Abstract
Urban heat stress is intensifying across mid-latitude Europe, yet districts-scale projections remain the missing link between regional climate modelling and actionable adaptation planning. This study presents a multi-method CMIP6 climate projection for Dunajská Streda, a rapidly urbanising lowland district in southwestern Slovakia, using two parallel downscaling approaches: the LARS-WG stochastic weather generator driven by a 16-model CMIP6 ensemble, and a bias-corrected direct ensemble of 18 CMIP6 general circulation models. Both approaches are applied under SSP2-4.5 and SSP5-8.5 scenarios for 2021–2070, and are independently evaluated against the ClimRisk.cz multi-model ensemble. Projections are derived from the Gabčíkovo meteorological station and represent peri-urban background conditions. Actual urban exposure will be higher owing to Urban Heat Island (UHI) amplification. Both approaches project robust, mutually consistent warming. Mean annual temperatures are projected to increase by +1.8 to +3.4 °C by 2051–2070 relative to each method’s observational baseline. Tropical days rise from 29–30 days/year at baseline to 53–70 days/year by 2051–2070 under SSP2-4.5 and SSP5-8.5, respectively. Tropical nights increase from 2 nights/year at baseline to 11–27 nights/year by 2051–2070, an approximately 5- to 14-fold amplification. Frost days decline by ~25–54% and icing days by up to ~83%. Combined with the district’s extensive impervious surfaces and limited urban vegetation, these trends point to substantially elevated UHI risk by mid-century and provide quantitative evidence for blue-green infrastructure investment. Because both scenarios project near-identical conditions through 2040, infrastructure commissioned today must be designed for a committed baseline regardless of emission pathway.