Monitoring drought over the Gavkhouni Lake using Terrestrial Water Storage data

Monitoring drought over the Gavkhouni Lake using Terrestrial Water Storage data Elahe Khesali* Department of Photogrammetry and Remote Sensing, Faculty of Geodesy and Geomatics Engineering, K. N. Toosi University of Technology, Tehran, Iran [email protected] Elham Ghasemi MUT Tehran, Iran [email protected] Fawziah Almutairi School of Earth and Planetary Sciences, Faculty of Science and Engineering Curtin University , Australia [email protected]. au Abstract— Drought is one of the major challenges in water resources and environmental management. Every year, this phenomenon has numerous detrimental effects on various sectors of countries, including the environment, economy, social conditions, and agriculture. Effective drought management can significantly reduce the damage caused by these impacts. Satellite data plays a crucial role in predicting and monitoring drought. In this study, the drought trend in the Gavkhouni Wetland region of Isfahan Province in Iran has been monitored and evaluated from 2000 to 2022. This paper uses the Drought Severity Index based on Terrestrial Water Storage (TWS) data derived from NASA Global Land Data Assimilation System (GLDAS). This index reflects the actual conditions of drought and reduces the impact of unnatural factors on drought estimation. The obtained results have been compared with the Standardized Precipitation Index (SPI) over different time periods. The results indicate a significant correlation between the indices and exhibit high accuracy in determining the severity of drought and its trends. Keywords— Drought, GLDAS, Standard Index, Gavkhouni Wetland, Terrestrial Water Storage I. INTRODUCTION Drought refers to a severe climatic condition characterized by a lack of moisture [1] and water shortages directly associated with decreased precipitation, lower surface runoff, and abnormal declines in the levels of lakes, groundwater, and reservoirs [6]. It is one of the most complex and widespread phenomena, and simultaneously one of the natural disasters in the world [2]. Drought primarily affects the environment, social and economic activities, and agriculture. Therefore, effective monitoring and evaluation of droughts are essential and unavoidable for preventing agricultural, economic, and social damages and for managing water resources to adopt appropriate strategies [3,4]. The agricultural and environmental sectors are directly affected, while the economic sector is indirectly impacted by drought. A critical and vital aspect in reducing the risk of drought is its monitoring [5]. Hydrological models and satellite remote sensing are powerful tools for monitoring drought conditions on regional and global scales [7]. These can provide an integrated perspective on drought indices on various scales over different and often remote regions [8]. Establishing a standardized drought index using these datasets that can comprehensively reflect changes in water storage greatly aids in better understanding the impacts of human activities and climate change. Nevertheless, drought has a complex nature and it depends on multiple parameters, which has led to a wide variety of drought monitoring indices [9]. Over the past few decades, especially after launching the Gravity Recovery and Climate Experiment (GRACE) various indices have been developed based on using TWS. These include but not limited to the GRACE-based Groundwater Drought Index (GGDI) [6], the Water Storage Deficit Index (WSDI) [10], and the Combined Climatological Deviation Index (CCDI) [11]. In addition to natural factors, drought indices can also be influenced by certain human factors. To overcome the weaknesses of current indices, the study [12] introduces a new index called the GRACE-based Drought Severity Index (GRACE-DSI) [10], based on the time series of TWS. This index can reflect the actual drought conditions associated with climate change. The present study aims to illustrate the overall drought conditions in the Gavkhouni Wetland using an adopted DSI [12]. Here, TWS is derived from the NASA Global Land Data Assimilation System (GLDAS), which takes advantage