A framework for automated and spatially-distributed modeling with the Agricultural Policy Environmental eXtender (APEX) model

Authors: PAN, FENG - Purdue University; FENG, QINGYU - Chinese Academy Of Sciences; MCGEHEE, RYAN - Purdue University; ENGEL, BERNARD - Purdue University; Flanagan, Dennis; CHEN, JINGQIU - Purdue University

Submitted to: Environmental Modelling & Software
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/15/2021
Publication Date: 7/19/2021
Citation: Pan, F., Feng, Q., McGehee, R., Engel, B.A., Flanagan, D.C., Chen, J. 2021. A framework for automated and spatially-distributed modeling with the Agricultural Policy Environmental eXtender (APEX) model. Environmental Modelling & Software. 144. Article 105147. https://doi.org/10.1016/j.envsoft.2021.105147.
DOI: https://doi.org/10.1016/j.envsoft.2021.105147

Interpretive Summary: Nonpoint source pollution of water leaving agricultural fields can contribute to serious downstream water body degradation. In particular, sediments from eroded soil and chemicals such as nitrogen and phosphorus from fertilizers and/or manures can cause increases in harmful algal blooms, decreases in water clarity/quality, and eutrophication. However, it is difficult and expensive to measure and monitor pollutant losses from large areas of cropland, thus computer simulation models are usually applied to estimate potential sediment and chemical losses for existing and alternative land management practices. Even with good natural resources models, it can be difficult and time-consuming to setup and simulate expansive agricultural areas. In this study, we developed a parallel modeling framework that makes the process of applying a water quality model more practical. We utilized the APEX (Agricultural Policy Environmental eXtender) model in the framework and tested it for small and large spatial scales in two case studies. The framework package includes national elevation, climate, land use, and soil databases, management scenarios, an APEX model executable file, a tutorial, and code documentation. The parallelized algorithms of the framework increased computational efficiency and reduced execution time significantly. The results impact scientists, university faculty, students, and conservation agency personnel that need to apply a natural resource model spatially across a landscape to evaluate effectiveness of various conservation and best management practices (BMPs).

Technical Abstract: Agricultural Best Management Practices (BMPs) are popular approaches to reduce nonpoint source (NPS) nutrient loadings. Hydrologic models that can simulate impacts of BMPs at the field-scale can help guide the selection of BMPs. High-performance computing techniques have significant potential for scaling spatial simulations and reducing model runtimes. In this study, a parallel modeling framework for the Agricultural Policy Environmental eXtender (APEX) was developed for large-scale, high-resolution, spatially-distributed model simulations. It provides a tool for conducting BMP evaluations at field-scale with distributed architecture and automatic model setup of APEX. Two case studies were conducted to demonstrate the capability of the framework for simulating different spatial scales, to compare the results between semi-distributed and distributed versions of the framework, and to illustrate the parallelization efficiency of the framework. This framework can help provide guidance for decision makers on agricultural BMPs with large-scale water quality assessments and NPS nutrient loading reductions.