Senthold Asseng

An overview of APSIM, a model designed for farming systems simulation

The Agricultural Production Systems Simulator (APSIM) is a modular modelling framework that has been developed by the Agricultural Production Systems Research Unit in Australia. APSIM was developed to simulate biophysical process in farming systems, in particular where there is interest in the economic and ecological outcomes of management practice in the face of climatic risk. The paper outlines APSIMs structure and provides details of the concepts behind the different plant, soil and management modules. These modules include a diverse range of crops, pastures and trees, soil processes including water balance, N and P transformations, soil pH, erosion and a full range of management controls. Reports of APSIM testing in a diverse range of systems and environments are summarised. An example of model performance in a long-term cropping systems trial is provided. APSIM has been used in a broad range of applications, including support for on-farm decision making, farming systems design for production or resource management objectives, assessment of the value of seasonal climate forecasting, analysis of supply chain issues in agribusiness activities, development of waste management guidelines, risk assessment for government policy making and as a guide to research and education activity. An extensive citation list for these model testing and application studies is provided.

Performance of the APSIM-wheat model in Western Australia

Abstract APSIM-wheat is a crop system simulation model, consisting of modules that incorporate aspects of soil water, nitrogen (N), residues, and crop development. The model was used to simulate above- and belowground growth, grain yield, water and N uptake, and soil water and soil N in wheat crops in Western Australia. Model outputs were compared with detailed field experiments from four rainfall zones, three soil types, and five wheat genotypes. The field experiments covered 10 seasons, with variations in sowing date, plant density, N fertiliser, deep ripping and irrigation. The overall APSIM model predictions of shoot growth, root depth, water and N uptake, soil water, soil N, drainage and nitrate leaching were found to be acceptable. Grain yields were well predicted with a coefficient of determination r2(1:1)=0.77, despite some underestimation during severe terminal droughts. Yields tended to be underestimated during terminal droughts due to insufficient pre-anthesis stored carbohydrates being remobilised to the grain. Simulation of grain protein, and depth to the perched water table showed limited accuracy when compared with field measurements. In particular, grain protein tended to be overpredicted at high protein levels and underpredicted at low levels. However, specific simulation studies to predict biomass, yield, drainage and nitrate leaching are now possible for wheat crops on the tested soil types and rainfall zones in Western Australia.

Multimodel ensembles of wheat growth: many models are better than one.

Crop models of crop growth are increasingly used to quantify the impact of global changes due to climate or crop management. Therefore, accuracy of simulation results is a major concern. Studies with ensembles of crop models can give valuable information about model accuracy and uncertainty, but such studies are difficult to organize and have only recently begun. We report on the largest ensemble study to date, of 27 wheat models tested in four contrasting locations for their accuracy in simulating multiple crop growth and yield variables. The relative error averaged over models was 24-38% for the different end-of-season variables including grain yield (GY) and grain protein concentration (GPC). There was little relation between error of a model for GY or GPC and error for in-season variables. Thus, most models did not arrive at accurate simulations of GY and GPC by accurately simulating preceding growth dynamics. Ensemble simulations, taking either the mean (e-mean) or median (e-median) of simulated values, gave better estimates than any individual model when all variables were considered. Compared to individual models, e-median ranked first in simulating measured GY and third in GPC. The error of e-mean and e-median declined with an increasing number of ensemble members, with little decrease beyond 10 models. We conclude that multimodel ensembles can be used to create new estimators with improved accuracy and consistency in simulating growth dynamics. We argue that these results are applicable to other crop species, and hypothesize that they apply more generally to ecological system models.

