Meteorological drought monitoring based on SPI and mRAI indices in the Urmia Lake basin

Extended Abstract



Introduction

Drought, a pervasive meteorological phenomenon, is driven by insufficient precipitation and linked to climatic factors. Its escalating frequency challenges natural resource management, water security, and mitigation efforts. This complex hazard is categorized into four types: meteorological, agricultural, hydrological, and socio-economic. Accurate drought monitoring relies on robust indicators, such as the Standardized Precipitation Index (SPI) and the modified Rainfall Anomaly Index (mRAI), both of which are precipitation-dependent. Acquiring reliable, spatially representative rainfall series remains a global challenge. The Global Precipitation Climatology Centre (GPCC) global dataset provides a crucial, quality-controlled, rain-gauge-based resource that addresses this data scarcity. Given Iran's limited and unevenly distributed rain gauge stations, global datasets like GPCC are indispensable. Previous research has confirmed the efficacy of the GPCC data in conjunction with the SPI and mRAI indices for comprehensive spatial and temporal drought analysis. Building on this, this research generates high-resolution drought severity maps for the Urmia Lake basin across multiple timescales using GPCC monthly precipitation data and both SPI and mRAI indices. This provides an essential tool for proactive drought monitoring in this ecologically sensitive, water-stressed region. The Urmia Lake basin, a vital ecological and economic area, has faced severe and prolonged drought, resulting in a dramatic decline in lake levels, increased dust storms, and significant socio-economic challenges. The urgent need for accurate and timely drought monitoring is thus highlighted. Effective monitoring is essential for developing sustainable water management strategies and mitigating adverse impacts on this vulnerable ecosystem. This research directly addresses this critical need by providing a substantial approach for informed management.



Materials and Methods

Drought analysis for the Urmia Lake Basin utilized monthly precipitation data from synoptic stations and the global GPCC dataset. The accuracy of GPCC data in estimating monthly precipitation at the studied synoptic stations was rigorously evaluated using R², RMSE, MAE, and PBIAS over the 1991–2020 period. Subsequently, raster maps of monthly precipitation for the Urmia Lake Basin were prepared using GPCC data for the same period. Drought severity maps of the Urmia Lake basin were then produced based on SPI and mRAI indices at 1-, 3-, 6-, 9-, and 12-month time scales, using GPCC precipitation data from 1991 to 2020. A comparative analysis identified critically vulnerable areas across drought severity classes (weak, moderate, severe, and very severe). Furthermore, the agreement between SPI and mRAI severity classes at different time scales was quantitatively assessed using the Kappa statistic and Cramer's V coefficient.



Results and Discussion

Evaluation results conclusively demonstrated the acceptable accuracy of GPCC data in estimating monthly precipitation at all studied synoptic stations (R² = 0.91, RMSE = 10.83). Analysis of the 30 years revealed consistently the highest average monthly rainfall in the western, southwestern, and southern regions of the basin across all investigated time scales. A direct comparison of drought index values at the Urmia and Tabriz stations revealed strong concordance between the SPI and mRAI indices at time scales of 3, 6, 9, and 12 months. Quantitatively, Kappa and Cramer's V coefficient values at these time scales across all stations were notably high: 0.851 and 0.837 (3 months), 0.863 and 0.847 (6 months), 0.867 and 0.850 (9 months), and 0.929 and 0.912 (12 months), respectively. These robust statistical measures confirm a significant relationship between the SPI and mRAI drought indices for assessing drought conditions at both individual and collective stations across varying time scales. Importantly, SPI-3 and mRAI-3 indices identified moderate drought affecting most areas of the Urmia Lake Basin in October 1995, 2001, and 2002; from October to December 2010; and in October and November 2019. Furthermore, based on SPI-6 and mRAI6 indices, the entire Urmia Lake basin experienced mild to severe drought, with extreme drought in localized regions, from October to November of the mentioned years (excluding December 2002).



Conclusion

The accuracy of GPCC data for monthly precipitation estimation at Urmia Lake Basin synoptic stations is confirmed. This validates GPCC as a foundational dataset for calculating SPI and mRAI drought indices at 3-, 6-, 9-, and 12-month time scales, enabling detailed delineation of drought severity zones across the basin (1991–2020). A critical finding is the consistent strong agreement between SPI and mRAI indices across different time scales, demonstrated by index value comparisons. High Kappa and Cramer's V coefficients further substantiate this correlation in drought class identification. Essentially, the spatial distribution of drought severity, as measured by both SPI and mRAI (at 3, 6, 9, and 12-month scales), shows a compelling agreement throughout the 1991–2020 period. Despite the utility of the GPCC data, potential estimation errors from synoptic stations necessitate rigorous calibration using advanced methods, such as linear and quantile regression. Given GPCC's 0.25-degree spatial resolution, future research should downscale this dataset, e.g., via geographically weighted regression, to identify finer-resolution drought zones, which is crucial for localized water resource management and planning in the Urmia Lake basin.