HR: 17:00h
AN: GC44A-05 INVITED [Abstracts]
TI: Investigating the Influence of Satellite-based Precipitation Uncertainty Estimate on Ensemble
Streamflow Forecasting
AU: * Moradkhani, H
EM: moradkha@uci.edu
AF: University of California, Irvine, E/4130 Engieering Gateway
Center for Hydrometeorology and Remote Sensing
Department of Civil and Environmental Engineering, Irvine, CA 91697
United States
AU: Hong, Y
EM: yanghong@uci.edu
AF: University of California, Irvine, E/4130 Engieering Gateway
Center for Hydrometeorology and Remote Sensing
Department of Civil and Environmental Engineering, Irvine, CA 91697
United States
AU: Hsu, K
EM: kuolinh@uci.edu
AF: University of California, Irvine, E/4130 Engieering Gateway
Center for Hydrometeorology and Remote Sensing
Department of Civil and Environmental Engineering, Irvine, CA 91697
United States
AU: Sorooshian, S
EM: soroosh@uci.edu
AF: University of California, Irvine, E/4130 Engieering Gateway
Center for Hydrometeorology and Remote Sensing
Department of Civil and Environmental Engineering, Irvine, CA 91697
United States
AB:
Precipitation is the key forcing environmental variable to any hydrometeorologic model and the emerging use of
satellite-based precipitation offers a high resolution spatio-temporal estimate as needed in the distributed
hydrometeorologic models. However, realistic estimation of precipitation uncertainty is crucial for ensemble generation
within the sequential data assimilation and ensemble forecasting paradigm. Owing to the multiplicative nature of
precipitation error, the model parameters and states estimation with their associated uncertainties are highly influenced by
the error of this model forcing field. In this work, we present the formalism for uncertainty estimation of satellite
precipitation product from PERSIANN-CCS system referenced on radar measurement and demonstrate how to propagate these errors
into hydrologic response. This method provides a more realistic error representation which is a function of time, space and
the spatio-temporal average of precipitation as opposed to the point measurement with lack of availability of some other
well-known limitations. The main step in data assimilation/ensemble forecasting system is the perturbation of these forcing
data to propagate the replicates of precipitation ensemble into the hydrologic model. The spread of the model ensemble is
directly related to the precipitation error; therefore, our major effort in this presentation goes into examining the impact
of such a probabilistic error characterization on the accuracy of state-parameter estimation of the hydrologic models. This
is done through forecasting-updating scheme using a sequential data assimilation technique.
DE: 1816 Estimation and forecasting
DE: 1853 Precipitation-radar
DE: 1855 Remote sensing (1640)
DE: 1860 Streamflow
DE: 1873 Uncertainty assessment (3275)
SC: Global Climate Change [GC]
MN: Fall Meeting 2005