Kenawy et al. (2026) Multiscalar and nonlinear controls of drought impacts in Spain revealed by media-reported data

Identification

• Journal: Natural hazards and earth system sciences
• Year: 2026
• Date: 2026-10-13
• Authors: Ahmed El Kenawy, Fernando Domínguez‐Castro, Santiago Beguerı́a, Javier Vela-Tambo, Amar Halifa‐Marín, María Adell-Michavila, Alex Crespillo, Borja Latorre, Magí Franquesa, Sergio Martín Vicente-Serrano
• DOI: 10.5194/nhess-26-4977-2026

Research Groups

• Instituto Pirenaico de Ecología, Consejo Superior de Investigaciones Científicas (IPE-CSIC), Zaragoza, Spain
• Laboratorio de Climatología y Servicios Climáticos (LCSC), CSIC–University of Zaragoza, Zaragoza, Spain
• Department of Geography, Mansoura University, Mansoura, Egypt
• Estación Experimental de Aula Dei, Consejo Superior de Investigaciones Científicas (EEAD-CSIC), Zaragoza, Spain
• Worcester Polytechnic Institute, Worcester, MA, USA

Short Summary

This study quantifies the multiscalar and nonlinear relationships between drought severity (SPI, SPEI) and media-reported hydrological and agricultural impacts in Spain (1976–2023), revealing that impacts are strongly linked to persistent moisture deficits, with SPEI showing a consistently stronger association than SPI, and reporting sensitivity increasing disproportionately during severe droughts.

Objective

• To quantify the multiscalar relationship between drought severity (SPI and SPEI) and media-reported hydrological and agricultural impacts across Spain (1976–2023).
• To assess the temporal coherence and timescale dependence of drought–impact relationships.
• To evaluate the role of atmospheric evaporative demand (SPEI vs. SPI) and the nonlinear sensitivity of impact reporting to drought severity.
• To identify the dominant timescales and spatial patterns associated with impact-reporting responses across sectors.

Study Configuration

• Spatial Scale: Spain (48 provinces), with gridded climate data at 1.1 km resolution.
• Temporal Scale: 1976–2023 (48 years), using monthly data and drought indices computed for accumulation periods from 1 to 48 months.

Methodology and Data

• Models used:
  • Standardized Precipitation Index (SPI)
  • Standardized Precipitation Evapotranspiration Index (SPEI)
  • Supervised classifier (Longformer model within SeqIA framework) for drought relevance filtering of news articles.
  • Schema-guided Generative Information Extraction (GenIE) approach (CienaLLM framework, gemma2:9b model) for impact extraction.
  • Principal Component Analysis (PCA) for spatial regionalization.
  • Ordinary Least Squares (OLS) log-log regression for drought–reporting sensitivity coefficient (β).
  • Logistic regression for the occurrence of drought-impact reporting.
• Data sources:
  • Media-reported drought impacts: Monthly frequencies of drought-related news articles from "El País" newspaper (1976–2023), standardized at the provincial scale for hydrological and agricultural impacts.
  • Drought indices: High-resolution gridded climate dataset (1.1 km) for Spain, used to compute SPI and SPEI for 48 provinces and national series.

Main Results

• Robust temporal coherence between drought conditions and media-reported impacts, with major drought episodes associated with anomalies exceeding +1 to +2 standard deviations.
• Drought–impact relationships strengthen significantly with accumulation timescale, reaching maximum correlations at 12–30 months (|r| ≈ 0.6–0.8), while short timescales (1–3 months) show weak associations (|r| < 0.3).
• Hydrological impacts are primarily associated with accumulated medium- to long-term moisture deficits (12–48 months).
• Agricultural impacts are associated with shorter accumulation timescales (3–12 months).
• For both sectors, dominant accumulation timescales lengthen as drought severity increases.
• SPEI consistently exhibits stronger associations than SPI across most conditions, particularly for agricultural impacts, suggesting a dominant contribution of atmospheric evaporative demand.
• Pronounced nonlinear responses are observed, with the frequency of media reports on impacts increasing disproportionately during severe drought conditions, peaking at 12–36 months.
• Spatial heterogeneity in drought-reporting sensitivity is evident, with SPEI-dominated responses prevailing over large areas, especially at longer timescales (12–36 months).
• A dominant large-scale signal (PC1) of drought-impact reporting is identified for both sectors, with regional contrasts (e.g., east–west for PC2).

Contributions

• Provides the first comprehensive multiscalar and nonlinear assessment of drought–impact relationships in Spain using media-reported data and advanced NLP/LLM techniques.
• Quantifies the critical role of atmospheric evaporative demand (SPEI) in amplifying drought impacts and media reporting, especially under warming conditions and for agricultural systems.
• Identifies distinct dominant accumulation timescales for hydrological (longer) and agricultural (shorter) impacts, demonstrating how these lengthen with increasing drought severity.
• Reveals the nonlinear nature of drought–reporting relationships, showing a disproportionate increase in reporting sensitivity during severe and extreme drought conditions.
• Offers a framework for improved impact-based drought monitoring and early warning systems by integrating climate indices with societal impact data.

Funding

• CSIC Open Access Publication Support Initiative through its Unit of Information Resources for Research (URICI).
• Research project PID2022-137244OB-I00, financed by the Spanish Ministry of Science.

Citation

@article{Kenawy2026Multiscalar,
  author = {Kenawy, Ahmed El and Domínguez‐Castro, Fernando and Beguerı́a, Santiago and Vela-Tambo, Javier and Halifa‐Marín, Amar and Adell-Michavila, María and Crespillo, Alex and Latorre, Borja and Franquesa, Magí and Vicente-Serrano, Sergio Martín},
  title = {Multiscalar and nonlinear controls of drought impacts in Spain revealed by media-reported data},
  journal = {Natural hazards and earth system sciences},
  year = {2026},
  doi = {10.5194/nhess-26-4977-2026},
  url = {https://doi.org/10.5194/nhess-26-4977-2026}
}