A high-resolution, multi-model analysis of Irish temperatures for the mid-21st century

Abstract

There is a paucity of dynamically downscaled climate model output at a high resolution over Ireland, of temperature projections for the mid-21st century. This study aims to address this shortcoming. A preliminary investigation of global climate model (GCM) data and high-resolution regional climate model (RCM) data shows that the latter exhibits greater variability over Ireland by reducing the dominance of the surrounding seas on the climate signal. This motivates the subsequent dynamical downscaling and analysis of the temperature output from three high-resolution (4–7 km grid size) RCMs over Ireland. The three RCMs, driven by four GCMs from CMIP3 and CMIP5, were run under different Special Report on Emissions Scenarios (SRES) and representative concentration pathway (RCP) future scenarios. Projections of mean and extreme temperature changes are considered for the mid-century (2041–2060) and assessed relative to the control period of 1981–2000. Analysis of the RCM data shows that annual mean temperatures are projected to rise between 0.4 and 1.8 °C above control levels by mid-century. On a seasonal basis, results differ by forcing scenario. Future summers have the largest projected warming under RCP 8.5, where the greatest warming is seen in the southeast of Ireland. The remaining two high emission scenarios (SRESs A1B and A2) project future winters to have the greatest warming, with almost uniform increases of 1.5–2 °C across the island. Changes in the bidecadal 5th and 95th percentile values of daily minimum and maximum temperatures, respectively, are also analysed. The greatest change in daily minimum temperature is projected for future winters (indicating fewer cold nights and frost days), a pattern that is consistent across all scenarios/forcings. An investigation into the distribution of temperature under RCP 8.5 shows a strong summer increase compounded by increased variability, and a winter increase compounded by an increase in skewness.