Climate models already predict a link between rising temperatures and more frequent and heavier rainstorms, but two recent studies that analysed historical weather data have confirmed the relationship.
The results should increase confidence in climate simulations but they also raise the uncomfortable possibility that models have underestimated the increasing frequency of extreme rainfall. This could have serious consequences for flood management.
The discrepancy may be particularly marked for extreme summer rain, like that which caused extensive flooding last year. However, more research will be needed to decide if the discrepancy is genuine.
According to basic physical laws, the frequency of extreme precipitation should increase by about 7% for each degree of warming. This is mostly because hot air can hold more water.
Geert Lenderink and Erik van Meijgaard, of the Royal Netherlands Meteorological Institute, looked at a 99-year rainfall and temperature record from De Bilt in the Netherlands. They found that above 12°C, hourly rainfall extremes went up by twice the expected 7% per degree.1 On the other hand, increases in daily extremes - which may be more relevant for flood planning - were less well defined.
Climate model scenarios broadly reproduced their observations, but performed less well above 20°C.
In the second study, satellite data on tropical rainfall and sea-surface temperature were compared to model simulations. Once again, high temperature rainfall extremes exceeded the 7% per degree increase.2 While simulations qualitatively reproduced these changes, models underestimated the rate of increase for the heaviest rain by two to three times.
The UK water industry is keen to make use of climate change scenarios for planning. In its second report on last summer’s floods, Water UK’s flooding review group acknowledged the likelihood that extreme rainfall will occur more often (see p 5 ). Working out just how much more often is important because sewers are designed to cope with a one-in-30-year storm event.
The UK Climate Impacts Programme is due to release updated climate scenarios in November. These new UKCIP08 scenarios are being calculated using several climate models and will present the full range of uncertainties for outcomes like rainfall, which should aid planning for extremes.
Simon Brown, climate extremes research manager at the Met Office, highlighted the differences between trends in daily and hourly extremes seen by Dr Lenderink. Although he acknowledges uncertainty in how well rainfall extremes are represented in the forthcoming UKCIP08, he said "I would not take these recent papers as undermining the current projected changes in extreme rainfall, for durations of greater or equal to one day."
Similarly Richard Allan, lead author of the second study, is clear that planning should not be based on his work’s results. "Differences can relate to deficiencies in the measurements, or the models used to predict future climatic change."