Calibrating The Apex Model For Simulations Of Environmental And Agronomic Outcomes On Dairy Farms In The Northeast Us: A Step-By-Step Example

This article documents the procedure used to calibrate the Agricultural Policy/ Environmental eXtender (APEX) model for a project that aimed to simulate crop yields, runoff erosion, and nutrient losses on dairy farms under climate change. We describe each step in the calibration process, presenting the model outputs, plots, and metrics that were found to be useful, and discussing the decisions that needed to be made along the way. Calibration improved the performance of the model relative to an uncalibrated "baseline" version, and 0.22 < NSE < 0.49 and PBIAS within +/- 12% were achieved for most of the model outputs. While the mean annual silage yield was correct (PBIAS = 2%), the model did not accurately capture year-to-year yield variations (NSE = -1.6), and results for nitrogen in runoff were poor (NSE = -0.04, PBIAS = -69%). We therefore also outline several ways in which the method could be improved. Other calibration methods exist, and the procedure presented here will not be applicable in all situations. However, fully documented APEX calibrations are rare in the literature, and the number of non-expert model users may be growing. We therefore anticipate that this paper will serve as a useful point of entry for those who are new to APEX. We also hope that this work contributes to the development of transparent and reproducible procedures for modeling studies that have real-world implications.