Improving the representation of agricultural management in land surface models.

摘要

To gain a better understanding of processes affecting crop yield, as well as two-way feedbacks between agricultural management and climate, a number of groups have recently incorporated croplands into regional and global land surface models. However, many aspects of agricultural management are still treated in a rudimentary way in these models. For my doctoral research, I have aimed to improve the representation of two key agricultural processes in land surface models: crop phenology and irrigation. In addition, I have investigated the effects of these processes on both crop yields and climate.;First, I assembled a dataset of global crop planting and harvesting dates for nineteen crops. I also investigated climatic and non-climatic factors that drive planting date decisions around the world. Second, I investigated trends and variability in crop planting dates and development progress across the U.S. I showed a trend to earlier planting of corn and soybeans, along with a trend to a longer crop growth period, and particularly a lengthening reproductive period in corn. In addition, I showed that growing degree days are a good predictor of the length of the vegetative period in corn, but less so for the reproductive period. Third, I used these observed trends along with the Agro-IBIS model to explore the implications of changes in crop phenology for both crop yields and fluxes of water and energy. I estimated that the trend to longer-season corn cultivars over the last three decades can account for 26% of the observed yield trend in the U.S. In addition, I found that earlier planting and longer-season cultivars shift the seasonality of water and energy fluxes, and have a small effect on annual-average fluxes. Finally, I investigated the effects of irrigation on climate, finding that this effect is significant in some large regions of the globe. Although the global-average temperature change was small, the large regional changes are important for both crop yields and the climate system.;The results of this dissertation will help to improve regional and global crop models, as well as improving our understanding of interactions between agricultural management, crop yields, and the climate system.