, often with ferruginous nodules and gravels or pebbles. These soils can have a shallow to very deep (0.3–1.5 m) effective rooting depth depending on depth to rocks, or amount of gravel and coarse fragments to result in highly variable AWCs (e.g. very low to high, for example, 10–110 mm). Large areas of cracking clay soils (SGG 9) are very slowly permeable, imperfectly drained, very deep (>1.5 m) and are self-mulching, cracking with gilgai mounds and depressions. The soils have grey light clay thin (<0.05 m) topsoils over dark structured light medium clay with silcrete pebbles and calcareous nodules. The soils have a deep (<1.3 m) effective rooting depth, although this depth may be restricted by saline/strongly sodic subsoil conditions; if not restricted, AWC can be high (>200 mm). Soils of the Doomadgee Plain physiographic unit This is a sandy plain overlying an old lateritic surface and as a result sandy and shallow soils have formed. Depending on the landscape position and other forming factors, soils may be brown, yellow and grey sandy soils (SGG 6.2) and loamy soils (SGG 4.2), or seasonally or permanently wet soils (SGG 3). The brown, yellow and grey sandy soils (SGG 6.2; Arenosols or Tenosols) dominate the plains. The soils are highly permeable, well drained, deep to very deep (1–1.5 m), commonly encountering ferricrete rock within 1 m depth. The soil is brown, single grain sand or clayey sand and has a moderately deep to very deep (0.6–1.5 m) effective rooting depth depending on the depth of the underlying rock. The soil has a very low AWC (24–60 mm), depending on soil depth. The brown, yellow and grey loamy soils (SGG 4.2; Kandosols) occur in the central part of the plains furthest from the coast along the road to Borroloola Road (Westmoreland Road). The soils are moderately permeable and poorly to moderately well drained. They have moderately thick to thick (0.2–0.6 m) loamy sand to sandy clay loam surface soils over brown or yellow, mottled, massive, sandy clay loam to sandy light clay subsoils. The soil has a moderately deep to very deep (0.6–1.5 m) effective rooting depth depending on the depth of the underlying rock. AWCs are typically low to moderate (70–150 mm), depending on soil depth. The seasonally or permanently wet soils (SGG 3; Hydrosols and Aquic Vertosols) occur on local alluvia along creeks and in swamps, especially between Lilly and Moonlight creeks. The soils are slowly permeable, very poorly to poorly drained with mottled grey clay subsoils. Soils of Donors Plateau physiographic unit Donors Plateau PU comprises several soil parent materials. The most elevated parts of the plateau have been deeply weathered, and laterisation with common ironstone and ferricrete is common. Where the lateritic surface has been removed, relatively fresher labile sandstones and siltstones occur. In the central and southern parts of the plateau, the sedimentary rocks of the Normanton Formation have been removed, creating small valleys, and more recently colluvium and alluvium has been deposited onto the floor of these valleys. Depending on the parent material a diverse range of soils have formed. In the small valleys cracking clay soils (SGG 9; Grey, Brown or Red Vertosols) and seasonally or permanently wet soils (SGG 3; Hydrosols) have formed. The cracking clay soils (SGG 9) are slowly permeable, imperfectly to moderately well drained, moderately deep (>0.5 m), are self-mulching and have sandstone pebbles on the surface and in the soil profile. Soils are grey, brown or red cracking clays with a sodic, light to light medium clay texture at depth. The wet soils (SGG 3) are moderately permeable, poorly drained, deep (<1.2 m) and contain nodular concretions. These soils have thick (<0.4 m), grey, massive, fine sandy clay loam to clay loam fine sandy, hard-setting surfaces over subsoils that are structured, grey mottled, sandy light to medium clay subsoils. They have a moderately deep (>0.85 m) effective rooting depth due to subsoil salinity and wetness, and a low AWC (>50 mm). The deeply weathered and labile sandstones have been removed from a large part of the southern part of the Donors Plateau PU north-west of Fourways, and sandy colluvium and outwash deposits have been superimposed onto this landscape. The sand and gravel parent material has developed into sand or loam over relatively friable brown, yellow and grey clay subsoils (SGG 1.2; Brown and Yellow Chromosols) and sandy soils (SGG 6; Yellow Tenosols). The Brown and Yellow Chromosols are moderately permeable, imperfectly to moderately well drained, very deep (>1.5 m), have moderately thick to thick (0.2–0.4 m), dark or grey, massive, coarse sand to sandy loam surface soils. These are above a subsoil of brown or yellow, occasionally mottled, massive or weakly structured course sandy light to light medium clay with abundant ferruginous nodules and quartz gravels. The soil has a very deep (>1.5 m) effective rooting depth and a moderate AWC (80– 110 mm) due to the abundant course fragments throughout the profile, as shown in Figure 3-22 and landscape setting in Figure 3-23. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-22 Soil profile of Yellow Chromosol (SGG 1.2) sampled on the Donors Plateau physiographic unit Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.5 m deep sample core was taken. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-23 Yellow Chromosol (SGG 1.2) landscape of woodland with spinifex, silver leaf box and broadleaf paperbark on the Donors Plateau physiographic unit The sandy soils (SGG 6) are highly permeable, well drained, very deep (>1.5 m) have thin (<0.1 m) brown loose sandy surfaces over moderately thick (<0.4 m) grey or light brown loose sands containing abundant nodules. These sands may overlay yellow or red loose sand with abundant nodules and ferricrete fragments. The soil has a very deep (>1.5 m) effective rooting depth but a very low AWC (<25 mm). In the Donors Plateau small areas of friable non-cracking clay or clay loam soils (SGG 2; Brown and Grey Dermosols) have formed from the labile sandstones and siltstones. The soils are moderately permeable, imperfectly to moderately well drained, very deep (>1.5 m) and have hard-setting surfaces, and ironstone gravels are common in the soil profile. Soils have moderately thick (<0.4 m) brown, massive structure, light medium to medium clay surface soils. This layer overlays brown, finely structured, light medium clay subsoils. The soil has a very deep (>1.5 m) effective rooting depth with a high AWC (>180 mm). On the deeply weathered elevated parts of the plateau shallow soils overlie ferricrete and other hard material (SGG 7; Red Kandosol). These soils are moderately permeable, well drained, shallow (<0.5 m), and have a massive structure with fragments of ferricrete and hard-setting surfaces with abundant ironstone pebbles in the profile. The soils have moderately thick (<0.4 m) brown sandy loam to sandy clay loam surface soils over red, sandy clay loam to sandy light clay subsoils. These soils have a shallow (<0.5 m) effective rooting depth and very low AWC (<30 mm). Soils of the Mornington Plateau physiographic unit The Mornington Plateau PU occurs on the elevated parts of the Wellesley Islands. The plateau is an old land surface that has been deeply weathered and lateralised where ironstone and ferricrete is common. Friable non-cracking clay or clay loam soils (SGG 2; Brown, Yellow and Grey Dermosols) and shallow soils (SGG 7; shallow and/or rocky soils) occur. Soils of the Gulf Fall physiographic unit The Gulf Fall has many soil parent materials that have developed from a diverse range of sedimentary rocks that are in situ, or severely dissected and removed. The landscape has been deeply weathered and lateritised as well as eroded. Steep scarps, flat to gently undulating mesas, plateaux, range tops and broad valley flats occur. Local alluvium from relatively recent erosion occurs along rivers and creeks as well as sandplains in the upper parts of the elevated areas. Despite this diversity in soil parent materials, relatively few soil types occur with shallow and/or rocky soils (SGG 7) being the dominate group. Sandy soils (SGG 6) occur in the Buddycurrawa, Breakfast, Running and Sandy creek subcatchments and the upper parts of the Nicholson River catchment. Loamy soils (SGG 4) occur in the upper parts of the Nicholson River catchment. Isolated cracking clay soils (SGG 9) and seasonally wet or permanently wet soils (SGG 3) occur at isolated claypans in the Buddycurrawa Creek subcatchment, at the Caulfield Clay Flats and in the upper parts of the Nicholson River catchment. The shallow and/or rocky soils (SGG 7; Rudosols) are highly permeable, well drained, very shallow (<0.25 m) brown or red, single grain or massive, sand to sandy clay loam with common to many ferric nodules over laterite, sandstone or siltstone. Hard rock outcrop is common and ironstone abundant on the surface. The soil has a very shallow (<0.3 m) effective rooting depth and very low AWC (<20 mm). The sandy soils (SGG 6) (Brown or Red Arenosols, or Tenosols) are highly permeable, moderately well to well drained, deep to very deep (>1.2 m), brown or red, single grain or massive, sand to sandy loam soils occasionally mottled at depth and with no course fragments or nodules. The soil has a very deep (>1.2 m) effective rooting depth but a low AWC (<60 mm). The loamy soils (SGG 4; Brown, Yellow and Red Kandosols) are moderately permeable, imperfectly to well drained depending on landscape position, moderately deep to very deep (0.6 to >1.5 m), massive structured soils with thin (<0.1 m) dark, grey or brown sandy loam to clay loam surface soils. These layers occasionally overlay thick (0.15–0.25 m), brown or yellow, clayey sands to sandy clay loam subsurface soils over occasionally mottled brown, yellow or red sandy clay loam to light clay subsoils with ferruginous nodules. The soil has a moderately deep to very deep (0.6–1.5 m) effective rooting depth, depending on the depth of the underlying rock. As a result the soils have a low to moderate AWC (50–115 mm). The seasonally or permanently wet soils (SGG 3; Hydrosols) are slowly to moderately permeable, poorly drained, deep (>1.5 m) with thin (<0.1 m), dark or grey, mottled, massive, sandy or silty clay loam surface soil over brown, mottled, structured, sandy or silty clay loam to light clay subsoils. The soil has a deep (>0.85 m) effective rooting depth and a moderate AWC (>100 mm). Cracking clay soils (SGG 9; Vertosols) are likely to have formed in the isolated lower clay plains in the elevated parts of the Gulf Fall. The soils are self-mulching, grey or brown, cracking clays, and moderately deep to very deep (1.2–1.5 m). The soils are slowly permeable and very poorly drained with a deep to very deep (1.2–1.5 m) effective rooting depth and a very high AWC (>220 mm). Soils of the Dissected Barkly Tableland physiographic unit This PU comprises rocks that had been covered by the swamp sediments of the Barkly Tableland that since have been exposed. As such, soils are relatively fresh despite their age. The most common rock is the Cambrian dolomites and limestones that outcrop extensively on this undulating landscape. Highly calcareous soils (SGG 10; Calcarosols) have formed on the dolomite and limestones – although much of the soils of these origins are shallow so are classified as shallow and/or rocky soils (SGG 7). The highly calcareous soils (SGG 10; Calcarosols) are moderately permeable, well drained, moderately deep to very deep (0.7–1.5 m) as shown in Figure 3-24, and a landscape setting in Figure 3-25. The soil surface is hardsetting with abundant pebbles and ironstone. The soil has a thin to moderately thick (0.05–0.2 m), red or dark, massive or structured, fine sandy clay loam to fine sandy light clay surface soil with calcareous nodules and course fragments. This layer overlies brown or red, structured, clay loam to light medium clay subsoils with calcareous segregations and course fragments. Depending on underlying rock, soils have moderately deep to deep effective rooting depth, and AWC is low to moderately high (30–150 mm). The soils are generally suitable for spray or trickle-irrigated cropping, particularly horticultural crops. However, nutrient availability can be compromised due to strong alkalinity. These soils tend to occur in small and fragmented patches, which may further reduce agricultural potential due to farm operation inefficiencies. The highly calcareous soils (SGG 10) often occur in association with the shallow and/or rocky soils (SGG 7). Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-24 Soil profile of Calcarosol (SGG 10) sampled in the Dissected Barkly Tableland physiographic unit Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 1.3 m deep sample core was taken. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-25 Calcarosol (SGG 10) landscape of spinifex and bull Flinders grass woodland with many limestone cobbles in the Dissected Barkly Tableland physiographic unit The shallow and/or rocky soils (SGG 7; Calcarosols, Kandosols, Rudosols) are moderately permeable, moderately well to well drained, very shallow (<0.25 m) with surface stones. These soils have a thin (<0.1 m), loam to light medium clay topsoil overlaying a light to light medium clay subsoil with abundant course fragments. The soil has a very shallow (<0.25 m) effective rooting depth and a very low AWC (<30 mm). The soil is shown in Figure 3-26 and its landscape setting in Figure 3-27. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-26 Soil profile of Rudosol (SGG 7) sampled in the Dissected Barkly Tableland physiographic unit Topsoil is lefthand of lower 1 m of core tray and the subsoil in the top core tray. A 0.8 m deep sample core was taken. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-27 Rudosol (SGG 7) landscape of spinifex grassland with snappy gum in the Dissected Barkly Tablelands Soils of the Karumba Plains physiographic unit Seasonally or permanently wet soils (SGG 3; Hydrosols and Aquic Vertosols) occur on tidal flats and wetlands and sandy soils (SGG 6.2; Grey or Brown Tenosols) found on the beach ridges define the PU. The wet soils (SGG 3) are slowly permeable, very poorly drained and very deep (>1.5 m). They may be black, brown or grey light to heavy clay with calcareous nodules in the upper part, and mottles at depth. They are structured and strongly alkaline, and saline at depths greater than 0.4 m, and are potential acid sulfate soils (ASS) at depth too. Beach ridges occur parallel to the coast and consist of fine sand forming a grey or brown sandy soil (SGG 6.2; Arenosols, Tenosols, Rudosols) that is highly permeable, moderately well to rapidly drained, very deep (<1 m), grey or brown, massive or single grain, fine sand with many shells at depth. The soil has a very deep (>1.5 m) effective rooting depth and a low AWC (90 mm). Soils of the Isa Highlands physiographic unit The Isa Highlands PU consists of north–south aligned hills and ridges, and narrow valleys. Soils have developed from a diverse range of parent materials derived from the oldest rocks (Precambrian Era) in the Assessment area. Despite the landscape diversity, most soils have stony surfaces (Figure 3-28) and/or are shallow (<0.5 m, Figure 3-29) in depth (SGG 7; belonging to a wide range of ASCs: Red Chromosols, Dermosols and Ferrosols; Leptic Rudosols and Tenosols; Calcarosols; Brown Kandosols and Dermosols). A typical landscape setting for these soils is shown in Figure 3-30. These soils have surface coarse fragments, rock outcrops and/or rock is encountered within 0.5 m depth across the PU. These soils tend to have a very low to low AWC (<70 mm) and may sometimes be found on eroded slopes or residual surfaces in gullied lands. Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-28 Rudosol (SGG 7) site of very abundant surface coarse fragments in the Isa Highlands physiographic unit Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-29 Shallow soil profile of Rudosol (SGG 7) profile sampled in the Isa Highlands physiographic unit Soil or landscape photo \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\2_Reporting\SGG_Photos For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-30 Rudosol (SGG 7) landscape of open woodland in the Isa Highlands physiographic unit Small occurrences of red sandy soil (SGG 6.1; Red Tenosols) are present on gently sloping plains of Quaternary sedimentary deposits. The soils are moderately permeable, well drained, moderately deep (<1 m), with surface gravel and cobbles (<50%) and have a moderately thick (<0.2 m), brown, massive, sandy loam surface soil over red subsoils with very few rocks. These soils have a moderately deep (<1 m) effective rooting depth with a low AWC (<50 mm). 3.3.4 General land suitability observations In addition to the quantified land evaluation completed using statistical sampling, DSM and land suitability analysis (sections 2.4 and 2.5), a number of qualitative land evaluation observations and notes were taken during expert field visits to the Assessment area discussed in Section 0. These observations and the computational land evaluation data have been used to identify larger tracts of land showing agricultural potential. These tracts are identified as A to F in Figure 3-31 and described in Table 3-7, which includes explanations of the agricultural potential of those lands. Of note: discussions are restricted to larger tracts, but this does not imply that there are not smaller tracts, which, under the right management, may be profitably farmed. Potential agriculture development map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-504_Pot_Dev_v4_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-31 Soil Generic Group map showing areas (A–F) referenced in Table 3-7. These locations identify the more extensive areas of potential agricultural development. Inset map shows the data reliability, based on the confusion index as described in Section 2.4.4 Table 3-7 Qualitative land evaluation observations of large tracts of land with agricultural potential in the Assessment area For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 3.4 Soil attribute data and maps Section 2.4 describes the production of 18 digital soil attribute maps for the Assessment. A selection of the more influential or sensitive of these in the land suitability analyses are discussed, including: soil thickness, soil texture class, available water capacity to 100 cm depth (AWC 100), permeability, surface pH and surface rockiness. Interpretation of the soil patterns references SGG mapping (Figure 3-1) and descriptions (Table 3-5), PU mapping (Figure 1-2) and descriptions (Table 1-2), and scorpan relationships from Section 3.2. 3.4.1 Soil thickness The deeper soils of Figure 3-32 dominate the lowlands and Wellesley Islands, and the western margins of the uplands. The lowland deeper soils are associated with cracking clay soils (SGG 9) of the Armraynald Plain and Cloncurry Plain PUs, whereas in the Doomadgee Plain PU there are significant areas of brown, yellow and grey sandy soils (SGG 6.2), and brown, yellow and grey loamy soils (SGG 4.2). Additionally, the Karumba Plain PU is dominated by seasonally or permanently wet soils (SGG 3) that are also generally deep. In the south-west of the upland areas there are extensive areas of cracking clay soils (SGG 9) on the Barkly Tableland and the Gulf Fall PUs. Similarly, there are large areas of deep red sandy soils (SGG 6.1) and brown, yellow and grey sandy soils (SGG 6.2) on the western margins of the Dissected Barkly Tableland and Gulf Fall PUs. Deep friable non-cracking clay or clay loam soils (SGG 2) dominate Mornington Island. Shallow and/or rocks soils (SGG 7) are extensive across the Isa Highland PU and the eastern areas of the Dissected Barkly Tableland and Gulf Fall PUs. Soil thickness mapping is least reliable along the costal fringes and Mornington Island and most reliable in the uplands, especially the Isa Highland PU. Modelled soil thickness map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-513_Soil_Thickness_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-32 Distribution of (a) soil thickness and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.2 Surface texture Surface soil texture class mapping is presented in Figure 3-33. The clayey surface soils represent a large part of the Assessment area, and dominate the Barkly Tableland, the Dissected Barkly Tableland, the Armraynald Plain, the Karumba Plain and the Donors Plateau PUs. These soils are mainly cracking clay soils (SGG 9) and seasonally or permanently wet soils (SGG 3). Sandy surface soils co-dominate the area and are mainly expressed as shallow and/or rocky soils (SGG 7) and the brown, yellow and grey sandy soils (SGG 6.2) of the Isa Highland and Doomadgee Plain PUs, and sandy textured surface soils are well represented on the Donors Plateau PU. Loamy soil surfaces are associated mainly with SGG 7 soils located in the Gulf Fall and Isa Highland PUs, although the contribution across the Assessment area is minor. Silty surface soils contribute a negligible area. Mapping reliability tends to be lower of the uplands area and dominated by the Isa Highland and Dissected Barkly Tableland PUs reflecting the scarcity of sites for DSM, whereas reliability is best in the Doomadgee Plain and Armraynald Plain PUs in the lowlands, and the Barkly Tableland in the uplands. Modelled soil surface texture map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-514_Text_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-33 Distribution of (a) surface texture class and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.3 Available water capacity (AWC 100) Patterns of AWC to 1 m (AWC 100) in Figure 3-34 closely correlate with soil thickness patterns in Figure 3-32 and surface soil textures in Figure 3-33. While these soil attributes do not definitively indicate subsurface soil textures, the correlation in Figure 3-34 shows the largest AWC values are found where soils are deep and are clay rich, especially the physiographic units of Armraynald Plain, Cloncurry Plain and Barkly Tableland PUs. Mapping reliability for AWC is generally most reliable in upland areas of the Gulf Fall and northern areas of Isa Highlands PUs. It is least reliable in the Barkly Tableland PU reflecting the limited spread of sites used in the DSM, and the short-range availability inherent in many rocky landscapes. Modelled soil Available Water Capacity to 1m map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-515_AWC_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-34 Distribution of (a) available water capacity (AWC) in millimetres to 100 cm depth and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.4 Soil permeability The soil permeability class mapping is presented in Figure 3-35. The Assessment area is dominated by moderately permeable soils, and to a lesser extent, slowly permeable soils. The latter correlate with patterns of cracking clay soils (SGG 9) of Armraynald Plain, Cloncurry Plain and the Barkly Tableland PUs. Highly permeable soils are associated with brown, yellow and grey loamy soils (SGG 6.2) of the Doomadgee Plain PU and areas of the Gulf Fall PU, and a part of the Cloncurry Plain PU on the eastern margin of the Assessment area dominated by shallow and/or rocky soils (SGG 7). Generally mapping reliability in the Assessment area is low and the best map reliability occurs on the Armraynald Plain PU where there are large areas of contiguous soils. Modelled soil permeability map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-511_Perm_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-35 Distribution of (a) soil permeability and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.5 Surface pH The surface pH mapping shown in Figure 3-36 shows results in the range pH 5.5 to 8.5, which is within the acceptable agronomic threshold (Peverill et al., 1999). Areas that are in the acid-to- neutral range (pH 5.5–7.0) are more associated with the sandier surface soils (Figure 3-33) dominated by shallow and/or rocky soils (SGG 7), red sandy soils (SGG 6.1) and brown, yellow and grey sandy soils (SGG 6.2). Sandier soils tend to be more acidic because of increased soil permeability (Figure 3-35) and higher leaching rates of neutralising soil components. Coarse textured soils also tend to lower buffering capacity by virtue of lower cation exchange capacity (CEC) supplied by clay minerals and organic matter. The more alkaline range soil (pH 7.0–8.5) patterns are associated with higher clay content soils, especially the cracking clay soils (SGG 9) and some areas of seasonally or permanently wet soils (SGG 3). Mapping accuracy is generally weak across the Assessment area, being weakest in the Karumba Plain, the Dissected Barkly Tableland and areas of the Gulf Fall PUs. Stronger mapping reliability is associated with the Doomadgee Plain PU. Modelled soil surface pH map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-512_Surf_pH_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-36 Distribution of (a) surface pH and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.6 Surface rockiness The surface rockiness mapping is displayed in Figure 3-37 and shows a strong correlation of the rocky soil surfaces to the shallow and/or rocky soils (SGG 7) seen in Figure 3-1. Generally, the reliability of mapping is high, although relatively localised areas of lower reliability are found in SGG 7 soil areas. Modelled soil rockiness map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-516_Rocky_2x1_v2_ArcGIS10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-37 Distribution of (a) surface rockiness and (b) the companion reliability mapping in the Southern Gulf catchments 3.4.7 Acid sulfate soils The potential acid sulfate soil (ASS) distribution in the Assessment areas is shown in Figure 3-38 and indicates the soils are found in the seasonally or permanently wet soils (SGG 3) of the Karumba Plain PU, on the mainland and also the Wellesley Islands where the lands are below 5 mAHD. The area of land affected by potential ASS is 545,000 ha, and these areas will significantly limit development opportunities for agriculture. ASS also affects built infrastructure due to seasonal or permanent wetness, natural salinity, and the requirement to manage structural degradation from ASS. However, these soils can be safely used for aquaculture with correct site management (e.g. lined ponds). Potential acid sulfate soils map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-505_ASS_v1_v11_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-38 Distribution of potential acid sulfate soils in the Southern Gulf catchments 3.5 Land suitability The following presents a selection of exemplar land suitability data and maps (with accompanying reliability mapping) for cropping and aquaculture in the Assessment area. Agricultural land versatility (i.e. cropping) is also discussed, along with methods to address landscape complexity that may impose additional farm management limitations on cropping land use options. 3.5.1 Land suitability distributions The following section presents irrigated and rainfed crop group suitability (Section 2.5) distributions for 14 ‘exemplar’ land uses chosen from the modelled 58 possibilities in Appendix C. The exemplar land use options have been selected to represent a realistic set of options (i.e. the crop groups by season by irrigation type are expected to align to growing conditions (land, soils and climate), market desirability and favourable growing experience from comparable growing conditions elsewhere in Australia). Throughout the discussion readers are referred to the catchment PUs in Section 1.1 and shown in Figure 1-2, SGG mapping in Figure 3-31 with references to larger areas expertly assessed to have agricultural potential, Table 3-6 documenting SGGs and their agricultural opportunity and soil landscape explanations in Section 3.3. Comment is also made on the reliability of the land suitability mapping from the methods in Section 2.4.4. Area calculations for the land suitability classes for the various land uses are presented in Figure 3-46. The first set of land suitability maps show crop group 7, ‘grain and fibre crops’ such as cotton or sorghum, grown using furrow irrigation in the wet season (Figure 3-39 (a)). The suitable area under this land use option is 775,618 ha (7.2% of the Assessment area), and all is class 3 suitability. The areas of suitability are consistent with the better drained cracking clay soils (SGG 9) of the Barkly Tableland and Armraynald Plain PUs. Figure 3-39 (b) shows the suitability distribution for the same crop group grown under wet-season, rainfed farming practices. This shows most of the Assessment area is unsuitable with only 363,396 ha (3.4%) in the area suitable as class 3, and these areas are restricted to cracking clay soils (SGG 9) from the Armraynald Plain and seasonally wet clay soils (SGG 3) of the Doomadgee Plain PUs. Land suitability map - cotton \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-525_Suit_Cotton, grains_Cotton, grains_v2.