Search

Digital agriculture is exciting attention because of an expectation that food systems will be disrupted by new digital technologies through improvements in precision, efficiency, volume, speed of p...

Bayesian networks were used to predict yield functions from three commercial oil palm estates. The networks were trained using a range of environmental, agronomic and management data routinely collected during plantation management. The Bayesian networks predicted fruit yield (FFB), average weight of fruit bunches (ABW) and average bunch number per hectare (BUNCH_HA). Comparing the predictions of most probable yield against observed data showed the Bayesian networks were highly accurate, with r2 values between 0.6 and 0.9. Predictions for attaining specific yield targets exceeded 75% accuracy for the FFB, 85% for the BUNCH_HA, and 90% for the ABW function. Supplementary analysis compared the precision of the Bayesian networks with artificial neural networks (ANNs), and demonstrated that the Bayesian networks gave equivalent or superior accuracy for every test. The utility of the networks were demonstrated by predicting the probability of achieving above average yield functions for each block across the three estates using a set of hypothetical rainfall and fertiliser input scenarios during the year prior to harvest. For the majority of blocks, the probability of exceeding the yield target depended on the level of fertiliser and rainfall inputs received, indicating that production from these blocks is greatly influenced by prior rainfall and fertilizer. However, some blocks in favourable areas showed a very high probability of exceeding the mean yields at all rainfall and fertiliser inputs, while a number of other blocks showed a consistently low probability of achieving the same productivity; production from these blocks will be resistant to the effects of historic rainfall and fertiliser inputs. The ability of Bayesian networks to represent future yield expectations will greatly assist managers under pressure to improve the economic and environmental sustainability of plantations. The demonstration that machine learning can extract important insight from complex datasets will have broad application in the analysis of big data collected from oil palm as well as other agricultural industries.

Selection of suitable urban stormwater harvesting sites and associated project planning are often complex due to spatial, temporal, economic, environmental and social factors, and related various other variables. This paper is aimed at developing a comprehensive methodology framework for evaluating of stormwater harvesting sites in urban areas using Multi Criteria Decision Analysis (MCDA). At the first phase, framework selects potential stormwater harvesting (SWH) sites using spatial characteristics in a GIS environment. In second phase, MCDA methodology is used for evaluating and ranking of SWH sites in multi-objective and multi-stakeholder environment. The paper briefly describes first phase of framework and focuses chiefly on the second phase of framework. The application of the methodology is also demonstrated over a case study comprising of the local government area, City of Melbourne (CoM), Australia for the benefit of wider water professionals engaged in this area. Nine performance measures (PMs) were identified to characterise the objectives and system performance related to the eight alternative SWH sites for the demonstration of the application of developed methodology. To reflect the stakeholder interests in the current study, four stakeholder participant groups were identified, namely, water authorities (WA), academics (AC), consultants (CS), and councils (CL). The decision analysis methodology broadly consisted of deriving PROMETHEE II rankings of eight alternative SWH sites in the CoM case study, under two distinct group decision making scenarios. The major innovation of this work is the development and application of comprehensive methodology framework that assists in the selection of potential sites for SWH, and facilitates the ranking in multi-objective and multi-stakeholder environment. It is expected that the proposed methodology will assist the water professionals and managers with better knowledge that will reduce the subjectivity in the selection and evaluation of SWH sites.

Narrowing the gap between actual and attainable yields in existing oil palm plantations is perceived as a key to fulfill the growing global demand for vegetable oil. To assess the scope for intensification we need robust estimates of attainable yields, which has been so far rarely done for perennial crops. For this purpose, we evaluated the complexities associated with estimating yield gaps (YGs) in oil palm (i.e. carry-over effect and aging), and adapted the existing framework for YG studies in annual crops. Based on this framework, we analyzed YGs for four sites within oil palm plantations located in Sabah (Malaysia), Central Kalimantan and North Sumatra (Indonesia) using a unique commercial yield data set covering an area of 38,300 ha. We assessed for each site at plantation scale water-limited potential yield using the PALMSIM simulation model, attainable yield determined by best performing blocks within the plantation as defined by 90th percentile of observed yields and actual yields (blocks representing the median yields). The water-limited potential yield did not differ very much; 35–39 t fresh fruit bunch (FFB) during the plateau phase, the most productive phase in the life time of a palm. This reflected the favorable environmental conditions found in many parts of Sumatra and Borneo for oil palm. Attainable yields were in the range of 26–31 t FFB/ha. The exploitable YG between attainable and actual yield ranged for the four sites from 5 to 7 t FFB/ha/year. For one site (Central Kalimantan), we assessed yield variability due to varying soil conditions at the block scale according to its dominant soil type. This suggested that they were indeed exploitable by management. If the plantation could close the gap between attainable and actual yield this could give about 21,000 t/yr higher FFB. This indicated the large scope for intensification oil palm offers in many parts of insular Southeast Asia.

