Analysis and Monitoring of Water Resources and Drought Using a Combination of GRACE, MODIS, and Landsat 8 Satellite Images (Case Study: Hamedan City)

Abstract
Objective: In recent years, the application of satellite-based techniques in the study of hydrological phenomena has been recognized as an effective tool for monitoring water resource conditions. Due to the scarcity of ground-based station data or the lack of sufficient in-situ observations, remote sensing provides a reliable means for collecting large-scale spatial and temporal data. Moreover, by applying appropriate algorithms, the impacts of drought distribution on rangelands and vegetation can be assessed using satellite imagery.
Method: Having access to information on drought spatial distribution within a region enables the identification of its adverse effects on vegetation cover and supports the implementation of mitigation and management strategies. In this study, a combination of Landsat 8, MODIS (Terra), and GRACE satellite imagery was used to investigate variations in drought indices, runoff, and groundwater storage changes in the Hamedan Plain over the period 2003-2016.
Results: The results indicate that the long-term mean runoff in the Hamedan Plain, derived from satellite data, was approximately 0.512 m³/s. In addition, groundwater level variations estimated from the three GRACE data processing centers-CSR, GFZ, and JPL-were 29.52, 31.05, and 35.11 cm, respectively.
Conclusion: Analysis of the NDVI-based drought distribution map revealed that the Hamedan Plain experienced predominantly moderate drought conditions, indicating a generally normal status of vegetation cover during the study period. 
 
 
 
English Extended Abstract
Introduction
Due to the special climatic conditions of arid and semi-arid regions such as Iran, groundwater is considered a vital and strategic resource. Population growth and increasing need for water, coupled with limited surface water resources, put severe pressure on groundwater resources. This can lead to irreversible consequences for ecosystems and hydrological cycles. Therefore, proper management and careful planning in watersheds are essential to ensure the sustainability of water resources and the preservation of ecosystems dependent on it. The spatial distribution of vegetation is strongly influenced by climatic conditions, especially precipitation and temperature. A drought is a prolonged dry period in the natural weather cycle that can occur anywhere in the world and is a slow-onset disaster characterized by a lack of rainfall and therefore a lack of water.
According to studies, it is estimated that 55 million people worldwide are affected by drought annually, and this drought is the most serious risk to livestock and agricultural products in almost every part of the world. Rising temperatures caused by climate change are making arid regions drier and humid areas wetter. In this study, this study aims to investigate surface and groundwater resources and the distribution and classification of drought in the study area of Hamadan Plain. For this purpose, using extracted images from the MODIS and Grace and Landsat 8 satellites, we investigated the trend of changes in each of the parameters in the period from 2003 to 2016.
 
Method
Landsat satellites with low spatial resolution and high temporal resolution make it possible to monitor atmospheric changes on a large scale. Climate satellites also play an important role in measuring atmospheric parameters. These data provide researchers with actual evapotranspiration, cumulative precipitation, and runoff. Groundwater is a vital and strategic resource for communities, and it requires careful and continuous management and monitoring. Grace satellite images are an efficient and effective method of monitoring groundwater level fluctuations. In this study, 168 Grace Satellite images from the Hamedan Plain from 2003 to 2016 were analyzed. The Grace satellite, an advanced dual-gravity satellite, was launched by NASA and DLR in 2002, and its scientific mission ended in 2017. The study examines and analyzes satellite images with special precision. This research includes the following stages:

Benefit from the advanced Google Earth Platform and JavaScript programming language for satellite image processing. Google Earth Engine, as a powerful and free tool, provides researchers with a wide range of possibilities.
Selection and processing of satellite images related to the study area (Hamedan Plain) in a specific period (2003 to 2016), and applying appropriate spatial and temporal filters for the desired area.
Analysis of gravitational images of Grace Satellites processed by different centers. Each center processes raw data with its own algorithms, but due to differences in methods, the resulting values also vary. Therefore, the use of multiple average values and algorithms increases the accuracy of the results.
Improving the quality of results by resampling processed data at high spatial resolution (e.g. 300 meters).

To conduct the present study, the Google Earth Engine system and the JavaScript programming language have been used. Google Earth Engine enables users to perform their calculations on large amounts of data without the need for powerful systems. After recalling the Modis product and applying spatial and temporal filters in the first place, in the second place, the cloud mask from the collection of images was taken. In the third place, after selecting the NDVI band, the NDVI average image was created for all months of the year. In the fourth place, after calculating the minimum and maximum values of the NDVI index, the VCI index for 2003 to 2016 was calculated. Finally, to analyze the values of this index and the vegetation status of the region, the VCI map was classified into three classes: no vegetation, medium vegetation, and rich vegetation, and the area of each class was calculated in terms of percentage.
 
Results
In the surface runoff section, the average discharge of Hamedan Plain from June 2003 to December 2016 was 0.512 m³/s and the highest runoff was related to February, March, and April of each year. Also, a decrease was shown in groundwater level in the Hamadan-Bahar plain by 31.05, 29.52, and 35.11 cm, respectively. Finally, according to the NDVI index, which is the basis for calculating the VCI index, the overall Geezi result was obtained as follows: The area is in a low to moderate vegetation state. Also, according to the results of the VCI classification, the worst conditions were related to June 17, 2008 with a rate of 4.1%, and the best conditions were related to April 23, 2010 with a rate of 86.2%. Also, by examining the indices of VCI, TCI, VHI, SVI, the drought severity of the study area was in the medium category according to the two indices of VCI and TCI and the results of the other two indices indicated severe drought.
 
Conclusion
This study aimed to investigate the changes in surface and subsurface water resources, considering the severity of drought from 2003 to 2016 in Hamadan city. In summary, based on the results of the Landsat satellite, the long-term average runoff value of the region was 0.512 m³/s, with the highest discharge in proportion to the highest rainfall months in the region in February, March, and April. In the analysis of the results of groundwater level changes based on three databases GFZ, CSR, and JPL, the distance between the maximum and the lowest level changes was 29.52, 31.05, and 35.11 cm, respectively. In the last part, TCI drought indices indicate moderate drought and based on vegetation indices, are classified as high drought intensity. Also, based on the classification map, most of the land in Hamadan city was classified as barren and with low vegetation.