Modeling Gross Primary Production of Midwest Maize and Soybean Croplands with Satellite and Gridded Weather Data

The gross primary production (GPP) metric is useful in determining trends in the terrestrial carbon cycle. Models that determine GPP utilizing the light use efficiency (LUE) approach in conjunction with biophysical parameters that account for local weather conditions and crop specific factors are beneficial in that they combine the accuracy of the biophysical model with the versatility of the LUE model. One such model developed using in situ data was adapted to operate with remote sensing derived leaf area index (LAI) data and gridded weather datasets. The model, known as the Light Use Efficiency GPP Model (EGM), uses a four scalar approach to account for biophysical parameters including temperature, water stress, light quality, and phenology. The model was calibrated for four locations (seven fields) in the northern Midwest and was driven using remotely sensed LAI data and gridded weather data for these locations. Results showed reasonable error estimates (RMSE = 3.5 g C m-2 d-1). However, poor gridded weather atmospheric pressure and incoming solar radiation inputs, increased climatic variation in the study sites and contributed to higher RMSE that observed when the model was applied exclusively to in situ data from the Nebraska sites (2.6 g C m- 2 d- 1). Additionally, the application of LAI algorithms calibrated using solely Nebraska sites to sites in Iowa, Minnesota, and Illinois without verification of their accuracy potentially lead to increased error. Despite this, the study showed there is good correlation between measured and modeled GPP using this model for the field years under study. As the ultimate objective of research is to develop regional estimates of GPP, the decrease in model accuracy is somewhat offset by the model’s ability to function with gridded weather datasets and remotely sensed biophysical data. Advisor: Elizabeth A. Walter-She