Farmers learn from their own experiences. However, they are rarely sure if an exceptionally good or bad outcome is due to chance effects or whether it is due to a given combination of management practices and environmental conditions. We surmised that, if each harvest event is adequately characterized and a large number of these events are analyzed together, it should be possible to associate crop response to management within a particular set of growth conditions. We tested this hypothesis using the perennial crop, oil palm. The characterization of the harvest events can be divided into factors the grower can control (CFs), and non-controllable factors (NCFs). Expert opinion, coupled with literature reviews, indicated that the most important NCFs for oil palm in the humid tropics were surplus water and water deficits in the three-year period before and including the year of harvest. Water deficit was assessed using a simplified water balance model with inputs on inherent soil characteristics and rainfall. Surplus water was evaluated from the rainfall, inherent soil characteristics and the topographical position of the block in question. Homologous events (HEs) with similar NCFs were determined from weather and soil data for 141 blocks covering > 6000 ha over the period 2007–2013 on a commercial plantation. The yield of fresh fruit bunches (FFB) of 262 blocks over the period 2009–2013 was analyzed on the conceptual basis that if HEs can be defined in terms of growing conditions and used to account for part of the yield variation (NCFs), then the remaining variation within these events can be attributable to controllable factors (CFs) or management practices. Inclusion of HEs for the three years before the harvest year improved models used to explain yield variation. The variations in yield were in accordance with the expected effects of the distinct HEs confirming their validity as an analytical tool with normal conditions giving the highest yields, either deficit or surplus water giving intermediate yields, and a combination of both deficit and excess water the lowest yields. We chose the CF of fertilizer response to associate variation of management practices and yield within and across HEs. The overall response to fertilizer at 12.8 kg FFB·kg − 1 , without including HEs in the model, was much greater than that obtained when HEs were included (5.9 kg FFB·kg − 1 ). As most data were from blocks with sup-optimal growing conditions, the response to fertilizer over the ranges used was small and under the poorest HEs was not detectable.

This article present a Bayesian probabilistic method to support out-scaling of technologies from pilot projects. The method is applied to aerobic rice, a water-saving technology with probable global potential. The method assumes that areas similar to pilot sites are more likely to adopt than those that are different or unfavourable. Similarity is defined from climate, landscape and socio-economic attributes. Favourability is further evaluated by project specialists. Scaling out is not a simple linear process, so the method is proposed as a complement to learning processes. Results can support prioritization and strategic planning over specific geographic areas.

On-farm experimentation (OFE) is a farmer-centric process that can enhance the adoption of digital agriculture technologies and improve farm profitability and sustainability. Farmers work with consultants or researchers to design and implement experiments using their own machinery to test management practices at the field or farm scale. Analysis of data from OFE is challenging because of the large spatial variation influenced by spatial autocorrelation that is not due to the treatment being tested and is often much larger than treatment effects. In addition, the relationship between treatment and yield response may also vary spatially. We investigate the use of geographically weighted regression (GWR) for analysis of data from large on-farm experiments. GWR estimates local regressions, where data are weighted by distance from the site using a distance-decay kernel. It is a simple approach that can be easily explained to farmers and their agronomic advisors. We use simulated data to test the ability of GWR to separate yield variation due to treatment from any underlying spatial variation in yield that is not due to treatment, show that GWR kernel bandwidth can be based on experimental design to accurately separate the underlying spatial variability from treatment effects, and demonstrate a step-wise model selection approach to determine when the response to treatment is global across the experiment or locally varying. We demonstrate our recommended approach on two large-scale experiments conducted on farms in Western Australia to investigate grain yield response to potassium fertiliser. We discuss the implications of our results for routine practical application to OFE and conclude that GWR has potential for wide application in a semi-automated manner to analyse OFE data, improve farm decision-making, and enhance the adoption of digital technologies.

This paper quantifies the effect of three policy levels on the water and energy futures of Melbourne, Australia. During a time of severe water shortages attributed to climate change, water strategies lacked consideration of energy consequences. Modeling, guided by urban metabolism theory, demonstrated that a compact urban form, reduced water consumption by 90 GL/a, compared with a sprawling city, and had greater water conservation impact than simulated demand management measures. Household water conservation, coupled with increased use of solar hot water systems, reduced grid energy use by some 30 PJ/a. Desalination, tripled water supply energy demand, growing to a total of 4.5 PJ/a, by 2045. While the increase is less than 1% of total Melbourne urban energy use, it contributes to a substantial increase in the energy bill for urban water provision. Importantly, the energy impact could be offset through demand management measures. Recommendations for the combined management of water and energy include improving energy characterization of the urban water cycle; impact-evaluation of regional plans; using total urban water and energy balances in analysis to provide context; and developing reporting mechanisms and indicators to help improve baseline data across the water and energy systems.

