Online

Mason, Rachel

Plant and Soil Science

2019

M.S.

This thesis aimed to model agricultural outcomes that are important to Vermont dairy farms and their surrounding communities -- runoff, erosion, nitrogen and phosphorus losses, crop yields, and timeliness of farm operations -- under a set of possible future climates. The Agricultural Policy/Environmental eXtender (APEX) model was used for this work, and the models were calibrated using data from a project that measured most of these outcomes on a set of local farms. The model setup and calibration methodology is thoroughly documented and may be a useful starting point for others who are new to agricultural modeling. Applied to two farms growing continuous corn, the future climate simulations showed that increasing temperatures by 2 C, combined with raising total precipitation or changing the seasonality of precipitation, had little effect on any outcome. Intense rainfall has increased greatly in recent decades, so a combination of higher temperatures and more intense precipitation was also simulated. This led to more runoff, more soil loss, and more nutrient losses. While median values were only modestly increased, the 95%-ile and total losses over the simulation period increased by a larger amount (as much as 53%, depending on the site). Management practices that can reduce runoff and soil/nutrient loss exist, but their effectiveness when a higher fraction of losses occur in large events is not well known. Crop yields changed by <10% in all simulations, and in some cases increased slightly. Other studies have warned of decreases in yields because of high summer temperatures and droughts. The pilot simulations in this thesis probed only a limited range of climate parameter space, so running the models for a wider range of scenarios may illuminate the circumstances in which particularly harmful and beneficial outcomes occur. Finally, APEX can in principle calculate the delays to corn planting that are expected if climate change leads to wetter conditions in the spring. However, the models consistently predicted that only harvest operations will be delayed. The reasons for this are not well understood, and it may be a useful avenue for future work. The present work is limited in a number of ways. Chief among these are somewhat mediocre model performance, and the narrow range of farming systems and climate scenarios investigated. Statistics describing the performance of the calibrated models were poorer than anticipated, and satisfactory results could not be obtained for some nutrient loss pathways. Only two farms were modeled, in just four hypothetical future climates; results for other relevant farming systems and climates may be quite different. Nonetheless, it is hoped that this thesis serves as a useful illustration of the potential and limitations of utilizing the APEX model in this context, and that it lays the groundwork for a more extensive investigation of agricultural outcomes under climate change in Vermont.