HR: 10:35h
AN: C12A-02 INVITED    [Abstracts]
TI: Implementation strategies for multi-sensor snow data assimilation
AU: * Andreadis, K M
EM: kostas@hydro.washington.edu
AF: Civil and Environmental Engineering, University of Washington, Wilson Ceramic Lab, Box 352700, Seattle, WA 98195-2700, United States
AU: Lettenmaier, D P
EM: dennisl@u.washington.edu
AF: Civil and Environmental Engineering, University of Washington, Wilson Ceramic Lab, Box 352700, Seattle, WA 98195-2700, United States
AB: Data assimilation provides a framework for optimally merging model predictions and remote sensing observations of snow properties (snow cover extent, water equivalent, grain size, melt state), ideally overcoming limitations of both. Remotely sensed snow-related observations to be assimilated can include visible (snow cover extent), passive microwave (brightness temperature), as well as IR (snow grain size). Although these observations can potentially provide more information about snow properties, they can also complicate the estimation problem due to differences in spatial/temporal scales between the model and the different types of observations, as well as different error structures associated with each data source. Different strategies of assimilating the three different types of observations (sequentially versus simultaneously) are explored, and simple models of the observation errors for each observation type are evaluated using measurements from the Cold Land Processes Experiment (CLPX). We examine how snow cover extent observations can be combined with passive microwave radiances to effectively downscale the latter observations (coarser scale). Additionally, the effects of forest and partial snow cover on the assimilation of microwave observations are shown, and simple error models that utilize ancillary datasets (e.g. land cover) are derived. Moreover, snow grain sizes estimated from IR sensors are evaluated in the context of simulating microwave brightness temperatures, and the implications of their simultaneous assimilation are discussed. These results are then applied to examples of SWE estimation using two data assimilation techniques: the Ensemble Kalman filter (EnKF) and the Multi-scale Ensemble Kalman filter (MSEnKF). The latter is especially attractive for multi-sensor assimilation applications as a result of its structure that accommodates multi-scale observations in a single tree structure.
DE: 0758 Remote sensing
DE: 3315 Data assimilation
SC: Cryosphere [C]
MN: 2007 Fall Meeting