Assessment of climate change impact on maximum precipitation in Spain

Abstract

AbstractThe potential effect of climate change on maximum precipitation is studied by considering several temporal intervals for its incorporation into flood risk management in Spain. The methodology used allows for a complete assessment of the effect of climate change on maximum precipitation by accounting for uncertainty, and may be applied to other case studies. Behaviour regarding observations, trends and changes in climate model maximum precipitation series are evaluated. In turn, multi‐model mean (and percentile) relative changes accounting for statistical significance, torrential rate changes, and relative changes in daily annual maximum precipitation accumulated along river network are estimated. Results show a large variability in climate model outcomes. In general, a suitable behaviour is found for the climate model ensemble regarding observations: it properly represents seasonality of daily annual maximum precipitation for all regions, and observation‐based statistics are found to be within range of simulation‐based statistics for most regions. As an exception, daily annual maximum precipitation is underestimated in Mediterranean regions, which may be attributed to known limitations of current climate models in adequately reproducing convective precipitation typical of those areas. Results also show an intensifying increase in trends and changes as precipitation duration decreases, a torrentiality rise, and relative changes related to accumulated maximum precipitation of up to 35%. These findings support the notion that the effect of climate change may be stronger for extreme precipitation associated with short precipitation durations. This may be attributed to a stronger scaling between change in extreme precipitation and mean surface temperature increase for shorter durations. The findings also highlight the importance of analysing statistical significance of relative changes in quantile for allowing its attribution to climate change. Statistical model choice is found to be essential for identifying significant changes.