of GRACE data assimilation and also allows for a longer study period. The latter is particularly important for drought analyses. The derived drought index is further evaluated against the SPI as one of the mostly applied index in literature. The remaining of this work is organised as follows: Section II describes the data used, the materials and methods employed are presented in Section III, and the final section discusses the results and compares them with other methods. II. STUDY AREA AND DATASETS A. Study Area In recent decades, wetlands have been severely impacted by human activities, lack of integrated management plans, and climate change [13]. Gavkhouni Wetland, depicted in Figure 1, is one of Iran's international and significant wetlands, located 130 kilometers southeast of Isfahan and in the Zayandeh Rud basin. Fig. 1. Map and Location of Gavkhouni Wetland [11] 2840979-8-3315-0810-4/25/$31.00 ©2025 IEEE IGARSS 2025 IGARSS 2025 - 2025 IEEE International Geoscience and Remote Sensing Symposium | 979-8-3315-0810-4/25/$31.00 ©2025 IEEE | DOI: 10.1109/IGARSS55030.2025.11242673 Authorized licensed use limited to: Universita degli Studi di Roma Tor Vergata. Downloaded on December 05,2025 at 17:54:17 UTC from IEEE Xplore. Restrictions apply. Gavkhouni Wetland is counted among the desert and hot regions in Iran, with an average daily temperature ranging between 18.3°C and 28.9°C in December and July, respectively. The elevation of the wetland above sea level is 1475 meters, and its area covers 417.7 square kilometers. The geographical coordinates of the wetland are approximately 52°52' East and 30°30' North [14]. The continuous and main source of water supply for this wetland is the Zayandehrud River, originating from the heights of the Zagros Mountains [15]. B. Datasets Two versions of GLDAS are used for TWS products, including GLDAS-2.0 [16] and GLDAS-2.2 [17] for the periods of 1980 to 2002 and 2003 to 2022, respectively. The data from 1980 is used due to the fact that extreme events, such as droughts, are often characterized by their duration, magnitude (or intensity), extent, and return period. A reliable estimation of these characteristics requires time series that are long enough and are also well representative of hydrometeorological characteristics of the regions of interest [18, 19]. GLDAS-2.0 is forced entirely with the Princeton meteorological forcing input data while GLDAS-2.2 uses GLDAS-2.0 Daily Catchment model simulation, forced with the meteorological analysis fields from the operational European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System. GLDAS-2.2 is particularly used here as the model takes advantage of assimilating the Gravity Recovery and Climate Experiment (GRACE). The TWS anomaly observation from GRACE and GRACE Follow-On (GRACE-FO) is assimilated into the model to enhance its simulations [17]. Due to the data agreement with ECMWF, this GLDAS-2.2 daily product does not include the meteorological forcing fields. Both model outputs, which are produced on 0.25 degree and daily scales, are averaged over the Gavkhouni Lake area at monthly temporal scale. III. PROPOSED METHODOLOGY A. Key Indices in Drought Monitoring Various indices have been defined for drought monitoring, combining different meteorological and climatic variables. The most important variable in defining drought severity and magnitude is precipitation. Indices introduced to describe drought severity include SPI, Deciles Index (DI), Percent of Normal Index (PN), Z-Index (ZSI), Modified China Z-Index (MCZI), and China Z-Index (CZI) [20]. Standardized Precipitation Index (SPI) SPI introduced by Mckee in 1993 for assessing climate variability and drought, is one of the most popular and widely used indices for drought monitoring. In this index, the probability density function of the incomplete gamma distribution is first fitted to the frequency distribution of precipitation data [21]. This index solely utilizes the precipitation variable over different time periods to determine drought or excessive precipitation in a region, disregarding other effects. It is defined for various monthly and yearly time scales, with output values ranging from -2.0 to 2.0. Negative values indicate excessive precipitation, while negative values denote dry events. The probability distribution function of precipitation series in this index generally conforms to the gamma distribution. Probability normalization obtained from the gamma distribution function is calculated using the inverse of the normal distribution. Table 1 illustrates the classification of this index based on different values. The method for calculating the SPI parameter is described in reference [22]. In summary, the SPI indicates the probability of precipitation occurrence within a specific time period in a particular region. This index has the advantages of simple calculation and stability, eliminating temporal and spatial differences in precipitation. SPI is sensitive to drought changes and is useful for monitoring drought and assessing climatic conditions on a monthly scale. The annual SPI plot for Lake Gavkhuni, calculated based on the classification provided in Table 1, is depicted in Figure 2. TABLE I. CLASSIFICATION OF SPI INDEX [21] Drought Severity Threshold Value Extremely Moist 2 and above Very Moist 1.0 to 2.99 Moderately Moist 0.5 to 0.99 Near Normal (Very Dry) -0.5 to -0.99 Normal -0.499 to 0.499 Near Normal (Slightly Dry) -0.5 to -0.99 Moderately Dry -1.0 to -1.49 Severely Dry -1.5 to -1.99 Extremely Dry -2 and below DSI is computed by decomposing the TWS time series (cf. Section IIB) into three components using the Loess time series method: long-term trends, seasonal signals, and residuals. The Seasonal-Trend decomposition using Loess (STL) method, commonly used for detecting nonlinear patterns in trend estimates, is a robust and computationally efficient approach. The calculation formula for this method is as follows: 𝑆total=Slong-term+Sseasonality+ Residuals  In this equation, S represents the primary TWS data, "Longterm trend" denotes the long-term trend component, "Seasonal signal" represents the seasonal signal component, and "Residual component" refers to the remaining portion [23]. 𝐷𝑆𝐼𝑖,𝑗 =𝑇𝑊𝑆𝑖,𝑗−𝑇𝑊𝑆𝑗 𝜎𝑗  In the above equation, i represents the year, and j denotes the month. Represents the mean, and denotes the standard deviation of the anomalies of total water storage reduced in month j. The index detects both moisture anomalies and droughts and it follows a semi-standardized normal distribution. The severity of DSI is categorized into five relative cumulative frequency classes using five threshold limits, as shown in Table 2. Fig. 2. Annual Plot of SPI (Standardized Precipitation Index) 2841 Authorized licensed use limited to: Universita degli Studi di Roma Tor Vergata. Downloaded on December 05,2025 at 17:54:17 UTC from IEEE Xplore. Restrictions apply. TABLE II. STANDARD CLASSIFICATION OF MOISTURE AND DRYNESS LEVELS BASED ON DSI INDEX VALUES DSI Value Drought Severity ≤ -2.00 Exceptional Drought -1.60 to -1.99 Severe Drought -1.30 to -1.59 Extreme Drought -0.80 to -1.29 Moderate Drought -0.50 to -0.79 Abnormally Dry -0.49 to 0.49 Near Normal 0.50 to 0.79 Slightly Moist 0.80 to 1.29 Moderately Moist 1.30 to 1.59 Very Moist 1.60 to 1.99 Extremely Moist ≥ 2.00 Exceptionally Moist Assessing drought primarily focuses on three drought variables that provide vital information for risk assessment and decision-making. These variables include the affected area by drought, drought severity, and drought duration. The area affected by drought is based on pixels defined as drought pixels over all pixels in an area. Drought severity refers to cumulative values of the DSI index during a drought event. Drought duration indicates the time duration of a drought period from onset to recovery. In this article, drought severity and duration are extracted from DSI time series. When more than 20% of pixels in the entire watershed are identified as drought pixels, a regional drought event occurs. If the DSI value for three consecutive months is less than -0.8, indicating a moderate drought threshold, a drought period has occurred. If the gap between two adjacent drought periods is one month, and the DSI value for that month is less than -0.8, the two adjacent drought events are merged into one event. Otherwise, two drought periods are considered independent of each other. The Theil-Sen method has also been used in this article to calculate the linear trend of DSI. This method is not sensitive to outliers and can provide a more accurate linear regression for skewed and heterogeneous data. It competes well with the least squares method for normally distributed data. The formula for this method is as follows: , ji XX Median j i ji           The trend of the Drought Severity Index (DSI) in the formula above is represented by β. 