STAR Satellite Rainfall Estimates
Self-Calibrating Multivariate Precipitation Retrieval (SCaMPR)
Enterprise Algorithm (SCaMPR) Precipitation Estimates
|Product||Global Images||CONUS Images||DC-area Images|
|Instantaneous rain rate||Current Images||Current Images||Current Images|
|1-hour accumulation||Current Images||Current Images||Current Images|
|3-hour accumulation||Current Images||Current Images||Current Images|
|6-hour accumulation||Current Images||Current Images||Current Images|
|12-hour accumulation||Current Images||Current Images||Current Images|
|24-hour accumulation||Current Images||Current Images||Current Images|
|48-hour accumulation||Current Images||Current Images||Current Images|
|72-hour accumulation||Current Images||Current Images||Current Images|
|Global netCDF files||Download||Global lat/lon file|
SCaMPR Precipitation Estimates
The Self-Calibrating Multivariate Precipitation Retrieval (SCaMPR) algorithm, also called the Enterprise Rainfall Rate algorithm, attempts to effort to combine the separate strengths of infrared (IR)-based and microwave (MW)-based estimates of rainfall. In particular, IR data are available from geostationary Earth orbit (GEO) platforms with high spatial resolution (2 km immediately below the satellite) and very frequent updates (every 10 min over the Western Hemisphere) and the observations are sent to the ground within minutes. However, rain clouds are opaque in the IR, so rainfall information has to be inferred from cloud-top properties like temperature and texture. Meahwhile, rain clouds are semitransparent at MW frequencies, so MW radiances are sensitive to the total amount of water and / or ice in a cloud which gives them stronger relationship with rain rates. However, unlike IR data, MW data are available only from low-Earth- orbit (LEO) platforms, which typically pass over a particular point only twice per day.
Different approaches have been taken to combine IR and MW data for rain rate estimation. SCaMPR estimates rain rates from GEO IR data in order to take advantage of the high resolution, rapid refresh rate, and short delay of IR data) but the relationships are calibrated using MW-based rain rates to optimize their accuracy). The algorithm calibrates the rain / no rain discrimination and the rain rate using two separate steps and using discriminant analysis for the first and linear regression for the second. The predictors are also transformed by regressing them against the MW rain rates in log-log space to get bettere regression fits. Since regression assumes a normal distribution and can distort data that does not have a normal distribution (like rain rates), a final processing step involves computing initial rain rates from the calibration data and matching the Cumulative Distribution Function (CDF) with that of the microwave rain rates to create a lookup table (LUT) that adjusts the distribution of the SCaMPR rain rates to match that of the calibration microwave rain rates. Finally, to account for the effects of evaporation of hydrometeors below the cloud base, the average relative humidity (RH) over the lower third of the atmosphere from the Global Forecast System (GFS) model is used to adjust the rainfall rates downward in dry environments.
In its current form, SCaMPR uses 5 bands from 5 different GEO platforms. The instruments are the Advanced Baseline Imager onboard GOES-16 and -18, the Advanced Himawari Imager (AHI) onboard Japan's Himawari-9, and the Spinning Enhanced Visible Infra-Red Imager (SEVIRI) onboard EUMETSAT's Meteosat-9 and -10. The five bands used correspond roughly to 6, 7, 8.5, 11, and 12 micrometers, with the exact wavelengths depending on the instrument. In addition, two values that describe local texture of the 11-micrometer field are also used. These predictors are calibrated against the Climate Prediction Center (CPC) combined microwave (MWCOMB) data set.
Additional details on SCaMPR can be found in the following references:
Kuligowski, R. J., 2002: A self-calibrating real-time GOES rainfall algorithm for short-term rainfall estimates J. Hydrometeor., 3, 112-130.
Kuligowski, R. J., 2010: Rainfall Rate (QPE) Algorithm Theoretical Basis Document, NOAA / NESDIS / STAR., 44 pp.
Kuligowski, R. J., Y. Li, and Y. Zhang, 2013: Impact of TRMM data on a low-latency, high-resolution precipitation algorithm for flash flood forecasting. J. Appl. Meteor. Cli., 52, 1379-1393.
Kuligowski, R. J., Y. Li, and Y. Hao, and Y. Zhang, 2016: Improvements to the GOES-R rainfall rate algorithm. J. Hydromet., 17, 1693-1703.