Hydrological evaluation of satellite precipitation products in La

Hydrological evaluation of satellite precipitation products in La Plata basin 1 Fengge Su, 2 Yang Hong, 3 William L. Crosson, and 4 Dennis P. Lettenmaier 1, 3 USRA at NSSTC/NASA/MSFC, Huntsville, AL of Oklahoma, Norman, OK 4 University of Washington, Seattle, WA 2 University

1. Objective • To evaluate the satellite-based precipitation estimates, and their utility in hydrologic predictions. • To provide insights into the potential utility of Global Precipitation Measurement mission (GPM) products for hydrologic applications. • To make a first step toward a hydrologically based global assessment framework for satellite-based precipitation products (particularly those using minimal in situ data).

2. Data sets and methodology Satellite-based precipitation: (1) The TRMM Multi-satellite Precipitation Analysis (TMPA) research product 3 B 42 V. 6 and real-time version 3 B 42 -RT (Huffman et al 2007). (2) The Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks (PERSIANN) (Sorooshian et al 2000). (3) The Climate Prediction Center morphing method (CMORPH) (Joyce et al 2004). Data resolution: 3 -hour time step and 0. 25° degree spatial resolution. Study period: January 1, 2003 ~ August 31, 2006 Study area: La Plata basin (-38 ° S ~ -14 ° S, -56 ° W ~ -41 ° W, 3. 1× 106 km 2) All three products are satellite only (no gauge-based post-processing, as in TMPA RP)

Gauge-based precipitation in La Plata basin South American gridded daily precipitation (0. 25 °) constructed from daily meteorological station data (Liebmann and Allured, 2005). Four sub-basins (highlighted) have relatively dense station coverage.

Methodology • Spatially-distributed and basin-wide precipitation estimates from satellite precipitation products are compared with gauge-derived precipitation at time scales from annual to daily for the period 2003 through 2005. • The Variable Infiltration Capacity (VIC) land surface hydrology model is forced by the daily satellite-based and gridded gauge precipitation estimates over La Plata basin. Simulated runoff fields and streamflow driven by satellite precipitation are compared with those driven by gauge precipitation inputs.

3. Precipitation evaluation- spatial fields Gauge TMPA V. 6 Annual mean precipitation over La Plata basin from gridded gauge, TMPA V. 6 (gauge adjusted), TMPA-RT, CMORPH, and PERSIANN mm/yr estimates for the years 2003 -2005. Difference of annual precipitation between satellite and gauge mm/yr estimates.

Basin-averaged precipitation TMPA V. 6 Basin 6682 Bias = 0. 17 Basin 6301 Basin 6598 Basin 3802 Scattergrams of daily basin- Bias = 0. 11 Bias = 0. 09 Bias = 0. 10 averaged precipitation from gauge and satellite estimates TMPA-RT CMORPH Bias = 0. 44 PERSIANN for four sub-basins in La Plata Bias = 0. 36 Bias = 0. 63 Gauged (mm/dy) Bias = 0. 08 for the period 2003 to 2005. Bias = 0. 24 Bias = 0. 42 Bias = 0. 31 Bias = 0. 49 Bias = 0. 19 Bias = 0. 31 Bias = 0. 16 Bias = -0. 03 Gauged(mm/dy) Gauged (mm/dy)

Bias relative to gauge estimates for each year ( 2003, 2004, and 2005). Basin 3802 0, 8 0, 5 0, 7 0, 4 0, 6 Bias Basin 6598 TMPA 0, 5 0, 3 RT 0, 4 CMORPH 0, 2 0, 3 PERSIANN 0, 2 0, 1 0 2002 2003 2004 2005 0 2002 2006 Basin 6301 0, 3 0, 2 TMPA Bias 0, 1 0 2002 -0, 1 -0, 2 -0, 3 -0, 4 RT 2003 2004 2005 2006 CMORPH PERSIANN 0, 9 0, 8 0, 7 0, 6 0, 5 0, 4 0, 3 0, 2 0, 1 0 2002 2003 2004 2005 2006 Basin 6682 2003 2004 The bias significantly decreased in 2005 over basins 3802 and 6598 for the three realtime satellite precipitation products.

4. Hydrological Modelling Model features: • multiple vegetation classes in each cell; • energy and water budget closure at each time step; • subgrid infiltration and runoff variability; • and non-linear baseflow generation. The VIC (Liang et al, 1994; 1996) model is a grid-based land surface scheme designed both for inclusion in GCMs, and for use as a stand-alone macroscale hydrological model.

Spatial fields of annual mean simulated runoff (2003 -2005) Gauge TMPA V. 6 Annual mean simulated runoff forced by the gauge, TMPA V. 6, TMPART, CMORPH, and PERSIANN estimates for the years 2003 -2005. mm/yr Difference of annual simulated runoff between satellite-driven and gaugemm/yr observation-driven simulations.

Daily simulated streamflow for basin 3802, Uruguay at Paso de los Libres (Area: 189, 300 km 2, Jan 2003~Aug 2006) Simulated with satellite Precip Simulated with gauge Precip Bias TMPA (%) 2003 20 2004 2005 RT (%) CMORP H (%) PERSIAN N (%) 110 105 32 120 142 73 22 39 55 17

Daily simulated streamflow for basin 6598, Iguazu at Estreito (Area: 63, 236 km 2, Jan 2003~Aug 2006) Bias TMPA (%) RT (%) CMOR PERSIA PH (%) NN (%) 2003 27 95 104 101 2004 28 56 80 35 2005 9 18 26 -3 Simulated with satellite Precip. Simulated with gauge Precip.

Daily simulated streamflow for basin 6301, Parana at Jupia (Area: 478, 000 km 2, Jan 2003 - Aug 30 2006) Bias TMPA (%) RT (%) CMORP PERSIAN H (%) N (%) 2003 22 45 34 -3 2004 25 8 66 -7 2005 35 34 36 7 Simulated with satellite Precip. Simulated with gauged Precip.

Summary • All satellite precipitation products captured the general precipitation regime of the La Plata basin, but with significant overestimation compared with gauge estimates over the middle part of the basin and slight underestimation in the northeastern part. • The four satellite products show different performance over different sub-basins, indicating the effects topography and possibly land cover on the precipitation retrieval algorithms. • Biases of the precipitation estimates significantly decreased in 2005 over the central subbasins, likely due to algorithm changes and increased coverage from microwave sensors. • The plausible performance of the satellite products in streamflow simulations in recent years suggests that even without the monthly gauge correction, there is considerable potential for hydrologic prediction using satellite precipitation estimates in data sparse areas. • Further evaluation is needed with more complete daily gauge data over other global river basins.

Thank You!