"As India's economy today relies heavily on rainfed agriculture, more frequent and severe monsoon failure would be an immediate danger to the livelihoods of a large share of its population," said Jabob Schewe of Potsdam Institute for Climate Impact Research, Germany. "Note, however, that in our model study, monsoon failure gets more frequent only in the more distant future; such a scenario, if deemed realistic, would probably rather have implications for long-term adaptation planning."
Together with Anders Levermann from the Potsdam Institute for Climate Impact Research and Potsdam University, Germany, Schewe used the MPI-ESM climate model to simulate the years 800–2200. The researchers used the IPCC SRES A1B scenario for carbon emissions from 2005 until 2100 and constant carbon-dioxide concentrations up to 2200.
Rainfall in South Asia is likely to be strengthened during much of the 21st century because atmospheric warming will enable more moisture to be held in the atmosphere, they found. But once the Walker circulation weakens, rainfall will drop off rapidly, falling short of the pre-industrial long-term mean roughly by the turn of the 22nd century. And from 2150 to 2200, dry years will be much more frequent than wet.
"In a nutshell, our results indicate that the response of monsoon rainfall to global warming can be complex, including periods of both strengthening and rapid weakening; and that nonlinear internal dynamics of the monsoon system may play an important role in its response to external changes," said Schewe.
Typically, summer-monsoon rain occurs in South Asia from June until September, when winds blow from the southwest Indian Ocean onto the continent. From about October the wind direction reverses. By modelling past data, the researchers found that there are two monsoon modes.
In the first, spring sees the troposphere over land being warmer than that above the ocean. Moist air moves over the continent and the resulting rainfall releases latent heat. This reinforces the tropospheric temperature contrast, and stabilizes the air circulation, in what the team dubbed the wet moisture-advection feedback.
In dry years, in contrast, less humid air in the upper troposphere tends to sink over both land and the Arabian Sea. This suppresses rainfall and keeps sea-level pressure anomalously high throughout the summer. As a result, the direction of the monsoon flow in the upper troposphere tends to change from generally westward to northward or even eastward. In the same way that rainfall sustains a wet monsoon regime, the team believes this anomalous circulation pattern may be a self-amplifying feedback that can sustain a dry-monsoon regime – the "dry-subsidence" feedback.
Incorporating these two feedbacks into a model enabled the team to statistically predict changes in the frequency distribution of seasonal mean all-Indian rainfall. The model was forced only by global mean temperature and the strength of the Pacific Walker circulation in spring.
The Walker circulation is a large overturning circulation in the atmosphere, oriented along the equator and spanning the Pacific and Indian Oceans, explained Schewe. Its interannual variations form the atmospheric part of the El Niño/Southern Oscillation (ENSO) phenomenon. As temperatures rise, the Walker circulation will, on average, bring more high pressure over India, suppressing the monsoon, as happens today in El Niño years.
"We plan to take the work forward by considering other climate models in order to test the robustness of the monsoon response found in our study, and also the applicability of the statistically predictive model," said Schewe. "Should it turn out that a nonlinear response of monsoon rainfall to global warming is indeed a robust feature of climate-change simulations, and that monsoon failure of the order of magnitude that we find in our study cannot be excluded in the future, then of course that would have far-reaching implications for adaptation strategies in India and other monsoon regions."
The team reported the study in Environmental Research Letters.