Now the scientists have produced data-based maps to show whether certain rainfall distributions can be reliably attributed to local climate change. Local climate changes are observable in parts of Scotland, the Dordogne in France, Tuscany in Italy and the Low Countries such as the Netherlands and Belgium, but not in other parts of Europe, they found. The results could be important for those seeking to prepare for changes in local climate.

"In some places there is a clear signal of what has changed [climatically], and in other places it is uncertain," said Sandra Chapman of the University of Warwick, UK. "It is as important to know where the change is uncertain as to know what the change is."

In recent years, scientists have placed more focus on local, rather than global climate change. The idea is to understand how changes will affect people, and what can be done in response.

"All of this is done from the observations, we do not use a model – that is what is different from existing methods," said Chapman. "The observations tell us what the uncertainty is."

Unfortunately, local datasets are limited in their historical reach. This raises the question of whether an apparently unusual season of weather in one year is due to climate change, or whether it is a result of natural variability.

Some recent studies have claimed to identify changes in temperature due to local climate change. But, as Chapman explains, rainfall is more changeable. "The difficulty with precipitation is that there are more larger events," she said. "These are harder to resolve statistically."

Chapman and her colleagues got around this problem by examining seasonal distributions of rainfall between the years 1950 and 2012 across Europe. Within this period, they compared rainfall distributions 45 years separate from each other. In this way, for every locale they could generate not just one but several estimates of a 45-year change in rainfall distribution.

Having several estimates meant that the researchers could tell whether the changes in distribution could be identified as climate change, or simply natural variation. They plotted their results on a map, with red regions showing where climate change was identifiable and white regions – the majority – showing areas where it was not, for a certain threshold of change.

Chapman said the method could help those deciding whether or not to respond to climate change in a certain area, for instance by building a dam. Her group is now planning to further the work by looking at other datasets.

Related links

Related stories