The use of site-calibrated models for predicting bedload transport in gravel-bed rivers remains relatively rare, despite advances in methodology and computing technology, and its notable advantages in terms of predictive accuracy. This article presents a new algorithm for site calibration of the Parker–Klingeman (1982) model, along with a detailed discussion of considerations that influence model selection and calibration methodology. New visualization techniques are explored to demystify the calibration process, using three examples with progressively more challenging calibration conditions. The new method is particularly well suited to streams with high sediment loads, or cases where extrapolation of transport function estimates is necessary.