There is the need to re-configure current urban water systems to achieve the objective of sustainable water sensitive cities. Stormwater represents a valuable alternative urban water source to reduce pressure on fresh water resources, and to mitigate the environmental impact of urban stormwater runoff. The selection of suitable urban stormwater harvesting sites is generally based on the judgement of water planners, who are faced with the challenge of considering multiple technical and socio-economic factors that influence the site suitability. To address this challenge, the present study developed a robust GIS based screening methodology for identifying potentially suitable stormwater harvesting sites in urban areas as a first pass for then more detailed investigation. The study initially evaluated suitability based on the match between harvestable runoff and demand through a concept of accumulated catchments. Drainage outlets of these accumulated catchments were considered as potential stormwater harvesting sites. These sites were screened and ranked under screening parameters namely demand, ratio of runoff to demand and weighted demand distance. The methodology described in this paper was successfully applied to a case study in Melbourne, Australia in collaboration with the local water utility. The methodology was found to be effective in supporting the selection of priority sites for stormwater harvesting schemes, as it provided the basis to identify, short-list and rank sites for further detailed investigation. The rapid identification of suitable sites for stormwater harvesting can assist planners in prioritising schemes in areas that will have the most impact on reducing potable water demand.

The impact of reforestation on water supplies is often considered in terms of impacts on water yields. In specific circumstances, reforestation will improve water quality, to the extent that previously unusable water can be utilised. Such is the case with salinisation, a process that threatens up to 17 million hectares of Australian farmland, major fresh water resources, biodiversity and built infrastructure. This paper highlights the value of bundling payments for environmental services (PES) from watershed restoration, including water quality improvement and carbon sequestration coupled with wood production, and compares the net returns with the existing agricultural land-use, using as an example the 408 000 ha Warren–Tone watershed (WT) in south-western Australia. The externalities of different land use systems are also taken into account. In this watershed 105 000 ha of the land was cleared for agriculture, with 25 000 ha subsequently reforested. A hydrological model (LUCICAT) was used to define the relationships between reforestation/deforestation and water yield and quality, thus providing a basis for valuing the hydrological benefits of reforestation. Various land-use change scenarios were examined, with modelling indicating more than 70% reforestation is required to restore stream salinity to a potable threshold of 500 mg/L total dissolved salts (TDS). Options that involve traditional agricultural land-uses or perennial pastures will not deliver potable water. A hydrological–land-use–economic (HLE) model was constructed, allowing the costs and benefits of different land-uses to be examined at watershed and sub-watershed scales. Reforestation was unprofitable when only wood revenues were considered with a discount rate of 9.5%, but was profitable at lower discount rates or with a discount rate of 9.5% and carbon prices of at least $22 t CO2-e. For the latter, reforestation activities driven by the economic benefits from selling wood and carbon, generate an externality benefit in the form of improved water quality without the need for any additional payment to landholders. Payments for activities that lead to improvements in water quality could represent a new, additional source of income for landholders on the proviso that there is sufficient reforestation to reach the potable threshold. Alternatively, costs could be imposed on those whose land-use practices cause the release of salt into waterways.

We describe the development of the algorithms that comprise the Spatial Decision Support System (SDSS) CaNaSTA (Crop Niche Selection in Tropical Agriculture). The system was designed to assist farmers and agricultural advisors in the tropics to make crop suitability decisions. These decisions are frequently made in highly diverse biophysical and socioeconomic environments and must often rely on sparse datasets. The field trial datasets that provide a knowledge base for SDSS such as this are characterised by ordinal response variables. Our approach has been to apply Bayes’ formula as a prediction model. This paper does not describe the entire CaNaSTA system, but rather concentrates on the algorithm of the central prediction model. The algorithm is tested using a simulated dataset to compare results with ordinal regression, and to test the stability of the model with increasingly sparse calibration data. For all but the richest input datasets it outperforms ordinal regression, as determined using Cohen’s weighted kappa. The model also performs well with sparse datasets. Whilst this is not as conclusive as testing with real world data, the results are encouraging.

Based on studies of water, poverty and livelihoods in nine river basins, this article reviews the role of institutions and organizations, focusing on their strengths and weaknesses, and to generalize as to why they fail to address the basin-wide issues of water, poverty and livelihoods. The authors show how a more comprehensive, integrated approach might change them to be more broadly relevant to basin-wide needs as well as to address the mismatch between development and the need to provide ecosystem services relevant to food, poverty, livelihoods and sustainable ecosystems.

SUMMARYGlobal food security is under threat by climate change, and the impacts fall disproportionately on resource-poor small producers. With the goal of making agricultural and food systems more climate-resilient, this paper presents an adaptation and mitigation framework. A road map for further agricultural research is proposed, based on the CGIAR Research Program on Climate Change, Agriculture and Food Security. We propose a holistic, integrated approach that takes into account trade-offs and feedbacks between interventions. We divide the agenda into four research areas, three tackling risk management, accelerated adaptation and emissions mitigation, and the fourth facilitating adoption of research outputs. After reviewing specific technical, agronomic and policy options for reducing climate change vulnerability, we acknowledge that science and good-faith recommendations do not necessarily translate into effective and timely actions. We therefore outline impediments to behavioural change and propose that future research overcomes these obstacles by linking the right institutions, instruments and scientific outputs. Food security research must go beyond its focus on production to also examine food access and utilization issues. Finally, we conclude that urgent action is needed despite the uncertainties, trade-offs and challenges.

