Batten down the hatches
Recent research from the Met Office Hadley Centre predicts British summers are likely to have more heavy downpours as a result of climate change. Elizabeth Kendon, Senior Climate Scientist at Met Office Hadley Centre, explains the findings of the recent study.
Our study shows the first evidence that summer downpours in the UK could become heavier with climate change. We used a very high-resolution model more typically used for weather forecasting to study changes in hourly rainfall. Unlike current climate models, it has a fine resolution and is able to represent hourly rainfall, enabling us to make future projections with some confidence.
We found that summers are likely to become drier overall by 2100, in a warming climate. But our results suggest that when it does rain, it will be heavier in short outbreaks. In particular, intense rainfall with the potential to cause serious flash flooding could become a more common occurrence.
The study provides a much more complete picture of how UK rainfall may change in the future. Climate models generally work at coarse resolutions, using grids of around 12 km square or larger. These have been able to accurately simulate winter rainfall, which generally comes from sustained, long-lasting periods of rain from large-scale weather systems. These models point toward wetter winters, with the potential for greater daily rainfall in the future.
But summer weather is harder to predict using such coarse models. It is changes on an hourly basis that are important, as rainfall tends to come in short but intense bursts during the summer - as seen during the Boscastle flooding of 2004 and "Toon Flood" in Newcastle in 2012. So far, climate models have lacked the resolution to accurately simulate the smaller-scale convective storms (intense showers formed by rising air) which cause this type of rain. To deal with this issue, our study uses the most high-resolution model ever used before in long climate simulations to examine rainfall change, based on a 1.5 km square grid (the same as the Met Office weather forecast model for the UK), leading to much higher accuracy.
We ran this model to simulate two 13-year periods; one based on the current climate and one based on the climate at the end of the century under a high-emissions scenario. The simulations were so computationally intensive that it took the Met Office's supercomputer - one of the world's most powerful - about nine months to run the simulations, and even then we could only run the model for the southern half of the UK, about as far north as Manchester.
The simulation showed increased hourly rainfall intensity during winter, consistent with the simulations for the future provided by coarser resolution models and previous studies looking at changes on daily timescales. However the finely grained model also revealed that short duration rain will become more intense during summer, something that the coarser model was unable to simulate.
This finding is of major importance due to the potential for flooding: a threshold of 30 mm per hour is used by the Met Office and Environment Agency Flood Forecasting Centre as guidance to indicate likely flash flooding. Our results suggest this may be exceeded more often (up to five times) and over a wide area in the future.
Our findings are only the results of one climate model so we need to wait for other similarly detailed simulations to see whether the results support these findings. However, an increase in summer storms in a warmer, moister environment is consistent with theoretical expectations, and with the limited observational studies we have of hourly rainfall.
This work is part of the joint Met Office and NERC-funded CONVEX project. The next steps are to see if the results are consistent with observations and predictions of hourly rainfall from climate models in other parts of the world, to be undertaken by the European Research Council-funded INTENSE project jointly run by Newcastle University academics in collaboration with the Met Office and other international scientists.