Comparison of Level 3 mean monthly GPROF products from GPM and TRMM microwave imager in estimating seasonal precipitation

Reliable and accurate measurement of precipitation plays a critical role in the studies of meteorology, hydrology and water resource management. In the past decades, satellite-based quantitative precipitation products have provided a suitable means to measure precipitation from space. An example is the TRMM (Tropical Rainfall Measuring Mission) microwave imager (TMI) products. It is well known that TMI products have an insensitivity to light precipitation and significant underestimation to heavy precipitation. In early 2014, National Aeronautics and Space Administration (NASA) and Japan Aerospace Exploration Agency (JAXA) successfully launched a new generation of satellite, Global Precipitation Measurement (GPM) Core Observatory. Onboard the GPM core satellite has a multichannel GPM microwave imager (GMI) which uses 13 different microwave channels to observe energy from the different types of precipitation through clouds for estimating everything from low to high precipitation. At present, several composite precipitation products with algorithms that related GMI and partner passive microwave sensors have been released, including Level 2 gridded Goddard Profiling GMI (2GPROF-GMI) and Level 3 mean monthly GPROF-GMI (3GPROF-GMI) products. It is therefore important to investigate whether these new GMI products are more reliable in estimating precipitation than those of TMI. In this study, differences between 3GPROF-GMI and 3GPROF-TMI products were compared over different surface types on a near-global scale for different seasons. The results show that a systematic difference (3GPROF-GMI > 3GPROF-TMI) for low and high precipitation. High positive correlation coefficient (CC) values are mainly in the Caribbean region, South Atlantic and over the African continent. Low CC values concentrate in equatorial regions, North Atlantic and Northwest Pacific. In addition, high mean absolute difference (MAD) and root mean square difference (RMSD) values dominate over the Tropics. By contrast, low negative MAD and RMSD values are found in the Inter-Tropical Convergence Zone (ITCZ). Histograms of both products are very similar; however higher frequencies for 3GPROF-GMI are found in the low and high precipitation ranges than those of 3GPROF-TMI. Statistics for different precipitation ranges reveal more details on the systematic differences over land and ocean. For low and relatively low precipitation, the MAD and RMSD values over land are slightly lower than those over oceans; on the other hand, the CC values over land are larger than those over oceans. For medium precipitation, the MAD, RMSD and CC values over oceans are in general higher than those over land in all seasons; but for relatively high precipitation, the opposite is found. The results of this study enhance the understanding of the difference between these two products, and provide a viable validation analysis for the updated GMI products.