Search

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.

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.

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.