SUMMARYThis paper seeks to establish the concept that the analysis of high temporal resolution meteorological data adds value to the investigation of the effect of climatic variability on the prevalence and severity of agricultural pests and diseases. Specifically we attempt to improve disease potential maps of root rots in common beans, based on a combination of inherent susceptibility and the risk of exposure to critical weather events. We achieve this using simulated datasets of daily rainfall to assess the probability of heavy rainfall events at particular times during the cropping season. We then validate these simulated events with observations from meteorological stations in East Africa. We also assess the utility of remotely sensed daily rainfall estimates in near real time for the purposes of updating the risks of these events over large areas and for providing warnings of potential disease outbreaks. We find that simulated rainfall data provide the means to assess risk over large areas, but there are too few datasets of observed rainfall to definitively validate the probabilities of heavy rainfall events generated using rainfall simulations such as those generated by MarkSim. We also find that selected satellite rainfall estimates are unable to predict observed rainfall events with any power, but data from a sufficiently dense network of rain gauges are not available in the region. Despite these problems we show that remotely sensed rainfall estimates may provide a more realistic assessment of rainfall over large areas where rainfall observations are not available, and alternative satellite estimates should be explored.

There is growing interest of international markets in differentiated agricultural products from the tropics. Coffee is a tropical crop of relatively high quality, whose value is increasing as consumer demand in developed countries for specialty coffee. Smallholders in emerging markets can benefit by capitalizing on the natural resource variability in their production system and from the knowledge that they have about this variability. The objective of this paper is to illustrate the benefits of systematically targeting management practices by coffee growers to improve attributes of their product. Data from case studies in Colombia and Mexico show statistically significant differences in beverage quality of coffees grown under different production conditions such as slope aspect, varieties, times of harvest, and shade levels. Possible intervention options can be selected by growers in terms of their ease of implementation, the likely improvement of quality that they achieve and the resource intensiveness they require. The conclusion is that optimum management is site specific so that it is not possible to make any blanket recommendations. Using continuous management cycles of implementation, observation, interpretation and evaluation the site specificity provides growers an opportunity to improve management over time to produce a higher quality product.

The authors analysed livelihood conditions in 10 river basins over three continents to identify generalizable links between water, agriculture and poverty. There were significant variations in hydrological conditions, livelihood strategies and institutions across basins, but also systematic patterns across levels of economic development. At all levels, access to water is influenced by local, regional or national institutions, while the importance of national versus local institutions and livelihood strategies vary with economic development. The cross-basin analysis suggests a framework for thinking about water–agriculture–poverty links that can inform future research and policy development.

Global population growth exerts stresses on river basins that provide food, water, energy and other ecosystem services. In some basins, evidence is emerging of failures to satisfy these demands. This paper assembles data from nine river basins in a framework that relates water and food systems to development. The framework provides a consistent basis for analysis of the water and food problem globally, while providing insight into specific conditions within basins. The authors find that successes occur when demand is met by increased productivity, while failure occurs when factors conspire to prevent development of land and water resources.

This paper presents a framework for the matching industry clusters to appropriate alternative water sources using an eco-industrial approach. Alternative water sources are increasingly being considered in new industrial developments to reduce demand on potable supply and improve sustainability performance. The suitability of alternative water sources for industrial development is influenced both by the demand profile of the industrial tenants and the limitations and opportunities presented by the local environment. This paper demonstrates an approach that integrates the principals of industrial ecology for considering the feasibility of alternative water services in a new industrial estate. The application of industrial ecology principles takes a site specific approach that considers the opportunities and limitations offered by the local environment and existing regional industries in moving towards sustainable water services. Industrial ecology facilitates opportunities for planning symbiosis between industries through by-product exchange, water cascading and cooperative infrastructure. The framework was applied to a new industrial development in Melbourne, Australia; with the assessment of alternative water sources part of a larger project that had the purpose of identifying regional by-product synergies based on materials, energy and water exchanges between businesses. The feasibility assessment detailed demonstrates that suitability of alternative water sources varies depending on the industry water demand profile. The paper demonstrates that the application of industrial ecology principles can be used to consider the feasibility of alternative water sources based on the regional context and the demand profile of industries locating to an eco-industrial park.