𝑋𝑖 and 𝑋𝑗 are the values of the DSI at time i,j. If β > 0, it represents an increasing trend, while if β < 0, it indicates a decreasing trend in DSI. IV. RESULTS AND DISCUSSION The results of the study indicate that for drought monitoring in the Gavkhuni Wetland, a multi-index approach has been used. As a single index may not adequately represent all aspects of drought, the study utilizes a multi-index approach. The precipitation levels in two 11-year periods, from 2000 to 2011 and from 2012 to 2022, are depicted in the graph in Figure 3. Comparing the two twelve-year periods of precipitation from 2000 to 2011 and from 2012 to 2022 shows a 12-year drought in 2008 and another 12-year drought in 2016 and 2021. These results are consistent with the report [24] published by the International Federation of Red Cross and Red Crescent Societies, presented by the United Nations Office for the Coordination of Humanitarian Affairs (OCHA). According to Figure 6, which represents general information about low, moderate, and high precipitation years, low precipitation has increased in the second twelve-year period. A comparison for years of high precipitation, low precipitation, and no precipitation in the two 12-year time periods in the Gavkhuni Wetland area is shown in Figure 4. The TWS data derived from the model outputs for the 23-year period from 2000 to 2022 is shown in Figure 6. According to this figure, there is a consistent pattern where the total water storage decreases in the mid-year, namely in the fifth and sixth months, and increases towards the end of the year, specifically in the ninth and tenth months. This decreasing trend could be attributed to seasonal factors such as snowmelt or water consumption during the warmer seasons. Despite the overall trend, some years exhibit more pronounced fluctuations while others show a more stable pattern, indicating variability possibly influenced by factors like weather, precipitation, and regional water availability. The comparison of the 23-year period of SPI and DSI indices is depicted in Figure 6. Positive SPI values indicate wetter conditions, while negative values indicate drier conditions. DSI represents the severity of drought obtained from satellite measurements of water storage changes (e.g., groundwater, soil moisture). Both indices exhibit similar patterns of variation over time. DSI, in addition to precipitation, considers factors such as evapotranspiration, soil moisture, and groundwater level. Both indices show fluctuations over time. Based on the plot, in most cases, when SPI indicates wetter Fig. 3. The precipitation levels in two 12-year periods: the first period from 2000 to 2011 and the second period from 2012 to 2022 in the Gavkhuni Wetland area. Fig. 4. Compares the years of high precipitation, low precipitation, and no precipitation in the two 12-year time periods in the Gavkhuni Wetland area. Fig. 5. The comparison of TWS over the 23-year period in the Gavkhuni basin using model outputs. 2842 Authorized licensed use limited to: Universita degli Studi di Roma Tor Vergata. Downloaded on December 05,2025 at 17:54:17 UTC from IEEE Xplore. Restrictions apply. conditions, DSI tends to follow the same trend, albeit with the inherent errors typically present in precipitation measurements and index calculations. The plot demonstrates a significant correlation between the two indices in the studied area. The comparison of monthly SPI and DSI values is further shows the relative decrease or increase in precipitation correlates with the changes in the DSI index. The graph indicates a positive trend of drought in recent years. Since the reduction in rainfall is an uncontrollable