Inter-annual Variations of Tropical Ocean RainfallDifferences Among Satellite Estimates

The TRMM/GPM era (1998-2020) is examined in terms of inter-annual patterns and total mean amplitude of tropical ocean rainfall (25N-25S). Because much of global precipitation is concentrated over tropical oceans, rainfall variations there on an inter-annual time scale are significant and are related primarily to the El Nino-Southern Oscillation (ENSO) phenomenon and its accompanying SST variations. Understanding the rainfall variations and their relation to the SST variations is important to understand the underlying processes and how well models simulate these variations. From an observational standpoint we are dependent on satellite-based estimates over the open ocean due to the lack of accurate surface-based estimates. During the past 30+ years passive microwave (PMW) observations (e.g., from SSMI and SSMIS) have provided the basis for such estimates, along with accompanying precipitation radar observations from TRMM and GPM for the past 20+ years. However, comparison of these routinely used microwave-based estimates, with different instruments and/or different algorithms, show significant differences among them. The objective here is to examine these differences and provide direction toward the best estimate of the magnitude and evolution of these variation. Although the patterns of the ENSO rainfall variations are usually similar among the estimates, characteristics related to the amplitudes vary among the estimates: The TRMM/GPM radar-based inter-annual amplitudes are smaller with different evolutions than the PMW (GPROF) estimates from the same satellites; Different PMW algorithms produce different amplitudes, even applied to same instrument (SSMI/SSMIS). Examples here are GPROF (larger) and METH (smaller, used in GPCP). GPROF applied to different instruments provides different amplitudes. GPROF (SSMI/SSMIS) has larger amplitude than GPROF (TMI/GMI), the latter agreeing better with METH (and GPCP). IMERG variations are irregular due to inhomogeneous input data sets. The different results estimated by the different algorithms/instruments are compared to available inter-annual estimates from water budget analyses as a possible constraining comparison to help arrive at a best estimate, which is important for incorporation into frequently used analyses such as GPCP.