Assessing the skill of satellite-based precipitation estimates in hydrologic applications

Ming Pan, Haibin Li, Eric F. Wood

Research output: Contribution to journalArticle

75 Citations (Scopus)

Abstract

An important application of global precipitation measurement rainfall products is providing forcing inputs for hydrologic applications, and the goal of this study is to assess how skillful they are for such applications. To do that, we force a land surface model with both satellite estimates and ground-based measurements and test how well they can predict hydrologic states and fluxes useful for water resource applications, i.e., soil moisture, evapotranspiration, and river streamflow. A number of satellite-based precipitation products ranging from retrievals based only on microwave measurements, combined microwave + infrared estimates, to gauge-corrected products are tested over the entire continental United States region. As a reference to the satellite retrievals, estimates from global and regional weather model reanalyses (the analysis fields from these models) are tested as well. It is found that the microwave + infrared combined estimates can match the skill of the coarse resolution European Center for Medium-Range Weather Forecasts global reanalysis but not the regional National Center for Environmental Predictions reanalysis. Gauge corrections to satellite products significantly enhance their skill by greatly reducing the bias in hydrologic predictions, especially over mountainous areas. Rainfall errors are shown to have strong impact on river streamflow predictions and column total soil moisture and relatively weak impact on near surface moisture and evapotranspiration. River streamflow experiments also suggest that satellite rainfall errors are highly correlated in space within the range of one storm system and thus do not reduce in magnitude with spatial scale (basin size).

Original languageEnglish (US)
Article numberW09535
JournalWater Resources Research
Volume46
Issue number9
DOIs
StatePublished - Oct 13 2010

Fingerprint

streamflow
rainfall
evapotranspiration
gauge
soil moisture
prediction
river
weather
ground-based measurement
land surface
water resource
moisture
product
basin
microwave
experiment
test
analysis
mountainous area
forecast

All Science Journal Classification (ASJC) codes

  • Water Science and Technology

Cite this

@article{7b2ff751bde049dfa37c36c15422a4ed,
title = "Assessing the skill of satellite-based precipitation estimates in hydrologic applications",
abstract = "An important application of global precipitation measurement rainfall products is providing forcing inputs for hydrologic applications, and the goal of this study is to assess how skillful they are for such applications. To do that, we force a land surface model with both satellite estimates and ground-based measurements and test how well they can predict hydrologic states and fluxes useful for water resource applications, i.e., soil moisture, evapotranspiration, and river streamflow. A number of satellite-based precipitation products ranging from retrievals based only on microwave measurements, combined microwave + infrared estimates, to gauge-corrected products are tested over the entire continental United States region. As a reference to the satellite retrievals, estimates from global and regional weather model reanalyses (the analysis fields from these models) are tested as well. It is found that the microwave + infrared combined estimates can match the skill of the coarse resolution European Center for Medium-Range Weather Forecasts global reanalysis but not the regional National Center for Environmental Predictions reanalysis. Gauge corrections to satellite products significantly enhance their skill by greatly reducing the bias in hydrologic predictions, especially over mountainous areas. Rainfall errors are shown to have strong impact on river streamflow predictions and column total soil moisture and relatively weak impact on near surface moisture and evapotranspiration. River streamflow experiments also suggest that satellite rainfall errors are highly correlated in space within the range of one storm system and thus do not reduce in magnitude with spatial scale (basin size).",
author = "Ming Pan and Haibin Li and Wood, {Eric F.}",
year = "2010",
month = "10",
day = "13",
doi = "https://doi.org/10.1029/2009WR008290",
language = "English (US)",
volume = "46",
journal = "Water Resources Research",
issn = "0043-1397",
publisher = "American Geophysical Union",
number = "9",

}

Assessing the skill of satellite-based precipitation estimates in hydrologic applications. / Pan, Ming; Li, Haibin; Wood, Eric F.

In: Water Resources Research, Vol. 46, No. 9, W09535, 13.10.2010.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Assessing the skill of satellite-based precipitation estimates in hydrologic applications

AU - Pan, Ming

AU - Li, Haibin

AU - Wood, Eric F.

PY - 2010/10/13

Y1 - 2010/10/13

N2 - An important application of global precipitation measurement rainfall products is providing forcing inputs for hydrologic applications, and the goal of this study is to assess how skillful they are for such applications. To do that, we force a land surface model with both satellite estimates and ground-based measurements and test how well they can predict hydrologic states and fluxes useful for water resource applications, i.e., soil moisture, evapotranspiration, and river streamflow. A number of satellite-based precipitation products ranging from retrievals based only on microwave measurements, combined microwave + infrared estimates, to gauge-corrected products are tested over the entire continental United States region. As a reference to the satellite retrievals, estimates from global and regional weather model reanalyses (the analysis fields from these models) are tested as well. It is found that the microwave + infrared combined estimates can match the skill of the coarse resolution European Center for Medium-Range Weather Forecasts global reanalysis but not the regional National Center for Environmental Predictions reanalysis. Gauge corrections to satellite products significantly enhance their skill by greatly reducing the bias in hydrologic predictions, especially over mountainous areas. Rainfall errors are shown to have strong impact on river streamflow predictions and column total soil moisture and relatively weak impact on near surface moisture and evapotranspiration. River streamflow experiments also suggest that satellite rainfall errors are highly correlated in space within the range of one storm system and thus do not reduce in magnitude with spatial scale (basin size).

AB - An important application of global precipitation measurement rainfall products is providing forcing inputs for hydrologic applications, and the goal of this study is to assess how skillful they are for such applications. To do that, we force a land surface model with both satellite estimates and ground-based measurements and test how well they can predict hydrologic states and fluxes useful for water resource applications, i.e., soil moisture, evapotranspiration, and river streamflow. A number of satellite-based precipitation products ranging from retrievals based only on microwave measurements, combined microwave + infrared estimates, to gauge-corrected products are tested over the entire continental United States region. As a reference to the satellite retrievals, estimates from global and regional weather model reanalyses (the analysis fields from these models) are tested as well. It is found that the microwave + infrared combined estimates can match the skill of the coarse resolution European Center for Medium-Range Weather Forecasts global reanalysis but not the regional National Center for Environmental Predictions reanalysis. Gauge corrections to satellite products significantly enhance their skill by greatly reducing the bias in hydrologic predictions, especially over mountainous areas. Rainfall errors are shown to have strong impact on river streamflow predictions and column total soil moisture and relatively weak impact on near surface moisture and evapotranspiration. River streamflow experiments also suggest that satellite rainfall errors are highly correlated in space within the range of one storm system and thus do not reduce in magnitude with spatial scale (basin size).

UR - http://www.scopus.com/inward/record.url?scp=77955169667&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955169667&partnerID=8YFLogxK

U2 - https://doi.org/10.1029/2009WR008290

DO - https://doi.org/10.1029/2009WR008290

M3 - Article

VL - 46

JO - Water Resources Research

JF - Water Resources Research

SN - 0043-1397

IS - 9

M1 - W09535

ER -