Using Sentinel satellite images to study the June 2023 Flood and vegetation indices in the Germi and Ungut counties

Introduction

Flooding, one of the most common natural disasters, poses a serious threat to ecosystems and human safety, with factors such as heavy rainfall, rapid snowmelt, dam failures, and poor management of water, soil, and vegetation resources exacerbating its impacts. Over the past two decades, approximately 2.4 billion people have been affected by natural disasters, resulting in economic damages amounting to $2.97 trillion. Among these, human-related factors such as reduced permeability, construction near riverbanks, and altering water flow paths have played a greater role in intensifying flood damage than natural factors. Accurate monitoring of floods and assessing affected areas are of significant importance. Flood zoning is a practical tool that plays a crucial role in managerial planning. Today, the combined use of optical and radar imagery can provide more comprehensive information. This study aims to effectively utilize Sentinel-1 and Sentinel-2 satellite images in conjunction with the Google Earth Engine (GEE) platform for flood mapping and vegetation indices in the Germi and Ungut counties, which have been impacted by flood events in May and June 2023. This innovative approach enhances the accuracy and speed of flood detection and offers a scalable solution for managing flood risks in vulnerable areas. By leveraging the latest advancements in remote sensing and cloud computing, this research contributes to developing more resilient strategies against flooding and is essential for regions like Ardabil Province, which are prone to irregular and heavy rainfall.



Materials and Methods

In this study, various methods were employed to identify and analyze flood-affected areas using Sentinel-1 and Sentinel-2 satellite data. Initially, Synthetic Aperture Radar (SAR) images from Sentinel-1 in the form of GRD (Ground Range Detected) products were utilized, which can capture images in all weather conditions and at any time of day. These data were derived from the C-band and Interferometric Wide Swath (IW) mode for flood mapping. To reduce speckle noise, the Refined Lee filter was applied, which removed random noise while preserving structural details. A threshold of 1.25 was set for VV polarization (vertical transmit, vertical receive) to identify flood-prone areas, with pixels having VV values above this threshold considered as flooded regions. To avoid misinterpretation, areas with permanent water bodies were excluded using JRC Global Surface Water data. SAR images of Sentinel-1 taken before, during, and after the flood (between April 1 and August 20, 2023) were compared for change detection. Additionally, Sentinel-2 images were used to calculate the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) to assess the impact of vegetation variability from April 1, 2019, to August 20, 2023. These methods effectively identified flood-affected areas and flood impacts on the vegetation cover of the two study counties.



Results and Discussion

In May and June 2023, the Germi and Ungut counties witnessed unprecedented rainfall, which led to severe flooding. The results of the interpretation of Sentinel-1 SAR images showed that about 185.83 km2 were affected by flooding, which is equivalent to 9.01 % of the total study area. Before the flood (April-May 2023), the mean NDVI and EVI respectively, were 0.25 and 0.18, indicating a significant decrease in vegetation cover compared to the reference year (April-May 2019). This could be due to land use change, drought, gradual destruction of dense vegetation, or other environmental stresses before the flood. At the time of the flood (June 2023), the mean NDVI and EVI reached 0.22 and 0.15, respectively, which are the lowest levels recorded up to that time. Only seven percent of the area had healthy vegetation during this period. This decrease could be due to the inundation of plants, destruction due to the intensity of the flood flow, or the covering of the soil surface by sediments. After the flood (August 2023), a more severe decrease in vegetation status was observed. The mean NDVI decreased to 0.15 and EVI to 0.1, and only about 2% of the area remained with healthy vegetation. This decrease may be due to severe soil degradation, erosion of the fertile topsoil, reduced vegetation cover due to seasonal conditions in the area.





Conclusion

This study mapped the June 2023 flood and vegetation indices in two counties of Germi and Ungut, located in the northwest of Iran, using Sentinel-1 SAR and Sentinel-2 data. In adverse weather conditions, SAR technology is an effective tool for mapping and monitoring flood-prone areas. The combination of SAR data, the Refined Lee filter, and the GEE platform facilitates crisis management and mitigates flood impacts. SAR's ability to penetrate cloud cover and provide rapid data access plays a critical role in the swift detection of floods. Integrating this technology with early warning systems enables quicker responses to floods, reducing both human casualties and financial losses. Analysis of the June 2023 Flood occurred in the two counties of Germi and Ungut, with an area of 2064 km2, showed that about 185.83 km2 (equivalent to nine percent of the total area) were directly affected by the flood. From an environmental perspective, the analysis of NDVI and EVI indices in four time periods (one reference period in 2019 and three time periods related to before, during, and after the flood of 2023) indicates a sharp decrease in the density and extent of vegetation cover in the region. The amount of desirable vegetation cover decreased from approximately 269 km² in the pre-flood period to only 41 km² in the post-flood period, indicating the destructive impact of the flood on the region's ecological structure. These findings emphasize the need for integrated watershed management and strategic land-use planning. To enhance community resilience, it is recommended to prioritize and protect vulnerable areas through strategic policies. Utilizing these results can help reduce flood damage and save time and costs in foundational studies.