The Indian summer monsoon feeds more than a billion people, yet it distributes rain unevenly: some regions receive over a meter of rainfall in a season, others three times less. The all-India rainfall index, which averages millimeters across the country, has been criticized for erasing this mosaic. Alongside this work in the archive lie studies on changing rainfall frequency and intensity, on moisture over Tibet, and on aircraft data. Our article stands somewhat apart: it does not examine trends or new data sources but reconsiders the tool itself, the all-India index, using the longest and most detailed observational record.
The authors took daily rainfall data from the India Meteorological Department from 1901 to 2020 on a 0.25-degree grid, derived from a network of about 2600 rain gauges per year. For each summer (June–September), they computed seasonal totals, the number of rainy days (threshold 1 mm/day), and mean rainfall intensity. They then constructed six alternative all-India metrics: normalized anomalies, the fraction of wet points, mean rainy-day count, mean intensity, mean anomaly over wet points only, and over dry points only. Links to El Niño and the Indian Ocean Dipole were assessed using the NINO3.4 and DMI indices, and the El Niño influence was removed via linear regression.
The all-India rainfall index correlates more strongly with the fraction of wet points and the mean number of rainy days than with rainfall intensity: correlation coefficients above 0.9 versus at most 0.77. The first mode of rainfall variability (18.2% of variance) is a quasi-uniform wetting or drying, closely tied to rainy-day count and El Niño. The second mode (9.3%) is a tripole: positive anomalies over the Western Ghats and Central Monsoon Zone, negative over the southeast. The Indian Ocean Dipole projects onto this tripole, so its correlation with the all-India index is nearly zero (r = 0.01). However, after removing the linear El Niño signal, the residual dipole's correlation with the index becomes significant: r = 0.21.
This work resolves three long-standing disputes. First, the all-India index indeed reflects the spatial extent of anomalies rather than their strength, explaining its high correlation with crop yields. Second, the use of coarse grids in the past distorted the picture: with detailed data, it is clear that the Western Ghats and the Central zone vary coherently, not independently. Third, the Indian Ocean Dipole does not ignore the overall index but acts through the tripole, and its contribution emerges once El Niño is accounted for. Compared with neighboring studies on rainfall trends, this article adds no new data but redefines the meaning of an old index on the longest record.
Ahead lies testing how intraseasonal oscillations (active and break phases) nonlinearly combine into seasonal anomalies. It is also important to understand why El Niño suppresses the number of rainy days rather than their intensity. This may be linked to increased static stability of the tropical atmosphere as the upper troposphere warms. Further research could apply the same frequency–intensity decomposition to seasonal forecasting.
The results directly affect agrometeorology and economic planning in India, where the all-India index serves as an indicator of droughts and floods. They also refine understanding of global teleconnections in the tropics.
The next step is to repeat the analysis on an ensemble of climate models to see whether they reproduce the identified relationships. Then one can build seasonal forecasts based on separate prediction of rainfall frequency and intensity.
For meteorology, the key unresolved problem is the predictability of intraseasonal monsoon variability, which remains chaotic. This work shows that at the seasonal scale, a link to slow ocean modes does exist, but its strength is limited. It remains an open question what fraction of rainfall variance is predictable in principle.
During the monsoon season in the Western Ghats, rain falls almost every day, up to 120 consecutive days, while in southeastern India the wet season arrives only in autumn, during the northeast monsoon.