Simulating in situ ammonia volatilization losses in the North China Plain using a dynamic soil‐crop model

• Published in: Journal of plant nutrition and soil science Vol. 179; no. 2; pp. 270 - 285
• Main Authors: Michalczyk, Anna, Kersebaum, Kurt Christian, Heimann, Lisa, Roelcke, Marco, Sun, Qin‐Ping, Chen, Xin‐Ping, Zhang, Fu‐Suo
• Format: Journal Article
• Language: English
• Published: Weinheim WILEY‐VCH Verlag 01-04-2016; WILEY-VCH Verlag; Wiley Subscription Services, Inc
• Subjects: Ammonia; corn; data collection; environmental factors; equations; fertilization; hydrolysis; intensive farming; management; NH3 volatilization; nitrification; nitrogen; nitrogen fertilizers; plant growth; simulation; simulation models; soil nutrient dynamics; soil solution; topsoil; urea; volatilization; Zea mays; China; China, People's Rep., North China Plain; China, People's Rep
• ISSN: 1436-8730; 1522-2624
• DOI: 10.1002/jpln.201400673

Ammonia (NH₃) volatilization is an important N loss pathway in intensive agriculture of the North China Plain (NCP). Simulation models can help to assess complex N and water processes of agricultural soil–crop systems. Four variations (Var) of a sub‐module for the deterministic, process‐based HERMES model were implemented ranging from simple empirical functions (Var 3 and 4) to process‐oriented approaches (Var 1 and 2) including the main processes of NH₃ volatilization, urea hydrolysis, nitrification from ammonium‐based N fertilizer, and changes in soil solution pH. Ammonia volatilization, plant growth, and changes in ammonium and nitrate pools in the soil over several winter wheat–summer maize double‐crop rotations at three locations in the NCP were simulated. Results were calibrated with two data sets (Dongbeiwang 1, Shunyi) and validated using two data sets (Dongbeiwang 2, Quzhou). They showed that the ammonia volatilization sub‐module of the HERMES model worked well under the climatic and soil conditions of N China. Although the simpler equations, Var 3 and 4, showed lower deviations to observed volatilization across all sites and treatments with a mean absolute error (MAE) of 1.8 and 1.4 in % of applied N, respectively, compared to process‐oriented approaches, Var 1 and 2, with a MAE of 2.2 and 1.9 in % of applied N, respectively. Environmental conditions were reflected better by the process‐oriented approaches. Generally, simulation results were satisfying but simulated changes in topsoil pH need further verification with measurements.