Virtual station rating curves derived from hydraulic models informed with UAS hydrometry and SWOT WSE

With the increasing frequency of extreme weather events, such as river flooding, there is a growing need for more cost-effective and efficient methods for hydrometric river monitoring. Moreover, traditional in-situ hydrometric surveys often face challenges when applied to remote or hard-to-access river locations. Therefore, we investigated the potential of using Unoccupied Aerial Systems (UAS) hydrometry surveys to develop a hydraulic model for extracting rating curves, which can then be used to derive discharge from satellite altimetry-based Water Surface Elevation (WSE) measurements.

This study employed UAS-borne Water Penetrating Radar (WPR) to map river bathymetry, while Digital Elevation Models (DEM) were used to extract the non-submerged portions along the WPR cross section. A Real-Time Kinematic (RTK) Global Navigation Satellite System (GNSS) receiver provided ground truth WSE measurements. Additionally, the pixel cloud data product of the Surface Water and Ocean Topography (SWOT) satellite mission was used to extract WSE along the river. Furthermore, based on the cross sectional geometry information, we constructed two one-dimensional hydraulic models, one using a steady gradually-varied flow solver and the other using the MIKE+ hydrodynamic solver.

The study site is located along the Torne River in northern Scandinavia, which forms part of the national border between Sweden and Finland. From September 3rd to 9th, 2024, surveys were conducted at 23 field sites distributed across two areas of interest along the river. Manning's numbers for the river reaches were calibrated against WSE observations derived from the SWOT pixel cloud dataset using the steady gradually-varied flow solver. Hydraulic models were employed to construct rating curves at chainage locations where observations from the Sentinel-3 satellite mission were available at two defined virtual stations: Övertorneå and Pello. These rating curves were subsequently used to convert WSE observations by the SWOT pixel cloud and Sentinel-3 to discharge, enabling the construction of a river discharge time series.