HR: 0800h
AN: IN31A-10    [Abstracts]
TI: Estimation of actual evapotranspiration through model coupling and data assimilation with remotely sensed land surface properties
AU: * Kovalskyy, V
EM: Valeriy.Kovalskyy@sdstate.edu
AF: Geographic Information Science Center of Excellence (GIScCE), 1021 Medary Ave, Wecota Hall Box 506B, Brookings, SD 57007, United States
AU: Henebry, G
EM: Geoffrey.Henebry@sdstate.edu
AF: Geographic Information Science Center of Excellence (GIScCE), 1021 Medary Ave, Wecota Hall Box 506B, Brookings, SD 57007, United States
AB: We report on preliminary results from the coupling of two models and satellite observations to track evapotranspiration (ET) dynamics in Northern Great Plains of the USA. The approach takes advantage of high- quality microclimate and irradiance/radiance measurements in a data assimilation scheme to estimate actual ET through a stepwise simulation of foliage dynamics, corrected by remotely sensed land surface properties. We used a recently developed VegET model that uses water balance principles and phenological constraints (Senay 2008) coupled with an event driven phenology model (EDPM) to simulate canopy dynamics unfolding in response to changing environmental conditions and disturbance events. We used NDVI derived from MODIS Collection 5 Nadir BRDF Adjusted Reflectance (NBAR; MCD43B4V5) to amend the outputs of the EDPM using one-dimensional Kalman filtering to achieve a better representation of changing canopy conditions. The model was trained on level 1 flux tower data from cropland sites at Mead, Nebraska and refined using similar records from Bondville, Illinois. Results from the test runs demonstrated the ability of EDPM to drive the phenological constrains of VegET with reasonable accuracy (RMSE 0.03-0.10 at Nebraska sites). Filtered and unfiltered results from the coupled model were compared with actual evapotranspiration recorded on flux towers and with tower NDVI (Wittich and Kraft 2008). Depending on vegetation type and location, Pearson correlation coefficients between model estimates and observed values ranged between 0.8 and 0.9.
DE: 1622 Earth system modeling (1225)
DE: 1631 Land/atmosphere interactions (1218, 1843, 3322)
DE: 1818 Evapotranspiration
DE: 1855 Remote sensing (1640)
SC: Earth and Space Science Informatics [IN]
MN: 2009 Joint Assembly