Submitted to: Hydrological Processes
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/12/2011
Publication Date: N/A
Citation: N/A Interpretive Summary: Improved estimates of soil moisture are critical in the areas of hydrologic, agricultural and environmental modeling because of its role in governing the transfer of energy between the land surface and the atmosphere. Due to current satellite technology, remotely sensed observations of surface soil moisture (5cm) have become increasingly available and often applied in various environmental studies. In this study, the USDA, Agricultural Research Service, Root Zone Water Quality Model (RZWQM) was used to estimate profile soil moisture in two agricultural fields in northern Indiana. The model simulation results were updated by applying two data assimilation techniques: the Ensemble Kalman Filter (EnKF) and the direct insertion method. Data assimilation has been traditionally used in meteorology to improve weather forecasting and more recently in an effort to improve predictions of soil moisture status in the soil profile. This approach is based on the assumption that assimilating (or updating) surface soil moisture with measured data, such as remotely sensed data, the hydrologic model would improve estimates of soil water content in the entire profile. Among the measured soil moisture data at four different depths (5cm, 20cm, 40cm and 60cm), only the top 5cm data were assimilated on a daily basis with simulated results validated using all of the measured data. Overall, the results of this study indicate that daily (or bi-daily) assimilation of surface soil moisture improves soil moisture estimation in the upper more dynamic layers (5 and 20cm) but has less affect at deeper layers (40 and 60cm). This work should help advance our efforts in using current and future satellite remotely sensed soil moisture products at both the farm and watershed scales.
Technical Abstract: Estimation of soil moisture has received considerable attention in the areas of hydrology, agriculture, meteorology and environmental studies because of its role in the partitioning water and energy at the land surface. In this study, the Ensemble Kalman Filter (EnKF), a popular data assimilation technique for non-linear systems was applied to the Root Zone Water Quality Model. Measured soil moisture data at four different depths (5cm, 20cm, 40cm and 60cm) from two agricultural fields (AS1 and AS2) in northeastern Indiana were used for assimilation and validation purposes. Through daily update, the EnKF improved all statistical results (correlation coefficient, Root Mean Square Error and Mean Bias Error) compared to the direct insertion method and model results without assimilation for 5cm and 20cm depth. Soil moisture estimates for deeper layers (40cm and 60cm) did not show significant improvement from assimilating surface soil moisture depending on the initial soil moisture simulation results. It is also demonstrated that more frequent update generally contribute to enhance the open loop simulation, but depending on the open loop simulation results frequent update might not be helpful. In addition, various ensemble sizes make little difference in the assimilation results. An ensemble of 100 members could produce results that were comparable to results from larger ensembles.