Climate models predicted the 2026 El Niño during early spring, just as they had in 2023. And the predictions have turned out to be accurate, with the robust evolution of a strong El Niño by the summer.
Models had predicted the 2023 monsoon to be slightly deficient, and the final seasonal total, at 94% of the long-period average, essentially validated the forecasts. Monsoon predictions this year have already called for a deficit of greater than 10% and it will also likely hold true.
But the real cause for concern is the erratic, and unprecedented, evolution of the monsoon across space and time in 2026, which the models have not been able to capture even a few days in advance.
The key question now is whether the challenges posed by the 2026 predictions were an aberration, and, if so, why. In other words, do the models need to do anything dramatic to ensure they do better in the coming years? Or is it just a trend to be expected for the monsoon system, which has only become more notorious for its vagaries?
Expectations of improvements
The El Niño is one of the most predictable modes of natural variability. But models being able to predict its occurrence to the tune of 80% still means they will be wrong one-fifth of the time. The failed La Niña prediction of 2024 is a good example. Models’ ability to predict the monsoon is much lower, hovering at around 60%, which means prediction failures will also occur more often. And this is for the all-Indian monsoon rainfall (AIMR); long-lead forecasting of monsoon onset and space-time evolution are even greater challenges.
This year’s monsoon onset over Kerala was predicted to be delayed, and it was, by a few days. But everything after that has been a whiplash. Rainfall in June was 40% below normal but July recovered quickly to 1% above normal. At the same time, these AIMR values hardly captured the heartache of farmers, who suffered much larger deficits over their arable land.
India’s multi-tiered prediction systems cover short (days 1-3), medium (days 3-10), and extended (weeks 2 and longer) range demands for weather information. Seasonal to interannual predictions of AIMR are relatively easier; long-lead predictions of spatial and temporal monsoon evolutions face irreducible uncertainties at present.
While one bad year is no reason to panic, it must also serve as an opportunity to identify the key issues that remain. For instance, are we going to have more years of such erratic monsoon evolution? Will combinations of the El Niño and global warming make the monsoon unpredictable at subseasonal and local scales? And how will models prepare for such a future, to better help the country maintain its progress towards becoming economically developed?
Knowns and unknowns
The monsoon domain is blessed with more than a century’s worth of data. Experts have extracted significant knowledge from this rich dataset of monsoon variabilities in space and time, and have made considerable advances in understanding the intrinsic timescales of active-break cycles, the onset becoming more erratic, delayed withdrawal, switching between extreme dry and wet events, more frequent extreme wet spells over Northwest India, and so on. However, their understanding of the mechanisms underlying these events remains incomplete.
The hope is that complete climate models with coupled land-ocean-atmosphere components can shed light on the missing links and help scientists make better, more skillful predictions. The multi-tiered approach taken to develop separate tools for short, medium, and extended range predictions has yielded many impressive advances in useful predictions for various sectors, including agriculture, water, energy, health, and so on.
At present, the most promising avenues for further advances in improving models and predictions are likely to come from using artificial intelligence (AI) tools. The dearth of mechanistic understanding is not a limitation for AI since these tools can extract patterns from the data and based on that make predictions, without any knowledge of the governing physics. That said, experts must overcome one critical limitation in the course of combining AI with climate models: they must ensure they collect sufficient data to cover all aspects of the monsoon and its drivers at all the relevant spatial and temporal scales. In fact, this data must also account for the local amplifiers of extreme weather, such as land use change, urbanisation, deforestation, and irrigation.
That said, given new technologies and platforms, the volume and value of data will only grow. And AI models are already advancing optimal and cost-effective observation networks and improving our knowledge of processes, models, and predictions. AI is also proving to be indispensable for bringing global predictions at the scale of several kilometres down to farm and neighbourhood scales, for sector-specific advisories and early warnings.
Public-private partnerships are also part of the landscape today when it comes to delivering weather and climate services.
Predictions and expectations
As predictions get better, so do the expectations. Specifically, advances in predictions related to the monsoon, including its onset, evolution, extreme dry/wet events, and withdrawal, also allow the users of these data to believe they can expect even better in future.
To this end, hybrid dynamic and AI models could deliver sector-specific predictions at the requisite space-time scales for users to better manage disasters, farm-related decisions, and water and energy use. Extensively engaging users with sustained feedback can pave the way for decision-support products to be co-produced with the people who make those decisions.
Predictions will never be perfect but they can continue to serve user needs by staying within the level of uncertainty a user can tolerate. In the coming years, policymakers, lawmakers, farmers, health workers, energy companies, and insurance providers can expect to be served bespoke forecasts. Global warming will keep throwing monkey wrenches into prediction systems but observations, dynamic models, and AI will continue to step up to the challenges.
Raghu Murtugudde is an Earth Systems Scientist and a visiting professor at IIT-Kanpur.
Published – August 26, 2026 09:00 am IST
