Satellite monitoring of summer heat waves in the Paris metropolitan area

Summer warming trends in Western Europe are increasing the incidence, intensity and duration of heat waves. They are especially deadly in large cities owing to population density, physical surface properties, anthropogenic heat and pollutants. In August 2003, for 9 consecutive days, the Paris metropolitan area experienced an extreme heat wave that caused 4867 estimated heat-related deaths. A set of 61 NOAA-AVHRR (advanced very high-resolution radiometer) images and one SPOT-high resolution visible (HRV) image were used to analyse the spatial variations of land surface temperature (LST) over the diurnal cycle during the heat wave. The LST patterns were markedly different between daytime and night-time. A heat island was centred downtown at night, whereas multiple temperature anomalies were scattered in the industrial suburbs during the day. The heat wave corresponded to elevated nocturnal LST compared to normal summers. The highest mortality ratios matched the spatial distribution of the highest night-time LSTs, but were not related to the highest daytime LSTs. LSTs were sampled from images at the addresses of 482 elderly people (half were deceased persons and half were control ones) to produce daily and cumulative minimal, maximal and mean thermal indicators, over various periods of time. These indicators were integrated into a conditional logistic regression model to test their use as heat exposure indicators, based on risk factors. Over the period 1-13 August, thermal indicators taking into account minimum nocturnal temperatures averaged over 7 days or over the whole period were significantly linked to mortality. These results show the extent of the spatial variability in urban climate variables and the impact of night-time temperatures on excess mortality. These results should be used to inform policy and contingency planning in relation to heat waves, and highlight the role that satellite remote sensing can play in documenting and preventing heat-related mortality. Copyright. (C) 2010 Royal Meteorological Society