Researchers assess heatwave risks associated with climate change
Combining climate and mortality data, researchers have estimated that 315 deaths in Greater London and 735 deaths in Central Paris can be strongly linked to the 2003 heatwave that set record-breaking temperatures across Europe. Taking their analysis a step further, they determine that 64 (± 3) deaths from the London dataset and 506 (± 51) deaths from the Paris dataset are attributable to anthropogenic climate change, which increased the risk of heat related mortality by 20% and 70%, respectively, in the two cities. The team, led by scientists from the University of Oxford and Public Health England, has reported its latest findings in the journal Environmental Research Letters.
Quantifying the impact of changes in climate on mortality will allow agencies to make more accurate emergency plans for extreme weather events and build a stronger case for longer term action, but there are hurdles to overcome.
"Deaths are usually recorded as being due to a specific cause such as cardiovascular disease or cancer, which makes it difficult to make a direct connection between deaths recorded during a heatwave and exposure to heat specifically," explained Clare Heaviside of Public Health England's Centre for Radiation, Chemical and Environmental Hazards. "So in order to calculate the number of deaths which can be attributed to the heatwave, we use a statistical relationship which relates increases in temperature above a certain threshold, to an increase in the number of deaths typically expected on a particular day of the year."
The temperature-mortality relationship depends on many factors. To improve the quality of the analysis, data is calculated separately for each region or city. The next step is to establish the coefficient relating high temperature and increased mortality and apply this expression to a period of hot weather (in this case the summer of 2003) to estimate how many deaths during this timescale can be attributed to heat.
To establish the proportion of these deaths that are due to human factors such as raised atmospheric greenhouse gas concentrations (GHGs), the scientists run climate models for the baseline year (2003) according to different scenarios reflecting 'actual' and 'natural' conditions. "Under the simulations with pre-industrial levels of GHGs, the probabilities of extreme temperatures are lower," commented Heaviside. "To determine the influence of changed atmospheric gas composition on heat-related mortality, we combine the climate projections with our statistical mortality relationship."
Thanks to a large ensemble of simulations the team is able to capture the statistical distribution of extreme temperature events in great detail, but the researchers couldn't complete this task without the help of the public. "Our analysis is based on a citizen science activity, where individuals donate spare computing power to perform simulations as a screensaver, returning projections to the University of Oxford once the climate model has finished running," said Daniel Mitchell of the Environmental Change Institute at the University of Oxford. "The total computing time to run all the simulations would be around 250 years if you had to perform the calculations on a single home PC."
Using the modelled data, the group has the potential to extend its analysis to include other cities, and also examine human factors in more detail. "In the future, we may investigate the partitioning of emissions between different sectors – for example, transport, industry and others, to see how the various sectors contribute to the effects of climate change," added Heaviside.