The 1951-2000 heavy-rain trend for central India matched the 2001-2025 IMD mean at 1-degree scale only with the 2005-2007 seasons; 0.25-degree counts stayed near the 1976-2000 level

Goswami et al. (2006) reported more heavy monsoon rain events over central India from 1951 to 2000. Lin and Huybers (2019) said that a station-network change after 1975 can cause part of this rise. We did an out-of-sample test with IMD 0.25-degree daily rainfall for 2001-2025. We used the Goswami definitions (June-September, 74.5-86.5E and 16.5-26.5N, 100 and 150 mm/day). We fitted a stationary mean, a log-linear trend and a step at 1976 to 1951-2000 only. For 1-degree box means, the 2001-2025 heavy-event mean was 0.97 of the trend forecast (95% season-bootstrap interval 0.80-1.17) and 1.39 of the step forecast (1.15-1.67). Without the three wet seasons 2005-2007, the ratios were 0.82 (0.72-0.92) and 1.18 (1.04-1.33). For 0.25-degree cells, the counts were near the step forecast and below the trend forecast. Very heavy cell events were 0.80 of the trend forecast (0.65-0.97), and 0.67 without 2005-2007. Inside 2001-2025, heavy counts had no clear trend (-8% per decade, -32% to +23%). Moderate events stopped their decline, and seasonal rain did not change. Box-mean extremes rose faster than cell extremes. A station-network proxy (local extrema in the interpolated fields) shows the 1971 station loss and a thinner network in 2005-2015. It explains about +3% to +12% of the +39% rise in 1-degree heavy events from 1976-2000 to 2001-2025. PERSIANN-CDR agrees with IMD seasonal totals (r = 0.92), but its daily extremes fall after 1997, so it cannot check the trend. Limitations: one gridded product, no public station counts, an indirect network proxy, and results that depend on a few seasons.

Indian monsoon · extreme rainfall · central India · IMD gridded rainfall · out-of-sample test · station network · heavy rain trend · PERSIANN-CDR · forecast verification · climate data homogeneity