Poster Number 208
Monday, 6 October 2008
George R. Brown Convention Center, Exhibit Hall E
Soil water is an important variable in agricultural environments as it contributes to yield response as well as areas of environmental concern including erosion, runoff, and N leaching. Crop models have been established as a method for simulating agricultural production and examining ecosystem responses. However, because all crop models are based on limited system information, models contain errors which increase uncertainty around their predictions. Field measurements are also useful in determining soil moisture status, but they are not predictive and their influence is limited due to the high spatial variability experienced in many agricultural areas. Data assimilation provides the opportunity to merge both model and observational data in order to obtain a better representation of the true physical system. Assimilation of near-surface soil moisture data has gained wide acceptance in remote sensing applications, yet its study in agricultural crop models is extremely limited. In this paper we will use the Kalman Filter to model the temporal patterns of the soil moisture profile, based on near-surface soil water content measurements of a maize field in Ames, IA. The objective of the experiment is to determine if data assimilation of near-surface soil moisture can lead to better predictions of deeper soil moisture. The motivating hypothesis behind this work is that accurate prediction of the soil water profile will (a) allow for improved calculations of the water balance and plant water stress and (b) potentially alter management decisions, thereby directly impacting plant growth and development. To accomplish our objective we will use The DSSAT Model, a crop model that has been thoroughly tested and calibrated for our research location. In our study we will assess the performance of the assimilation algorithms under different observational frequencies and model and observational errors that one might realistically expect to obtain in the field.