1  Thirty years of APSIM: from crop model to systems platform

Over three decades, the Agricultural Production Systems SIMulator (APSIM) has become one of the most influential modelling frameworks for understanding agricultural systems (Fig. 1). Its development reflects a growing need to move beyond single-trial analyses towards integrated, systems-based approaches that capture interactions among soils, climate, crops, genotypes and management practices (Fig. 2). As reviewed by Keating (2024) in this Virtual Issue, APSIM emerged from a culture of interdisciplinary collaboration involving agronomists, soil scientists, crop physiologists, software engineers, and modelers, united by the objective of improving both scientific understanding and practical decision-making in agriculture. This combination of scientific rigor and practical relevance has remained a defining characteristic throughout APSIM’s evolution.

Fig. 1
Fig. 1
Full size image

Schematic of APSIM. Figure adapted from Chenu et al. (2017).

Fig. 2
Fig. 2
Full size image

APSIM publications and citations A over time since 1996, B in different fields, and C in different countries. In A, number of publications and citations referred in Web of Science (WoS) for APSIM in May 2026. In B, estimates of the purpose of APSIM use based on papers with APSIM as a Topic in the WoS database for 2000, 2005, 2010, 2015 and 2020 years of publication. In C, country of use data derived from on WoS papers that identify APSIM as a Topic (based on a sample of 5 years of publication, namely 2000, 2005, 2010, 2015 and 2020, n = 270). Figures 2B and 2 C are from Keating (2024).

Early versions of APSIM focused primarily on simulating crop growth and production under varying environmental and management conditions (Keating et al. 2003). However, as agricultural systems became increasingly complex and new scientific questions emerged, the model expanded well beyond crop production. Advances in soil water and nutrient dynamics, residue management, crop physiology, genetics, crop rotations, livestock integration, economics, and environmental assessment progressively transformed APSIM into a broader agricultural systems platform. Improvements in computing capacity, software engineering, and data availability enabled increasingly sophisticated analyses spanning larger spatial scales and longer planning horizons (Holzworth et al. 2014; Keating 2024 in this Virtual Issue).

A key factor underpinning APSIM’s longevity has been its modular design philosophy. As highlighted by Pasley et al. (2022) in this Virtual Issue, successful crop models must balance biological realism, complexity and usability while remaining fit for purpose across diverse applications. APSIM was designed around these principles, allowing individual components representing soils, crops, management, and micro-environmental processes to be developed, tested and improved independently while remaining connected within a coherent modelling framework. This modularity has facilitated the continual incorporation of new scientific knowledge while preserving the flexibility required to address diverse agricultural questions.

The transition to APSIM next generation represented a major step in this evolution (Holzworth et al. 2018). By modernizing the software architecture and improving usability, APSIM next generation addressed challenges associated with maintaining increasingly complex legacy code. An important innovation has been the plant modelling framework (PMF), which enables physiological processes to be configured and extended without requiring extensive modification of source code (Brown et al. 2014). This capability has accelerated model development, facilitated testing of alternative process representations, and enhanced APSIM’s capacity to respond to evolving scientific priorities.

Today, APSIM is used globally by researchers, agronomists, farmers, plant breeders, consultants, industry, and government agencies to address challenges ranging from field-scale management decisions to national assessments of agricultural productivity, sustainability, and climate adaptation. The papers assembled in this Virtual Issue illustrate the breadth of these applications across Africa, Asia, Australia, and North America, and demonstrate how APSIM continues to evolve in response to new scientific opportunities and societal needs.

2 Applications across scales: prediction, decision support, and adaptation

One of APSIM’s key strengths is its ability to operate across multiple biological, spatial, and temporal scales. Applications range from physiological and genetic studies (e.g., Amin et al. 2025) to field-scale and farm management (e.g., Rodriguez et al. 2014; Verburg et al. 2022 in this Virtual Issue), regional production systems (e.g., Lawes et al. 2022 in this Virtual Issue), economic analyses (e.g., Thorburn et al. 2024 in this Virtual Issue; Muleke et al. 2022), insurance evaluations (e.g., Barratt et al. 2025), environmental assessments (Heidariask et al. 2026; e.g., Diancoumba et al. 2024 in this Virtual Issue), and policy analyses (e.g., Keating 2024 in this Virtual Issue). While APSIM remains widely used at the field scale (Fig. 2B; Robertson et al. 2015), its increasing integration with large datasets, remote sensing, and advanced computational approaches is enabling applications at unprecedented spatial scales (Lawes et al. 2022; Meier et al. 2023; Sadeh et al. 2024, all three in this Virtual Issue). This capability is increasingly important as agriculture confronts interconnected challenges associated with food security, resource constraints, environmental sustainability, and climate change.