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-39 Modelled land suitability (a) for crop group 7, ‘grain and fibre crops’ such as cotton or sorghum, grown using furrow irrigation and (b) grown during wet season, relying on rainfall Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. The suitability for crop group 10, ‘pulse crops (food legumes)’ such as mungbean or soybean grown under dry-season furrow irrigation is shown in Figure 3-40 (a). There is 1,666,038 ha (15.4%) of land available classified as suitability class 3 for this land use. The suitable areas are centred on the cracking clays (SGG 9) in the Barkly Tableland and Armraynald Plain PUs, with smaller areas available on alluvium of the Donors Plateau PU. The suitability of the same crop group (i.e. 10) under wet-season rainfed (non-irrigated) land use is presented in Figure 3-40 (b). This shows considerably less suitable land compared to the dry- season furrow irrigated option in that 360,776 ha (3.3%) is suitable as class 3. These areas are associated with pockets in the Armraynald Plain with better drained cracking clays soils (SGG 9) and on the Doomadgee Plain PU, brown, yellow and grey loamy soils (SGG 4.2) and seasonally wet clay soils (SGG 3) adjacent to rivers and creeks. Many of these soils have a large AWC. Land suitability map – pulses \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-526_Suit_Pulses_Pulses_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-40 Modelled land suitability (a) for crop group 10, ‘pulse crops (food legumes)’ such as mungbean or soybean, grown using furrow irrigation in the dry season and (b) wet-season rainfed Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. The land that is suitable for crop group 14, ‘grass hay and forage (perennial)’ such as perennial Rhodes grass, grown using spray irrigation is shown in Figure 3-41 (a). Under this land use option there is just under half (47.1%) of the Assessment area suitable; in terms of class 2, there is 1,699,483 ha (15.7%), and for class 3, 3,403,611 ha (31.4%). These areas coincide with red loamy soils (SGG 4.1), brown, yellow and grey loamy soils (SGG 4.2), red sandy soils (SGG 6.1), and brown, yellow and grey sandy soils (SGG 6.2) of the Doomadgee Plain PU, cracking clay soils (SGG 9) of the Armraynald Plain, Cloncurry Plain and Donors Plateau PUs in the lowlands. These PUs also host much of the class 2 suitability land in the Assessment area. In the uplands there are areas of class 2 and 3 coinciding with red sandy soils (SGG 6.1) and brown, yellow and grey sandy soils (SGG 6.2) that occur in the Barkly Tableland, the Dissected Barkly Tableland and the Gulf Fall PUs. Figure 3-41 (b) shows the suitability distributions for crop group 12, ‘grass hay and forage (annual)’, under dry-season spray irrigation. Again, this shows a significant proportion of the Assessment area to be suitable (43.5%). This breaks down to 893,938 ha (8.2%) of class 2 lands and 3,820,857 ha (35.2%) of class 3 lands. The patterns are similar to those of the spray irrigation option (Figure 3-41 (a) below) with clay soils (e.g. SGG 9), sandy soils (i.e. SGG 6.1 and 6.2), and the loamy soils (i.e. SGG 4.1 and 4.2) being suitable. Land suitability map – forage \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-527_Suit_Rhodes_Annual_Forages_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-41 Modelled land suitability (a) for crop group 14, ‘hay and forage (perennial)’ such as Rhodes grass, grown using spray irrigation and (b) for crop group 12, ‘hay and forage (annual)’ such as sorghum (forage) and maize (silage), grown using spray irrigation in the dry season Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. The land suitability distributions for crop group 13, ‘legume and forage (annual)’ such as lablab, grown using dry-season furrow irrigation is displayed in Figure 3-42 (a) and shows that there is 1,799,551 ha (16.6% of the Assessment area) of class 3 land suitable. These areas are strongly associated with the cracking clay soils (SGG 9). These soils correspond with the Barkly Tableland and Armraynald Plain PUs. Figure 3-42 (b) depicts the suitability class distributions for crop group 1 ‘tree crops/horticulture (fruit)’ such as mango and lychee, grown using trickle irrigation. There are 3,855,045 ha (35.5%) available as suitable for the land use, of which a small area (103.4 ha; <1%) is class 1, 973,124 ha (9%) is class 2 and 2,881,817 ha (25.6%) is class 3. The class 1 and 2 areas are strongly associated with the sandy soils (i.e. red sandy soils; SGG 6.1 and brown, yellow and grey sandy soils; SGG 6.2) of the Gulf Fall and Doomadgee Plain PUs. Class 3 areas are associated with the cracking clay soils (SGG 9), the friable non-cracking clay or clay loam soils (SGG 2) predominantly in the Armraynald Plain PU, areas of the Cloncurry Plain and Doomadgee Plain PUs in the lowlands, and the Barkly Tableland PU in upland areas. Land suitability map – legume feed and mango \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-528_Suit_Lablab_Mango_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-42 Modelled land suitability (a) for crop group 13, ‘hay and forage (annual)’ such as lablab, grown using furrow irrigation in the dry season and (b) crop group 1 ‘tree crops/horticulture (fruit)’ such as mango and lychee, grown using trickle irrigation Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. Twenty-four percent (2,658,022 ha) of the Assessment areas is suitable for growing crop group 2, ‘citrus tree crops/horticulture (fruit)’ such as lime with trickle irrigation (Figure 3-43 (a)). Of these suitable lands, 518,088 ha (4.8%) is class 2 and these areas tend to coincide with brown, yellow and grey sandy soils (SGG 6.2) of the Doomadgee Plain. Class 3 areas comprise 2,139,934 ha (19.7%) of the Assessment area, and these areas are mostly associated with the moderately or highly permeable soils (Figure 3-35) of the Doomadgee Plain and Gulf Fall PUs. These soils include brown, yellow and grey sandy soils (SGG 6.2) and better drained cracking clay soils (SGG 9). Other class 3 areas occur in the red sandy soils (SGG 6.1), brown, yellow and grey sandy soils (SGG 6.2) and brown, yellow and grey loamy soils (SGG 4.2) of the Gulf Fall PU. Figure 3-43 (b) presents the land suitability distributions for group 19, ‘oilseeds’ such as sunflower or sesame under wet-season spray irrigation. Modelling shows that 3,112,699 ha (28.7%) is suitable, with 344,428 ha (3.2%) being class 2 and 2,768,271 ha (25.5%) being class 3. The class 3 areas coincide with the better drained cracking clay soils (SGG 9), the brown, yellow and grey loamy soils (SGG 4.2) of the lowland Armraynald Plain and Doomadgee Plain PUs, and the red sandy soils (SGG 6.1), brown, yellow and grey loamy soils (SGG 4.2) and cracking clay soils (SGG 9), predominantly in the upland Barkly Tableland and Gulf Fall PUs. Class 2 areas are associated with the Doomadgee Plain PU with brown, yellow and grey sandy soils (SGG 6.2). Land suitability map – citrus and oilseed \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-529_Suit_Citrus_SunflowerSes_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-43 Modelled land suitability (a) for crop group 2, ‘tree crops/horticulture (fruit)’ such as citrus, grown using trickle irrigation and (b) group 19, ‘oilseeds’ such as sunflower or sesame using wet-season spray irrigation Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. Figure 3-44 (a) shows the land use suitabilities for crop group 9, ‘small-seeded crops’ such as chia, grown using dry-season spray irrigation. A significant proportion of the Assessment area is suitable, amounting to 4,757,510 ha (43.9%). Of this, 939,628 ha (8.7%) is class 2 and 3,817,882 (35.2%) class 3. The suitability patterns for crop group 3 ‘intensive horticulture’ (vegetables, row crops) such as cucurbits using dry-season trickle irrigation are shown in Figure 3-44 (b), indicating that 757,003 ha (7.0%) is class 2 and 4,161,014 ha (38.4%) is class 3. Given the strong similarity in patterns for both of these land use options, the pattern descriptions are combined for this discussion. Class 2 suitability closely aligns with distributions of red loamy soils (SGG 4.1) and brown, yellow and grey sandy soils (SGG 6.2) of the Doomadgee Plain and Armraynald Plain PUs, and in the Armraynald Plain PU some areas of cracking clay soils (SGG 9). There are also minor areas in the Donors Plateau PU. Class 3 areas are more closely associated with cracking clay soils (SGG 9) of the Armraynald Plain and Cloncurry Plain PUs, and brown, yellow and grey loamy soils (SGG 4.2) of the Doomadgee Plain PU. Other significant areas of class 3 lands are found on the Barkly Tableland PU dominated by cracking clay soils (SGG 9), and the Gulf Fall PU on red sandy soils (SGG 6.1), brown, yellow and grey sandy soils (SGG 6.2), brown, yellow and grey loamy soils (SGG 4.2), and cracking clay soils (SGG 9). Land suitability map – Quinoa and cucurbits \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-530_Suit_ChiaQuinPop_Cucurbits_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-44 Modelled land suitability (a) for crop group 9, ‘small-seeded crops’ such as chia, grown using dry-season spray irrigation and (b) for crop group 3 ‘intensive horticulture (vegetables, row crops)’ such as cucurbits using dry- season trickle irrigation Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. Figure 3-45 (a) presents the land suitabilities for crop group 6, ‘root crops’ such as sweet potato, grown using dry-season spray irrigation. In terms of class 2 suitability, there is 897,510 ha (8.3%) available, and class 3, 2,754,380 (25.4%) available, combining to make 3,651,891 (33.7%) suitable for this land use. Class 3 areas are most associated with the better drained cracking clay soils (SGG 9) of the Armraynald Plain, and brown, yellow and grey loamy soils (SGG 4.2) and brown, yellow and grey sandy soils (SGG 6.2) of the Doomadgee Plain PU. There are also areas of class 3 in red sandy soils (SGG 6.1) and brown, yellow and grey sandy soils (SGG 6.2) of the Gulf Fall and cracking clays of the Barkly Tableland PUs. Class 2 areas are found in the Doomadgee Plain PU where brown, yellow and grey sandy soils (SGG 6.2) and brown, yellow and grey loamy soils (SGG 4.2) are dominant. There are also areas of class 2 in areas of cracking clay soils (SGG 9) of the Donors Plateau PU. The land suitabilities for crop group 15, ‘silviculture/forestry’ such as Indian sandalwood grown using trickle irrigation are shown in Figure 3-45 (b). In many ways, patterns echo those of crop group 6 grown using dry-season spray irrigation discussed immediately before. The modelling indicates a very minor incidence of class 1 (103.4 ha; <1% of the Assessment area), 679,294 ha (6.3%) of class 2 and 3,101,764 ha (28.6%) of class 3. Combined, 3,781,162 ha is suitable, representing 34.9% of the Assessment area. The most significant difference to the suitability patterns of this land use compared to those of the immediately preceding group 6 under dry- season spray irrigation is that there is less class 2 area in the loamy (SGG 4.2) and sandy (SGG 6.2) soils of the Doomadgee Plain PU. Land suitability map – root crops and forestry \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-531_Suit_Root_Sandal_v1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-45 Modelled land suitability for (a) crop group 6, ‘root crops’ such as sweet potato, grown using dry-season spray irrigation and (b) crop group 15, ‘silviculture/forestry’ such as Indian sandalwood grown using trickle irrigation Insets illustrate reliability of land suitability mapping. Note that these land suitability maps do not take into consideration flooding, risk of secondary salinisation or availability of water. More detail for the crop groups can be found in Table 2-5. The bar chart in Figure 3-46 gives a comparative overview of the 14 exemplar land uses areas presented above and shows that crop group 14 (perennial grass Rhodes) under spray irrigation is the most suitable in the Assessment area with 5,103,095 ha that is suited (i.e. classes 1 to 3). Suitable areas equate to 47.1% of the Assessment area. Other prospective land uses include crop group 3, intensive horticulture (e.g. cucurbits) under dry-season trickle irrigation (4,918,110 ha; 45.4%), crop group 9, small-seeded crops under dry-season spray irrigation (4,757,511 ha; 43.9%), and crop group 12, annual forage under dry-season spray irrigation (4,714,796 ha; 43.5%). The least prospective land uses are crop group 7 (e.g. cotton) grown under wet-season furrow irrigation (775,619 ha; 7.2%), crop group 7 (e.g. cotton) grown under rainfed, wet-season growing conditions (363,436 ha; 3.4%), and crop group 10 (e.g. pulses) grown under rainfed wet-season practices (360,813 ha; 3.3%). For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Figure 3-46 Area (ha) of the Southern Gulf catchments mapped in each of the land suitability classes for the 14 exemplar land use options A description of the five land suitability classes is provided in Table 2-4. More detail on the 21 crop groups, and example crops, is found in Table 2-5 and Section 2.5.1. The full suite of 58 land suitability outputs is presented in Appendix G. 3.5.2 Versatile agricultural land Figure 3-47 shows the versatility of the agriculture index under four irrigation management systems namely, spray (a), trickle (b), furrow/flood (c) irrigation methods, plus (d), rainfed agricultural versatility from the methods in Section 2.5.7. Figure 3-47 (a) shows that the most versatile lands with respect to spray irrigation are generally associated with the brown, yellow and grey sandy soils (SGG 6.2), the brown, yellow and grey loamy soils (SGG 4.2), and some cracking clay soils (SGG 9) of the lowlands. These tracts are predominantly found in the lowlands on the Doomadgee Plain, the Armraynald Plain, the Donors Plateau and Cloncurry Plain PUs, and in the highlands, the Barkly Tableland and areas of the Gulf Fall PUs where the better drained cracking clay soils (SGG 9), red sandy soils (SGG 6.1) and brown, yellow and grey sandy soils (SGG 6.2) occur. These areas of greatest versatility under spray irrigation are consistently deeper soils (Figure 3-32) enabling deep drainage, reducing risk of waterlogging and creating larger AWC (Figure 3-34). Areas of low or no versatility are associated with the shallow and/or rocky soils (SGG 7) of the Isa Highland and seasonally or permanently wet soils (SGG 3). The former have very shallow soils (Figure 3-32) and/or are rocky (Figure 3-37) so have a low AWC (Figure 3-34), whereas the latter will be prone to waterlogging and/or salinity. The versatility under trickle irrigation is presented in Figure 3-47 (b) and shows similar patterns to spray irrigation in Figure 3-47 (a) for similar reasons, including the positive soil attributes of deep soils and large AWC, and the negative attributes of shallow soils, rockiness and proneness to waterlogging and/or salinity. Visually, spray and trickle irrigation show a sizable proportion of the Assessment area as being moderately to largely versatile. Furrow irrigation versatility is shown in Figure 3-47 (c) and visually shows that most of the Assessment has zero versatility for this irrigation type. Versatile areas are restricted to areas of cracking clay soils (SGG 9) and the limited areas of friable non-cracking clay or clay loam soils (SGG 2). These soils have characteristics for more efficient furrow irrigation, namely being slowly permeable (Figure 3-35) to help water flows along furrows, and are deep (Figure 3-32) for greater water storage (AWC; Figure 3-34). These soils are mostly found on the Armraynald Plain and the Barkly Tableland PUs. Rainfed versatile areas are limited in extent (Figure 3-47 (d)) across the Assessment area and are restricted to lowland areas where lands are less arid (Figure 2-4 (d)). The small areas of versatility are evident in the brown, yellow and grey loamy soils (SGG 4.2) of the Doomadgee Plain and the cracking clay soils (SGG 9) of