Analysis of water- and nitrogen-use efficiency of wheat in a Mediterranean climate

Water-use efficiency (WUE [g grain yield m−2 mm−1 ET]) and nitrogen-use efficiency (NUE [Δ g grain yield g−1 Napplied]) are important measures that can affect the productivity of crops in different environmental systems. However, measurement and interpretation of WUE and NUE in the field are often hampered by the high degree of complexity of these systems due to season-to-season variability in rainfall, the variation in crop responses to soil types and to agronomic management. To be able to guide agronomic practice, experimentally-derived measurements of WUE and NUE need to be extrapolated across time and space through appropriate modelling. To illustrate this approach, the Agricultural Production Systems Simulator (APSIM), which has been rigorously tested for wheat (Triticum aestivum L.) in a Mediterranean environment, was used to estimate and analyse the WUE and NUE of wheat crops in the Mediterranean-climatic region of the central Western Australian agricultural zone. The APSIM model was run for three locations (average annual rainfall of 461 mm [high rainfall zone], 386 mm [medium] and 310 mm [low]) and two soil types that had contrasting plant-available water-holding capacities in the rooting zone (sand: 55 mm, clay soil: 109 mm). Simulations were carried out with historical weather records (82–87 years) assuming current crop management and cultivars. The modelling analyses highlighted the inherently high degree of seasonal variability in yield, WUE and NUE of wheat, depending on soil type, N fertiliser input, rainfall amount and, in particular, rainfall distribution. The clay soil tended to be more productive in terms of grain yield, WUE and NUE in the high and medium rainfall zones, but less productive in most years in the low rainfall zone. The sandy soil was less productive in the high rainfall zone due to the high nitrate leaching potential of this soil type, but more productive than the clay in the low rainfall zone due to poorer pre-anthesis growth and less water use, less water loss by soil evaporation and relatively more water use in the post-anthesis phase. When a wheat crop was sown early on clay soil in the low rainfall zone, it yielded as high as in the other rainfall zones in seasons when rainfall was above average or there was a good store of water in the soil prior to sowing. The simulations confirmed findings from a limited number of field experiments and extended these findings both qualitatively and quantitatively across soil types, rainfall regions and crop management options. Furthermore, by using long-term historical weather records, the simulations extended the findings across the wide range of climatic scenarios experienced in mediterranean-climatic regions.

Analysis of the benefits to wheat yield from assimilates stored prior to grain filling in a range of environments

Grain yields of rainfed agriculture in Australia are often low and vary substantially from season to season. Assimilates stored prior to grain filling have been identified as important contributors to grain yield in such environments, but quantifying their benefit has been hampered by inadequate methods and large seasonal variability. APSIM-Nwheat is a crop system simulation model, consisting of modules that incorporate aspects of soil water, nitrogen (N), crop residues, crop growth and development. Model outputs were compared with detailed measurements of N fertilizer experiments on loamy soils at three locations in southern New South Wales, Australia. The field measurements allowed the routine for remobilization of assimilates stored prior to grain filling in the model to be tested for the first time and simulations showed close agreement with observed data. Analysing system components indicated that with increasing yield, both the observed and simulated absolute amount of remobilization generally increased while the relative contribution to grain yield decreased. The simulated relative contribution of assimilates stored prior to grain filling to grain yield also decreased with increasing availability of water after anthesis. The model, linked to long-term historical weather records was used to analyse yield benefits from assimilates stored prior to grain filling under rainfed conditions at a range of locations in the main wheat growing areas of Australia. Simulation results highlighted that in each of these locations assimilates stored prior to grain filling often contributed a significant proportion to grain yield. The simulated contribution of assimilates stored prior to grain filling to grain yield can amount to several tonnes per hectare, however, it varied substantially from 5–90% of grain yield depending on seasonal rainfall amount and distribution, N supply, crop growth and seasonal water use. High N application often reduced the proportion of water available after anthesis and decreased the relative contribution of remobilization to grain yield as long as grain yields increased, particularly on soils with greater water-holding capacity. Increasing the capacity or potential to accumulate pre-grain filling assimilates for later remobilization by 20% increased yields by a maximum of 12% in moderate seasons with terminal droughts, but had little effect in poor or very good seasons in which factors that affect the amount of carbohydrates stored rather than the storage capacity itself appeared to limit grain yield. These factors were, little growth due to water or N deficit in the weeks prior to and shortly after anthesis (when most of the assimilates accumulate for later remobilization), poor sink demand of grains due to low grain number as a result of little pre-anthesis growth or high photosynthetic rate during grain filling. Increasing the potential storage capacity for remobilization is expected to increase grain yield especially under conditions of terminal drought.