The studies assembled in this Virtual Issue illustrate four major application domains that have emerged as particularly important for contemporary agricultural systems science: improving resource-use efficiency and farming systems performance; integrating process-based models with digital agriculture technologies; understanding genotype × environment × management (G×E×M) interactions; and evaluating climate adaptation and mitigation strategies.

2.1 Improving resource-use efficiency and farming systems performance

Improving the efficiency with which water, nutrients, and other resources are used remains a central challenge for sustainable agriculture. Several studies in this Virtual Issue showcase how APSIM can be used to understand the complex interactions that determine crop productivity and environmental outcomes, providing insights that are difficult to obtain through experimentation alone.

Nitrogen management provides a particularly important example. Thorburn et al. (2024) in this Virtual Issue demonstrated that targeted crop yield alone is often a poor predictor of nitrogen fertilizer requirements. Their simulation analysis showed that nitrogen demand emerges from interactions among cultivar characteristics, soil properties, management practices, and especially weather conditions. Their findings challenge conventional fertilizer recommendation approaches and highlight the value of systems modelling for improving nutrient management decisions. Similarly, in this Virtual Issue, Verburg et al. (2022) showed that the agronomic and environmental benefits of enhanced-efficiency fertilizers depend strongly on environmental and management conditions, helping explain why field responses are often inconsistent and identifying situations in which these technologies are most likely to improve productivity and reduce losses.

Nutrient management challenges are not limited to nitrogen. Das et al. (2023) in this Virtual Issue investigated phosphorus-use efficiency in subtropical dryland systems and found that the dominant drivers of phosphorus utilization differ between low- and high-rainfall environments. Their findings illustrate the value of systems modelling for disentangling complex soil–plant–atmosphere interactions and identifying opportunities for more targeted management.

Beyond the management of individual crops, APSIM has been widely applied to evaluate crop sequences and farming system design. Chaki et al. (2022) in this Virtual Issue showed that introducing mungbean into rice–wheat systems in the Eastern Gangetic Plains can simultaneously increase productivity and reduce environmental impacts when combined with optimized irrigation and fertilizer management.

Intercropping systems provide another example of innovation supported by APSIM. Zhang et al. (2022) in this Virtual Issue demonstrated that maize–soybean strip intercropping can improve land productivity in rainfed systems, with benefits becoming increasingly pronounced under recent, warmer climatic conditions.

Together, these studies demonstrate APSIM’s capacity to integrate biophysical processes within a systems framework and support management strategies that improve productivity, resource-use efficiency, and environmental sustainability simultaneously.

2.2 Digital agriculture: integrating models, data, and artificial intelligence

The expansion of APSIM applications from individual fields to regional and national scales reflects major advances in data availability, computational capacity, and digital agriculture technologies.

One of the principal constraints to large-scale crop modelling has long been the limited availability of reliable soil information. Recent advances in georeferenced soil databases are helping overcome this barrier. Mamassi et al. (2022) in this Virtual Issue, for example, demonstrated how Morocco’s Fertimap database can be used to derive APSIM soil parameters related to soil organic carbon, pH, and labile phosphorus, facilitating model deployment across broad geographic regions. The importance of improved soil information extends beyond model parameterization. Feng et al. (2022) in this Virtual Issue showed that soil organic matter and water-holding capacity strongly influence maize productivity under warming conditions, underscoring the importance of soil management for climate adaptation.

Another significant development has been the integration of APSIM with remote sensing technologies. Satellite observations provide continuous spatial information that complements process-based modelling by enabling the monitoring of crop growth, environmental conditions, and management activities such as sowing and harvesting across large areas (Sadeh et al. 2024 in this Virtual Issue). Sadeh et al. (2024) in this Virtual Issue demonstrated the potential of combining APSIM simulations with satellite imagery to predict both within-field yield variability (Fig. 3) and paddock-scale productivity across diverse Australian environments several months before harvest. Their framework achieved accurate predictions without site-specific calibration, highlighting the scalability and operational potential of this approach. Similarly, Lawes et al. (2022) in this Virtual Issue integrated APSIM, remote sensing and machine learning to evaluate crop rotation opportunities across more than 20 million hectares of the Western Australian wheatbelt (Fig. 4).

Fig. 3
Fig. 3
Full size image

Observed and predicted yield map of a wheat field near Birchip, Australia. In A, yield map produced by the combine harvester. In B, map of yield estimated by combining APSIM simulations and satellite imagery, following the methodology of Sadeh et al. (2024). In C, comparison of observed and estimated yield (RMSE = 525 kg/ha and R2 = 0.81). Figures from Sadeh et al. (2024).

