Kalin et al. (2025) Sub-daily rainfall extremes in Croatia: a basis for improved warning thresholds

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2025
• Date: 2025-10-13
• Authors: Ksenija Cindrić Kalin, Irena Nimac, Leonardo Patalen, Tanja Renko, Petra Mikuš Jurković, Zoran Pasarić
• DOI: 10.1007/s00704-025-05809-0

Research Groups

• Faculty of Science, Department of Geophysics, University of Zagreb, Zagreb, Croatia
• Wegener Center for Climate and Global Change, Graz, Austria
• Croatian Meteorological and Hydrological Service, Zagreb, Croatia
• European Severe Storms Laboratory, Wessling, Germany

Short Summary

This study provides the first comprehensive climatology of sub-daily rainfall extremes across Croatia (1961–2020) to establish an improved basis for rainfall warning thresholds. It reveals distinct regional and seasonal patterns of extreme rainfall and proposes new warning thresholds based on stationary Generalized Extreme Value (GEV) return levels.

Objective

• To provide a comprehensive climatology of sub-daily rainfall extremes across Croatia and estimate return values, considering potential climate change effects.
• To address the need for revising Croatia’s rainfall warning thresholds, which are currently based on limited historical data, by proposing new, evidence-based criteria.

Study Configuration

• Spatial Scale: 54 meteorological stations across Croatia, covering continental mainland, Adriatic coast, hinterlands, and highlands.
• Temporal Scale: Hourly rainfall records from 1961–2020. Sub-daily durations analyzed: 3 hours, 6 hours, 12 hours, and 24 hours.

Methodology and Data

• Models used:
  • Generalized Extreme Value (GEV) distribution (stationary and non-stationary models).
  • Stationary GEV (sGEV) with constant shape, scale, and location parameters.
  • Non-stationary GEV (nGEV) models: nGEV1 (time-varying location parameter), nGEV2 (exponentially time-varying scale parameter), and nGEV3 (time-varying location and scale parameters).
  • Parameter estimation: Maximum Likelihood (ML) method and L-moments method.
  • Model selection criteria: Akaike’s Information Criterion (AIC) and Likelihood Ratio Test (LRT).
• Data sources:
  • Hourly rainfall records from 54 meteorological stations operated by the Croatian Meteorological and Hydrological Service (DHMZ).
  • Data collected using classical Hellman pluviographs, recorded at 5-minute intervals.
  • Quality control involved comparing ombrographic data with daily data from classical rain gauges.

Main Results

• Spatial Patterns: Lowest rainfall extremes (e.g., 3-hour median 15.2 mm) occur in eastern continental areas, while the Adriatic region (coast, hinterlands, highlands) experiences the highest intensities (e.g., 3-hour median up to 57.8 mm).
• Seasonal Patterns: Rainfall maxima predominantly occur in summer (June–August) over the mainland and in autumn (September–November) along the coast, consistent with regional climate drivers.
• Non-stationarity Assessment: The stationary GEV model was preferred by 54–61% of stations. While 9–17% of stations showed statistically significant trends in location parameters (nGEV1), no statistically significant differences were found between stationary and non-stationary return values for practical applications.
• Warning Thresholds Proposal: The study proposes revising rainfall warning thresholds by adopting the 2-year, 5-year, and 10-year return levels as reference for moderate (yellow), severe (orange), and extreme (red) rainfall warnings, respectively. This represents a significant increase in thresholds (16–58% for severe, 8–33% for extreme) compared to previous criteria.

Contributions

• Provides the first comprehensive, nationwide climatological analysis of sub-daily rainfall extremes in Croatia using an extensive, quality-controlled dataset (1961–2020, 54 stations).
• Establishes a robust climatological foundation for revising Croatia's rainfall warning thresholds, moving from limited historical data to a more comprehensive, evidence-based system.
• Assesses the impact of climate change on sub-daily rainfall extremes using stationary and non-stationary GEV models, concluding that stationary models remain reliable for current operational practice despite some observed trends.
• Proposes specific, updated return-level based thresholds (2-, 5-, and 10-year return levels) for moderate, severe, and extreme rainfall warnings, enhancing the European Meteoalarm framework in Croatia.

Funding

• Croatian Meteorological and Hydrological Service
• Croatian Science Foundation (project HRZZ-IP-2022-10-4144)

Citation

@article{Kalin2025Subdaily,
  author = {Kalin, Ksenija Cindrić and Nimac, Irena and Patalen, Leonardo and Renko, Tanja and Jurković, Petra Mikuš and Pasarić, Zoran},
  title = {Sub-daily rainfall extremes in Croatia: a basis for improved warning thresholds},
  journal = {Theoretical and Applied Climatology},
  year = {2025},
  doi = {10.1007/s00704-025-05809-0},
  url = {https://doi.org/10.1007/s00704-025-05809-0}
}