Potential deep drainage under wheat crops in a Mediterranean climate. I. Temporal and spatial variability

High rates of deep drainage (water loss below the root-zone) in Western Australia are contributing to groundwater recharge and secondary salinity. However, quantifying potential drainage through measurements is hampered by the high degree of complexity of these systems as a result of diverse soil types, a range of crops, different rainfall regions, and in particular the inherent season-to-season variability. Simulation models can provide the appropriate means to extrapolate across time and space. The Agricultural Production Systems Simulator (APSIM) was used to analyse deep drainage under wheat crops in the Mediterranean climate of the central Western Australian wheatbelt. In addition to rigorous model testing elsewhere, comparisons between simulated and observed soil water loss, evapotranspiration, and deep drainage for different soil types and seasons confirmed the reasonable performance of the APSIM model. The APSIM model was run with historical weather records (70–90 years) across 2 transects from the coast (high rainfall zone) to the eastern edge of the wheatbelt (low rainfall zone). Soils were classified as 5 major types: deep sand, deep loamy sand, acid loamy sand, shallow duplex (waterlogging), and clay soil (non-waterlogging). Simulations were carried out on these soil types with historical weather records, assuming current crop management and cultivars. Soil water profiles were reset each year to the lower limit of plant-available water, assuming maximum water use in the previous crop. Results stressed the high degree of seasonal variability of deep drainage ranging from 0 to 386 mm at Moora in the high rainfall region (461 mm/year average rainfall), from 0 to 296 mm at Wongan Hills in the medium rainfall region (386 mm/year average rainfall), and from 0 to 234 mm at Merredin in the low rainfall region (310 mm/year average rainfall). The largest amounts of drainage occurred in soils with lowest extractable water-holding capacities. Estimates of annual drainage varied with soil type and location. For example, average (s.d.) annual drainage at Moora, Wongan Hills, and Merredin was 134 (73), 90 (61), and 36 (43) mm on a sand, and 57 (64), 26 (43), and 4 (18) mm on a clay soil, respectively. These values are an order of magnitude higher than drainage reported elsewhere under native vegetation. When not resetting the soil each year, carry-over of water left behind in the soil reduced the water storage capacity in the subsequent year, increasing long-term average deep drainage, depending on soil type and rainfall region. The analyses revealed the extent of the excess water problem that currently threatens the sustainability of the wheat-based farming systems in Western Australia.

Productivity, sustainability, and rainfall-use efficiency in Australian rainfed Mediterranean agricultural systems

Mediterranean environments are characterised by hot, dry summers and cool, wet winters. The native vegetation in Mediterranean-climatic regions is predominantly perennial shrubs and trees intermixed with annual forbs. In south-western Australia, the spread of agriculture has seen the well adapted perennial vegetation replaced by rainfed annual crops and pastures. This has increased waterlogging and secondary salinity, thereby causing loss of productivity in ~10% of the cleared land area. To reduce deep drainage and make the agricultural systems environmentally sustainable requires the re-introduction of perennial vegetation in the form of belts of trees or shrubs, and phase-farming systems with perennials such as lucerne replacing annual pastures between the cropping years. To be economically viable, agricultural productivity needs to increase by at least 3% per annum. Yields of dryland wheat, the predominant crop in the Mediterranean agricultural regions of Australia, have increased at ~1%/year for the century preceding the 1980s and since then by nearly 4%/year. Increases have arisen from both genotypic and agronomic improvements. Genotypic increases have arisen from selection for earliness, early vigour, deep roots, osmotic adjustment, increased transpiration efficiency, improved disease resistance, and an improved harvest index from high ear weight (grain number) at flowering and high assimilate storage and remobilisation. Agronomic increases have arisen from early sowing that has been enabled by minimum tillage, increased fertiliser use, especially nitrogen, weed control, and rotations to improve weed control, minimise disease risk, and increase nitrogen availability. Evidence is presented suggesting that the rapid increase in yield of wheat in the last two decades has likely arisen from the rapid adoption of new technologies. For productivity to be maintained in the face of the increasing requirement to be environmentally sustainable will be a challenge and will require better integration of breeding and agronomy.