the Armraynald Plain PUs. In addition to the reduced aridity, these soils are deep (Figure 3-32) and with larger AWCs (Figure 3-34) and together with rainfall, combine environmental attributes to be suitable for rainfed agriculture. Ag versatility by irrigation map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-533_Versatility_2X2_v1_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-47 Agricultural versatility for (a) spray, (b) trickle and (c) furrow irrigation types, and (d) rainfed Higher index values indicate greater versatility for each irrigation option. Displays are not suited to inter-irrigation type comparisons. Figure 3-48 shows the versatility index map of agricultural lands derived from the 14 exemplar land uses presented in Section 3.5.1 and from methods in Section 2.5.7. The most versatile lands are strongly influenced by the patterns of the spray and trickle irrigation versatilities (Figure 3-47 (a) and (b)) and show that much of the Assessment area is versatile to at least some extent regarding the combinations of irrigation and rainfed land uses. Areas with no versatility are found in the Isa Highland PU where shallow and/or rocky soils (SGG 7) dominate, combining shallow soils (Figure 3-32) with surface rockiness (Figure 3-37). The other area with zero versatility is associated with the Karumba Plain featuring seasonally or permanently wet soils (SGG 3) where waterlogging and/or salinity will be an issue. The areas of largest versatility in the lowlands are strongly associated with the Doomadgee Plain PU dominated by brown, yellow and grey loamy soils (SGG 4.2) and brown, yellow and grey sandy soils (SGG 6.2), and the Armraynald Plain PU areas with the better drained cracking clay soils (SGG 9). In the uplands, the most versatile soils are on the Barkly Tableland PU with cracking clay soils (SGG 9), and Gulf Fall with red sandy soils (SGG 6.1), brown, yellow and grey sandy soils (SGG 6.2), and smaller tracts of cracking clay soils (SGG 9). Parts of the Wellesley Islands – Mornington Island in particular – are shown to be versatile on the Mornington Plateau PU where friable non-cracking clay or clay loam soils (SGG 2) are common. The versatility maps help to identify land where types of irrigation investment may be best targeted, or to guide where land can be most flexibly used if and as markets and technologies shift. These are likely to be the lands where farming resilience can be achieved by virtue of the variety of land use choices that are available. The scale of mapping presented here is not suitable for identifying the potential of small parcels of land that may be sufficiently large enough on their own, or sufficiently closely clustered, to be viable for farming on a case-by-case basis. Ag versatility map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-532_AgVers14_v4_v11_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-48 Agricultural versatility index map combining 14 unique land use options 3.5.3 Aquaculture land suitability The land suitability for aquaculture considers access to seawater (i.e. coastal strip and the major tidal river estuaries) of ≤2 km for marine species and no proximity consideration for freshwater species. Soil and land limitations for lined and earthen ponds vary, for example pH relating to the physiological tolerances of species, sodicity for integrity of soil for impoundment maintenance and longevity, and permeability for water retention, whereas limitations like slope, soil thickness and rockiness are pertinent to both earthen and lined ponds. The land suitability frameworks for aquaculture are presented in Appendix E. In discussions that follow reference is made to PUs (Figure 1-2 and Table 1-2) and SGGs (Section 3.3.2 and Figure 3-1). The mapped land suitability classes for lined marine aquaculture are shown in Figure 3-49 (a). This shows that 300,206 ha (2.8%) of land is suitable for marine aquaculture land use. Of this area 86,000 ha (0.8%) is class 2. Suitable areas are restricted to the Karumba Plain PU (Figure 1-2) where SGG 3 soils (seasonally or permanently wet) dominate (Figure 3-1). The suitable areas extend into some of the most downstream areas of the Armraynald Plain PU where tidal influence is still felt and where those places coincide with the presence of SGG 9 soils (cracking clays). These soils have no surface rocks Figure 3-37. Land suitability patterns for earthen marine aquaculture impoundments are shown in Figure 3-49 (b) and resemble patterns of lined marine aquaculture above given the land uses are restricted to the same PUs and soil attributes. There is only class 3 land, which represents 193,600 ha (1.8%) of the Assessment area. The possibility for earthen impoundments is dependent on soil factors including sufficient depth (Figure 3-32), low soil permeability (Figure 3-35) and heavier surface textures (Figure 3-33). Aquaculture marine map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-535_Aquaculture_marine_v1_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-49 Land suitability for marine aquaculture in (a) lined ponds and (b) earthen ponds proximal to coastal areas Figure 3-50 shows land suitability classes for (a) lined freshwater and (b) freshwater earthen aquaculture. These land uses are not restricted by proximity to marine or tidal water sources, hence they may be possible anywhere in the Assessment area where there is access to reliable fresh water (e.g. groundwater, surface water storage, or from creeks or rivers). Viability of freshwater sources are not considered in this activity, only soil and land attributes. Reference to Figure 3-50 (a) illustrates that a majority of the Assessment area is suited to freshwater lined aquaculture (i.e. 6,265,400 ha (57.9%)). Of this area, 5,154,300 ha (47.6%) is class 2, with only a minor proportion (40,300 ha; 0.4%) class 1 and 1,070,800 ha (9.9%) class 3. Unsuitable areas are associated with the shallow and/or rocky soils (SGG 7) and surface rockiness (Figure 3-37), with all other SGGs being suitable. Freshwater, earthen impoundment aquaculture shown in Figure 3-50 (b) is suitable for 2,408,273 ha (22.3%) of the Assessment area, and of this, only 170 ha falls under class 2. The suitable areas match the cracking clay soils (SGG 9) distribution (Figure 3-1) as these soils provide the necessary soil conditions including depth (Figure 3-32), slower permeability (Figure 3-35) and heavier textured surface soils (Figure 3-33). Aquaculture freshwater map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-534_Aquaculture_fresh_v2_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-50 Land suitability for freshwater aquaculture in (a) lined ponds and (b) earthen ponds 3.5.4 Landscape complexity Methods were tested (Section 2.5.6) to assess the contiguousness of parcels of suitable cropping lands. The methods were based on natural distributions of soil and land variability to address operational farming constraints imposed by parcels of suitable land being too small according to natural variability of land, or physical limits on suitable farming land parcel sizes caused by land dissection through anabranching. Examples of analytical results are presented below. Contiguous suitable areas – an example The final product of the contiguous suitable area analysis is shown in Figure 3-51 (c), and the plates (a) to (d) illustrate the workflow to address contiguous areas in situations where spatial variability of suitability class distributions is high as shown in plate (a). The following illustrated discussion focuses on changes occurring in the fixed position ellipsoids in the illustration plates. Plate (b) presents an aggradation of the suitable classes (i.e. classes 1 to 3) as green, and non- suitable (classes 4 and 5) as white. Next, plate (c) shows the result of geographic information system (GIS) spatial filtering to aggregate units into minimum on-ground areas of 25 ha. At some locations filtering causes non-suitable land to be included in suitable lands (compare plates (c) and (a)), and conversely, suitable lands included in non-suitable areas (plate (d)). While this approach means localised losses of suitable land, incorporation of non-suitable lands into the enlarged suitable land units is likely to be of overall benefit to the farmer because of efficiencies gained through operating at scale. Contiguous areas method map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-506_ContigAreas_2x2_v1_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-51 Contiguous suitable areas workflow outputs from the Doomadgee Plains Floodplain stream dissection – an example The results of the floodplain stream dissection analysis demonstrates that while there are widespread areas of stream dissection beyond the main river channels in much of the alluvial areas in the catchment, most channels are less than 1 m deep thus do not make the land unsuitable from a land dissection perspective. The example shown in Figure 3-52 shows image overlayed Light Detection and Ranging (LiDAR) analysis of a section of Gregory River floodplain showing channels greater than 1 m deep in red. This shows that the majority of channels are greater than 1 m deep, so according to the dissected land criteria, most of the land is dissected here. While this analysis has not been applied throughout the Assessment area, the example serves to illustrate its use as part of a suite of preliminary land-development analyses to assess on- ground viability and planning for a new farming enterprise. Stream dissection map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-518_Floodplain_v1_v11_Arc10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-52 Satellite image of land south of the Gregory and Nicholson river confluence showing an extensive network of streams on the Gregory River floodplain. Streams deeper than one metre below land surface are highlighted in red 3.5.5 Propagation of artefacts in digital soil mapping and land suitability data As a footnote to the land suitability analysis, there will be instances of visible spatial artefacts in some mapped attribute and suitability data. Such artefacts are typically observed as unnaturally crisp edges in maps with no clear physiographic rationale and can be inherited from DSM covariates (Section 2.4.2). In this study, there are likely to be three sources of artefacts from: • binary covariates (e.g. vector-based geological mapping) • the decision tree algorithm reflecting data decision point threshold values in continuous covariates (Section 2.4.4) • raster-based covariates with cultural and natural ephemeral land patterns derived from remote sensing. For example, cultural patterns due to land use such as road corridors, urban/rural interfaces, and paddock boundaries with different crops on each side showing up in the satellite imagery. Natural ephemeral patterns include burn scars showing up in the satellite imagery. These artefacts are generally not significant in the Assessment area, but in places may be locally conspicuous. While mapped artefacts may draw the eye they are not real-world data features and overall mapping quality is best judged against the quantitative data suite (i.e. statistical error and reliability mapping discussed in sections 2.4.4 and 2.4.4). Due diligence by prospective land developers involving on-ground assessments prior to decision making can either identify or explain mapped artefacts. An example of DSM artefact propagation from input covariates is shown in the panels in Figure 3-53 (a), (b) and (c) of the same ground area. Figure 3-53 (a) shows the Landsat Thematic Mapper image for the area showing land features such as drainage areas and vegetation patterns. Figure 3-53 (c) of the same area represents the Multi Resolution Valley Bottom index (MrVBF) covariate comprising discrete values showing responses to lower elevation landscape features like drainage areas and plains, and higher landscape features including low hills and rises. Figure 3-53 (b) shows the surface exchangeable sodium percentage (ESP) DSM (Section 2.4, Table 3-3), which has inherited the crisp boundaries from the MrVBF in the mapping. Such artefacts are also propagated in the land use suitability mapping. Figure 3-53 (d) shows an example of a sharp divide between land suitability classes 3 and 4 for perennial citrus under trickle irrigation caused by a numeric threshold in the land suitability framework, in this case instigated by the heat stress climate data (i.e. number of days over 40 °C). Artefact map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-508_Artefact_v1_10_8.png For more information on this figure please contact CSIRO on enquiries@csiro.au Figure 3-53 Example propagation of apparent artefacts For the panels above: (a) Landsat Thematic Mapper satellite image showing part of the Doomadgee Plain landscape north of Doomadgee; (c) the MrVBF covariate exhibiting the effect of a categorical input dataset for the same area; (b) surface exchangeable sodium percentage mapping with the artefacts from the input covariate data (c) as distinct features; and (d) a distinct line feature in the perennial trickle citrus suitability data an artefact from decision points for threshold splits in continuous covariates exist, in this case heat stress climate data (number of days >40 °C). 