Cheng, Si , Alexander, Lisa V. , Sherwood, Steven , Blanco, Joaquin E.
2026-01-01 JOURNAL OF CLIMATE 2026 39(卷), 1(期), (281-296页)
Understanding changes in global oceanic precipitation remains challenging due to observational limitations and model deficiencies, particularly in representing cloud and precipitation properties over oceans. In this paper, climatologies and trends in oceanic precipitation are examined using a collection of 27 state-of-the-art satellite and reanalysis data-sets available on a uniform daily 1 degrees 3 1 degrees resolution from the Frequent Rainfall Observations on Grids (FROGS) database between 2001 and 2020. Reanalysis datasets generally report higher annual-mean daily precipitation than satellite datasets. The humid tropics exhibit the greatest absolute discrepancies in precipitation rates, while arid regions such as the southeast Pacific and Atlantic show substantial relative differences among products. An upward ocean-mean trend is observed in most of latest-version satellite products, whereas reanalyses suggest declining trends. We also assess the precipitation changes in different locations against the "wet gets wetter, dry gets drier" (WWDD) hypothesis. Reanalyses on average show a pronounced decrease over the intertropical convergence zone, whereas the latest-version satellite products on average more closely follow the WWDD pattern, with agreements over more than half of the oceanic regions. The precipitation trend in averaged reanalyses also exhibits the weakest consistency with sea surface temperature trends in wet regions (34.2%), compared with dry regions in the averaged reanalyses (53.4%) and both wet (59.6%) and dry (58.5%) regions in averaged latest-version satellite products. Overall, the latest-version satellite trends, while diverse across products, tend to align better on average with this simple hypothesis than reanalyses. We recommend using the latest-version satellite products for investigating global oceanic precipitation while exercising greater caution when utilizing reanalysis datasets.