Changing extremes at the extremes: how will climate change impact hydroclimatic extremes in the Himalayas?

The livelihoods of over 1 bn people inhabiting the Himalayan region and the downstream river basins are heavily dependent on agriculture, making them highly vulnerable to water shortages and flood events. In addition, hydropower is a major contribution to energy production. Mitigation against climate- and hydrology-related risks caused by climate change and exploitation of potential benefits of hydrological changes requires comprehensive climate and glacio-hydrological projections for the region.

However, providing these projections is still a major challenge, partly due to systematic biases in the output of global and regional climate models, which in turn can lead to unrealistic hydrological projections. Current projections for precipitation changes in the Himalayan region during the 21st century show a large spread over some areas, with possibilities across the full spectrum from wetter to drier conditions. The question is whether some of these projections are more likely than others. Although some of the spread may reflect fundamental and non-reducible uncertainties caused by random, natural climate variability, a large part of it may be due to differing sytematic errors in the climate models, which in principle can be reduced.

The aim of this project is to reduce the uncertainties in hydroclimatic projections for the Himalayan region by only using models that represent the key large-scale mechanisms well, rather than using a skill-independent full ensemble approach. This ‘emergent constraints’ approach will:

• identify the specific aspects of large- and meso-scale processes such as Indian Summer Monsoon and wintertime westerly disturbances that drive regional precipitation variability, with a focus on extremes;
• define suitable performance measures for these aspects to evaluate climate models against observations;
• drive glacio-hydrological models with the full ensemble of climate simulations, as well as with the well-performing subset, and thus understand and reduce the spread in hydroclimatic predictions;
• analyse whether the well-performing subset has smaller temperature biases and a better distinction between snow and rain.