phenomenon, ensuring sustainable development requires planning agricultural policies and water usage based on the available water resources. The graph demonstrates a strong correlation between the overall water anomaly based on the TWS data and the SPI index. The results of this study highlight a strong correlation between the DSI derived from TWS and the SPI. This alignment suggests that the proposed method is effective in capturing hydrological drought trends in the Gavkhouni Wetland. However, there are some important limitations to consider. The DSI is sensitive to seasonal variations, and may also be influenced by human-induced changes such as groundwater extraction and land-use changes. These factors can distort the natural drought signal. Furthermore, a comparative analysis with other commonly used drought indices, such as the Palmer Drought Severity Index (PDSI) or the Vegetation Condition Index (VCI), could help further validate the robustness and advantages of the proposed approach. This would provide a clearer understanding of how the TWS-based DSI performs under various climatic and environmental conditions. The observed drought conditions in the Gavkhouni Wetland are likely driven by both climatic and anthropogenic factors. Reduced precipitation and increased temperatures have intensified evapotranspiration, leading to a persistent water deficit. Additionally, unsustainable groundwater extraction and upstream water usage for agriculture have significantly contributed to declining water levels. These changes threaten the ecological stability of the wetland, impacting biodiversity and accelerating desertification. Sustainable water management strategies, along with improved drought forecasting systems, are critical to mitigate these impacts and support long-term ecosystem resilience. V. CONCLUSION In this paper, precipitation data related to the Gavkhoni wetland area in Isfahan province from 2000 to 2022 were analyzed to investigate two important drought monitoring indices, namely SPI and DSI. These indices were calculated as strong indicators for identifying the trend and severity of drought using TWS data. Comparison of the results indicates a strong correlation between the two indices in drought monitoring, capable of detecting drought conditions in the studied area. This study provides a comprehensive estimation of hydrological drought. The results demonstrate that the employed indices can effectively represent drought conditions in the studied area and serve as useful tools for monitoring it. Through result analysis, the most severe droughts were identified in the years 2008, 2016, and 2021. Although this study focuses on the Gavkhouni Wetland, the methodology presented here can be adapted to other arid and semi-arid regions globally, providing a valuable tool for monitoring and managing drought conditions under climate change. REFERENCES [1] S. Rawat, A. Ganapathy, and A. Agarwal, "Drought characterization over Indian sub-continent using GRACE-based indices," Scientific Reports, vol. 12, no. 1, p. 15432, 2022. [2] S. Ali et al., "Constructing high-resolution groundwater drought at spatio-temporal scale using GRACE satellite data based on machine learning in the Indus Basin," Journal of Hydrology, vol. 612, p. 128295, 2022. [3] J. Lu et al., "Quantification and assessment of global terrestrial water storage deficit caused by drought using GRACE satellite data," IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 15, pp. 5001-5012, 2022. [4] X. Liu et al., "Performance of different drought indices for agriculture drought in the North China Plain," Journal of Arid Land, vol. 10, pp. 507-516, 2018. [5] A. M. Seka et al., "Hydrological drought evaluation using GRACE satellite-based drought index over the lake basins, East Africa," Science of The Total Environment, vol. 852, p. 158425, 2022. [6] K. Satish Kumar et al., "Monthly and seasonal drought characterization using grace-based groundwater drought index and its link to teleconnections across south Indian river basins," Climate, vol. 9, no. 4, p. 56, 2021. [7] A. C. Thomas et al., "A GRACE ‐based water storage deficit approach for hydrological drought characterization," Geophysical