Introduction:
In recent decades, population growth, urbanization, and climate change have increased pressure on water resources, making drought a major crisis in cities like Tehran. Tehran’s semi-arid climate and geography lead to frequent droughts that reduce the quality of life and threaten sustainable development. Hydrological and meteorological indices, especially the Standardized Precipitation Index (SPI), based on recorded precipitation remain the standard for estimating drought over various timescales, enabling short  and long term forecasts. Although short statistical periods limit analyses, they are standard in remote sensing and do not compromise accuracy. Reliable drought assessment requires high quality ground data; scarcity or errors can skew drought timing, severity, and impact estimates. This study validates Google Earth Engine satellite data against ground SPI measurements (r = 0.29) and applies pixel-level analysis to retain spatial detail. Building on composite indicator findings from urban contexts, the current study has addressed research gaps in Tehran by deriving drought indices via remote sensing. These cost effective, repeatable, and scalable methods could overcome meteorological data limitations and link meteorological drought to vegetation health indicators.
Methods:
This study evaluates drought indicators in Tehran by integrating remote sensing and meteorological data from January 2000 to December 2022, covering a 23-year period to approach the recommended 30-year climatological baseline. The meteorological data includes precipitation from local stations and NASA''s GPM product, as well as soil moisture data from the Global Land Data Assimilation System (GLDAS) and local monitoring networks. The Standardized Precipitation Index (SPI) was calculated using rainfall data at six temporal scales (1, 3, 6, 9, 12, and 24 months) to quantify meteorological drought and serve as a reference for evaluating remote sensing indicators. Satellite-derived variables were accessed via Google Earth Engine, including the Normalized Difference Vegetation Index (NDVI) and land surface temperature (LST), both of which were corrected for atmospheric and geometric distortions. Monthly NDVI composites were created to align with meteorological data, facilitating the analysis of vegetation dynamics and temperature variations in Tehran. The datasets were co-registered to a common 1 km grid for pixel-level analysis, allowing for a direct comparison between SPI and remote sensing indices. This comprehensive framework enhances our understanding of urban drought monitoring by evaluating the strengths and weaknesses of both meteorological and remote sensing methods. The findings offer beneficial perspectives on urban planning and environmental management, assisting stakeholders in making informed decisions about drought response and resource management in Tehran.
Results:






نسخه پیش نویس





The analysis of composite drought indices across various timescales (monthly, seasonal, and annual) ed significant sensitivities and patterns in response to climate variability. Monthly fluctuations in these indices highlighted their responsiveness to short-term changes, with the Vegetation Health Index (VHI) ranging from 47% in 2022 to 65% in 2020. Such variations may indicated vegetation stress stemming from rising temperatures, alterations in irrigation, or changes in water management practices. Similarly, the Standardized Drought Condition Index (SDCI) exhibited sharp shifts between 20% and 70%, reflecting rapid transitions between wet and dry periods, a trend also observed in other composite indicators. At the seasonal level, distinct cycles were apparent. The Optimized Vegetation Drought Index (OVDI), which aligns with plant growth patterns, peaked in spring and summer while declining in autumn and winter. The SDCI also displayed seasonal patterns but with greater year-to-year variability, suggesting that meteorological factors can diverge from vegetation responses. This divergence underscores the influence of sudden weather changes, such as unexpected droughts or heavy rainfall, on how vegetation adapttions. Understanding these dynamics is crucial for developing effective agricultural strategies and managing water resources in the context of climate change. On an annual scale, long-term trends came into focus. From 2000 to 2012, the SDCI showed a slight downward trend, indicating a prolonged drought, before gradually improving. In contrast, the OVDI has been steadily increasing since 2012, suggesting vegetation recovery and improved hydrological conditions. Comparing these indices highlights the complex nature of drought, revealing instances where vegetation-focused metrics diverge from those based on meteorological data.
Conclusion:
The temporal trend of droughts periodic patterns, with the years 2000–2002 classified  mainly as the moderate drought category, while 2003–2006 showed a relative improvement and a trend towards mild or no drought conditions. The return to moderate to severe droughts in 2007–2010, followed by a gradual improvement from 2011 to 2020 (with exceptions such as 2013 and 2015, which experienced moderate droughts), emphasizes the need for continuous monitoring. On a monthly and seasonal scale, the months of February, July, and November, and the summer season, showed the highest drought intensity, while April and October, and the winter season, were more favorable. Mann-Kendall trend analysis revealed that SDCI showed significant decreasing trends at seasonal (P=0.049) and annual (P=0.018) scales, indicating drought intensification. Conversely, vegetation indices (VCI, VHI, NDVI) exhibited significant increasing trends, suggesting vegetation recovery despite meteorological drought conditions.Correlation analysis also highlighted the pivotal role of composite indices. These indices not only had strong correlations with independent factors (such as precipitation and vegetation), but also provided the ability to accurately reflect changes at different time scales. Precipitation indices such as SPI and PCI, with their perfect correlation, are ideal tools for assessing droughts caused by rainfall deficiency, while the role of temperature (exacerbation) and vegetation (modification) emphasizes the importance of a multidimensional approach.