4 Synthesis The Assessment area spans the NT and Queensland border and comprises the Settlement Creek, Gregory – Nicholson, Leichhardt and Morning Inlet river catchments, and the selected larger islands of the adjacent Wellesley Island group in the Gulf of Carpentaria. Combined, these lands cover 108,200 km2 and is dominated by extensive cattle grazing. This land suitability activity produced a suite of outputs to inform a broad appraisal of land intensification options based on the Assessment area’s soil and land characteristics. This information is useful for various stakeholders including land managers and developers, land use policy makers, and the wider Australian community. This has been achieved by creating 18 soil and landscape attribute maps and datasets and applying these to a land suitability framework to show the potential for new agriculture and aquaculture land use options. Fifty-eight agriculture and four aquaculture options were generated. Two main tasks were completed. Firstly, new land and soil attribute maps and datasets were created using digital soil mapping (DSM) employing new soils data from 97 sampling sites visited during the project, which were augmented by soil data from 712 pre-existing sampling sites and sites drawn from other sources. Secondly, the new attribute maps were applied in a computer land suitability analysis framework, based on the conventional land suitability assessment approach (FAO, 1976; 1985), to test and map land suitability for an expertly selected set of crops grown under plausible management options (irrigated and non-irrigated) for the Assessment area. An analogous procedure was followed to map land suitability for freshwater and marine aquaculture. As with the preceding assessments, namely the Flinders and Gilbert Agricultural Resource Assessment (Bartley et al., 2013; Harms et al., 2015; Thomas et al., 2015), the Northern Australia Water Resource Assessment (Thomas et al., 2018a; Thomas et al., 2018b), the Roper River Water Resource Assessment (Thomas et al., 2022), and most recently, the Victoria River Water Resource Assessment (Thomas et al., 2024), procedures have been incrementally adapted. Despite the changes, the outputs in this report remain consistent and comparable with previous assessments in northern Australia. The land suitability activity is framed around crop ‘limitations’ relating to land, soil and climate attributes. Ready-to-go climate attribute datasets were accessed from public sources while all land and soil attribute maps were generated with DSM. A tailored selection of 18 were generated that included soil thickness, the soil’s available water capacity (AWC), surface pH, surface texture class and surface rockiness. The tailored selection was guided by the needs of land suitability models for crop groups. Individual crops were not modelled, rather 21 groups of crops (‘crop groups’) were used, grouped according to similar traits and growing needs (i.e. limitation thresholds). The land suitability framework matches instances of crop group by season by irrigation type. While not an irrigation type per se, rainfed agriculture is included because it may be viable in some instances. Accordingly, 58 land use scenarios – comprising the 21 crop groups by season (wet, dry and perennial) by irrigation type (17 furrow/flood, 23 spray and 10 trickle, and 8 rainfed) – were modelled (Appendix A). Following the Food and Agriculture Organization of the United Nations (FAO) land suitability system (FAO, 1976; 1985), land use options are presented using a 5-class ranking system: class 1 means lands are highly suitable land with negligible limitations for the land use option (i.e. there is a very strong match between conditions and land use), whereas class 5 represents unsuitable land with extreme limitations (i.e. extremely poor match so effectively making the land use impossible). Classes 1 through to 3 are deemed suitable – with increasing levels of land and crop management (i.e. cost), while classes 4 and 5 are deemed unsuitable. Appendix Gpresents the maps of the 58 land suitability options. Reliabilities of DSM attribute and land suitability maps were estimated using statistical measures. These measures are presented in map form as companions to DSM and land suitability mapping. These give users objective, visual views of mapping quality to support decisions. According to these and the other statistical measures, the quality of results are generally acceptable, and the mapped estimates show the distributions to be variable. In an extension of land suitability outputs, a methodology was employed to create indices capturing land use versatility across the Assessment area. These combine land suitability mapping and, for example, highlight areas where the suitable land use classes intersect showing which areas are highly versatile lands and those that are not. Two types of versatility maps were generated: (i) the first showing crop versatility for 14 ‘exemplar’ land uses, and (ii) the second showing irrigation type versatility for all 58 modelled land uses. These versatility products may be used to guide land use policy or for targeting agricultural development in the Assessment area. Tracts with high versatility warrant further investigation by developers. Summarising outputs of the agricultural versatility maps from Section 3.5.2, the most versatile lands based on the combination of the 14 exemplar land uses (i.e. covering spray, trickle and furrow irrigation, and rainfed) are found on the Doomadgee Plain physiographic unit (PU) (Figure 1-2) dominated by brown, yellow and grey loamy soils (Soil Generic Group (SGG) 4.2) and brown, yellow and grey sandy soils (SGG 6.2) (SGGs are shown in Figure 3-1 and described in Table 3-5) and the better drained areas of the Armraynald Plain PU bearing cracking clay soils (SGG 9) in the lowlands. In the uplands, the most versatile soils are on the Barkly Tableland PU with cracking clay soils (SGG 9), and Gulf Fall PU with red sandy soils (SGG 6.1), brown, yellow and grey sandy soils (SGG 6.2), and smaller tracts of cracking clay soils (SGG 9). Parts of the Wellesley Islands (especially Mornington Island) of the Mornington Plateau PU where friable non-cracking clay or clay loam soils (SGG 2) are common. Section 3.5.1 includes detailed summaries of each of the 14 exemplar land use suitability areas and the links to soils (SGGs) and PUs. The outputs of the aquaculture land suitability discussed in Section 3.5.3 show the scale of opportunity for marine and freshwater species grown under different management (i.e. lined or earthen impounded ponds). The area of opportunity for marine species is limited by access to tidal marine waters, and so limited to coastal margins and inland along the major rivers. Of the Assessment area, 300,206 ha (2.8%) is suited to marine line aquaculture, and 193,600 ha (1.8%) for earthen ponds. With freshwater aquaculture that is potentially fed by ground or surface waters, the whole Assessment area is considered, and results show that most lands of the Assessment area are suited to freshwater lined aquaculture (i.e. 6,265,400 ha (57.9%)), and a smaller proportion for freshwater with earthen ponds (i.e. 2,408,273 ha (22.3%)). This Assessment gives no consideration of water supply, rather only considers land and soil attributes. It has been noted above that outputs of this Assessment inform users of the reliability of the maps generated (DSM and land suitabilities) and enables users of the data to set their own confidence level when using the Assessment products. It is recommended that the maps and products generated by the Assessment be used at a printed map scale of approximately 1:250,000, thus reflecting a low intensity or reconnaissance-type land evaluation scale. It is therefore important for users to be aware that the information provided characterises land suitability over a broad area and thus best lends itself to regional-scale overview and appreciation of opportunity. Additional, detailed on-ground land and soil investigations must be followed for land developments at the scheme or property scale, and assessment and management made in line with the correct jurisdictional guidance or legislation (e.g. relating to ASS in Queensland, Dear et al., 2014). Limitations not considered in this Assessment may have a bearing on the viability of an enterprise. These include: • economics and finance, including subsidies and grants, commodity, fertiliser and fuel prices, etc. • proximity to produce processing facilities (e.g. cotton gins, abattoirs), transport networks, service hubs and markets • risk of flooding • land management-induced secondary salinity • conservation areas and reserves • proximity to available irrigable water. The methodologies have been developed with no consideration with respect to climate uncertainty. Policy and land tenure considerations are not imposed in the land suitability modelling in recognition that these socio-economic and political attributes of the landscape may shift as economic and technologies change, and community sentiments and aspirations shift along with community values. Some of these factors are reported with other activities in the Assessment. Caution must be followed if using Assessment outputs for planning purposes without wider consideration of these limitations. Finally, the land suitability frameworks (crops and aquaculture) offer a systematic, quantitative framework to analyse land and water development opportunities in the Assessment area. These outputs inform the other activities undertaken in the Assessment, including surface water hydrology; agriculture and socio-economics; surface water storage; Indigenous water values, rights, interests and development goals; and ecology. Outputs of this activity permit realistic trade-offs to be made between the types and size of development opportunities (and limitations) before development should continue. Should conditions change in the Assessment area, modifications can be made to the framework and analyses re-run and updated. Modifications to the framework can include changed thresholds to reflect new crop varieties, policy shifts, changing climate or other environmental conditions, or availability of newer and better datasets. For example, access to finer scale covariates or increased soil sampling intensities for DSM will allow finer scale assessments with added reliability. All data generated through this Assessment, including land suitability and soil attribute maps, and companion reliability maps, are publicly available from the CSIRO Data Access Portal (CSIRO Data Access Portal ). References Abrecht D and Bristow K (1996) Coping with soil and climatic hazards during crop establishment in the semi-arid tropics. Australian Journal of Experimental Agriculture 36(8), 971-983. DOI: https://doi.org/10.1071/EA9960971. Aldrick JM and Wilson PL (1990) Land Systems of the Southern Gulf Region, Northern Territory. Technical Report No. 42. CSIRO Division of Land Research, Palmerston. Allen RG, Pereira LS, Raes D and Smith M (1998) Crop evapotranspiration - Guidelines for computing crop water requirements. Food and Agriculture Organisation, United Nations, Rome. Andrews K and Burgess J (2021) Soil and Land Assessment of the Southern Part of Flying Fox Station for Irrigated Agriculture. Part B: Digital Soil Mapping and Crop Specific Land Suitability. Department of Environment, Parks and Water Security, Northern Territory Government, Darwin, N.T. Baker DE and Eldershaw VJ (1993) Interpreting soil analysis for agricultural land use in Queensland. Department of Primary Industries, Brisbane. Bartley R, Thomas M, Clifford D, Philip SR, Brough D, Harms B, Willis R, Gregory L, Glover M, Sugars M, Eyre L, Smith D, Hicks W and Petheram C (2013) Land suitability: technical methods. A technical report to the Australian Government from the CSIRO Flinders and Gilbert Agricultural Resource Assessment, part of the North Queensland Irrigated Agriculture Strategy. Water for a Healthy Country and Sustainable Agriculture Flagships, CSIRO Brisbane. Behrens T and Scholten T (2007) Digital soil mapping in Germany - a review (vol 169, pg 434, 2007). Journal of Plant Nutrition and Soil Science-Zeitschrift Fur Pflanzenernahrung Und Bodenkunde 170(1), 181-181. DOI: 10.1002/jpln.200790000. Biggs AJW, Kidston EM, Searle RD, Wilson PL, Slater BK, Grundy MJ, Heiner IJ, Smith DC and Seweeny C (2000) SALI - The Soil and Land Information system for Queensland. In: NZSSS/ASSSI Soil Conference 2000. NZSSS/ASSSI, 2. Breiman L (1996) Bagging predictors. Machine Learning 24(2), 123-140. DOI: 10.1023/a:1018054314350. Breiman L (2001) Random forests. Machine Learning 45(1), 5-32. DOI: 10.1023/a:1010933404324. Brooks AP, Shellberg JG, Knight J and Spencer J (2009) Alluvial gully erosion: an example from the Mitchell fluvial megafan, Queensland, Australia. Earth Surface Processes and Landforms 34(14), 1951-1969. DOI: 10.1002/esp.1883. Brungard CW and Johanson J (2015) The gate's locked! I can't get to the exact sampling spot! .... can I sample nearby? , Pedometrician. Bryant KB, Wilson PR, Biggs AJW, Brough DM and Burgess JW (2008) Soil Indicators of Queensland Wetlands: State-wide Assessment and Methodology. Brisbane. Budget Strategy and Outlook (2024) Budget 2024-25. A future made in Australia. Viewed 11 September 2024, https://budget.gov.au/content/factsheets/download/factsheet-fmia.pdf. Bui E (2007) A review of digital soil mapping in Australia. In: Lagacherie P, McBratney AB and Voltz M (eds), Digital Soil Mapping: An Introductory Perspective. Bui EN, Searle RD, Wilson PR, Philip SR, Thomas M, Brough D, Harms B, Hill JV, Holmes K, Smolinski HJ and Van Gool D (2020) Soil surveyor knowledge in digital soil mapping and assessment in Australia. Geoderma Regional 22, e00299. DOI: https://doi.org/10.1016/j.geodrs.2020.e00299. Bui EN, Simon D, Schoknecht N and Payne A (2007) ADEQUATE PRIOR SAMPLING IS EVERYTHING: LESSONS FROM THE ORD RIVER BASIN, AUSTRALIA. In: Lagacherie P, McBratney AB and Voltz M (eds), Digital Soil Mapping: An Introductory Perspective. Bureau of Meteorology (2023) Climate Data Online. BOM. Viewed 20/10/2023, . Burgess J, McGrath N, Andrews K and Wright A (2015) Agricultural Land Suitability Series, Report 1. Soil and Land Suitability Assessment for Irrigated Agriculture in the Larrimah Area, Sturt Plateau. Department of Land Resource Management, Darwin, NT. Burgess J, McGrath N, Andrews K, Wright A, Firby L, Carnavas M and andNorthern Territory. Department of Environment PaWS (2022) Soil and Land Suitability Assessment for Irrigated Agriculture, Northern Barkly Region, Northern Territory. Northern Territory Government, Palmerston NT. Available online: . Burrough PA, van Gaans PFM and Hootsmans R (1997) Continuous classification in soil survey: spatial correlation, confusion and boundaries. Geoderma 77(2), 115-135. DOI: https://doi.org/10.1016/S0016-7061(97)00018-9. Christian CS, Noakes LC, Perry RA, Slatyer RO, Stewart GA, Traves DM and Csiro (1954) Survey of the Barkly Region, Northern Territory and Queensland, 1947-48. Land Research Series No.3. Christian CS and Stewart GA (1953) General Report on Survey of Katherine-Darwin Region, 1946. CSIRO, Melbourne. Clifford D, Payne JE, Pringle MJ, Searle R and Butler N (2014) Pragmatic soil survey design using flexible Latin hypercube sampling. Computers & Geosciences 67, 62-68. Cohen J (1960) A COEFFICIENT OF AGREEMENT FOR NOMINAL SCALES. Educational and Psychological Measurement 20(1), 37-46. DOI: 10.1177/001316446002000104. Cresswell HP and Hamilton GJ (2002) Bulk density and pore space relations. In: McKenzie DC, Coughlan K and Cresswell HP (eds), Soil Physical Measurement and Interpretation for Land Evaluation. CSIRO Publishing, Collingwood. Dear SE, Ahern CR, O'Brien LE, McElnea AE, Dobos SK, Moore NG and Watling KM (2014) Queensland Acid Sulfate Soil Technical Manual: Soil Management Guideline. Department of Science, Information Technology, Innovation and the Arts, Queensland Government, Brisbane. Department of Environment Parks and Water Security (2000) Soil and Land Information System (SALInfo). Department of Environment PaWS, Darwin, (http://www.ntlis.nt.gov.au/metadata/export_data?type=html&metadata_id=2DBCB771214A06B6E040CD9B0F274EFE). Department of Natural Resources and Mines and Department of Science, Information Technology, Innovation and the Arts (2013) Regional Land Suitability Frameworks for Queensland. Queensland Government. Available online: . Department of State Development, Manufacturing, Infrastructure and Planning (2019) North west Queensland economic diversification strategy 2019. Viewed 12 September 2024, https://www.statedevelopment.qld.gov.au/regions/regional-priorities/a-strong-and- prosperous-north-west-queensland/north-west-queensland-economic-diversification- strategy Edgoose C (1987) The Land Resources of Mittebah Station (1996). Technical Memorandum 96/3. Alice Springs. Fältmarsch R (2006) Effect of toxic metals mobilised from Finnish acid sulphate soils on terrestrial and aquatic biota and human health: A literature review. 