Extended Abstract (44K)3.13
Seasonal Soil Moisture Prediction Using a Climate-Plant-Soil Coupled Agroecosystem Water Management Model
Z. Pan, Iowa State Univeristy, Ames, IA; and R. Horton, M. Segal, E. Takle, D. Herzmann, D. Todey, D. Flory, and J. Roads
Long-term soil moisture projection is a critical factor for economic decision-making in agriculture. The National Outlook is given at grid resolution that is too coarse and that cannot include locally observed soil information as input initial conditions. By coupling models of regional climate, surface hydrology, and crop development, we are establishing a two-way agroecosystem water modeling system that provides soil moisture (and other hydrological variables) projections four months in advance. Lateral boundary conditions were provided by the Experimental Climate Prediction Center's weekly 16-week global forecasts at 6-hour intervals. Initial soil moisture is continuously updated using Iowa Mesonet observations.
A prototype of the fully coupled forecasting system was tested for the spring of 2002. The weekly forecasts form a time-lagged ensemble forecast. Preliminary results show that (1) the model has some forecast skill in seasonal precipitation total and soil moisture status, with the model often being able to keep track of the observed precipitation time series in the first 7-10 days; (2) the weekly time-lagged ensemble members differ quite significantly after about two weeks, without the tendency of forecasts based on later weeks being better; and (3) the particular implementation of the model tends to underpredict precipitation, and thus gives drier soils.
Session 3, Land-Atmosphere Interactions 2: Process Representation and Evaluation
Tuesday, 11 February 2003, 1:30 PM-5:30 PM
Previous paper Next paper