HR: 0800h
AN: C21A-1077 [Abstracts]
TI: Feasibility of distributed snowpack characterization during the Cold Land Processes Field Experiment
(CLPX) using a multi-scale multi-frequency data assimilation approach
AU: * Durand, M
EM: durand@seas.ucla.edu
AF: Department of Civil and Environmental Engineering,
University of California, Los Angeles, 5731 Boelter Hall
Box 951593, Los Angeles, CA 90095-1593
United States
AU: Margulis, S A
EM: margulis@seas.ucla.edu
AF: Department of Civil and Environmental Engineering,
University of California, Los Angeles, 5731 Boelter Hall
Box 951593, Los Angeles, CA 90095-1593
United States
AB:
Snowpack is a critical freshwater reservoir and plays an important role in the global water cycle. In addition, snow plays a
major role in the energy balance. Accordingly, much work has been done to estimate snow mass and extent. Under the prospect
of climate change, the task of characterizing snowpack has become more urgent: studies show that Western U.S. snowpack has
steadily declined for the past half century.
These facts have motivated inversion of remote sensing observations and development of prognostic models to estimate snow
water equivalent (SWE). Most retrieval algorithms, based on passive microwave (PM) observations, suffer from problems related
to the coarseness of the observations, deep/wet snowpacks, attenuating vegetation cover, and the many-to-one relationship
with SWE. In modeling approaches, SWE estimates are degraded by significant uncertainty in forcing data (especially
precipitation) available in mountainous regions.
The best prospect for snowpack characterization requires merging multiple data streams of disparate nature. Data assimilation
(DA) is an ideal framework for merging multi-scale multi-frequency remote sensing observations and model predictions because
it provides a means of weighing the uncertainty of meteorological data and other model inputs against the uncertainty of
remote sensing observations.
The objective of this study is to explore the potential of a DA methodology to characterize SWE over a Meso-cell Study Area
(MSA) of the CLPX. Synthetic multi-frequency MODIS visible/near-infrared measurements generated at 1 km and AMSR-E PM
measurements generated at 25 km are merged with a spatially distributed land surface model forced with real data using the
Ensemble Kalman Filter (EnKF). The EnKF follows a Monte Carlo simulation approach to obtain an estimate of the uncertainty in
the state variables, which is weighed against the measurement uncertainty to optimally update the state variables. Results
are presented that test the premise that the DA approach and the synergy between the multi-scale multi-frequency observations
provide robust estimates of basin-scale SWE under conditions which generally limit retrieval and modeling approaches.
DE: 0736 Snow (1827, 1863)
DE: 0758 Remote sensing
DE: 1855 Remote sensing (1640)
DE: 1863 Snow and ice (0736, 0738, 0776, 1827)
DE: 1895 Instruments and techniques: monitoring
SC: Cryosphere [C]
MN: Fall Meeting 2005