Table 1.

Timeline of key events that shaped the development and use of agricultural system models.

Year	Event	Impacts
1940s–1950s	de Wit (1958) and van Bavel (1953) develop early computational analyses of plant and soil processes; Development of nutritional requirement tables for cattle (NRC, 1945)	Foundation established for the application of simulation and operations research optimization in plant-soil systems research and for modeling farm animal responses to nutrients
1950–1970s	Demand for policy analysis of rural development	Representative farm optimization models were developed and applied by Heady and students at Iowa State University, thus establishing use of linear programming methods for agricultural production
1960–1970	Pioneers in soil water balance modeling (WATBAL) [(Slatyer, 1960, Slatyer, 1964, Keig and McAlpine, 1969; Ritchie, 1972; McCown, 1973)]	Water balance models proved to be useful in the evaluation of climatic constraints to agricultural development. Foundations for linking soil and plant models established.
1964–1974	International Biological Program	Strong emphasis on large scale ecological and environmental studies led to development of grassland ecosystem models; laid foundation for ongoing work today
1965	UK releases nutrient requirement tables for ruminants (ARC, 1965, first work since the 50s)	Very influential publication; subsequent development of feeding systems models throughout Europe.
1965–70	Early crop modeling pioneers develop photosynthesis and growth models (C. T. de Wit, W. G. Duncan, R. Loomis)	Captured imagination of many crop and soil scientists. Prompted many to follow in their steps.
1969–75	S-69 Cotton Systems Analysis Project (Bowen et al., 1973, Stapleton et al., 1973, Jones et al., 1974, Jones et al., 1980, Baker et al., 1983)	Prompted development of several cotton models (W. G. Duncan, J. D. Hesketh, D. Baker, J. Jones, J. McKinion)
1971	Creation of the Biological System Simulation Group (BSSG)	Led to self-supported annual workshops aimed at advancing cropping system and other biological system models, continuing through 2014
1970s and early 80s	Development of early herd dynamics simulation models (Freer et al., 1970, IADB, 1975, Davis et al., 1976, ILCA, 1978, Sanders and Cartwright, 1979, Konandreas and Anderson, 1982)	Established in the developed world but some early examples in the developing world. Crucial for the advancement of whole livestock farm modeling and for representing disease and reproductive impacts
1970s	Gordon Conway develops concept of IPM in Malaysia. Huffaker Integrated Pest Management (IPM) Project begins in USA, evolves into the Consortium for IPM, ending in 1985. Global emphasis on reducing pesticide use, due to major increases in pesticide use globally and resistance in target pest populations.	Insect and disease models developed and used to help establish economic thresholds and to predict timing of threshold exceedance; some pest models were linked with crop models
Mid 1970s	Discovery of chaos in ecological systems by Robert May (May, 1976) and related advances in theoretical population ecology	Led to new approaches to modeling predator-prey, host-disease interactions
1972–74	Soviet Union purchase of US wheat reserves, causing major price spike (see Pinter et al., 2003)	US Government created LACIE, AGRISTARS projects to develop and use crop models with remote sensing to obtain strategic crop forecasts. Led to development of CERES-Wheat and CERES-Maize models (first published in 1986)
1974–1978	FAO development of Land Evaluation Framework in 1974 and an automated Agro-Ecological Zoning (AEZ) in 1978. (FAO, 1976, FAO, 1978–81)	Provided first methodology for land evaluation on a global basis, integrating soil, climate, vegetation, and socio-economic factors, leading to many applications and efforts to improve integrated assessment approaches
1975–1982	Early pioneers in computer simulation based decision support — SIROTAC and Australian Cotton Industry (CSIRO, 1980); S-107 Project on soybean modeling in the US	The Australian cotton modeling was the first major initiative to put crop and pest models in the hands of farmers for decision support. The soybean project in the US led to development of two major soybean models SOYGRO (Wilkerson et al., 1983) and GLYCIM (Acock et al.,1985).
