Search

Agriculture is weather dependent, and changes in climate can have a drastic impact on our ability to feed, fuel, and clothe the world's population. Climate change is causing more frequent and unprecedented extreme weather events that are already negatively affecting agriculture. We need to assess the effects of extreme temperatures and rainfall on agriculture. Patterns of short‐term extreme weather events, such as elevated temperatures, flooding, and strong winds, are not predictable enough to design field experiments around. Process‐based crop and soil simulation models allow us to explore new management options and thus provide whole‐system‐based knowledge and management guides for different locations over variable climate conditions. By using crop simulation models, researchers can test different adaptation strategies and assess their effectiveness in reducing the impacts of climate change on agricultural production. In this paper, we discuss the development of crop models and how they have been used to assess the effects of a changing climate on agricultural productivity and propose methods for agriculture to adapt to those changes. We describe potential applications of crop models to assess regional issues such as irrigation demand, greenhouse gas emissions, and policy decisions. Better understanding of how weather and climate forecasts at various scales are provided and the reliability of these forecasts is important for using crop models as a planning tool. Different approaches for simulating long‐term climate change impacts on crop yield and seasonal yield forecasting are discussed. The use of ensemble models to better assess climate change impacts is also discussed. Core Ideas: Assessment of climate change impacts on agriculture requires a systems approach.Agriculture is weather dependent, and changes in climate can have a drastic impact on food production.Crop models can extend the results of physical experiments to better assess long‐term climate change effects on agriculture.Crop models are critical in assessing climate change impacts and developing adaptation and mitigation strategies. [ABSTRACT FROM AUTHOR]

• The APEX model is used in a wide range of locations to evaluate conservation practices on farm land. • Of the several approximations to calculate Curve Number (CN) method for estimating runoff, the Variable Daily CN with Soil Moisture Index provided the best simulations of water content. • The diagnostic APEX performance in the Optimizing Production Inputs for Economic and Environmental Enhancement (OPE3) site suggested reasonably accurate Evapotranspiration (ET), soil water calculations, and corn yields. • APEX performance in the OPE3 site varied by season (pre-plant, growing and post-plant). The Agricultural Policy/Environmental Extender (APEX) is a complex mechanistic simulation model designed to evaluate agricultural management strategies and their effects on yields, water and soil quality, and resource use. It is one of the main models used in the Conservation Effects Assessment Project (CEAP) assessment program and has been employed to evaluate the effects of conservation practices in the Chesapeake Bay Watershed (CBW). Calibration of model parameters is an important process in order to apply a simulation model to a particular area. The APEX model is normally parameterized for each application using selected datasets from the region. However, the APEX application area is usually very large and all areas may not be well represented by the calibration. Thus, additional tests of APEX without site-specific calibration for crop yield and soil water content can provide information on the stability of the parameters related to these variables in the CBW. The objective of this study is to evaluate the performance of APEX for corn yields and water balance in a new location using parameters previously developed for the CBW. This study used observations from the OPE3 site (Optimizing Production Inputs for Economic and Environmental Enhancement) located at the US Department of Agriculture, Henry A. Wallace Agricultural Research Center, Beltsville, Maryland, USA. Data from the OPE3 site consists of water content at various layers, corn yields, and ET covering the period from 1998 to 2013. The results for total water in the soil profile (up to 100 cm deep) showed that the simulated water amounts were about 4.4 cm higher than the observed values, but the trends were very similar to the observed trends. The statistical error analysis of R-square (RSQ), Nash-Sutcliffe (E NS) efficiency coefficient and Root Mean Square Error (RMSE) are 0.59, 0.51 and 0.21 cm•day−1, respectively. The average simulated yields were 0.3 t•ha−1 (4%) lower than observed yields with an average absolute deviation between APEX and OPE3 of 1.3 t•ha1 where the average observed yields over the period was 5.18 t•ha−1, The year-to-year trends were very similar to the measured trends. The monthly evapotranspiration data also showed reasonable trends between observed data and model results with the model RSQ, E NS and RMSE being 0.56, 0.41 and 4.49 mm•month−1, respectively. These results suggest that APEX parameters for yield and water use simulations calibrated for the CBW can be considered applicable to other areas of the CBW where calibration data are not available. [ABSTRACT FROM AUTHOR]

Accurate predictions of crop yield are critical for developing effective agricultural and food policies at the regional and global scales. We evaluated a machine-learning method, Random Forests (RF), for its ability to predict crop yield responses to climate and biophysical variables at global and regional scales in wheat, maize, and potato in comparison with multiple linear regressions (MLR) serving as a benchmark. We used crop yield data from various sources and regions for model training and testing: 1) gridded global wheat grain yield, 2) maize grain yield from US counties over thirty years, and 3) potato tuber and maize silage yield from the northeastern seaboard region. RF was found highly capable of predicting crop yields and outperformed MLR benchmarks in all performance statistics that were compared. For example, the root mean square errors (RMSE) ranged between 6 and 14% of the average observed yield with RF models in all test cases whereas these values ranged from 14% to 49% for MLR models. Our results show that RF is an effective and versatile machine-learning method for crop yield predictions at regional and global scales for its high accuracy and precision, ease of use, and utility in data analysis. RF may result in a loss of accuracy when predicting the extreme ends or responses beyond the boundaries of the training data. [ABSTRACT FROM AUTHOR]

