Loading...

Improving River Discharge Estimation by Considering the Impact of Water Quality via Optical Remote Sensing Data

Tayebi Alashti, Amir Hossein | 2025

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 58250 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Danesh Yazdi, Mohammad
  7. Abstract:
  8. Recent years have seen the development of various remote sensing-based methods for river streamflow estimation. Among them, the C/M approach estimates streamflow by analyzing surface reflectance at two benchmark pixels, namely dry and wet. However, its efficiency heavily depends on the expert-selected dry and wet pixels, which cannot be chosen randomly. In this study, we introduced the Normalized Difference Streamflow Index (NDQI) as a novel proxy for streamflow estimation. This index, which utilizes the spectral bands of the Sentinel-2 satellite, is sensitive both to soil moisture changes along riverbanks and turbidity variations during high-flow events. Unlike the C/M approach, NDQI does not rely on dry pixels, yet it proves to be a robust predictor of streamflow. We also proposed an automated framework for selecting wet pixels that maximizes surface reflectance sensitivity to streamflow fluctuations. Additionally, to demonstrate the superiority of NDQI against the C/M approach, we developed an automated dry pixel selection method. We examined the applicability of the above framework at six gauging stations along the Mississippi River, where continuous and simultaneous streamflow were available. Furthermore, to assess the improved performance of NDQI, turbidity data were used as as auxiliary information to identify appropriate bands and pixels for extracting river water quality’s spectral behavior. The results showed that NDQI at shallow riverbank pixels with high short-wave infrared reflectance yields reliable streamflow predictions, evidenced by NSE ranging from 74% to 92%. The findings also indicated that the spectral band used in NDQI can successfully reflect changes in the water quality. However, due to bathymetric variability and potential chemical pollutants along different sections of the river, no fixed secrion of the river can be assumed to maintain a consistent correlation between spectral features and turbidity
  9. Keywords:
  10. Remote Sensing ; Sentinel Satellite ; Normalized Difference Stream Flow Index (NDQI) ; Stream Flow Rate ; Automated Dry and Wet Pixel Selection

 Digital Object List

 Bookmark

...see more