1976	Launch of the first issue of Agricultural Systems, edited by C. R. W. Spedding (Spedding, 1976)	This journal helped legitimize agricultural system modeling, providing a place for scientists to publish their agricultural systems modeling and analyses as well as a collection of scholarly work in this area. This journal continues today with impact factor of about 2.5
1979	E.R. Orskov establishes the ‘Dacron bag technique’ for measuring the degradability of feed in the rumen (Orskov and McDonald, 1979)	Very influential method developed for characterizing the nutritional value of feeds, opening possibilities of new types of models; a new era of dynamic feed characterization started, leading to better animal models
1980	Soil and Water Resources Conservation Act analysis for 1980, mandate to develop a model to predict impacts of soil erosion on crop productivity	The comprehensive soil-cropping system model, (EPIC, the Environmental Policy Integrated Climate model), was developed to estimate soil productivity as affected by erosion
1980s	Growth of CGIAR Centers creates demand for assessment of economic returns to investments in agricultural research	Market surplus methods developed for estimating economic returns to investments, demonstrating high returns to agriculture research investment
1981–1984	Personal computer (PC) revolution led by IBM introduction of its Model 5150 personal computer and the first Apple Mac computer in 1984	These new PCs led to major increases in individual access to computer power; many agricultural models began appearing on PCs
1981	Development of the first soil nitrogen (N) model for predicting crop responses under both water and N limiting conditions (Seligman and van Keulen, 1981)	This model was the foundation for future soil N models in APSIM, DSSAT, and other suites of crop models
1980s through early 1990s	Development and growth of the Internet that began to connect computers globally	Ushered in new era of global communication and information technologies that has affected all areas of our lives, including agricultural system model development and use
1982 to 1986	CERES Models (Maize and Wheat) and GRO (SOYGRO and PNUTGRO) models were developed (Wilkerson et al., 1983, Boote et al., 1986)	The CERES models linked soil water, soil nitrogen and crop growth and yield together in a comprehensive fashion for the first time. They stimulated interest and activity in crop modeling in many parts of the world.
1980s	Development of duality theory and advances in nonlinear optimization via development of GAMS by World Bank	Led to advances in applications of econometric methods for production model estimation and to national and regional policy analysis models; use of new entropy methods reduced data requirements for the models
1980–1990	Influential developments in pasture modeling (Hurley pasture model — Johnson and Thornley, 1983 and the SAVANNA model (Coughenour et al., 1984)	Led to a proliferation of pasture models for intensive temperate and tropical grasslands and savanna systems. These models simulated herbage mass and accounted for sward components, which led to a more sophisticated representation of grazing processes.
1983–1993; DSSAT continuing today	USAID funded international IBSNAT project for facilitating technology transfer using systems approaches and crop and soil models	This led to the creation of the DSSAT suite of crop models that combined the CERES family of models with the SOYGRO and PNUTGRO models. The availability of the IBSNAT guidelines for data collection for crop modeling strengthened the crop model testing effort around the world.
1984 –continuing today	Dutch Government funding of the SARP (Systems Analysis of Rice Production) project at IRRI in the Philippines.	Development of a dynamic rice model that later was named ORYZA, which is still widely used today (Penning de Vries et al., 1991)
1985–1992	Earliest application of crop-soil systems models in a developing country “research for development” context — Kenya-Australia Dryland Farming Systems Project (McCown et al., 1992, Keating et al., 1991)	First PC used in agricultural research in Kenya running CERES Maize (influenced strongly by the IBSNAT minimum data set guidance) in 1985. Formed the foundation for modeling low input subsistence agricultural systems and exploring development opportunities. This experience went on to strongly influence the evolution of the APSIM farming systems simulator.
1986	Launch of the IGBP (International Geosphere-Biosphere Program) by the International Council for Science (ICSU)	Brought attention to the planet under pressure, including climate change, and helped coordinate research at regional and global scales on interactions of Earth's biological, chemical, physical, and human systems, including influence on ecosystem modeling
1970s–1980s	Development of optimization and econometric methods for application to production risks	Broadened analysis of production to include risk management behavior (see Anderson et al., 1977, Just and Pope, 1978, Antle, 1983, Antle, 1987)
1980s until now	Modeling herd replacement decisions with dynamic programming (Van Arendonk and Dijkhuizen, 1985)	As computer power increased, more complex applications attempting to optimize intensive and industrial livestock production occurred.
1990	Publication of the first Intergovernmental Panel on Climate Change (IPCC, 1990) Assessment Report	Led to first use of crop and economic models for climate change impact assessments on crops at field to global-scales (e.g., Curry et al., 1990, Rosenzweig and Parry, 1994); led to broad use of agricultural and ecological models that estimate GHG emissions and carbon dynamics and economic models for assessing impacts of climate change on agriculture
1990s until now	The era of livestock systems model integration (Herrero et al., 1996, Herrero et al., 1999, Freer and Donnelly, 1997)	Many soft ‘modular’ couplings of simulation models of individual animal performance, herd dynamics, pasture and crop models happened at this time.
1990–1994	First studies on global impacts of potential climate change on agricultural systems (Rosenzweig and Parry, 1994)	These were the first studies making broad use of crop and economic models for global impacts. These studies paved the way for many other national and global impact studies of climate change impacts and adaptation.
1991–continuing today	Australian governments develop a new APSRU group for modeling agricultural systems for practical uses	This led to the now widely used APSIM (McCown et al., 1996, Keating et al., 1991, Keating et al., 2003) suite of cropping system models which drew on early experience with CERES, EPIC and PERFECT models but re-engineered the “farming systems” foundations.
1992	Comprehensive, model-based scenario analysis funded by the European Union for policy decisions	Grounds for Choice published (Netherlands Scientific Council for Government Policy, 1992). Grounds for Choice.