The struggle to find sustainable formal insurance for droughts in developing countries captures the attention of many in the development community for good reason. Droughts disrupt the development process, and the impacts of drought are exacerbated by an unwillingness on the part of the poor to invest combined with a lack of access to credit. This paper first examines the problems associated with traditional approaches to formal drought insurance and goes on to examine the potential of index insurance that is event-driven. A combined weather-generation and crop simulation modelling approach is used to estimate site-specific risks. The method is demonstrated in a case study for dry bean production in Honduras.

We compare water availability, water use, water productivity and poverty across the diverse river basins studied by the CGIAR Challenge Program on Water and Food. Water productivity tends to be higher in drier areas and where livestock grazing is integrated with rainfed crop production. We find that links among water, food security and poverty are best understood within a historical perspective. We identify opportunities to reduce poverty through water-related interventions. The way in which water-related investments affect poverty is influenced by changes in demography, climate, and rural society. In most basins, these trends involve trade-offs that require good governance at local, regional and basin scales.

Conflicting demands for food and water, exacerbated by increasing population, increase the risks of food insecurity, poverty and environmental damage in major river systems. Agriculture remains the predominant water user, but the linkage between water, agriculture and livelihoods is more complex than “water scarcity increases poverty”. The response of both agricultural and non-agricultural systems to increased pressure will affect livelihoods. Development will be constrained in closed basins if increased demand for irrigation deprives other users or if existing agricultural use constrains non-agricultural activities and in open basins if agriculture cannot feed an expanding or changing population or if the river system loses capacity due to degradation or over-exploitation.

Providing the water needed to produce food for more than 9 billion people by 2050 seems simple: agriculture must produce more food with less water. However, three complex issues are involved: First, water, food production and rural development do not have a simple correlation. Second, there are interactions between processes at local, basin and global scales. Third, change involves people in complex networks of institutions. The Challenge Program on Water and Food brings together agriculturalists, hydrologists and development specialists in a global-to-local programme that focuses on change through institutions. We believe that this scale, complexity and involvement are necessary to deliver plausible change.

All systems have causes and effects that can be appreciated at different spatial scales. Understanding and representing the complexity of multi-scale patterns in maps and spatial models are key research objectives. We describe the use of three types of correlation analyses: (1) a standard Pearson correlation coefficient, (2) a 'global' multi-scale correlation, and (3) local geographically weighted correlation. These methods were applied to topographic and vegetation indices in a small catchment in Honduras that is representative of the country's hillsides agro-ecosystem which suffers from severe environmental degradation due to land-use decisions that lead to deforestation, overgrazing, and unsustainable agricultural. If the geographical scale at which topography matters for land-use allocation can be determined, then integration of knowledge systems can be focused. Our preliminary results show that: (1) single-scale correlations do not adequately represent the relationship between NDVI and topographic indices; (2) peaks in the global multi-scale correlations in agricultural areas coincided with the median farm size, but there was no evidence of any community or larger-scale land-use planning or optimization; and (3) local multi-scale correlations varied considerably from the global results at all scales, and these variations have a strong spatial structure which may indicate local optimization of land use.

This article explores the question of how scientific information can improve local agronomic management using concepts of uncertainty classification and uncertainty management. Information and data on local management of soil fertility based on a local classification system of soil quality were collected from a small watershed in Cauca (Colombia). The analyses suggest that farmers hold local knowledge about soils at two levels. The first is based on empirical observations and refers to local knowledge about soils and landscape, which shows that the classes identified in the local soil quality classification are consistent with results obtained using measured soil parameters. At the second level, farmers have some awareness of ecological processes and the appropriate use of relationships between key soil characteristics and management options. It is argued that local knowledge is not sufficient to cope with uncertainty introduced by a rapidly changing agriculture, including, for example, increasing land pressure, unpredictable market forces and climate change. We have suggested how scientific knowledge can contribute to the solution, based on an analysis that relates Cohen's (Heuristic reasoning about uncertainty: an artificial intelligence approach. Pitman London, 1985) and Rowe's (Risk Analysis 14, 743–750, 1994) uncertainty concepts to local knowledge.

Part II analyses approach C experiments. Field-scale experiments were applied to four wheat fields in the Western Australian wheat belt. Different experimental designs were used two two-dimensional sine-waves, a chequerboard, and a two-factor strip arrangement. In each experiment, the yield associated with a particular treatment was predicted by kriging to where the other treatments were located. Different forms of kriging were investigated. Co-located cokriging, using the previous-season yield map as a covariate, was the most promising. The kriged data were then modelled with polynomial yield response functions. The outcome was a map for each field that described the optimum application of experimental input. The requirements varied continuously across the field, and could justify future site-specific crop management. The two-factor strip experiment was the most successful of those presented; the field on which it was used showed relatively strong responses to the applied inputs. The other sites were affected by lack of rain and/or design flaws. The underlying philosophy is sound, but the method proposed is time-consuming and inefficient. We hope that this paper can stimulate further research on the subject.

