Low-Altitude Remote Sensing Inversion of River Flow in Ungauged Basins

Abstract

Runoff is closely related to human production, the regional environment, and hydrological characteristics. It is also an important basis for water cycle research and regional water resource development and management. However, obtaining hydrological information for uninformed river sections is complicated by harsh environments, limited transportation, sparse populations, and a low density of hydrological observation stations in the inland arid zone. Here, low-altitude remote sensing technology was introduced to combine riverbed characteristics through unmanned aerial vehicle (UAV) inversion with classical hydraulic equations for ungauged basins in the middle and lower reaches of the Keriya River, northwest China, and investigate the applicability of this method on wide and shallow riverbeds of inland rivers. The results indicated that the estimated average error of the low-altitude remote sensing flow was 8.49% (ranging 3.26–17.00%), with a root mean square error (RMSE) of 0.59 m3·s−1 across the six selected river sections, suggesting that this method has some applicability in the study area. Simultaneously, a method for estimating river flow based on the water surface width– and water depth–flow relationship curves for each section was proposed whereas the precise relationships were selected based on actual section attributes to provide a new method for obtaining runoff data in small- and medium-scale river areas where information is lacking.