High-resolution projections of ambient heat for major European cities using different heat metrics

Abstract. Heat stress in cities is projected to strongly increase due to climate change. The associated health risks will be exacerbated by the high population density in cities and the urban heat island effect. However, impacts are still uncertain, which is among other factors due to the existence of multiple metrics for quantifying ambient heat and the typically rather coarse spatial resolution of climate models. Here we investigate projections of ambient heat for 36 major European cities based on a recently produced ensemble of regional climate model simulations for Europe (EURO-CORDEX) at 0.11∘ spatial resolution (∼ 12.5 km). The 0.11∘ EURO-CORDEX ensemble provides the best spatial resolution currently available from an ensemble of climate model projections for the whole of Europe and makes it possible to analyse the risk of temperature extremes and heat waves at the city level. We focus on three temperature-based heat metrics – yearly maximum temperature, number of days with temperatures exceeding 30 ∘C, and Heat Wave Magnitude Index daily (HWMId) – to analyse projections of ambient heat at 3 ∘C warming in Europe compared to 1981–2010 based on climate data from the EURO-CORDEX ensemble. The results show that southern European cities will be particularly affected by high levels of ambient heat, but depending on the considered metric, cities in central, eastern, and northern Europe may also experience substantial increases in ambient heat. In several cities, projections of ambient heat vary considerably across the three heat metrics, indicating that estimates based on a single metric might underestimate the potential for adverse health effects due to heat stress. Nighttime ambient heat, quantified based on daily minimum temperatures, shows similar spatial patterns to daytime conditions, albeit with substantially higher HWMId values. The identified spatial patterns of ambient heat are generally consistent with results from global Earth system models, though with substantial differences for individual cities. Our results emphasise the value of high-resolution climate model simulations for analysing climate extremes at the city level. At the same time, they highlight that improving the predominantly rather simple representations of urban areas in climate models would make their simulations even more valuable for planning adaptation measures in cities. Further, our results stress that using complementary metrics for projections of ambient heat gives important insights into the risk of future heat stress that might otherwise be missed.