Fig. 4
Fig. 4
Full size image

Map of the impact on wheat yield of adopting a wheat-canola instead of a wheat-cereal rotation across the Western Australian Wheatbelt in 2017 (left) and 2018 (right). Figure adapted from Lawes et al. (2022).

These studies reflect a broader transition in agricultural modelling, where process-based models are combined with remote sensing and artificial intelligence in complementary ways. APSIM provides mechanistic understanding, scenario-testing capability and the capacity to extrapolate beyond observed conditions, while remote sensing enables spatial monitoring and machine learning enhances predictive performance and pattern recognition. Together, these approaches are creating new opportunities for operational decision support, digital agronomy, large-scale forecasting, and climate-resilient agricultural planning across scales.

2.3 Harnessing genotype × environment × management interactions for crop adaptation and improvement

Understanding genotype × environment × management (G×E×M) interactions remains a central challenge in agricultural science. Crop performance emerges from complex interplays among genetic characteristics, environmental conditions, and management decisions, making integrated approaches essential for crop adaptation and improvement (Holzworth et al. 2014; Chenu et al. 2018).

APSIM provides a powerful platform for exploring these interactions by explicitly representing physiological processes while accounting for environmental variability and management interventions. Tirfessa et al. (2023) in this Virtual Issue illustrated the value of this approach by evaluating sorghum adaptation strategies across diverse Ethiopian environments and quantifying productivity–risk trade-offs associated with alternative management options.

Diancoumba et al. (2024) in this Virtual Issue used APSIM to characterize drought stress patterns experienced by sorghum crops across contrasting production environments in Mali and to investigate the adaptive value of photoperiod sensitivity. Their study generated insights relevant to both crop improvement and agronomic management, while demonstrating how crop models can help characterize target environments for breeding.

The potential of APSIM to support crop design is further illustrated in this Virtual Issue by Liu et al. (2022), who simulated thousands of virtual wheat genotypes to identify potential trait combinations suited to later sowing within rice–wheat systems in southern China, an adaptation strategy that is becoming increasingly important under climate change. Their study illustrates how process-based models can explore G×E×M interactions to identify potential ideotypes and trait combinations to improve adaptation for future climatic and management conditions.

Together, these studies illustrate the growing role of APSIM as a bridge between agronomy, physiology and breeding. By enabling exploration of G×E×M interactions across environments and management scenarios that would be impossible to evaluate experimentally, crop models are increasingly supporting predictive breeding, ideotype design, and the development of climate-resilient cropping systems (Chenu et al. 2017).

2.4 Climate change adaptation and mitigation

Climate change is altering the biophysical conditions under which agricultural systems operate, with rising temperatures, more frequent heat stress, changing rainfall patterns, and elevated atmospheric CO₂ increasingly affecting crop growth, resource use, and production risk (Wang et al. 2023; He et al. 2024). Process-based models such as APSIM are essential for evaluating adaptation strategies because they provide a mechanistic framework for exploring environments and management options that have not yet been observed experimentally (Chenu et al. 2017; Keating and Thorburn 2018; Keating 2024 in this Virtual Issue).

Several studies in this Virtual Issue illustrate how APSIM can be used to identify adaptation opportunities. Feng et al. (2022) showed that soil properties strongly influence crop responses to warming, while Elli and Archontoulis (2023) demonstrated how weather and management affect water-table dynamics and crop responses to excess moisture under present and future climate scenarios in the US Corn Belt. Together, these studies highlight the importance of considering soils, hydrology, and management when evaluating agricultural adaptation to climate change across the full spectrum of water-related stresses, from drought to waterlogging.

APSIM is also increasingly used to evaluate mitigation opportunities. Agriculture is both a contributor to greenhouse gas emissions and a potential source of climate solutions through improved resource-use efficiency, carbon sequestration, and changes in management that reduce emissions intensity. In this Virtual Issue, Meier et al. (2023), for example, coupled APSIM with machine learning to identify management practices capable of increasing soil organic carbon while reducing net greenhouse gas emissions across Australian grain-producing regions.

Collectively, the studies assembled in this Virtual Issue illustrate the value of APSIM as a platform for climate-smart agriculture, where productivity, resilience, and environmental performance are evaluated jointly rather than as separate objectives. By integrating knowledge across soils, crops, climate, and management, APSIM provides a framework for assessing trade-offs and synergies among adaptation, mitigation, and food production, while helping identify management strategies that are both agronomically effective and environmentally robust.

