The Effect of Multiple Environmental Changes on Crop Model Response and Potential Improvements of Dynamical Land Surface Models

Abstract

Agriculture has become a dominant form of land cover through changes in land use, and is increasingly being considered an important part of land surface and general circulation models. The objective of this study is to analyze crop models' responses to multiple changes in environmental conditions and to provide sources for potential improvements of dynamical land surface models. We address these goals by evaluating a crop model's response to changes in observed climate and future projections from a regional climate model (RCM), and discerning the effect of multiple environmental changes on C3 and C4 plants.

After successful validation of the model CROPGRO (soybean), we modified prescribed variations in solar radiation (R), precipitation (P), temperature (T), in the observed climate for a field experiment's ambient and enhanced carbon dioxide (CO2) treatments. We found that the impact of changes in radiation and precipitation is affected by water stress, while temperature effects differ greatly for varying water-stress conditions and CO2 concentrations.

We then analyzed the model's responses to data from an RCM simulation for current and transient increase in atmospheric CO2 levels. Using model data and calculated anomalies, we found that higher temperatures had a negative impact on crops. We found that higher CO2 reduced the impact of water stress.

Finally, we investigated the effect of individual versus simultaneous changes in R, P, and T on plant response in a C3 (soybean) and a C4 (maize) plant. Using CROPGRO/SOYGRO and CERES-maize, we found that soybean and maize respond differently for R, P, and T and maize is more sensitive. The results also show that simultaneous changes in variables do not necessarily agree with individual changes.

Our findings suggest that there is good potential for using the crop models within dynamical land surface modeling systems for current and doubled CO2 scenarios. Further, our results indicate that additional considerations for ozone and process-level formulation that account for radiation changes should be added to the model.