272
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
12.1 INTRODUCTION
Precipitation, an important input for land surface processes such as the hydrologic
cycle and vegetation growth, is characterized by high spatial and temporal variability. Traditionally, precipitation measurements are available at rain gauge points,
which are usually too sparsely distributed to capture spatial variability; therefore,
these point data need to be interpolated to estimate the spatial distribution of precipitation. Development of methods to interpolate precipitation data from sparse
networks of rain gauge stations has been a focus of past research (e.g., Phillips
et al. 1992; Hasenauer et al. 2003). In recent years, the U.S. National Weather
Service (NWS) installed a network of (approximately 160) Weather Surveillance
Radar—1988 Dopplers (WSR-88Ds) radar stations as part of a Next Generation
Radar (NEXRAD) program that began implementation in 1991 (Young et al. 2000;
Hardegree et al. 2008). The NEXRAD products, located in the Contiguous United
States (CONUS) at approximately 4 × 4 km 2 resolution, provide nominal coverage of
96% of the country (Crum et al. 1998). The ability of NEXRAD to provide spatially
distributed precipitation estimates makes it one important source of precipitation
information for hydrologists and natural resources managers. The NEXRAD precipitation products have been used for multiple purposes in hydrologic modeling and
agricultural and rangeland management (e.g., Diak et al. 1998; Krajewski and Smith
2002; Zhang et al. 2004; Hardegree et al. 2008).
Two major issues arise concerning the application of NEXRAD. One is the lack
of a NEXRAD geoprocessing and georeferencing tool (Hardegree et al. 2008).
Digital, distributed precipitation NEXRAD products in binary-coded format can be
obtained from NWS; however, a few user-friendly software or analysis tools exist in
the public domain to facilitate accessibility of radar precipitation products. Although
ideas for practical application of NEXRAD precipitation in agricultural and water
resources management have been derived, implementation has been relatively slow
(Hardegree et al. 2008).
A second issue is accuracy of estimates (Krajewski and Smith 2002). In general, traditional rain gauges are able to provide more accurate measurements of
precipitation than NEXRAD, because they physically measure the depth of precipitation. Many previous studies evaluated the accuracy of the NEXRAD data using
rain gauge data and reported substantial discrepancies (Krajewski and Smith 2002).
12.5.2 Overall Assessment of NEXRAD Products .................................... 293
12.5.3 Visual Inspection of Precipitation Maps Obtained by Different
Methods ............................................................................................ 293
12.5.4 Comparing the Performances of Different Calibration Methods ..... 295
12.5.4.1 Overall Performance Assessment ..................................... 295
12.5.4.2 Performance Comparison for Daily Spatial
Precipitation Prediction ..................................................... 296
12.6 Conclusions .................................................................................................. 298
Acknowledgments ................................................................................................. 299
References ............................................................................................................. 299
Multiscale Hydrologic Remote Sensing: Perspectives and Applications
12.1 INTRODUCTION
Precipitation, an important input for land surface processes such as the hydrologic
cycle and vegetation growth, is characterized by high spatial and temporal variability. Traditionally, precipitation measurements are available at rain gauge points,
which are usually too sparsely distributed to capture spatial variability; therefore,
these point data need to be interpolated to estimate the spatial distribution of precipitation. Development of methods to interpolate precipitation data from sparse
networks of rain gauge stations has been a focus of past research (e.g., Phillips
et al. 1992; Hasenauer et al. 2003). In recent years, the U.S. National Weather
Service (NWS) installed a network of (approximately 160) Weather Surveillance
Radar—1988 Dopplers (WSR-88Ds) radar stations as part of a Next Generation
Radar (NEXRAD) program that began implementation in 1991 (Young et al. 2000;
Hardegree et al. 2008). The NEXRAD products, located in the Contiguous United
States (CONUS) at approximately 4 × 4 km 2 resolution, provide nominal coverage of
96% of the country (Crum et al. 1998). The ability of NEXRAD to provide spatially
distributed precipitation estimates makes it one important source of precipitation
information for hydrologists and natural resources managers. The NEXRAD precipitation products have been used for multiple purposes in hydrologic modeling and
agricultural and rangeland management (e.g., Diak et al. 1998; Krajewski and Smith
2002; Zhang et al. 2004; Hardegree et al. 2008).
Two major issues arise concerning the application of NEXRAD. One is the lack
of a NEXRAD geoprocessing and georeferencing tool (Hardegree et al. 2008).
Digital, distributed precipitation NEXRAD products in binary-coded format can be
obtained from NWS; however, a few user-friendly software or analysis tools exist in
the public domain to facilitate accessibility of radar precipitation products. Although
ideas for practical application of NEXRAD precipitation in agricultural and water
resources management have been derived, implementation has been relatively slow
(Hardegree et al. 2008).
A second issue is accuracy of estimates (Krajewski and Smith 2002). In general, traditional rain gauges are able to provide more accurate measurements of
precipitation than NEXRAD, because they physically measure the depth of precipitation. Many previous studies evaluated the accuracy of the NEXRAD data using
rain gauge data and reported substantial discrepancies (Krajewski and Smith 2002).
12.5.2 Overall Assessment of NEXRAD Products .................................... 293
12.5.3 Visual Inspection of Precipitation Maps Obtained by Different
Methods ............................................................................................ 293
12.5.4 Comparing the Performances of Different Calibration Methods ..... 295
12.5.4.1 Overall Performance Assessment ..................................... 295
12.5.4.2 Performance Comparison for Daily Spatial
Precipitation Prediction ..................................................... 296
12.6 Conclusions .................................................................................................. 298
Acknowledgments ................................................................................................. 299
References ............................................................................................................. 299