This is a two-part paper concerned with the role of field-scale experiments in site-specific crop management (SSCM). Part I is a general introduction to experimental design for SSCM, while Part II focuses on applied analysis. All references are listed at the end of Part II. In Part I, we list the goals of SSCM-experimentation, in addition to the classes of valid experimental designs. Three general approaches are proposed for choosing experimental designs: approach A is for experiments concerning management classes, while approaches B and C are alternatives for continuous management. Approaches A and B are comparative and can be analysed with an appropriately modified ANOVA.

EXPECTOR is a method of combining data and ‘expert’ knowledge within a Geographic Information System to provide information on the occurrence of spatially distributed attributes. It was developed to predict soil property values from spatially variable input data. Although initially developed to provide soil surveyors with a quantitative soil mapping method, it also has applications in land evaluation, land capability assessment, geological mapping and in precision agriculture. It operates on the basis that the state of a particular property, which may be difficult to measure directly, can be inferred from other (more measurable) entities and a knowledge of their inter-relationships. The method has been implemented as a stand-alone ‘Knowledge Editing’ module for the PC that can be linked to raster GIS packages. This paper describes the basis of the method and illustrates its use with an example describing the production of a surface clay content map for a small catchment in south-western Western Australia.

Decisions about fertiliser applications are fraught with uncertainty. Uncertainty about the outcome of an application of fertiliser is caused by unknown or uncontrolled variation about the condition of the soil to which it is applied, its fate, and the demand from the crop. Uncertainty can be eased by providing information which reduces ignorance about the likely outcomes of applying fertiliser, thereby increasing the decision‐maker's chances of success. Such information could include analyses of soil and plant tissue, prior information about crop performance, and predictions of climate and prices. This information could be used to improve the rate, location and timing of applications. Precision agriculture technology greatly enhances our ability to acquire and manage more of this information. However, information costs money, which must be traded‐off against the greater likelihood of success. This trade‐off is very difficult to evaluate, and in practice depends on a range of factors, including t...

Acknowledgements. Preface Asit K. Biswas Part 1: The Basin Focal Projects 1. Introduction to Part One 2. The Andes Basins: Biophysical and Developmental Diversity in A Climate of Change 3. The Indus and The Ganges: River Basins under Extreme Pressure 4. The Karkheh River Basin: The Food Basket of Iran Under Pressure 5. Vulnerable Populations, Unreliable Water and Low Water Productivity: A Role for Institutions in the Limpopo Basin 6. The Mekong: A Diverse Basin Facing the Tensions of Development 7. Water, Agriculture and Poverty in The Niger River Basin 8. The Nile Basin: Tapping the Unmet Agricultural Potential of Nile Waters 9. Farming Systems and Food Production in the Volta Basin 10. Yellow River Basin: Living with Scarcity Part 2: Cross-Basin Analysis and Synthesis 11. Water, Food and Poverty: Global - and Basin-Scale - Analysis 12. Water Availability and Use Across the Challenge Program on Water and Food (CPWF) Basins 13. Producing More Food with Less Water in a Changing World: Assessment of Water Productivity in 10 Major River Basins 14. The Resilience of Big River Basins 15. The Nature and Impact of Climate Change in the Challenge Program on Water and Food (CPWF) Basins 16. Connections between Poverty, Water and Agriculture: Evidence from 10 River Basins 17. Institutions and Organizations: The Key to Sustainable Management of Resources in River Basins

Conventional soil mapping is limited in its capabilities in that it presents a summary of the soil surveyor's conceptual view of soil variation. As such, the method conveys little regarding what is known about the variation of individual soil properties, or the quantitative nature of their variation. We developed a new method for soil mapping, based on the concepts employed in the PROSPECTOR mineral exploration system, which builds on existing soil surveyor knowledge to construct quantitative statements about individual soil properties via the development of a network of rules. These rules operate within a system of Bayesian inference to assign the varying probability of occurrence of a soil property of interest within an area, given evidence that relates to it in a known way. Permissible evidence includes the range of attributes normally used by a soil surveyor, such as landform, vegetation, land use, or parent material, and can also include remotely sensed digital data. Evidence is weighted according to the uncertainty associated with it, and combined to produce a single estimate of probability of a given attribute. The relationship between the evidence and prediction is stated explicitly at each stage of the procedure and is thus repeatable in a consistent manner. The system has the advantage that while it does not discard the evidence and knowledge used in conventional soil survey, it produces quantitative estimates of the distribution of soil properties, which can be used for a wide range of applications. The data produced is amenable to storage in geographic information systems and related data bases. As such, it can be updated or enhanced as new information or knowledge becomes available.

