Asymmetric Warming over Croatia: Quantile Trends in Temperature Distribution and Cumulative Heat Intensity of Compound Dry-Hot Extreme Events

This study examines asymmetric warming in Croatia by analysing changes in the distribution of maximum daily air temperature and the characteristics of cumulative heat intensity in compound dry and hot (DH) extreme events in the JJA season. The analysis covers the period 1963–2025 and is based on observations from the main meteorological stations in Croatia, grouped into mountain, inland, and coastal regions.

Compound DH events are defined as periods of consecutive dry days (daily precipitation < 1 mm) and the maximum of maximum daily air temperatures (Tx) during these dry spells, while extremes of such events are defined as the simultaneous exceedance of the 90th percentile for these parameters, based on the 1981–2010 reference period. Event intensity is quantified using cumulative heat intensity (CHI), defined as the sum of positive daily temperature anomalies during an event. To further characterise high-intensity events, two additional metrics are analysed: the number of days with positive temperature anomalies (CHI_ndays) contributing to the CHI and the mean temperature anomaly (CHI_meanTa) of those days.

The results reveal pronounced regional differences. At coastal stations, high-CHI events are mainly associated with longer durations of moderately elevated temperatures, reflected in higher CHI_ndays. In contrast, inland stations show high-CHI events characterised by fewer days than at coastal stations, but with stronger temperature anomalies, indicated by higher CHI_meanTa.

Trends in temperature distribution during the JJA season are analysed using quantile trends based on 30-year sliding windows. Inland stations exhibit approximately linear increases across percentiles, with the largest trends at the highest percentiles, indicating a stronger amplification of the upper tail of the distribution. Coastal stations show increasing trends up to the lower-to-middle percentiles, followed by a relative stabilisation of trends at higher percentiles, although trends remain positive.

The same analysis applied to other seasons reveals geographically consistent patterns with varying magnitudes. Overall, the results indicate strong spatial and seasonal variability in temperature distribution changes and provide additional insight into the characteristics of cumulative heat intensity in compound DH extreme events in Croatia.