Comparison of TRMM and Rain Gage Rain Rates over New Mexico

After the successful launch of the Tropical Rainfall Measuring Mission (TRMM) in November 1997, close to three years of data have been collected. Of the three rain instruments, the Precipitation Radar (PR), the TRMM Microwave Imager (TMI), and Visible Infrared Scanner (VIRS), standard rain products are derived from PR, TMI and from a combination of the PR and TMI (TRMM Combined Instrument, or TCI). The major reprocessing (version 5) of TRMM data was completed in May 2000.

Preliminary results indicate that while the differences between PR and TCI rain rate are small, the difference between TMI and TCI is about 10-15% globally. As part of an effort to promote the use of remote sensing data to the land application communities, we compare the monthly mean rain rates derived from TMI and TCI over New Mexico. There is a systematic difference between TCI and TMI rain rate. We also found a substantial improvement from the version 4 to the version 5 of TMI rain rate. This difference is mostly attributed to an algorithm change. Other rain rate parameters, such as fractional rain rate, conditional and unconditional rain rates and their monthly variance are also computed. There is a high correlation between the fractional rain rate and unconditional rain rate, consistent with an established relation between areal rainfall and rain area and the area time integral (ATI) technique in radar meteorology. With an inclination of 35 degrees, New Mexico is situated at the northern edge of the TRMM satellite swath, and hence sampling is another source of error.

A network of automatic weather stations was set up by Earth Data Analysis Center (EDAC) in the Sevilleta NWR/LTER area in New Mexico. This network consists of 28 stations instrumented with tipping rain gages and temperature sensors. The network is designed to capture rain variability within a TMI and TCI pixel, thus allowing the examination of the within pixel variability. The rain data are time-recorded by event and temperature data are recorded at 5-minute intervals. Preliminary results show that the spatial pattern of accumulated rainfall is highly dependent on topography. We will examine the spatial variability of rain rate and compare the gage with TRMM overpasses. Statistics such as probability of detection, false and non-detection rates will be computed. The high spatial variability and dependency of rainfall on elevation has strong implication for future satellite microwave rainfall estimation over land, such as the Advanced Microwave Scanning Radiometer (AMSR) on board the Aqua satellite and microwave sensor to be flown by the Global Precipitation Mission (GPM).