18th World Congress of Soil Science. Brisbane. Fanning DS and Fanning MCB (1989) Soil: morphology, genesis and classification. John Wiley & Sons Inc., New York, NY, USA. FAO (1976) A Framework for Land Evaluation. Food and Agriculture Organisation of the United Nations, Rome. Available online: . FAO (1985) Guidelines: Land evaluation for irrigated agriculture. Food and Agriculture Organization of the United Nations, Rome. FAO (2007) Land evaluation: Towards a revised framework. FAO, Rome. Fitzpatrick EA (1971) Pedology : a systematic approach to soil science. Edinburgh : Oliver & Boyd, Edinburgh. Fitzpatrick EA (1980) Soils: their formation, classification and distribution. Longman Group Ltd., London. Gallant J and Austin J (2012) Topographic Wetness Index derived from 1" SRTM DEM-H. 2. CSIRIO, Canberra. Gallant JC and Austin JM (2015) Derivation of terrain covariates for digital soil mapping in Australia. Soil Research 53(8), 895-906. DOI: 10.1071/sr14271. Grimes KG (1974) Mesozoic and Cainozoic geology of the Lawn Hill, Westmoreland, Mornington, and Cape Van Diemen 1:250 000 Sheet areas, Queensland. . Canberra. Available online: . Grundy MJ, Rossel RAV, Searle RD, Wilson PL, Chen C and Gregory LJ (2015) Soil and Landscape Grid of Australia. Soil Research 53(8), 835-844. DOI: 10.1071/sr15191. Grundy MJ, Searle R, Meier EA, Ringrose-Voase AJ, Kidd D, Orton TG, Triantafilis J, Philip S, Liddicoat C, Malone B, Thomas M, Gray J and Bennett JM (2020) Digital soil assessment delivers impact across scales in Australia and the Philippines. Geoderma Regional 22, e00314. DOI: https://doi.org/10.1016/j.geodrs.2020.e00314. Harms B, Brough D, Philip S, Bartley R, Clifford D, Thomas M, Willis R and Gregory L (2015) Digital soil assessment for regional agricultural land evaluation. Global Food Security 5(0), 25-36. DOI: http://dx.doi.org/10.1016/j.gfs.2015.04.001. Hartemink AE, Hempel J, Lagacherie P, McBratney A, McKenzie N, MacMillan RA, Minasny B, Montanarella L, de Mendonça Santos ML, Sanchez P, Walsh M and Zhang G-L (2010) GlobalSoilMap.net – A New Digital Soil Map of the World. In: Boettinger JL, Howell DW, Moore AC, Hartemink AE and Kienast-Brown S (eds), Digital Soil Mapping: Bridging Research, Environmental Application, and Operation. Springer Netherlands, Dordrecht. Hartemink AE, Krasilnikov P and Bockheim JG (2013) Soil maps of the world. Geoderma 207– 208(0), 256-267. DOI: http://dx.doi.org/10.1016/j.geoderma.2013.05.003. Harwood T (2019) 9s climatology for continental Australia 1976-2005: BIOCLIM variable suite. 1. CSIRO, Canberra. Hewitt AE, McKenzie NJ, Grundy MJ and Slater BK (2008) Qualitative survey. In: McKenzie NJ, Grundy MJ, Webster R and Ringrose-Voase AJ (eds), Guidelines for Surveying Soil and Land Resources. CSIRO Publishing, Melbourne. CSIRO Publishing, Collingwood. Hubble GD and Beckmann GG (1957) The Soils of some Western Queensland Properties of Australian Estates Co. Divisional Report. CSIRO, Division of Soils, Adelaide. Hudson BD (1992) The Soil Survey as Paradigm-based Science. Soil Sci. Soc. Am. J. 56(3), 836-841. DOI: 10.2136/sssaj1992.03615995005600030027x. Ingram JA (1972) 1:250,000 Geological Series - Explanatory Notes, Donors Hill, Queensland, Sheet SE/54-10 International Index. Canberra. Irvin S, Coman G, Musson D, Doshi A and Stokes C (2018) Aquaculture viability. A technical report to the Australian Government from the CSIRO Northern Australia Water Resource Assessment, part of the National Water Infrastructure Development Fund: Water Resource Assessments, Australia. CSIRO, Canberra. Isbell RF and National Committee on Soil and Terrain (2021) The Australian Soil Classification. CSIRO Publishing, Melbourne. Jacquier D, Wilson PL, Griffin E and Brough D (2012) Soil Information Transfer and Evaluation System (SITES)—Database design and exchange protocols, Version 2.0. CSIRO Land and Water, Canberra, A.C.T. Jeffrey SJ, Carter JO, Moodie KB and Beswick AR (2001) Using spatial interpolation to construct a comprehensive archive of Australian climate data. Environmental Modelling & Software 16(4), 309-330. DOI: https://doi.org/10.1016/S1364-8152(01)00008-1. Jenny H (1941) Factors of soil formation, a system of quantitative pedology. McGraw-Hill, New York. Karssies L, Jacquier DW, Wilson P and Ringrose-Voase A (2011) CSIRO National Soil Archive Manual. ACLEP, Canberra. Available online: . Kidd D, Malone BP, McBratney AB, Minasny B and Webb M (2015) Operational sampling challenges to digital soil mapping in Tasmania, Australia. Geoderma Regional 4, 1-10. DOI: http://dx.doi.org/10.1016/j.geodrs.2014.11.002. Kidd D, Searle R, Grundy M, McBratney A, Robinson N, O'Brien L, Zund P, Arrouays D, Thomas M, Padarian J, Jones E, Bennett JM, Minasny B, Holmes K, Malone BP, Liddicoat C, Meier EA, Stockmann U, Wilson P, Wilford J, Payne J, Ringrose-Voase A, Slater B, Odgers N, Gray J, van Gool D, Andrews K, Harms B, Stower L and Triantafilis J (2020) Operationalising digital soil mapping – Lessons from Australia. Geoderma Regional 23, e00335. DOI: https://doi.org/10.1016/j.geodrs.2020.e00335. Kidd DB, Malone BP, McBratney AB, Minasny B and Webb MA (2014) Digital mapping of a soil drainage index for irrigated enterprise suitability in Tasmania, Australia. Soil Research 52(2), 107-119. DOI: 10.1071/sr13100. Kuhn M (2015) caret: Classification and Regression Training. Ljung K, Maley F, Cook A and Weinstein P (2009) Acid sulfate soils and human health—A Millennium Ecosystem Assessment. Environment International 35(8), 1234-1242. DOI: http://dx.doi.org/10.1016/j.envint.2009.07.002. Lymburner L, Tan P, Mueller N, Thackway R, Thankappan M, Islam A, Lewis A, Randall L and Senarath U (2011) The National Dynamic Land Cover Dataset - Technical report. Geoscience Australia, Canberra. Mackenzie L, Heijnis H, Gadd P, Moss P and Shulmeister J (2017) Geochemical investigation of the South Wellesley Island wetlands: Insight into wetland development during the Holocene in tropical northern Australia. The Holocene 27(4), 566-578. DOI: 10.1177/0959683616670219. Malone BP, Kidd DB, Minasny B and McBratney AB (2015) Taking account of uncertainties in digital land suitability assessment. Peerj 3, 21. DOI: 10.7717/peerj.1366. McBratney AB, Santos MLM and Minasny B (2003) On digital soil mapping. Geoderma 117(1-2), 3- 52. DOI: 10.1016/s0016-7061(03)00223-4. McGrath N, Andrews K, Firby L, Walton S and Burgess J (2019) Soil and Land Suitability Assessment for Irrigated Agriculture on NT Portion 4478 and 4663 in the Larrimah Township area, Sturt Plateau. Department of Environment and Natural Resources, Northern Territory Government, Darwin, N.T. McKenzie NJ, Coughlan K and Cresswell H (2002) Soil physical measurement and interpretation for land evaluation. CSIRO Publishing, Collingwood, Australia. McKenzie NJ and Grundy MJ (2008) Approaches to land resource survey. In: McKenzie NJ, Grundy MJ, Webster R and Ringrose-Voase AJ (eds), Guidelines for Surveying Soil and Land Resources. CSIRO Publishing, Melbourne. CSIRO Publishing, Collingwood. McKenzie NJ, Grundy MJ, Webster R and Ringrose-Voase AJ (2008) Guidelines for surveying soil and land resources. CSIRO Publishing, Collingwood, Vic. McKenzie NJ and Ryan PJ (1999) Spatial prediction of soil properties using environmental correlation. Geoderma 89(1-2), 67-94. McKenzie NJ and Ryan PJ (2008) Measuring soil. In: McKenzie NJ, Grundy MJ, Webster R and Ringrose-Voase AJ (eds), Guidelines for Surveying Soil and Land Resources. CSIRO Publishing, Melbourne. Minasny B and McBratney AB (2006) A conditioned Latin hypercube method for sampling in the presence of ancillary information. Computers & Geosciences 32(9), 1378-1388. DOI: 10.1016/j.cageo.2005.12.009. Minty B, Franklin R, Milligan P, Richardson M and Wilford J (2009) The Radiometric Map of Australia. Exploration Geophysics 40(4), 325-333. DOI: 10.1071/eg09025. Minty BRS (1997) Fundamentals of airborne gamma-ray spectrometry. AGSO Journal of Australian Geology & Geophysics 17(2), 39 - 50. Moore G (2001) Soil Water. In: Moore G (ed.), Soilguide (Soil guide): a handbook for understanding and managing agricultural soils. Department of Agriculture and Food, Perth. National Committee on Soil and Terrain (2009) Australian Soil and Land Survey Field Handbook. CSIRO Publishing, Collingwood, Vic. National Native Title Tribunal (2022) Queensland Native Title Claimant Applications and Determination Areas, As per the Federal Court (30 September 2022). Neldner VJ, Niehus RE, Wilson BA, McDonald WJF, Ford AJ and Accad A (2021) The Vegetation of Queensland. Descriptions of Broad Vegetation Groups. Northern Territory Herbarium (2015) FloraNT Northern Territory Flora Online. Department of Land Resource Management. NT Government (2023) Territory Water Plan. A plan to deliver water security for all Territorians, now and into the future. Viewed 6 September 2024, https://watersecurity.nt.gov.au/__data/assets/pdf_file/0003/1247520/territory-water- plan.pdf. Pachepsky YA and Rawls WJ (2003) Soil structure and pedotransfer functions. European Journal of Soil Science 54(3), 443-452. DOI: https://doi.org/10.1046/j.1365-2389.2003.00485.x. Perry RA, Sleeman JR, Twidale CR, Prichard CE, Sleeman JR, Lazarides M and Collins FH (1964) General Report on the Lands of the Leichhardt - Gilbert Area, Queensland, 1953-54. Melbourne. Petheram C, Hughes D, Rustomji P, Smith K, Van Niel T and Yang A (2009) River modelling for Northern Australia. Water for a Healthy Country Flagship, CSIRO, Canberra. Petheram C, Watson I and Stone P (2013) Agricultural resource assessment for the Flinders catchment. A report to the Australian Government from the CSIRO Flinders and Gilbert Agricultural Resource Assessment, part of the North Queensland Irrigated Agriculture Strategy. CSIRO Water for a Healthy Country and Sustainable Agriculture flagships, Canberra. Available online: . Peverill KI, Sparrow LA and Reuter DJ (eds) (1999) Soil analysis: an interpretation manual. CSIRO Publishing, Collingwood, Victoria. Pillans B (1997) Soil development at snail's pace: evidence from a 6 Ma soil chronosequence on basalt in north Queensland, Australia. Geoderma 80(1), 117-128. QDNRME (2004) Gulf and Mitchell Agricultural Land and Water Resource Assessment Report. Brisbane. Queensland Herbarium (2007) Pre-clearing broad vegetation groups - Queensland. The State of Queensland, Brisbane. Queensland Government (2023) Queensland Water Strategy. Water. Our life resource. Viewed 11 September 2024, https://www.rdmw.qld.gov.au/qld-water-strategy/strategic-direction. R Core Team (2018) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Randal MA (1966) 1 : 250,000 Geological Series, Explanatory Notes, Rankin, N.T. Sheet SE/53-16. Canberra. Raupach M (1957) Investigations into the nature of soil pH. CSIRO, Melbourne. Rayment GE and Lyons DJ (2011) Soil Chemical Methods: Australasia. CSIRO Publishing, Melbourne. Reimann C, de Caritat P, Team GP and Team NP (2012) New soil composition data for Europe and Australia: Demonstrating comparability, identifying continental-scale processes and learning lessons for global geochemical mapping. Science of the Total Environment 416, 239-252. DOI: 10.1016/j.scitotenv.2011.11.019. Renard KG, Foster GR, Weesies GA and Porter JP (1991) RUSLE: Revised universal soil loss equation. Journal of soil and Water Conservation 46(1), 30-33. Roberts D, Wilford J and Ghattas O (2019) Exposed soil and mineral map of the Australian continent revealing the land at its barest. Nature Communications 10(1), 5297. DOI: 10.1038/s41467-019-13276-1. Rosewell CJ and Loch RJ (2002) Estimation of the RUSLE soil erodibility factor. In: McKenzie DC, Coughlan K and Cresswell HP (eds), Soil Physical Measurement and Interpretation for Land Evaluation. CSIRO Publishing, Collingwood. Rossiter DG (1996) A theoretical framework for land evaluation. Geoderma 72(3–4), 165-190. DOI: http://dx.doi.org/10.1016/0016-7061(96)00031-6. Roudier P (2022) Conditioned Latin Hypercube Sampling. Roudier P, Beaudette DE and Hewitt AE (2012) A conditioned Latin hypercube sampling algorithm incorporating operational constraints. In: Minasny B, Malone BP and Baudette AB (eds), Digital soil assessments and beyond. CRC Press, Taylor & Francis Group, Boca Raton. Searle R, McBratney A, Grundy M, Kidd D, Malone B, Arrouays D, Stockman U, Zund P, Wilson P, Wilford J, Van Gool D, Triantafilis J, Thomas M, Stower L, Slater B, Robinson N, Ringrose- Voase A, Padarian J, Payne J, Orton T, Odgers N, O’Brien L, Minasny B, Bennett JM, Liddicoat C, Jones E, Holmes K, Harms B, Gray J, Bui E and Andrews K (2021) Digital soil mapping and assessment for Australia and beyond: A propitious future. Geoderma Regional 24, e00359. DOI: https://doi.org/10.1016/j.geodrs.2021.e00359. Smith JW and Roberts HG (1972) Mount Drummond, Northern Territory: 1:250,000 Geological Series - Explanatory Notes. Record SE5312. Canberra. Stephen J (2005) Aridity Indexes. In: Oliver JE (ed.), Encyclopedia of World Climatology. Springer Netherlands, Dordrecht. Stewart GA (1954) Geomorphology of the Barkly Region. CSIRO Land Research Series 3, 42-58. Sullivan LA, Fitzpatrick RW, Bush RT, Burton ED, Shand P and Ward NJ (2010) The classification of acid sulfate soil materials: further modifications. Southern Cross University, Lismore, NSW, Australia. Taylor C (2002) Recognising channel and floodplain forms. Water Rivers Commission, Perth. Thomas M, Brough D, Bui EN, Harms B, Hill JV, Holmes K, Morrison D, Philip S, Searle R, Smolinski H, Tuomi S, Van Gool D, Watson I, Wilson P and Wilson P (2018a) Digital Soil Mapping A technical report from the CSIRO Northern Australia Water Resource Assessment to the Government of Australia. CSIRO, Canberra. Thomas M, Clifford D, Bartley R, Philip S, Brough D, Gregory L, Willis R and Glover M (2015) Putting regional digital soil mapping into practice in Tropical Northern Australia. Geoderma 241– 242, 145-157. DOI: http://dx.doi.org/10.1016/j.geoderma.2014.11.016. Thomas M, Gregory L, Harms B, Hill JV, Holmes K, Morrison D, Philip S, Searle R, Smolinski H, Van Gool D, Watson I, Wilson P and Wilson P (2018b) Land Suitability Analysis A technical report from the CSIRO Northern Australia Water Resource Assessment to the Government of Australia. CSIRO, Canberra. Thomas M, Philip S, Stockmann U, Wilson PR, Searle R, Hill J, Bui E, Gregory L, Watson I, Wilson PL and Gallant J (2022) Soils and Land Suitability for the Roper River Catchment, Northern Territory. A technical report from the CSIRO Roper River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Thomas M, Philip S, Stockmann U, Wilson PR, Searle R, Hill J, Gregory L, Watson I and Wilson PL (2024) Soils and land suitability for the Victoria catchment, Northern Territory. A technical report from the CSIRO Victoria River Water Resource Assessment for the National Water Grid. CSIRO, Australia. Thorburn PJ and Shaw RJ (1987) Effects of different dispersion and fine-fraction determination