Putting mechanisms into crop production models

Crop growth models dynamically simulate processes of C, N and water balance on daily or hourly time-steps to predict crop growth and development and at season-end, final yield. Their ability to integrate effects of genetics, environment and crop management have led to applications ranging from understanding gene function to predicting potential impacts of climate change. The history of crop models is reviewed briefly, and their level of mechanistic detail for assimilation and respiration, ranging from hourly leaf-to-canopy assimilation to daily radiation-use efficiency is discussed. Crop models have improved steadily over the past 30-40 years, but much work remains. Improvements are needed for the prediction of transpiration response to elevated CO₂ and high temperature effects on phenology and reproductive fertility, and simulation of root growth and nutrient uptake under stressful edaphic conditions. Mechanistic improvements are needed to better connect crop growth to genetics and to soil fertility, soil waterlogging and pest damage. Because crop models integrate multiple processes and consider impacts of environment and management, they have excellent potential for linking research from genomics and allied disciplines to crop responses at the field scale, thus providing a valuable tool for deciphering genotype by environment by management effects.

Response of wheat growth, grain yield and water use to elevated CO2 under a Free-Air CO2 Enrichment (FACE) experiment and modelling in a semi-arid environment

Abstract The response of wheat crops to elevated CO 2 (eCO 2) was measured and modelled with the Australian Grains Free‐Air CO 2 Enrichment experiment, located at Horsham, Australia. Treatments included CO 2 by water, N and temperature. The location represents a semi‐arid environment with a seasonal VPD of around 0.5 kPa. Over 3 years, the observed mean biomass at anthesis and grain yield ranged from 4200 to 10 200 kg ha−1 and 1600 to 3900 kg ha−1, respectively, over various sowing times and irrigation regimes. The mean observed response to daytime eCO 2 (from 365 to 550 μmol mol−1 CO 2) was relatively consistent for biomass at stem elongation and at anthesis and LAI at anthesis and grain yield with 21%, 23%, 21% and 26%, respectively. Seasonal water use was decreased from 320 to 301 mm (P = 0.10) by eCO 2, increasing water use efficiency for biomass and yield, 36% and 31%, respectively. The performance of six models (APSIM‐Wheat, APSIM‐Nwheat, CAT‐Wheat, CROPSYST, OLEARY‐CONNOR and SALUS) in simulating crop responses to eCO 2 was similar and within or close to the experimental error for accumulated biomass, yield and water use response, despite some variations in early growth and LAI. The primary mechanism of biomass accumulation via radiation use efficiency (RUE) or transpiration efficiency (TE) was not critical to define the overall response to eCO 2. However, under irrigation, the effect of late sowing on response to eCO 2 to biomass accumulation at DC65 was substantial in the observed data (~40%), but the simulated response was smaller, ranging from 17% to 28%. Simulated response from all six models under no water or nitrogen stress showed similar response to eCO 2 under irrigation, but the differences compared to the dryland treatment were small. Further experimental work on the interactive effects of eCO 2, water and temperature is required to resolve these model discrepancies.

Impact of Spatial Soil and Climate Input Data Aggregation on Regional Yield Simulations

We show the error in water-limited yields simulated by crop models which is associated with spatially aggregated soil and climate input data. Crop simulations at large scales (regional, national, continental) frequently use input data of low resolution. Therefore, climate and soil data are often generated via averaging and sampling by area majority. This may bias simulated yields at large scales, varying largely across models. Thus, we evaluated the error associated with spatially aggregated soil and climate data for 14 crop models. Yields of winter wheat and silage maize were simulated under water-limited production conditions. We calculated this error from crop yields simulated at spatial resolutions from 1 to 100 km for the state of North Rhine-Westphalia, Germany. Most models showed yields biased by <15% when aggregating only soil data. The relative mean absolute error (rMAE) of most models using aggregated soil data was in the range or larger than the inter-annual or inter-model variability in yields. This error increased further when both climate and soil data were aggregated. Distinct error patterns indicate that the rMAE may be estimated from few soil variables. Illustrating the range of these aggregation effects across models, this study is a first step towards an ex-ante assessment of aggregation errors in large-scale simulations.