3 Challenges and future directions

The studies assembled in this Virtual Issue demonstrate both the maturity of APSIM and its continuing scientific evolution. They also highlight several important challenges that will shape the next phase of agricultural systems modelling.

First, improving the representation and parameterization of soils, genotypes, and environmental conditions remains essential. The increasing availability of georeferenced soil and climatic databases, alongside genomic information and phenotypic data from high-throughput phenotyping methods, is expanding opportunities for model application across scales. However, the dynamic nature of agricultural systems that APSIM seeks to represent, combined with uncertainty in soil properties and climatic inputs (e.g., variability in frost occurrence within fields or the prediction of extreme events under future climates), challenges in genotype parameterization, and the rapid turnover of cultivars in commercial production, continues to limit model accuracy and application in many contexts. Further improvements in data quality, accessibility, and standardization will be essential for reducing uncertainty and enhancing predictive performance, together with the development of robust and efficient methods for cultivar parameterization.

A second challenge concerns the representation of interacting stresses under changing climates. While APSIM has been widely used to evaluate climate adaptation strategies, confidence in future projections depends on the ability of models to represent increasingly complex interactions among heat stress, drought, waterlogging, nutrient limitations, and elevated atmospheric CO₂. Emerging stress combinations remain difficult to simulate accurately (Richetti et al. 2025; Severini et al. 2026), while model development and validation are constrained by the scarcity of experimental datasets that capture the range of conditions likely to occur in future environments (Gawinowski et al. 2025). Addressing these knowledge gaps will be critical for improving the robustness of climate impact assessments and adaptation analyses.

A third frontier lies in the integration of process-based modelling with rapidly expanding data streams. Remote sensing, sensor networks, environmental monitoring platforms, and high-throughput phenotyping systems are generating unprecedented volumes of information on crop performance and environmental conditions. As illustrated by several studies in this Virtual Issue (Lawes et al. 2022; Sadeh et al. 2024), APSIM increasingly serves as a mechanistic framework for supporting decision-making across a wide range of environments and management scenarios. Future developments will likely focus on improved data assimilation, near-real-time forecasting, and dynamic model updating, enabling more responsive and operational decision-support systems.

The rapid emergence of artificial intelligence and machine learning presents both opportunities and challenges. Machine learning methods can enhance parameter estimation, uncertainty quantification, and predictive performance, while process-based models provide biological realism, causal understanding, and the capacity to extrapolate beyond observed conditions. These approaches are complementary, and hybrid modelling frameworks that combine mechanistic understanding with data-driven learning are likely to become an important component of next-generation agricultural decision-support systems.

Finally, APSIM’s continued success will depend on its ability to integrate scientific advances while maintaining an appropriate balance between complexity, usability, and transparency. As emphasized by Pasley et al. (2022) in this Virtual Issue, models must remain fit for purpose while evolving in response to new scientific knowledge and user needs. The modular architecture of APSIM provides an important mechanism for managing this balance. Equally important will be the interdisciplinary collaboration among agronomists, soil scientists, crop physiologists, breeders, climatologists, software engineers, and data scientists that has characterized APSIM’s development over the past three decades (Keating 2024 in this Virtual Issue).

4 Conclusion

Over the past three decades, APSIM has evolved from a cropping systems simulation framework into a comprehensive platform for analyzing agricultural systems. Throughout this evolution, its development has been guided by a commitment to scientific rigor, practical relevance, and interdisciplinary collaboration.

The papers assembled in this Virtual Issue illustrate the breadth of APSIM applications across scales, from field-level management decisions to regional assessments of agricultural productivity, sustainability, and climate adaptation. They demonstrate how APSIM is increasingly being used to improve resource-use efficiency, evaluate farming systems, support crop improvement, assess climate adaptation, and mitigation strategies, and integrate emerging technologies such as remote sensing, machine learning, and artificial intelligence.

Collectively, these studies highlight a broader transformation occurring within agricultural systems science. Process-based models are no longer used solely to simulate crop growth; they are increasingly used to connect biological understanding with diverse sources of data and support decision-making across scales. In this emerging digital agriculture landscape, APSIM provides a mechanistic foundation to interpret observations and evaluate solutions to complex agricultural challenges.

As agriculture faces the challenge of producing more food, feed, fiber, and fuel while adapting to climate change and reducing environmental impacts, APSIM is well positioned to continue supporting innovation in research, policy, and practice. Its enduring strength lies not only in the scientific foundations that underpin the model, but also in its capacity to integrate knowledge across disciplines, scales, and technologies. The contributions presented in this Virtual Issue demonstrate that APSIM remains at the forefront of agricultural systems science and is well positioned to contribute to the next generation of productive, resilient, and environmentally sustainable agricultural systems.