methods on the results of routine particle-size analysis. Australian Journal of Soil Research 25(4), 347-360. DOI: 10.1071/sr9870347. Twidale CF (1956) The physiography of northwest Queensland. Geographic Studies 3(1). Viscarra-Rossel RA, Chen C, Grundy MJ, Searle R, Clifford D and Campbell PH (2015) The Australian three-dimensional soil grid: Australia's contribution to the GlobalSoilMap project. Soil Research 53(8), 845-864. DOI: 10.1071/sr14366. Viscarra Rossel RA (2011) Fine-resolution multiscale mapping of clay minerals in Australian soils measured with near infrared spectra. Journal of Geophysical Research: Earth Surface 116(F4). DOI: https://doi.org/10.1029/2011JF001977. Webb AA, Beeston GR and Hall TJ (1974) The soils and vegetation of part of the Mayvale Land System in the Gulf of Carpentaria. Agric. Chem. Lab. Br. Tech. Rep(5). Whitehouse FW, Atherton DO and Johnston DG (1947) The gulf country between the Gregory and Gilbert Rivers. Report of the committee to recommend a site for an agricultural experiment station. Queensland. Wilford J (2012) A weathering intensity index for the Australian continent using airborne gamma- ray spectrometry and digital terrain analysis. Geoderma 183, 124-142. DOI: DOI 10.1016/j.geoderma.2010.12.022. Wilford JR (1995) Airborne gamma-ray spectrometry as a tool for assessing landscape activity and weathering development of regolith, including soils., AGSO Research Newsletter. Williams J (1983) Physical properties and water relations: Soil hydrology. In: Division of Soils C (ed.), Soils: an Australian viewpoint. CSIRO: Melbourne / Academic Press: London, Melbourne. Wilson P, Ringrose-Voase A, Jacquier D, Gregory L, Webb M, Wong M, Powell B, Brough D, Hill J and Lynch B (2009) Land and soil resources in Northern Australia. Northern Australia Land and Water Science Review. Northern Australia Land and Water Taskforce, Canberra. Wright M and Ziegler A (2015) ranger: A fast implementation of random forests for high dimensional data in C++ and R. arXiv preprint arXiv:1508.04409. Wright MN, Wager S and Probst P (2019) A fast implementation of random forests. R package version 0.11 2. Appendices Available water capacity estimation method This pedotransfer function for available water capacity (AWC) is based on work originally begun by Williams (1983) who collated field texture class and soil water retained (mm cm–1) for a wide range of soils from sands to self-mulching clays (Table 32.2 in Williams, 1983). Moore (2001) added samples from course-textured soils measured in the field in WA to this table of soil samples. Moore also accounted for: • presence of course fragments in the profile, reducing AWC proportionately • soil pedality (structured verses weakly structured or massive soils); structured soils have a higher AWC • differentiated between light sandy clay loam and sandy clay loam, increasing the range of field texture classes considered. Moore’s work has been applied to this Assessment to determine AWC for all soil profiles that have a morphological description. AWC is assessed for three depth ranges (0–0.6; 0–1.0; 0–1.5 m). Where profiles encounter rock or dense soil within these depth ranges, the corresponding AWC is calculated just to the root limiting layers, upper depth. Hence some profiles have the same AWC for two or three of the above depth ranges. AWC is determined as follows: 𝜃𝜃𝐴𝐴𝐴𝐴𝐴𝐴=Σ𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶𝑆𝑆𝑆𝑆𝑆𝑆 𝑖𝑖=0 𝑖𝑖𝑛𝑛 (3) where: θ AWC = available water capacity over a given soil depth (mm) i = specific interval in the soil profile (m) with a unique combination of texture, pedality and course fragments n = number of intervals d = specific interval thickness (m) CF = amount of course fragments in soil profile. Both ‘course fragments’ and ‘segregations’ are considered. 𝐶𝐶𝐶𝐶 =Σ(𝐶𝐶𝐶𝐶%+𝑆𝑆𝑆𝑆𝑆𝑆%)/100 (𝑤𝑤ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝐶𝐶𝐶𝐶 > 1,𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑡𝑡𝑡𝑡 1) (4) 𝐶𝐶𝐶𝐶%= ∫(𝐶𝐶𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ,𝑚𝑚𝑚𝑚𝑚𝑚 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝐶𝐶𝐶𝐶% 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟) (5) Apx Table A-1 Soil profile course fragment abundance class, equivalent percentage range and midpoint of percentage range For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au ∫(𝑆𝑆𝑆𝑆𝑆𝑆𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 ,𝑚𝑚𝑚𝑚𝑚𝑚 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 % 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟) (6) Apx Table A-2 Soil profile segregation abundance class, equivalent percentage range and midpoint of percentage range For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au 𝜃𝜃𝑆𝑆𝑆𝑆𝑆𝑆∫(𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐,𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞𝑞,𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝,𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝) (7) where: 𝜃𝜃𝑆𝑆𝑆𝑆𝑆𝑆 = soil water retained (mm m–3) Texture class as per Texture qualifier: ‘-’ (Light) or ‘+’ (Heavy) Pedality: • Pedal – all soils with a ‘moderate’ or ‘strong’ structure grade. • Apedal – all soils with a ‘weak’ structure grade or structure type ‘massive’ or ‘single grain’. Vertic properties: ‘Present’ or ‘Not Present’. Assessed using the ‘soil surface condition’ of the described soil profile. Profiles with a ‘self-mulching’ or ‘cracking’ surface condition are assumed to have vertic properties (Present) (National Committee on Soil and Terrain, 2009). For each combination of ‘Texture class’, ‘texture qualifier’, ‘pedality’ and ‘vertic properties’ a corresponding ‘soil water retained’ value (𝜃𝜃𝑆𝑆𝑆𝑆𝑆𝑆) has been collated in the lookup table created after Moore (2001). Apx Table A-3 Estimated soil water retained based on soil texture class, texture qualifier, pedality and vertic property For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au The pedotransfer function has been implemented in R code to enable rapid assessment of many profiles simultaneously. Supplementary data in digital soil mapping Collecting or creating extra data at locations that are not in a statistical-based sampling plan is a legitimate activity in digital soil mapping (DSM). These ‘extra data’ points are attribute specific and do not describe the soil fully, they are equivalent to Level A soil description sites (McKenzie and Ryan, 2008). As the activity was unable to collect a proportion of the DSM statistical sites and there exists a knowledge-base and conceptual understanding of the soil–landscapes, ‘extra data’ point-based observation datasets were created and used in the modelling of some of the soil attributes as described below. Permeability Soil permeability (saturated hydraulic conductivity of the least permeable layer in the soil) was observed across the Assessment area using a qualitative 4-class assessment of permeability that takes into account observations of soil structure, soil texture, soil porosity, cracks and shrink–swell properties (National Committee on Soil and Terrain, 2009). As this is a qualitative attribute (not measured) it was possible to make additional observations based on landscape knowledge of expected soils. Extra data points were created on the Karumba Plain and Wellesley Islands, specifically on the beach ridges that run parallel to the coastline. These distinctive features were located by satellite imagery and Google Earth. Sandy soils are expected on beach ridges and these soils were considered to be highly permeable (class 4). Drainage Soil drainage was observed across the Assessment area using a qualitative 6-class assessment of drainage that takes into account observations of soil permeability, soil water-holding capacity, soil colour, soil segregations, landscape position and slope and native vegetation indicator species (National Committee on Soil and Terrain, 2009). As this is a qualitative attribute (not measured) it was possible to make additional observations based on published vegetation information. Using Broad Vegetation Groups mapped in Queensland (Queensland Herbarium, 2007) (Apx Table B-1) and specific species extracted from the Northern Territory Vegetation Site Database (Northern Territory Herbarium, 2015) (Apx Table B-2), it was possible to estimate drainage. Extra sites were located using a geographic information system (GIS) and vegetation spatial data and given a drainage class 1 (Very poorly drained) or class 2 (Poorly drained) based on the description of the vegetation group (Neldner et al., 2021). Extra data was also created based on satellite imagery interpretation combined with geology (Apx Table B-3). Apx Table B-1 Classification of extra drainage sites based on vegetation mapping in Queensland (Queensland Herbarium, 2007) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table B-2 Vegetation species extracted from the NT Vegetation Site Database (Northern Territory Herbarium, 2015) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table B-3 Classification of extra drainage sites based on satellite imagery and geology For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Surface condition Soil surface conditions are well-known surface characteristics that affect plant establishment and growth. The condition is assessed in the field when soil is dry or interpolated to a dry state if the soil is wet based on one or more described conditions (National Committee on Soil and Terrain, 2009). Certain soils are expected to have certain conditions, for example cracking clay soils often have self-mulching surfaces or surface soils with a sandy texture can be expected to be loose or soft when dry. Because of these relationships, ‘extra data’ points have been estimated on the Karumba Plain where site observation density is low and certainty high of expected surface condition (Apx Table B-4). Apx Table B-4 Classification of extra soil condition sites based on satellite imagery For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Surface structure The structure of the surface soil horizon contributes to a number of limitations in the land suitability framework. The type of soil ped is assessed in the field and recorded based on standard structure descriptions (National Committee on Soil and Terrain, 2009). Certain soils are expected to have certain surface structures, for example sands nearly always have single grain structure, or if more organic matter, massive structure. Extra data points were made on the Karumba Plain on the sandy beach ridges that run parallel with the coastline, which were identified using Google imagery. Surface texture Surface soil texture contributes to a number of limitations in the land suitability framework. Surface texture is assessed in the field and recorded based on standard field texture descriptions (National Committee on Soil and Terrain, 2009). These field textures are simplified into just four categories for this Assessment (sandy, loamy, silty and clayey). Certain landforms consistently have certain soils and consequently certain field textures. For example, beach ridges consistently have sandy soils with field texture from sand to sandy loam, which matches one of our simplified categories, sandy. Extra data points were made on the Karumba Plain on beach ridges and tidal flats identified using Google imagery. The beach ridges are expected to have sandy soils and the tidal flats clayey surface soils. Microrelief Microrelief as applied here is small relief on the land surface of gilgai (natural mound and depressions). For this Assessment only gilgai depressions that are deeper than 0.3 m are considered a limitation to cropping. The limitation is categorised as either present or absent. Gilgai can be seen from the air and on satellite imagery. Gilgai is associated with cracking clay soils and as such is only expected to occur on the Armraynald Plain and Barkly Tableland. Both these physiographic units (PUs) have a low density of soil observations. To improve the accuracy of the mapping, extra data points were compiled across the Armraynald Plain and Barkly Tableland in a systematic way using Google satellite imagery and referencing existing data as training data, recording presence or absence of ‘Gilgai deeper than 0.3 m’. Extra data points were made along north–south transects spaced 10 km apart with a 5-km interval on the Barkly Tableland and 16 km apart with a 10-km interval on the Armraynald Plain. Gilgai can only be seen at a certain zoom level based on the GIS layout scale. To ensure consistent observations, the observation scale was set using actual field observed sites with gilgai >0.3 m. The observation scale on the GIS was consistent with a 1:10,000 mapping scale. Depth of A horizon The depth in metres of the A horizon was mapped using data from soil descriptions. The deepest occurrence of any horizons classified as A1, A2 or A3, according to Australian soil layer designation from National Committee on Soil and Terrain (2009), was taken as the depth of the A horizon. In rocky areas, A horizon and also soil depth is expected to be very shallow. Extra data points were made where it was obvious using Google imagery that rock outcropped. These points were randomly allocated a depth between 0.001 and 0.02 m. Soil thickness The depth in metres of soil thickness was mapped using data from soil site descriptions. The deepest occurrence of any horizons classified as A1, A2, A3, B1, B2 or B3 was taken as the thickness of the soil. In rocky areas, sites have generally not been described due to low agricultural potential with soil depth expected to be very shallow. Extra data points were made to augment the existing data where it was obvious using Google imagery that rock outcropped. These points were randomly allocated a depth between 0.001 and 0.02 m. Soil Generic Groups Soil Generic Groups (SGGs) are broadly defined groups of soils that assist the non-technical communication of soil and land resources and agricultural potential. The SGGs are aligned, where practical, to the Australian Soil Classification (ASC) (Isbell and National Committee on Soil and Terrain, 2021). Some PUs distinctly lacked data such as the Wellesley Islands and the predicted SGGs were not plausible. Because SGGs are broad soil groups and there are only 13 groups, it is possible to make a reasonable prediction of which SGGs occur and where to find them. Extra data points were made on the islands based on similar soil conditions on Donors Plateau, Google satellite imagery interpretation and published information (Mackenzie et al., 2017) (Apx Table B-5). Apx Table B-5 Classification of Soil Generic Groups based on satellite imagery and other information (Mackenzie et al., 2017) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Available water capacity Estimates of a soil’s water-holding capacity (AWC) were made using a pedotransfer function methodology developed by Moore (2001), which is based on soil texture and course fragments. AWC is dependent on soil depth and is estimated for the 0.5, 1.0 and 1.5 m intervals. Extra data points with estimates of AWC to 0.5, 1.0 and 1.5 m were