Evaluating a Hybrid Soil Temperature Model in Agricultural Systems.
S. Feng1, Fernando Salvagiotti2, Marty Schmer3, Ana Wingeyer4, and A. Weiss1. (1) Univ of Nebraska, School of Natural Resources, 34th and Fair Streets, Lincoln, NE 68583-0728, (2) Univ of Nebraska - Lincoln, Department of Agonomy and Horticulture, 243 Keim Hall, Lincoln, NE 68503, (3) Univ of Nebraska - Lincoln, 3935 X St., Lincoln, NE 68503, (4) Univ of Nebraska, 1400 N. 33 Street #12, Lincoln, NE 68503
Soil temperature is an important factor that helps regulate seedling growth and survival, root growth and turnover, mycorrhiza colonization and development, soil organic matter decomposition, nutrient dynamics, and soil respiration. There are various approaches to simulate soil temperature, which are in part due to the time scale of interest and the purpose for simulating these temperatures. These approaches range from regression relationships to numerical solutions of the energy balance equations. A hybrid soil temperature simulation model that combines elements of mechanistic and empirical approaches to simulate soil temperatures can be of value when incorporated into a crop simulation model. Soil temperature was measured at 0.1m, 0.2m, 0.3m, and 0.5m in an irrigated and rain fed corn-soybean rotations over the period 2001-2002. A hybrid mechanistic empirical soil temperature model, initially developed for forest applications, was evaluated in this production system. Simulated and observed soil temperatures were compared using the mean bias error, the mean absolute error, and the root mean square error. The mean square error was decomposed into systematic (trends) and unsystematic (random) errors. Modifications to improve this hybrid mechanistic empirical soil temperature simulation model will be presented.