Global climate change (GCC) presents serious challenges to agricultural systems as they strive to meet increasing demand over the coming decades. Within these systems, food value chains are increasingly recognised as being vital for development, yet relatively little is known about the vulnerability of such chains to GCC, or their capacity to adapt. This paper provides a framework to examine how food value chains are affected by the uncertain impacts posed by GCC and what this implies for adaptation. While intervention models mostly deal with measures for producers, the authors hypothesise that adaptation strategies must include the entire chain to achieve the scale needed to tackle GCC. The authors propose that comprehensive situation assessment is necessary to examine both behaviour and assets—two key attributes for adaptive capacities. This framework examines three sets of attributes: (1) The general setup and nature of the supply chain; (2) rural livelihoods’ and food supply systems’ vulnerability to GCC, including downscaled crop suitability modelling to assess precise impacts of GCC, and (3) the behaviour of people and entities involved in value creation and the institutions mediating them. The framework provides decision-makers with a scale, crop and site independent road map to design and evaluate adaptation strategies to changing climate conditions. The authors test the framework in three case studies with differing supply chain characteristics and geographical contexts.

Creating detailed soil maps is an expensive and time consuming task that most developing nations cannot afford. In recent years, there has been a significant shift towards digital representation of soil maps and environmental variables and the associated activity of predictive soil mapping, where statistical analysis is used to create predictive models of soil properties. Predictive soil mapping requires less human intervention than traditional soil mapping techniques, and relies more on computers to create models that can predict variation of soil properties. This paper reports on a multi-disciplinary collaborative project applying advanced data-mining techniques to predictive soil modelling for Honduras. Gaussian process models are applied to map continuous soil variables of texture and pH in Honduras at a spatial resolution of 1,km, using 2472 sites with soil sample data and 32 terrain, climate, vegetation and geology related variables. Using split sample validation, 45% of variability in soil pH was explained, 17% in clay content and 24% in sand content. The principle variables that the models selected were climate related. Gaussian process models are shown to be powerful approaches to digital soil mapping, especially when multiple explanatory variables are available. The reported work leverages the knowledge of the soil science and computer science communities, and creates a model that contributes to the state of the art for predictive soil mapping.

The question has to be asked why – given the substantial advances in quantitative techniques over the years – ‘full’ Digital Soil Mapping has not been mainstreamed further and harnessed to the problems soil information can help address. This paper suggests some reasons for a slow adoption, causes for optimism for a wider adoption than at present and – using a case study – demonstrates the ease of further development at national scale. Finally, we propose how a major effort of digital soil mapping could support development in Africa, outlining the opportunities and obstacles that await contributors.

Recent concern over food prices has triggered a renewed interest in agricultural production systems. While attention is focused mainly on cropping, a complete analysis of food production systems should recognise the importance of livestock as major consumers of resources – in particular water – and as providers of food and other products and services. We propose that there is a need to examine not just food systems in isolation, but combined food and water systems, both of which are described as in a critical condition. From this broader perspective, it appears even more important to understand livestock systems because first, a total evaluation of agricultural water productivity – the gain from water consumed by agriculture – cannot be made without understanding the complexities of livestock-containing systems and; second, because in most tropical river basins, livestock systems are the major consumers of water. To identify total water productivity of livestock-containing systems, we describe concepts of agricultural water productivity and review the complexities of tracking the flow of water through livestock-containing systems: from inputs as evapotranspiration (ET) of forage and crops to outputs of valued animal products or services. For the second part, we present preliminary results from water use accounts analysis for several major river basins, which reveal that for Africa at least, livestock systems appear to be the major water consumers. Yet, little is known about the fate of water as it passes through these systems. We propose that livestock-containing systems offer substantial scope for increasing total water productivity and that there is considerable merit in improving the capacity to analyse water consumption and water productivity through such systems. Without removing this major source of uncertainty, the potential for systemic improvement to meet the world food and water crisis remains undefined and hence under-acknowledged.

The management of excess water (recharge or deep drainage) is one of many important issues required to address dryland salinity. To provide information on site-specific risk of deep drainage at a landholder scale, spatially variable deep drainage beneath an annual crop was estimated for a catchment of approximately 25 000 ha. The method coupled the results from one-dimensional crop simulation modelling (Agricultural Production Systems Simulator) with probabilistic soil mapping (Expector). Annual deep drainage over the catchment averaged 18 mm, representing 5% of rainfall. The annual deep drainage for the most probable soil types averaged 25 mm for acid loamy sand, 19 mm for yellow duplex, 14 mm for loamy sand, and 12 mm for red duplex. Groundwater modelling (Flowtube) highlighted that the groundwater table below the flat valley is not likely to be lowered with the introduction of perennial vegetation upslope. However, scenario spatial modelling indicated that replacing vegetation in areas of the most drainage-prone soil types, acid loamy sand (22% of catchment) and acid loamy sand plus yellow duplex (together at 48% of catchment area), was predicted to reduce annual average deep drainage by 38 and 60%, respectively. Mapping deep drainage by this method is of value to land managers who want to reduce deep drainage by locating priority areas for the establishment of perennial vegetation.