made on the Karumba Plain based on similar soils near the town of Karumba. Rockiness Land is considered rocky or not, based on the percentage and size of course fragments and soil segregations or proportion of rock outcrop at the site. To augment the available data, rockiness observations were made of locations that met the criteria for rockiness while driving between field sites and when viewing Google satellite imagery. The known association between Acacia shirleyi (lancewood) and rocky soils and the available vegetation mapping in Queensland (Queensland Herbarium, 2007) and the Northern Territory Vegetation Site Database (Northern Territory Herbarium, 2015) was also used to make extra data points classed as rocky in the Assessment area. Surface salinity Soil that is affected by salt is a limitation to cropping and was observed during fieldwork. Only presence or absence of surface salinity is assessed. Observations of surface salinity were made during fieldwork and when using Google satellite imagery mainly on the Karumba Plain. Extra data points were also made using the vegetation mapping in Queensland (Queensland Herbarium, 2007) (Apx Table B-6). Apx Table B-6 Classification of extra salinity sites based on vegetation For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Land use combinations for crop groups and suitability analyses To enable ease of compilation, the land use combinations for the land suitability analysis are presented in a coded form in the land suitability rules in Appendix D. The expanded forms are presented below. The structure of the code is ‘crop group’ then underscore ‘season’ then underscore ‘irrigation type’ (e.g. land use combination code ‘CropGrp3_D_S’ is ‘Crop group 3 dry- season spray-irrigated’). The ‘crop’ list below is from the Roper River Water Resource Assessment (Thomas et al., 2022) and carried into this Assessment. For the full list of crops in the Assessment refer to the crop groups in Table 2-5. Apx Table C-1 Land use combinations for crop suitability analyses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Land suitability rules for land uses Climate – frost Low temperatures (<2 °C) affect frost-sensitive crops and reduce crop yields through damage to flowers and fruits. Generally, there are few frost- prone areas in northern Australia, but they are known in some inland areas, some higher elevated locations and may be localised along low-lying creeks and drainage lines. Apx Table D-1 Climate – frost – wet-season land uses not included For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Climate – heat stress Excessive heat damages crops impacting on seedlings, fruit, flowers and leaves. Parts of northern Australia are noted for exceptionally hot temperatures that occur over long periods. Apx Table D-2 Climate – heat stress, table 1 of 2 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-3 Climate – heat stress, table 2 of 2 CODE DESCRIPTION SUITABILITY SUBCLASSES FOR LAND USES G H I J K L Ch1 Low heat stress (<5 35 °C days) – Dry-season 1 1 1 1 1 1 Ch2 Moderate heat stress (5 to 50 35 °C days) – Dry- season 1 1 2 1 2 2 Ch3 Severe heat stress (≥50 35 °C days) – Dry-season 2 3 3 2 2 3 Ch4 Low heat stress (<5 40 °C days) – Wet-season 1 1 1 Ch5 Moderate heat stress (5–50 40 °C days) – Wet- season 1 1 2 Ch6 Severe heat stress (≥50 40 °C days) – Wet-season 2 3 3 CropGrp11_P_F CropGrp10_W_R CropGrp9_W_R CropGrp3_D_S CropGrp7_D_F CropGrp5_D_T CropGrp11_P_S CropGrp3_D_T CropGrp7_D_S CropGrp9_D_F CropGrp11_P_R CropGrp4_D_S CropGrp12_D_F CropGrp9_D_S CropGrp4_D_T CropGrp12_D_S CropGrp10_D_F CropGrp6_D_S CropGrp18_D_F CropGrp8_D_F CropGrp18_D_S CropGrp8_D_S CropGrp19_D_F CropGrp10_D_S CropGrp19_D_S CropGrp13_D_F CropGrp13_D_S Climate – annual rainfall – rainfed land uses only The amount of rainfall that falls during the growing season has a significant impact on the suitability for rainfed cropping (i.e. grown without supplementary irrigation). The suitability subclasses shown below identify the different rainfall zones and assume the soils have a high soil water storage capacity (i.e. available water capacity (AWC) >180 mm to 1.0 m soil thickness). Apx Table D-4 Climate – annual rainfall, rainfed land uses only For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Climate – temperature variation Northern Australia generally experiences warm daytime temperatures, but overnight minimums can drop regularly by 15 to 20 °C, particularly during the dry season in inland locations. While some crops (e.g. chickpeas and lychees) require cool temperatures for seed/fruit set, other crops (e.g. cassava) do not prefer such conditions. Apx Table D-5 Climate – temperature variation For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Gilgai microrelief – all land uses Severe gilgai microrelief affects machinery use and irrigation efficiency. Apx Table D-6 Gilgai microrelief – all land uses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Acid sulfate soil potential – all land uses Potential for soil sulfides to oxidise to sulfates (forming sulfuric acid) from site disturbance and soil drying. Apx Table D-7 Acid sulfate soil (ASS) potential – all land uses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Surface salinity – all land uses Seed establishment is hindered due to high levels of salt in the soil surface. Apx Table D-8 Surface salinity For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Irrigation efficiency – furrow and flood-irrigated land uses Soil infiltration characteristics need to deliver water evenly and efficiently down furrows and across paddocks to minimise water loss. Inefficiencies arise from high infiltration rates and waterlogging at upper end of furrows if furrows are too long. Apx Table D-9 Irrigation efficiency – furrow and flood-irrigated land uses For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Irrigation efficiency – high application method irrigated land uses (spray, trickle, mini-spray) Soil infiltration characteristics need to deliver water effectively from high application rate irrigation methods to wet up the soil profile. Rapid to moderately high infiltration is desirable as more water can enter the soil profile in a shorter period. Quick movement of irrigation infrastructure may also be required to cover large areas with repeat applications to top-up the root zone. Apx Table D-10 Irrigation efficiency – other high application method irrigated land uses (spray, trickle, mini-spray) and rainfed For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Soil water availability – irrigated land uses Available water capacity (AWC) estimates the capacity of a soil to store water for plant use (volumetric soil water between field capacity and wilting point). Subclasses relate to irrigation efficiency, that is the frequency of water applications required during the period of maximum water demand. Apx Table D-11 Soil water availability – irrigated land uses available water capacity (AWC) to 1.0 m, table 1 of 2 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-12 Soil water availability – irrigated land uses available water capacity (AWC) to 1.0 m, table 2 of 2 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-13 Soil water availability – irrigated land uses available water capacity (AWC) to 0.6 m (shallow-rooted crops) For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Nutrient balance Surface soil pH affects the availability of nutrients for plant use. Strong acidity or alkalinity may lead to certain nutrient deficiencies and/or toxicities. Apx Table D-14 Nutrient balance For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Physical restrictions – soil surface condition Soil surface condition can cause problems with a range of management activities, especially seedbed preparation, germination and crop establishment and the fruiting/harvesting of root crops. Apx Table D-15 Physical restrictions – soil surface condition For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Physical restrictions – surface infiltration Silty and surface sealing (hard-setting) soils have reduced infiltration of rainfall and irrigation water. Apx Table D-16 Physical restrictions – surface infiltration For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Physical restrictions – soil surface texture Factors relating to soil surface texture and the type of soil affect crop growth in a range of different ways, for example the recoverability (harvest difficulties) and condition of root crops, and the establishment of tree crops (vertic effects). Soils with a sodic subsoil and only a thin surface soil (A horizon) are difficult to manage for all cropping applications and also pose a significant land degradation hazard. Apx Table D-17 Physical restrictions – soil surface texture For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Rockiness Surface rockiness affects machinery and harvesting operations and reduces crop growth. Surface gravel, stone and rock outcrop can interfere significantly with planting, cultivation and harvesting machinery used for root crops, small crops, annual forage crops and sugarcane. Sites were assigned as being rocky or not based on the thresholds below, or where the combined total of any of the field observations had an abundance greater than 50% at the surface or in the top 0.1 m of soil: (i) rock outcrop or boulders >2%; (ii) cobbles or stones (60–600 mm) >20%; (iii) coarse gravel (20–60 mm) >50%; (iv) medium gravel (6–20 mm) >90%; and (v) hard segregations >50%. Apx Table D-18 Rockiness For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Soil thickness Soil thickness generally relates to the requirements for plants for physical support, in supporting plant root development and structural growth. Additional soil thickness is required to fulfil the requirements for certain crops (e.g. avocado, African mahogany). Additional soil thickness is required for efficient harvesting of root crops. Apx Table D-19 Soil thickness For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Water erosion Soil loss from water erosion needs to be minimised to reduce land degradation risk and productivity decline. Apx Table D-20 Water erosion For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Wetness Site and soil conditions that result in poor soil aeration. Excess water on the soil surface or in the soil profile caused from inadequate site drainage reduces crop growth and quality and restricts machinery use. Crops grown entirely in the dry season are less affected by this limitation as they will not generally experience very wet conditions. Apx Table D-21 Wetness, table 1 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-22 Wetness, table 2 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-23 Wetness, table 3 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Soil water availability – rainfed land uses Available water capacity (AWC) estimates the capacity of a soil to store water for plant use (volumetric soil water between field capacity and wilting point). For rainfed cropping, suitability subclasses are determined by a combination of annual rainfall and AWC to various depths. Three rainfall zones have been identified in the Assessment area. Apx Table D-24 Soil available water capacity (AWC) – rainfed land uses, table 1 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-25 Soil available water capacity (AWC) – rainfed land uses, table 2 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Apx Table D-26 Soil available water capacity (AWC) – rainfed land uses, table 3 of 3 For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Land suitability rules for aquaculture Apx Table E-1 Land suitability rules for aquaculture For more information on this figure, table or equation please contact CSIRO on enquiries@csiro.au Soil data sites used in digital soil mapping Plates show the contribution and location of site soil data collected in the activity (Assessment sites) and pre-existing site soil data for modelling each attribute. Supplementary input data (Appendix B) used to augment digital soil mapping (DSM) attribute modelling is not shown. Sites input to attribute models map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-509_3x3_DSM_ModelSites_10_8_Sheet1.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure F-1 Sites of available water capacity (AWC) estimations Sites input to attribute models map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-510_3x3_DSM_ModelSites_10_8_Sheet2.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure F-2 Sites of available water capacity (AWC) estimations Maps of land suitability options Full suite of land suitability maps for crop groups by season by irrigation type from the rules in Appendix C. The following land suitability maps do not consider economics and finances (e.g. subsidies and grants, commodity prices, fertilisers and fuel costs, etc.), land tenure, conservation area exclusions or factors such as flooding, secondary salinisation risk or availability of irrigable water. A quantitative assessment of the reliability of the suitability data, although not shown here, is available for each land use. All data including land suitability and the companion reliability maps are publicly available from the CSIRO Data Access Portal (CSIRO Data Access Portal ). Land suitability thumbnail map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-520_Suit_thumbnails_sheet0_20231009_300.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure G-1 Suitability of Crop group land use options 1 to 12 from Appendix A Land suitability thumbnail map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-521_Suit_thumbnails_sheet1_20231009_300.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure G-2 Suitability of Crop group land use options 13 to 24 from Appendix C Land suitability thumbnail map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-522_Suit_thumbnails_sheet2_20231009_300.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure G-3 Suitability of Crop group land use options 25 to 36 from Appendix C Land suitability thumbnail map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-523_Suit_thumbnails_sheet3_20231009_300.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure G-4 Suitability of Crop group land use options 37 to 48 from Appendix C Land suitability thumbnail map \\FS1-CBR.nexus.csiro.au\{lw-rowra}\work\3_Land_suitability\4_S_Gulf\1_GIS\1_Map_docs\1_Exports\LL-S-524_Suit_thumbnails_sheet4_20231009_300.png For more information on this figure please contact CSIRO on enquiries@csiro.au Apx Figure G-5 Suitability of Crop group land use options 49 to 58 from Appendix C As Australia’snational scienceagency andinnovation catalyst, CSIRO is solving the greatestchallenges through innovativescience and technology. CSIRO. Unlocking a better futurefor everyone. Contact us 1300 363 400+61 3 9545 2176csiroenquiries@csiro.aucsiro.au For further informationEnvironment Dr Chris Chilcott+61 8 8944 8422chris.chilcott@csiro.au Environment Dr Cuan Petheram+61 3 6237 5669cuan.petheram@csiro.au Agriculture andFood Dr Ian Watson+61 7 4753 8606 Ian.watson@csiro.au