273
Precipitation Estimate Using NEXRAD Ground-Based Radar Images
Although errors exist in NEXRAD precipitation products, radar estimates remain
a viable source of precipitation data, especially as radar algorithms are improved
and denser rain gauge networks are created for radar validation (Habib et al. 2009).
Nevertheless, previous efforts to improve the accuracy of NEXRAD showed the
potential of calibrating NEXRAD data using rain gauge data (e.g., Seo et al. 1990;
Steiner et al. 1999; Haberlandt 2007; Li et al. 2008; Zhang and Srinivasan 2010) to
provide more accurate spatial precipitation.
The aims of this chapter are to (1) briefly review the NEXRAD precipitation
image products from NWS and its validation and calibration using rain gauge observations and (2) introduce and illustrate the application of NEXRAD Validation and
Calibration (NEXRAD-VC) (Zhang and Srinivasan 2010), a geographic information
system (GIS)-based, user-friendly software, for processing, validating, and calibrating NEXRAD data.
12.2 LITERATURE REVIEW
12.2.1 nexRad PReciPitation PRoductS
The production of NEXRAD precipitation products involves several major procedures and various “stages” of processing by the NWS (Anagnostou and Krajewski
1998; Fulton et al. 1998). First, a radar system measures the reflectivity of a volume
of air by scanning over a fixed polar grid with a radial resolution of 1° in azimuth
by 1 km in range. The relationship between these reflectivities and precipitation is
expressed in the so-called Z–R relationship. The NEXRAD precipitation algorithms
utilize a power law Z–R, which is formulated as
R = aZ b ,
(12.1)
where R is the precipitation rate (in millimeter per hour), a and b are adjustable
parameters, and Z is the radar reflectivity factor and is expressed in linear units (in
millimeter to the sixth power per cubic meter). The default values of a and b are 0.017
and 0.714, respectively. Deriving a single equation accurate for every storm type and
intensity is often not possible, leading scientists to generate different relationships
case by case (e.g., the convective Z–R relationship and the Rosenfeld tropical Z–R
relationship; for more details, see http://www.roc.noaa.gov/ops/z2r_osf5.asp) for converting the reflectivities into precipitation rates contingent on the precipitation type.
The first precipitation estimates are referred to as Stage I data. Next, Stage II data are
produced through correcting Stage I data using bias adjustment (BA). Finally, Stage
III mosaics the data from multiple radar systems for the areas under the umbrella
of more than one radar unit. Based on several years of operational experience with
Stages II and III, much of the software was overhauled in 2000 and redeveloped into
the multisensor precipitation estimator (MPE) and enhanced multisensor precipitation
estimator (http://www.nws.noaa.gov/oh/hrl/dmip/stageiii_info.htm), which incorporates the precipitation measurements from gauges and precipitation estimates from
NEXRAD and geostationary operational environmental satellites (Wang et al. 2008).
Most of the NWS cooperative observers’ data have been used as a quality control
Precipitation Estimate Using NEXRAD Ground-Based Radar Images
Although errors exist in NEXRAD precipitation products, radar estimates remain
a viable source of precipitation data, especially as radar algorithms are improved
and denser rain gauge networks are created for radar validation (Habib et al. 2009).
Nevertheless, previous efforts to improve the accuracy of NEXRAD showed the
potential of calibrating NEXRAD data using rain gauge data (e.g., Seo et al. 1990;
Steiner et al. 1999; Haberlandt 2007; Li et al. 2008; Zhang and Srinivasan 2010) to
provide more accurate spatial precipitation.
The aims of this chapter are to (1) briefly review the NEXRAD precipitation
image products from NWS and its validation and calibration using rain gauge observations and (2) introduce and illustrate the application of NEXRAD Validation and
Calibration (NEXRAD-VC) (Zhang and Srinivasan 2010), a geographic information
system (GIS)-based, user-friendly software, for processing, validating, and calibrating NEXRAD data.
12.2 LITERATURE REVIEW
12.2.1 nexRad PReciPitation PRoductS
The production of NEXRAD precipitation products involves several major procedures and various “stages” of processing by the NWS (Anagnostou and Krajewski
1998; Fulton et al. 1998). First, a radar system measures the reflectivity of a volume
of air by scanning over a fixed polar grid with a radial resolution of 1° in azimuth
by 1 km in range. The relationship between these reflectivities and precipitation is
expressed in the so-called Z–R relationship. The NEXRAD precipitation algorithms
utilize a power law Z–R, which is formulated as
R = aZ b ,
(12.1)
where R is the precipitation rate (in millimeter per hour), a and b are adjustable
parameters, and Z is the radar reflectivity factor and is expressed in linear units (in
millimeter to the sixth power per cubic meter). The default values of a and b are 0.017
and 0.714, respectively. Deriving a single equation accurate for every storm type and
intensity is often not possible, leading scientists to generate different relationships
case by case (e.g., the convective Z–R relationship and the Rosenfeld tropical Z–R
relationship; for more details, see http://www.roc.noaa.gov/ops/z2r_osf5.asp) for converting the reflectivities into precipitation rates contingent on the precipitation type.
The first precipitation estimates are referred to as Stage I data. Next, Stage II data are
produced through correcting Stage I data using bias adjustment (BA). Finally, Stage
III mosaics the data from multiple radar systems for the areas under the umbrella
of more than one radar unit. Based on several years of operational experience with
Stages II and III, much of the software was overhauled in 2000 and redeveloped into
the multisensor precipitation estimator (MPE) and enhanced multisensor precipitation
estimator (http://www.nws.noaa.gov/oh/hrl/dmip/stageiii_info.htm), which incorporates the precipitation measurements from gauges and precipitation estimates from
NEXRAD and geostationary operational environmental satellites (Wang et al. 2008).
Most of the NWS cooperative observers’ data have been used as a quality control