In this study, 9 Western Australian fields sown to dryland wheat were monitored using precision agricultural techniques, in order to understand implications for processing quality of the grain. Four fields received variable inputs of N fertiliser, and 6 received variable seed rates. In most cases, there was a very large variation within-field in grain yield, protein content, and protein quality, and such variation was related to variation in soil acidity, soil nitrate, and soil organic carbon, and specific weed and pest problems in some fields, as well as variation in the inputs. Grain protein content was positively correlated with soil nitrate levels in 6 of the 7 fields for which soil analyses were carried out. For several of the larger fields, separate harvesting of zones within the fields having differences in grain protein content would have enabled a greater proportion of the grain to be in a higher return quality grade. In 7 of the 9 fields, variation in protein quality (sodium dodecyl sulfate-sedimentation volume, SDS-SV) was greater than variation in protein content. The different measures of protein quality (SDS-SV, polymeric protein (glutenin) content, and glutenin molecular weight distribution) sometimes followed similar spatial trends, but in many cases did not. Therefore, total protein estimates are probably suitable measures for predicting within-field variation in protein quality. In none of the 9 fields was there overall a negative relationship between grain yield and protein content, although limited moisture availability can affect such relationships. The results suggest that the use of precision agricultural methods to manipulate inputs such as fertiliser, lime, or seed rates to increase yield does not have a negative effect on protein content. Farmers, therefore, can use precision agricultural methods along with other approaches to maximise wheat yield and grain protein content/quality at the same time.

The intensely weathered nature of Western Australian cropping soils and the long history of potassium depletion by the farming system has resulted in increased incidence of potassium deficiency in wheat. There is currently no scientifically based method for potassium recommendation in Western Australia. This paper describes the use of site-specific plot-scale field trials carried out in 1995–98 and a crop response model to develop a generally applicable potassium recommendation system. Geographic information system technology was used to extend the uniform potassium recommendation system into a system for mapping spatially variable potassium requirement that takes account of crop demand and soil available potassium. The field trials were carried out on a range of soil types and showed that wheat response to potassium can be described by the Mitscherlich equation. The size of the response was dependent on the soil test value for plant available potassium and the yield of the crop. The latter is mainly dependent on rainfall in the water-limited Mediterranean environment of Western Australia. The relationships between the maximum achievable yield, crop response and soil available potassium values were quantified in order to allow the decision support system to be developed for uniform whole-paddock fertiliser recommendation. Both soil available potassium and yield are very spatially variable in Western Australia and for wheat, the coefficient of variation of yield within the paddock is often of the order of 30%. Soil property variation can be of a similar order. Maps of soil available potassium values and of spatially variable target yield determined either from (i) farmer’s estimate, (ii) yield monitors and (iii) remotely sensed normalised difference vegetation index measurements allow this decision system to map spatially variable potassium requirement. Comparison of the map of potassium requirement with measured spatially variable response to potassium showed that the decision support system performed satisfactorily.

A field-scale experiment was conducted to determine the ability of a deterministic model developed for Western Australian wheat farmers to guide site-specific applications of nitrogen fertiliser. The results indicated that site-specific information of achievable yield improved the prediction accuracy much more than information about soil nitrogen — even though the latter was more costly to acquire. When applied together, these sources of information improved the prediction accuracy of the model markedly, explaining about half of the variation of yield response to nitrogen. However, the model failed to explain a substantial portion of site-specific variation, even with this intensity of information. This failure indicates the difficulty of representing complex biological systems with simple functional models.

Summary. Precision agriculture is the term given to crop management methods which recognise and manage within-paddock spatial and temporal variations in the soil–plant–atmosphere system. This paper reviews the principles, practice and perceived benefits of precision agriculture. The objective of precision agriculture is to improve the control of input variables such as fertiliser, seed, chemicals or water with respect to the desired outcomes of increased profitability, reduced environmental risk or better product quality. The practice can be viewed as comprising 4 stages: information acquisition; interpretation; evaluation; and control. Much of the technology to acquire information and control machinery is available or at a late stage of development. However, methods of interpretation are less well developed.

Variations of naturally emitted gamma radiation have been used in geological prospecting for over 20 years to detect anomalies associated with exploitable ore deposits. We examined their ability to detect spatial variation of soil material by comparing simultaneous ground and airborne measurements of gamma emissions with ground observations over a catchment in south-western Australia. Measurements were taken in the spectral windows for 40K, 238Uand 232Th. Variations of gamma radiation corresponded with the distribution of soil-forming materials over the landscape, and were used to distinguish between highly weathered residuum and fresh material from granitic outcrops. Gamma radiometric data also discriminated clearly between doleritic, lateritic and granitic soil parent materials. Airborne data indicated the distribution of these materials through the catchment, with the exception of dolerite dikes, which were too narrow to be detected using pixels greater than 20 m wide. It is concluded that gamma radiometric data can provide valuable insights into the spatial distribution of soil-forming materials but, given their limitations to provide direct information of pedological alteration, such data are likely to prove most valuable to soil survey when considered jointly with other information such as terrain models or aerial photography.