Quality evaluation of the precipitation over the tropical oceans in the <scp>JRA</scp> ‐ <scp>3Q</scp> reanalysis

Abstract The Japan Meteorological Agency (JMA) has conducted the third Japanese global atmospheric reanalysis named the Japanese Reanalysis for Three Quarters of a Century (JRA‐3Q) using the JMA operational data assimilation system that has been upgraded and improved since the Japanese 55‐year Reanalysis (JRA‐55) was completed. One of the issues identified with JRA‐55 was an overestimation of precipitation especially over the tropical oceans. Hence we investigated the characteristics associated with daily precipitation bias, using the observational precipitation datasets, focusing primarily on the period of 2001–2010. Compared with JRA‐55, JRA‐3Q shows a closer spatial correlation with the observations, and also significantly reduces the overestimation of precipitation over both the eastern Pacific and the Indian Oceans. Composite analysis of the extreme overestimation cases in the eastern Pacific revealed a clear El‐Niño‐like distribution; i.e., a warmer sea surface temperature (SST), wetter conditions in the lower troposphere, and low sea level pressure (SLP) anomalies. In this region, both JRA‐3Q and JRA‐55 have a tendency to show a larger precipitation bias with increasing specific humidity at 925 hPa when the specific humidity exceeds 14.0 × 10 −3 kg·kg −1 . In contrast, over the Indian Ocean, no distinct relationship was evident between the precipitation bias and SST or lower‐tropospheric specific humidity compared with those in the eastern Pacific. Instead, the precipitation bias over the Indian Ocean is rather dependent on the intraseasonal oscillation (ISO) phases: extreme overestimation (extreme underestimation) tends to occur when the ISO in the region is in the inactive (active) phase. In particular, JRA‐3Q tends to overestimate precipitation during the boreal summers.