River water extent is essential for river hydrological surveys. Traditional methods for river water mapping often result in significant uncertainties. This paper proposes a support vector machine (SVM)-based river water mapping method that can quantify the extraction uncertainties simultaneously. Five specific bands of Landsat-8 Operational Land Imager (OLI) data were selected to construct the feature set. Considering the effect of terrain, a widely used terrain index called height above nearest drainage, calculated from the 1 arc-second Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM), was also added into the feature set. With this feature set, a posterior probability SVM model was established to extract river water bodies and quantify the uncertainty with posterior probabilities. Three river sections in Northwestern China were selected as the case study areas, considering their different river characteristics and geographical environment. Then, the reliability and stability of the proposed method were evaluated through comparisons with the traditional Normalized Difference Water Index (NDWI) and modified NDWI (mNDWI) methods and validated with higher-resolution Sentinel-2 images. It was found that resultant probability maps obtained by the proposed SVM method achieved generally high accuracy with a weighted root mean square difference of less than 0.1. Other accuracy indices including the Kappa coefficient and critical success index also suggest that the proposed method outperformed the traditional water index methods in terms of river mapping accuracy and thresholding stability. Finally, the proposed method resulted in the ability to separate water bodies from hill shades more easily, ensuring more reliable river water mapping in mountainous regions.

Quantifying river discharge is a critical component for hydrological studies, floodplain ecological conservation research, and water resources management. In recent years, a series of remote sensing-based discharge estimation methods have been developed. An example is the use of the near infrared (NIR) band of optical satellite images, with the principle of calculating the ratio between a stable land pixel for calibration (C) and a pixel within the river for measurement (M), applying a linear regression between C/M series and observed discharge series. This study trialed the C/M method, utilizing the Harmonized Landsat and Sentinel-2 (HLS) surface reflectance product on relatively small rivers with 30~100 m widths. Two study sites with different river characteristics and geographic settings in the Murray-Darling Basin (MDB) of Australia were selected as case studies. Two independent sets of HLS data and gauged discharge data for the 2017 and 2018 water years were acquired for modeling and validation, respectively. Results reveal high consistency between the HLS-derived discharge and gauged discharge at both sites. The Relative Root Mean Square Errors are 53% and 19%, and the Nash-Sutcliffe Efficiency coefficients are 0.24 and 0.69 for the two sites. This study supports the effectiveness of applying the fine-resolution HLS for modeling discharge on small rivers based on the C/M methodology, which also provides evidence of using multisource synthesized datasets as the input for discharge estimation.

Monitoring of river discharge is a key process for water resources management, soil and water conservation, climate change, water cycling, flood or drought warning, agriculture and transportation, especially for the sustainable development of rivers and their surrounding ecological environment. Continuous and comprehensive discharge monitoring was usually impossible before, due to sparse gauges and gauge deactivation. Satellite remote sensing provides an advanced approach for estimating and monitoring river discharge at regional or even global scales. River morphology is generally considered to be a direct factor that affects the accuracy of remote sensing estimation, but the specific indicators and the extent to which it affects the estimation accuracy have not yet been explored, especially for medium to small rivers (width < 100 m). In this paper, six sites with hydrological gauges in the upper Heihe River Basin (HRB) of northwestern China and the Murray Darling Basin (MDB) of southeastern Australia were selected as the study cases. River discharge was estimated from Landsat imagery using the C/M method accordingly. River gradient, sinuosity, and width were obtained from Digital Elevation Model data for each site. Global Surface Water Dataset (GSWD) was also employed for indicating the dynamic status of river morphology. A series of methods were applied to analyze the influence of river morphology on estimation accuracy qualitatively and quantitatively, based on which we established inference about the theory of selecting satellite virtual gauges (SVGs). The results confirm the feasibility of the C/M method for discharge estimation, with the accuracy affected by multiple river morphological indicators. Among them, river width was found to be the most significant one. Moreover, water occurrence and water extent extracted from GSWD also have impact on the discharge estimation accuracy. Another independent river section in MDB was set as an example to demonstrate the reasonability of the established theory. It is anticipated that this study would promote the application of remote sensing for discharge estimation by providing practical guidance for establishing appropriate SVGs.