1992	The Cornell Net Carbohydrate and Protein System is launched (Russell et al., 1992)	The CNCPS became the first commercially available dynamic model of digestion in ruminants. Its development influenced the current livestock performance models in many parts of the world.
1993–2011	International Consortium for Agricultural Systems Applications (ICASA), formed in 1993, ended in 2011	Helped crop modelers collaborate to develop standards for input data for crop models (Hunt et al., 1994), leading later to the ICASA data dictionary and data standards (White et al., 2013), now used in harmonizing model inputs in AgMIP project (White et al., 2013).
1998	Initiation of open source software movement, leading to more collaborative software development	Led to interest in providing open-source versions of widely-used crop simulation models; now being done by some ag system modelers (e.g., APSIM, DSSAT).
1999	The Livestock Revolution study (Delgado et al., 1999)	Key analysis explaining projected growth of livestock sector showing that ‘as people get richer and societies urbanize they consume more livestock’. Led to acknowledgement of need for increased understanding of livestock sector for agricultural development.
1980s–1990s	Interest in trade liberalization	Led to quantitative analysis of trade policies and development of national and global agricultural trade policy models.
1990s–2010s	The molecular genetics revolution: Genome sequencing technological advances and advances in understanding of the functions of crop and animal genes; ability to genotype new lines and breeds	Led to still evolving efforts by various public crop modeling groups and by seed companies to connect ecophysiological crop models for plant breeding and management purposes (e.g., see White and Hoogenboom, 1996, Hoogenboom and White, 2003, Hammer et al., 2006, Messina et al., 2006).
1990s–2000s	Sustainable agriculture movement; greater concern on environmental consequences of agriculture	Led to incorporation of biophysical processes into farm household, econometric and programming approaches; also led to development of “tradeoff analysis” approach; spatial data and tools increasingly used to develop spatially explicit biophysical and economic models
Late 1990s–2000s	Construction and release of global datasets of cropping areas, sowing dates and yields (Ramankutty and Foley, 1999, Ramankutty et al., 2008)	Allowed researchers to run simulations at finer resolution over greater model domains with more clearly documented assumptions and inputs.
2000s	Increasing interest in greenhouse gas (GHG) mitigation and the importance of ecosystem services	Led to models for analysis of mitigation of GHG in agriculture via soil C sequestration, afforestation, reduced livestock emissions; also led to linkages of economic models with crop, livestock, hydrology, and ecosystem models.
2001–2003	European Society Agronomy meeting hosts special session on modeling cropping systems. Published as Volume 18 European Journal Agronomy	This meeting led to a special issue of European Journal of Agronomy (vol 18) in which comprehensive papers on the major modeling systems, namely DSSAT, APSIM, CROPSYST, STICS, Wageningen models. Over 2000 citations for models in this publication.
2006	Representation of CO2 effects in crop model simulations challenged by Long et al. (2006)	Opened a debate between plant experimenters and modelers on the skill of crop models for yield prediction in future climates; prompted interest in more evaluations of CO2 effects interacting with temperature, other factors
2005–2009	European Union funding of the System for Environmental and Agricultural Modeling: Linking European Science and Society (SEAMLESS)	This led to major collaboration across Europe for developing models for use across scales, from field to farm, country, and EU levels.
2005–2010	Development of Earth system models, components of general circulation models (GCMs)	Led to new methods for coupling crop simulation models to land surface schemes of numerical climate models (Challinor et al., 2004).
2006	FAO Livestock's Long Shadow report (Steinfeld et al., 2006)	Demonstrated the large environmental footprint of livestock leading to programs for assessing and reducing the environmental impacts of livestock. Most of this work was done through modeling.
Mid 2005s onwards	Development of global livestock models (Bouwman et al., 2005, FAO, 2013, Herrero et al., 2013)	Global integrated assessment of livestock systems now possible at high resolution including land use, emissions, economics, biomass use and others (Havlik et al., 2014, Cohn et al., 2014, PBL Netherlands Environmental Assessment Agency, 2013, Bouwman et al., 2013 and others) and their links to other sectors (crops, forestry, energy, etc.).
2010	Creation of the Agricultural Model Intercomparison and Improvement Project (AgMIP), a global program and community of agricultural scientists	This initiative led to model comparisons and initiatives for improving models, capturing the imagination and interest of agricultural modelers worldwide (Rosenzweig et al., 2013a, Rosenzweig et al., 2013b, Asseng et al., 2013).
2010s	Increasing interests by the private sector in agricultural system models	Some companies create their own crop modeling teams, others start working in public-private collaborations.
2010s	With the food price shock of 2008/2010, a realization of the need to increase food production to meet needs of 10 billion by 2050, including challenges of climate change and sustainable natural resources	This realization is leading to greater interest in use of new ICT developments (e.g., cloud computing, smart phones, app stores, mobile computing, use of UAVs for agricultural management) and agricultural system models to help guide investments and development and to greater interest by the private sector.