Minea et al. (2026) Coupled evolution of meteorological and hydrological drought until 2100 based on changes in climate scenarios

Identification

• Journal: Journal of Environmental Management
• Year: 2026
• Date: 2026-03-01
• Authors: Ionuţ Minea, Daniel Boicu
• DOI: 10.1016/j.jenvman.2026.129072

Research Groups

• Department of Geography, Faculty of Geography and Geology, “Alexandru Ioan Cuza” University of Ias¸i, Ias¸i, Romania
• Research Center with Integrated Techniques for Atmospheric Aerosol Investigation in Romania, RECENT AIR, Laboratory of Interdisciplinary Research in Geo-Chemistry of Rural Areas, Environmental Quality Monitoring Station for Geographic Research, Madarjac, Iasi, Alexandru Ioan Cuza University of Iasi, Iasi, Romania

Short Summary

This study analyzed the coupled evolution of meteorological and hydrological droughts in Eastern Romania from 1971-2100 using historical data and future climate scenarios, revealing strong correlations between drought types and a projected increase in severe and extreme hydrological droughts by the end of the century.

Objective

• To statistically analyze the relationship and propagation mode between meteorological (SPI, SPEI) and hydrological (SDI) droughts across multi-scale temporal levels in Eastern Romania, for both historical (1971-2020) and future (2021-2100) climate scenarios, to understand their coupled evolution and predict future hydrological drought periods.

Study Configuration

• Spatial Scale: Eastern Romania, specifically the Moldavian Plateau, covering the counties of Botos¸ani, Ias¸i, Vaslui, and Galați.
• Temporal Scale: Historical period (1971-2020) and future climate scenarios (2021-2100), with analyses conducted for 1, 3, 6, 9, and 12-month time intervals, defined according to the hydrological year (October to September).

Methodology and Data

• Models used:
  • Standardized Precipitation Index (SPI) for meteorological drought.
  • Standardized Precipitation-Evaporation Index (SPEI) for meteorological drought.
  • Streamflow Drought Index (SDI) for hydrological drought.
  • Bravais-Pearson correlation coefficients for assessing relationships between drought indices.
  • Hybrid neural network architecture combining Nonlinear Autoregressive Network with Exogenous Inputs (NARX) and Long Short-Term Memory (LSTM) for projecting future SDI.
  • Linear scaling bias correction for climate scenario data.
• Data sources:
  • Historical meteorological data: Monthly average air temperature (°C) and precipitation (mm) from four meteorological stations (Botosani, Cotnari, Vaslui, Barlad) provided by the National Meteorological Administration (1971-2020).
  • Historical hydrological data: Monthly river flow (m³/s) from 12 hydrometric stations provided by the Prut-Bârlad National Water Administration (1971-2020).
  • Climate scenario data: Daily precipitation (mm) and monthly mean temperature (°C) from the EURO-CORDEX archive, based on 10 bias-corrected regional climate models (RCMs) at a spatial resolution of 0.11° (approximately 12.5 km x 12.5 km). Data aggregated to monthly level and framed in two Representative Concentration Pathways (RCPs): RCP4.5 (medium) and RCP8.5 (extreme) for the period 2021-2100.

Main Results

• Historical Droughts (1971-2020): SPI and SPEI analyses revealed high variability with a concentration of dry periods in the last two decades. Extreme drought periods (SPI 12) were clearly marked between November 1986 and May 1988, followed by frequent moderate-severe droughts. Cumulative frequencies of extreme and severe droughts (SPI 12) ranged from 7.7% to 9.2%. SPEI showed an increasing frequency of high-intensity droughts over the last two decades, peaking at 11.9% in 2011-2020. SDI analysis identified three major historical hydrological drought periods: 1986-1995, 2000-2007, and 2012-2020, with cumulative decadal frequencies of extreme and severe hydrological droughts exceeding 6-9% in some stations.
• Correlation: Strong links were found between meteorological and hydrological droughts (Bravais-Pearson r ≈ 0.43–0.68), particularly at 9 and 12-month temporal scales. The hybrid NARX-LSTM model achieved high accuracy in predicting SDI from SPI/SPEI, with R² values up to 0.90.
• Future Droughts (2021-2100):
  • SPI 12: Severe to extreme droughts are predicted in the next decade (June-October), with a slight increase in intensity after mid-century. Cumulative periods of severe and extreme drought are estimated at 65-73 months for RCP4.5 and 71-77 months for RCP8.5, indicating up to 8-10 more months of severe/extreme drought compared to the historical period.
  • SPEI 12: Remains relatively stable in the immediate future, with a possible slight decrease in drought trend by mid-century. However, a sharp increase in severe and extreme droughts is observed towards the end of the century, with cumulative shares exceeding 8% (RCP4.5) and 18-21% (RCP8.5) at the decadal level.
  • SDI: The frequency of hydrological droughts will remain relatively constant in the medium term. Under RCP4.5, a decrease in severe/extreme drought frequency by mid-century is projected, followed by an increase towards the end of the century (e.g., Stefanesti station: 6.5% extreme droughts). Under the extreme RCP8.5 scenario, a continuity of extreme hydrological drought is expected from the early century, with sequences lasting 3-5 years. By the end of the century, extreme hydrological droughts are projected to become permanent, lasting more than 5 consecutive years, particularly in central and northern regions. The highest frequency of extreme hydrological droughts is projected for Dorohoi and Stefanesti hydrometric stations (over 13%), with cumulative moderate-to-extreme droughts exceeding 40% in 9 of the 12 analyzed stations.

Contributions

• Provides the first comprehensive analysis of the coupled evolution of meteorological and hydrological droughts in Eastern Romania until 2100, integrating both historical data and future climate scenarios.
• Highlights the significant vulnerability of Eastern Romania's water resources and human communities to increasing drought frequency and intensity under climate change.
• Utilizes a multi-scale temporal approach (1, 3, 6, 9, and 12 months) for standardized drought indices (SPI, SPEI, SDI), offering a detailed understanding of drought propagation.
• Employs a novel hybrid NARX-LSTM neural network architecture for robust projection of hydrological drought (SDI) based on meteorological drought indices (SPI, SPEI) under future climate conditions.
• Offers valuable insights for developing innovative decision support systems and regional water resource management strategies to mitigate the impacts of future extreme climate events.

Funding

• Department of Geography, Faculty of Geography and Geology, “Alexandru Ioan Cuza” University of Ias¸i.
• Operational Program Competitiveness 2014–2020, Axis 1, under POC/448/1/1 Research infrastructure projects for public R&D institutions/ Sections F 2018, through the Research Center with Integrated Techniques for Atmospheric Aerosol Investigation in Romania, RECENT AIR.

Citation

@article{Minea2026Coupled,
  author = {Minea, Ionuţ and Boicu, Daniel},
  title = {Coupled evolution of meteorological and hydrological drought until 2100 based on changes in climate scenarios},
  journal = {Journal of Environmental Management},
  year = {2026},
  doi = {10.1016/j.jenvman.2026.129072},
  url = {https://doi.org/10.1016/j.jenvman.2026.129072}
}