Enhanced ammonia volatilization prediction with EPIC: Model description and testing of different fertilizers

Biogeochemical models are promising cost-effective tools to evaluate ammonia (NH₃) volatilization reduction strategies at large spatial scales provided they capture all the relevant processes that regulate universally the nitrogen (N) cycle in agroecosystems. This work aims to enhance the Environmental Policy Integrated Climate (EPIC) model to improve its ability to simulate NH₃ volatilization from both organic and mineral N fertilizers in agricultural fields. Extant algorithms of NH₃ volatilization in EPIC were replaced with a mechanistic submodel that, operating at hourly steps, effectively simulated NH₃ volatilization by capturing processes of ammonium (NH₄⁺) adsorption, urea hydrolysis, soil pH-based partitioning of total ammoniacal N into NH₃ and NH₄⁺, and mass transfer. The new EPIC submodel was calibrated and validated using data from two different locations (Legnaro, NE Italy; Eatonton, Georgia, USA), including several combinations of fertilizer types and application methods in different pedo-climatic conditions. Results showed that the new submodel provided more accurate estimates of cumulative NH₃ loss than the original one (validated R² = 0.79 vs. 0.50, validated RMSE = 10.6 vs 17.8 kg N ha⁻¹). Furthermore, the implementation now enables more accurate simulation of fertilizer types and management, incorporating the effect of fertilizer pH, depth and method of application, and infiltration of the liquid fraction. In conclusion, the updated EPIC reported here becomes an effective tool to evaluate and select best agricultural practices capable of reducing NH₃ volatilization.