Palacios-Gutiérrez et al. (2025) Identification of extreme temperature and precipitation patterns in Spain based on multiscale analysis of time series

Identification

• Journal: Natural Hazards
• Year: 2025
• Date: 2025-01-21
• Authors: Arnobio Palacios-Gutiérrez, Jose Luis Valencia-Delfa, María Villeta
• DOI: 10.1007/s11069-024-07082-2

Research Groups

• Faculty of Statistical Studies, Complutense University of Madrid, Spain
• Group Valoración y Aprovechamiento de la Biodiversidad, Technological University of Chocó, Quibdó, Chocó, Colombia

Short Summary

This study develops a novel multiscale analysis methodology to identify patterns of change in extreme maximum and minimum temperatures and precipitation in Spain from 1951 to 2021, revealing twelve distinct climate zones with varying warming trends, precipitation decreases, and increased drought duration and magnitude.

Objective

• Develop a novel multiscale analysis methodology, adapted from the financial field, to identify and characterize patterns of change in extreme maximum temperature (Tmax), minimum temperature (Tmin), and precipitation (PRCP) across Spain.
• Regionalize Spain into distinct climate zones based on these identified patterns of change, providing a granular understanding of climate variability.

Study Configuration

• Spatial Scale: Spain (mainland, Balearic Islands, Ceuta, and Melilla), analyzed on a 5 × 5-kilometer grid, comprising 16,156 grid points.
• Temporal Scale: 1951 to 2021 (71 years), using monthly extreme data, with multiscale analysis windows of 36 months and 72 months.

Methodology and Data

• Models used:
  • Novel multiscale analysis methodology for dimensionality reduction and similarity measurement, adapted from Shi et al. (2021) to account for multivariate climate data and asymmetry (using median instead of mean).
  • Clustering algorithms: Partitioning Around Medoids (K-medoids, PAM), Hierarchical Agglomerative (HA) using Ward method, and Self-Organizing Maps (SOM).
  • Internal validation indices for optimal cluster number: C Index, Baker-Hubert Gamma Index, McClain-Rao Index, and Ray-Turi Index.
  • Trend analysis: Non-parametric Mann–Kendall test (with pre-whitening for autocorrelation) and Theil-Sen slope estimator.
  • Drought analysis: Standardized Precipitation Index (SPI) calculated on a 12-month scale.
• Data sources:
  • Monthly extreme maximum temperature (Tmax), minimum temperature (Tmin), and precipitation (PRCP) time series provided by the Climatological Development Service of the Spanish Meteorological Agency (AEMET).
  • Data obtained through spatial interpolation using kriging on a 5 × 5-kilometer grid.

Main Results

• Spain was regionalized into twelve distinct climate zones, showing clear geographical differences based on altitude, latitude, longitude, and proximity to the sea.
• A significant warming trend was observed across all zones of Spain, with the most pronounced increases during summer months.
  • Extreme maximum annual temperatures increased by up to 0.49 °C per decade, with Cluster 3 (Eastern Andalusia) showing the highest rate.
  • Annual average maximum temperatures increased by 0.14 °C to 0.26 °C per decade.
• A reduction in extreme cold events was detected, particularly in mountainous zones (Cluster 4, 5) and Atlantic wetlands (Cluster 11, 12).
• Minimum temperatures increased by 0.10 °C to 0.22 °C per decade, with the most notable changes in mountainous zones.
• Annual precipitation significantly decreased across central and southern zones, ranging from 11 to 25 millimeters per decade (lower than 5% per decade).
• In contrast to the overall decline, maximum daily rainfall (intense rainfall events) increased by 2% to 5% per decade across most zones, suggesting fewer rainy days but higher maximum precipitation values.
• The duration and magnitude of droughts increased by approximately 10% per decade, particularly in Extremadura, Andalusia, and mid-mountain zones (Clusters 2, 3, 7, and 4). Cluster 3 showed an increase of six drought months per decade.

Contributions

• Introduces a novel multiscale analysis methodology for climate regionalization, adapted from the financial field, which is robust for simultaneously analyzing multivariate climate variables (Tmax, Tmin, PRCP) over an unusually large number of small territorial areas (16,156 5x5-kilometer grid points).
• Provides a more detailed and granular climate regionalization of Spain (12 distinct zones) compared to previous studies, enabling the detection of microclimate differences in small neighboring areas.
• Incorporates the median instead of the mean for dimensionality reduction to better account for the common asymmetry present in climatic variables.
• Offers a comprehensive spatiotemporal analysis of extreme temperature and precipitation patterns and drought characteristics across Spain, highlighting differential rates of climate change and the need for region-specific adaptation strategies.

Funding

• Colombian Ministry of Science
• Technological University of Chocó (for Arnobio Palacios's doctoral formation)
• Agencia Estatal de Investigación (PID2019-106433GB-I00 / AEI / https://doi.org/10.13039/501100011033), Spain.

Citation

@article{PalaciosGutiérrez2025Identification,
  author = {Palacios-Gutiérrez, Arnobio and Valencia-Delfa, Jose Luis and Villeta, María},
  title = {Identification of extreme temperature and precipitation patterns in Spain based on multiscale analysis of time series},
  journal = {Natural Hazards},
  year = {2025},
  doi = {10.1007/s11069-024-07082-2},
  url = {https://doi.org/10.1007/s11069-024-07082-2}
}