Streamflow calculation for medium-to-small rivers in data scarce inland areas

Inland streamflow estimation is essential in global water supply and environment protection. In data-scarce areas a highly efficient way of estimating streamflow is through remote sensing methods. However, high requirement of most previous methods on ground-measured data hinder their wide use in data-scarce areas. Therefore, this paper presented a new framework for estimation of streamflow in medium-to-small rivers with few ground measurements by using high-resolution unmanned aerial vehicle (UAV) imagery. A new Virtual Hydraulic Radius (VHR) method was proposed to complement AMHG (at-many-stations hydraulic geometry), a method not requiring any ground measurements when global parameters are used (global-AMHG) in large-scaled rivers but yielding great uncertainties in smaller scaled rivers, thus creating a VHR-AMHG method for medium-to-small rivers. The accuracy verification of the proposed method was performed by comparing it to field measurement data and the global parameters of the original AMHG (global-AMHG). Results showed that the root mean square error calculated from VHR-AMHG was 32.15 m³/s, while that from global-AMHG was 305.65 m³/s, indicating that the VHR-AHRG method yields a significantly higher accuracy for streamflow estimation for medium-to-small rivers. We found that regardless of the size of the river, AMHG is not applicable for rivers having excessively small b values in the equation w = aQ^b (low-b rivers). For medium-to-small rivers with b < 0.25, AMHG is not recommended. The accuracy of the original AMHG method is limited by the initial value of the model parameters and the condition that the congruent discharge (Q_c) has to be within the range of observational discharge. The initial value setting of the model parameters significantly impacts the calculation accuracy. The VHR-AMHG method is able to overcome the deficiencies of the original AMHG, i.e. being overly dependent on the initial value setting with long-series known discharge data. It also eliminates the limitation of the Q_c condition, as it achieves a higher accuracy for rivers in which Q_c does not satisfy the condition compared to using global-AMHG on rivers that actually meet the condition, thus greatly expanding its usage scope. Thus VHR-AMHG method can provide detailed data on the spatial and temporal distribution of regional and national streamflow for governments and stakeholders, and offer scientific data support for wisely making water supply polices and sustainably protecting eco-environment.