The Eastern Seaboard region (ESR) of the United States is densely populated and depends on imported food. Agricultural systems are vulnerable to uncertainties such as environmental conditions, climate change, and transportation costs. Local populations could benefit from regional food systems as a way to provide security; however, the potential production capacity of the region would first need to be quantified. Potential production capacity for a specific crop, potato (Solanum tuberosum L.), was explored in two ways: expansion of the harvested land area and closing the yield gap between observed and potential yield. Potato production was assessed from Maine to Virginia for current land use (land under potato cultivation) and potential land use (other cropland). Simulations were based on two water availability scenarios: limited and nonlimited. A geospatial model implementing the explanatory model SPUDSIM estimated crop production (crop yield, water use, and N uptake) based on spatially variable input data (weather, soil, and management). Potato production was simulated in 35 potato-producing counties and in 346 counties with cropland. Under water-limited conditions, the response surface of production showed greater yield in the northern ESR states (median 28.24Mgha-1 ) than in the southern states (median 15.41 Mg ha-1 ). Resource requirements (water and N) and biophysical constraints (climate and soil) to production were also evaluated. In general, potato yield was negatively correlated with higher average seasonal temperatures and denser soil profiles. The results from this study will be valuable for regional policy planners to assess the capacity of the regional ESR food system. [ABSTRACT FROM AUTHOR]

Mechanistic crop models capable of representing realistic temperature responses of key physiological processes are necessary for enhancing our ability to forecast crop yields and develop adaptive cropping solutions for achieving food security in a changing climate. Leaf growth and phenology are critical components of crop growth and yield that are sensitive to climate impacts. We developed a novel modeling approach that incorporates a set of nonlinear functions to augment traditional thermal time methods (e.g., growing degree days) for simulating temperature responses of leaf expansion and phenology in maize or corn (Zea mays L.). The resulting leaf expansion and phenology models have been implemented into a new crop model, MAIZSIM, that simulates crop growth based on key physiological and physical processes including C4 photosynthesis, canopy radiative transfer, C partitioning, water relations, and N dynamics for a maize plant. Coupled with a two-dimensional soil process model, 2DSOIL, MAIZSIM was applied to simulate leaf growth, phenology, biomass partitioning, and overall growth of maize plants planted at two field sites on the Eastern Shore of Maryland and in Delaware for 3 yr of data. The model parameters were estimated using data from outdoor sunlit growth chambers and the literature. No calibration was performed using the field data. The MAIZSIM model simulated leaf area, leaf addition rate, leaf numbers, biomass partitioning and accumulation with reasonable accuracy. Our study provides a feasible method for integrating nonlinear temperature relationships into crop models that use traditional thermal time approaches without sacrificing their current structure for predicting the climate change impacts on crops. [ABSTRACT FROM AUTHOR]

Simulation models are valuable tools to evaluate soil processes, crop growth, and production responses under varied agroclimatic and input-management production conditions. In this study, an upgraded potato (Solanum tuberosum L.) crop growth simulation model (CSPotato) was integrated with a multi-year, multi-crop simulation model (CropSystVB). The integrated CropSystVB-CSPotato model facilitated prediction of soil processes, and growth and yields of different crops in a potato rotation system under center pivot irrigation. The integrated model was validated using two years (2001-2002) of field data on 'Ranger Russet' cultivar grown in a Quincy fine sand in the Pacific Northwest Columbia Basin production region under different nitrogen-management practices. This study showed good agreement between the measured and predicted yields as well as N uptake across different N management practices in both years. The predicted water as well as nitrogen drainage below the potato rooting depth (0.6 m) was greater in 2002 than that in 2001. This study demonstrated that the upgraded potato model integrated with CropSystVB can be used as a valuable decision tool to predict the crop yields, fate, and transport of N in irrigated potato rotation cropping systems. [ABSTRACT FROM AUTHOR]

Abstract: Crop models are used to predict yield and resource requirements as well as to evaluate different climate or management scenarios at a specific site. However, problems involving land use or global climate change encompass larger, more diverse, spatial scales and would benefit from simulating over broader areas using high-resolution, spatially-distributed data. A geospatial interface was developed to combine the explanatory potato crop model SPUDSIM with the geographic information system (GIS) software ArcGIS using the scripting language Python. Multiple geospatial input data layers were incorporated, including weather, soil, management, and land use. Modeling units (MUs) were defined as homogeneous field-scale areas created by the intersection of the input layers. Crop production was simulated for each unique combination of climate, soil, and management for MUs classified as cropland. The outputs (crop yield, water use, and nitrogen uptake) were mapped to show the spatial distribution within each county and aggregated to the county-level over the region of interest. An example was provided for potato production in Maine and illustrates how potential crop yield varies spatially over the state. The geospatial crop model showed evidence of both spatial and temporal variability of crop yield at the county level. The interface was designed to be flexible and easy to apply to applications such as evaluating crop production capacity and response under different scenarios. [Copyright &y& Elsevier]