Publication Date: 2024
پژوهشنامه مدیریت حوزه آبخیز (22516174)15(1)pp. 147-158
Background: Soil erosion and sedimentation are among the major problems and environmental challenges in watersheds. The sediment load in rivers causes numerous issues, such as sedimentation in dam reservoirs, changes in river courses due to sedimentation in their beds, reduced water carrying capacity in waterways and water transfer facilities, and changes in water quality for drinking and agriculture. These problems are of great concern to researchers and water resource managers as they can investigate sedimentation, erosion, and potential effects on biological processes. Therefore, it is essential to investigate and evaluate suspended sediment concentrations to determine water quality and hydrological functions. In this regard, the use of accurate, extensive, and cost-effective techniques, such as remote sensing, is invaluable for improving sedimentation estimation and, consequently, water quality assessment. Understanding the spatial relationships between upstream vegetation and sedimentation is crucial for the effective control and optimal management of water resources and soil. The present study aimed to determine the relationship between vegetation cover and suspended sediment concentrations in the Doab Mereg and Gamasiab watersheds. Methods: In this research, the Sentinel-2 satellite data were investigated in the first step. If there were cloudy conditions, dust particles, or other radiometric problems, they were removed from the calculations. Then, factors such as band type (visible and infrared bands), river width (higher than the image pixel-40 meters at the Doab station and 80 meters at the Polchehr station), separation power, and sensor location (10 and 20 meters) were considered to select the appropriate image pixel for obtaining the spectral reflectance of water. The pixel corresponding to the hydrometric station and its surroundings was selected to extract the spectral reflectance. Then, the suspended sediment concentration statistics of the Doab Mereg station located in the Qarasu River and the Polchehr station located in the Gamasiab River in the five-year period (2016 to 2020) were used simultaneously to investigate the correlation between the spectral reflectance of the Sentinel-2 image bands and sediment concentration. Next, the Normalized Difference Vegetation Index (NDVI) for the two May and June seasons was extracted using Sentinel-2 images. The relationship between vegetation cover and suspended sediment concentration recorded at the station and extracted from the images was estimated separately. Results: The results of the correlation analysis of suspended sediment concentration showed that the best result for the Doab and Polchehr stations belonged to band 4 (R2 = 0.86) and band5 (R2 = 0.83), respectively. The suspended sediment concentrations varied from 0.17 to 76.45 and 0.44 to 118.86 mg/liter in the Doab Mereg and Polchehr stations, respectively. In the Doab station, the depth of the power state had the highest correlation coefficient between the observational data (recorded in the station) and the data extracted from the images. In the Polchehr station, it had a high correlation coefficient in both polynomial and exponential modes. The best values of the coefficient of determination (R2) of the normalized vegetation cover difference index for the Doab and Polchehr stations were 0.98 and 0.64, respectively. This means that the amount of sediment decreases with an increase in vegetation cover. The average values of the vegetation index for the Mereg watershed (0.35) and the Gamasiab watershed (0.28) show the relatively sparse vegetation in the area. The lowest average values of vegetation cover in the studied season (spring) were respectively equal to 0.11 and 0.21 in the Mereg watershed in June and the Gamasiab watershed at the end of May. The results of the regression test showed a strong and significant relationship between the density of vegetation and the amount of suspended sediment concentration recorded in the hydrometric stations of the two basins. Conclusion: The results showed that six models were extracted for the studied area, which had acceptable and suitable R2 and error values. Among the obtained models, better results were obtained in the single-band model than in the case of using the band ratio. The highest correlations belonged to bands B2, B3, B4, and B5 in the Doab Mereg station, and to B4 and B5 in the Pol Cheher station. The highest R2 values obtained for the two stations were 0.86 and 0.83, respectively, in the exponential model. The results of this study show that Sentinel-2 can be used as a suitable tool to estimate suspended sediment concentrations with acceptable accuracy in small-scale basins and flood conditions, which has been confirmed in a number of similar studies. Sediment rate refers to the positive effect of vegetation on soil protection and the reduction of sediment production and transport within watersheds. In general, the results demonstrate that vegetation has been effective in the quantity and quality of the spatial changes in the sedimentation rate of the basins. In fact, the NDVI vegetation index, as a representative of the vegetation, can be successfully used to create a statistical model of the changes in the sedimentation rate. The revitalization of vegetation should be included in development plans. © 2024 Parvizi et al.
