All the bankfull stages are calculated! WinXSPro worked like a dream.

Some of the streams have impressive variability between cross-sections, but overall, reasonably consistent values for reach-scale bankfull discharge can be computed. And the results so far are about what should have been expected: for instance, bankfull discharge magnitude below the large logjam on the Albert River (where degradation was occuring due to reduction in sediment supply by the jam) was about twice the magnitude of Qbf above the jam, where aggradation was occurring.

The unexpected result was that many, but not all of the streams have Qbf return periods significantly higher than the Q2. Explaining this should make a nice paper.

Right now I am working on estimating the effective discharges. I'm using, or trying to use, the Sedimetrics Digital Gravelometer. Having some problems generating results: the program keeps stalling in the report generation process. David Graham at U of Loughborough has suggested some potential fixes, which I am currently trying.