Mirzaei et al. (2026) Integrated global to regional atmosphere predictors for drought modeling in Iran

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2026
• Date: 2026-04-10
• Authors: Faezeh Mirzaei, Abolfazl Rezaei, Mahdi Vasighi
• DOI: 10.1007/s00704-026-06124-y

Research Groups

• Department of Earth Sciences, Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran
• Center for Research in Climate Change and Global Warming (CRCC), Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran
• Department of Computer Science and Information Technology, Institute for Advanced Studies in Basic Sciences (IASBS), Zanjan, Iran

Short Summary

This study developed a comprehensive statistical framework integrating 27 global and regional atmospheric–oceanic predictors to model and forecast meteorological drought indices (SPI and SPEI) across 18 homogeneous regions of Iran. The framework demonstrated high predictive skill, explaining 55–97% of SPI and 55–92% of SPEI variance, showing that combining global teleconnections with regional marine indicators substantially enhances drought prediction.

Objective

• To develop a comprehensive statistical framework incorporating 27 climatic predictors (13 large-scale atmospheric–oceanic indices and Sea Surface Temperature/Sea Level Pressure fields from seven surrounding seas/gulfs) to model and forecast Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) across 18 homogeneous climatic regions of Iran.
• To assess the relative roles of global teleconnections and nearby marine–atmospheric drivers in Iran's drought variability.
• To evaluate regional differences in drought predictability and provide physical insights for adaptation planning and climate model improvement.

Study Configuration

• Spatial Scale: Iran (1.648 million km²), divided into 18 climatically homogeneous regions. Gridded data at 0.5° × 0.5° spatial resolution. Surrounding water bodies include the Caspian Sea, Red Sea, Black Sea, Mediterranean Sea, Arabian Sea, Persian Gulf, and Bay of Bengal.
• Temporal Scale: Monthly data from January 1957 to December 2020. Predictor series (1957–2019) were partitioned into training-validation (1957–2017; 80/20 split) and prediction (2018–2019; 24 months). Continuous Wavelet Transform (CWT) was used to extract statistically significant interannual variability within the 0.5–4 year band.

Methodology and Data

• Models used:
  • Multiple Linear Regression (MLR) for drought modeling and forecasting.
  • k-shape clustering for delineating 18 homogeneous climatic regions.
  • Empirical Orthogonal Function (EOF) analysis for deriving new regional climate indices from Sea Surface Temperature (SST) and Sea Level Pressure (SLP) anomalies of surrounding seas.
  • Continuous Wavelet Transform (CWT) for isolating dominant, statistically significant interannual variability (0.5–4 years) in drought and climate indices.
  • Variable Clustering (VARCLUS) and Variance Inflation Factor (VIF) for multicollinearity assessment.
• Data sources:
  • Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) datasets (monthly, 0.5° × 0.5°): National Oceanic and Atmospheric Administration (NOAA).
  • 13 established large-scale ocean–atmosphere climate indices (AMO, AO, DMI, ENSO, NAO, PDO, PNA, SAM, TPI, PW, SCAND, EAWR): NOAA and National Center for Atmospheric Research (NCAR) Climate Data Guide.
  • Monthly SST and SLP fields for Caspian, Arabian, Red, Black, Mediterranean seas, and Persian Gulf: Copernicus Climate Data Store (used for EOF analysis to derive regional indices).

Main Results

• The MLR framework, using CWT-filtered predictors, demonstrated robust performance in reconstructing and predicting monthly drought indices.
• During the 2018–2019 prediction period, the models explained 55–97% (77% on average) of the variance in monthly SPI and 55–92% (81% on average) of the variance in monthly SPEI across diverse climatic zones.
• For SPI, regional Sea Surface Temperature (SST) variability in the Red Sea, Mediterranean Sea, and Persian Gulf were the most influential predictors, collectively surpassing any single global index. Among global indices, Pacific modes (ENSO, NPI, PW) ranked highest.
• For SPEI, the Red Sea SST was the single strongest predictor, followed by the Atlantic Multidecadal Oscillation (AMO), Persian Gulf Sea Level Pressure (SLP), Pacific–North America (PNA) pattern, and Arctic Oscillation (AO). This highlights the increased importance of temperature-linked large-scale modes and circulation-driven evaporative demand for SPEI.
• When Pacific indices were evaluated simultaneously, ENSO, NPI, and PW emerged as more influential for SPI drought variability than the Tripole Index (TPI) and Pacific Decadal Oscillation (PDO).
• Predictor importance showed spatially heterogeneous teleconnection footprints, with the weakest predictive skill observed in hyper-arid eastern and southeastern regions due to sparse and irregular precipitation.

Contributions

• This study provides the first systematic nationwide modeling for Iran that concurrently incorporates both established large-scale teleconnection indices and newly derived regional SST and SLP variability from surrounding seas.
• It introduces tailored regional climate indices derived from EOF analysis of SST and SLP fields of nearby water bodies, filling a gap in existing literature.
• The research quantifies the relative roles of global teleconnections versus regional marine–atmospheric drivers for both precipitation-based (SPI) and evapotranspiration-inclusive (SPEI) drought metrics across 18 homogeneous climatic regions.
• The integrated framework offers physically interpretable insights into the complex, multi-source controls on Iran's hydroclimate, advancing scientific understanding and practical forecasting capabilities.
• The methodological approach, particularly the explicit inclusion of EOF-derived regional indices and CWT-based filtering, provides a transferable template for drought modeling in other drought-prone regions influenced by combined global–regional climate processes.

Funding

None

Citation

@article{Mirzaei2026Integrated,
  author = {Mirzaei, Faezeh and Rezaei, Abolfazl and Vasighi, Mahdi},
  title = {Integrated global to regional atmosphere predictors for drought modeling in Iran},
  journal = {Theoretical and Applied Climatology},
  year = {2026},
  doi = {10.1007/s00704-026-06124-y},
  url = {https://doi.org/10.1007/s00704-026-06124-y}
}