Research Letters, vol. 41, no. 5, pp. 1537-1545, 2014. [8] L. Longuevergne et al., "GRACE water storage estimates for the Middle East and other regions with significant reservoir and lake storage," Hydrology and Earth System Sciences, vol. 17, no. 12, pp. 4817-4830, 2013. [9] M. Zhao et al., "A global gridded dataset of GRACE drought severity index for 2002–14: Comparison with PDSI and SPEI and a case study of the Australia millennium drought," Journal of Hydrometeorology, vol. 18, no. 8, pp. 2117-2129, 2017. [10] S. Azadi et al., "The Gavkhouni wetland dryness and Its impact on air temperature variability in the eastern part of the Zayandeh-Rud River Basin, Iran," Water, vol. 14, no. 2, p. 172, 2022. [11] D. Sinha, T. H. Syed, and J. T. Reager, "Utilizing combined deviations of precipitation and GRACE-based terrestrial water storage as a metric for drought characterization: A case study over major Indian river basins," Journal of Hydrology, vol. 572, pp. 294-307, 2019. [12] X. Liu et al., "GRACE satellite-based drought index indicating increased impact of drought over major basins in China during 2002– 2017," Agricultural and Forest Meteorology, vol. 291, p. 108057, 2020. [13] A. R. Pirali Zefrehei, M. Kolahi, and J. Fisher, "Modeling wetland restoration scenarios in Gavkhooni International Wetland," Restoration Ecology, vol. 31, no. 7, p. e13721, 2023. [14] H. Aalamifar et al., "A Comparative Study on Effects of Different Salinities on the Reproduction rates of two parthenogenetic species of Artemia (Gaav Khooni wetlands of Isfahan and ponds around the Urmia lake) from Iran," International Journal of Aquatic Biology, vol. 2, no. 2, pp. 91-98, 2014. [15] S. Eslamian et al., "Climate change impacts on some hydrological variables in the Zayandeh-Rud River Basin, Iran," Reviving the Dying Giant: Integrated Water Resource Management in the Zayandeh Rud Catchment, Iran, pp. 201-217, 2017. [16] B. Li, M. Rodell, J. Sheffield, E. Wood, and E. Sutanudjaja, "Longterm, non-anthropogenic groundwater storage changes simulated by three global-scale hydrological models," Scientific Reports, vol. 9, 10746, 2019, doi:10.1038/s41598-019-47219-z. [17] B. Li, H. Beaudoing, and M. Rodell, "GLDAS_CLSM025_DA1_D. Version 2.2. GLDAS Catchment Land Surface Model L4 daily 0.25 x 0.25 degree GRACE-DA1 V2.2," NASA/GSFC/HSL, Greenbelt, Maryland, USA, 2020. Available: https://doi.org/10.5067/TXBMLX370XX8,https://disc.gsfc.nasa.gov/d atacollection/GLDAS_CLSM025_DA1_D_2.2.html. [18] A. Cancelliere and J. D. Salas, "Drought length properties for periodicstochastic hydrologic data," Water Resour. Res., vol. 40, no. 2, W02503, 2004, doi:10.1029/2002WR001750. Fig. 6. Comparison of SPI and DSI indices in the Gavkhouni Basin. SPI DSI 2843 Authorized licensed use limited to: Universita degli Studi di Roma Tor Vergata. Downloaded on December 05,2025 at 17:54:17 UTC from IEEE Xplore. Restrictions apply. [19] E. Forootan, M. Khaki, M. Schumacher, et al., "Understanding the global hydrological droughts of 2003–2016 and their relationships with teleconnections," Science of The Total Environment, vol. 650, part 2, pp. 2865-2882, 2019, doi:10.1016/j.scitotenv.2018.09.231. [20] detection using GRACE," Journal of Geophysical Research: Solid Earth, vol. 103, no. B12, pp. 30205-30229, 1998. [21] T. B. McKee, N. J. Doesken, and J. Kleist, "The relationship of drought frequency and duration to time scales," in Proceedings of the 8th Conference on Applied Climatology, California, 1993. [22] C. Liu et al., "Spatiotemporal drought analysis by the standardized precipitation index (SPI) and standardized precipitation evapotranspiration index (SPEI) in Sichuan Province, China," Scientific Reports, vol. 11, no. 1, p. 1280, 2021. [23] I. Bergmann, G. Ramillien, and F. Frappart, "Climate-driven interannual ice mass evolution in Greenland," Global and Planetary Change, vol. 82, pp. 1-11, 2012. [24] I. H. Worldwide, "Iran: Drought - Jul 2021," 2021. [Online]. Available: https://reliefweb.int/disaster/dr-2021-000089-irn 2844 Authorized licensed use limited to: Universita degli Studi di Roma Tor Vergata. Downloaded on December 05,2025 at 17:54:17 UTC from IEEE Xplore. Restrictions apply.