Description
In the current design, each HRU drains to a single 'downhill' unit and flow is computed as a function of the length of the shared boundary. When HRUs are delineated independent of routing considerations, there are cases when a single HRU can flow to multiple 'downhill' units. A simple example is a large HRU wrapping around the upper reaches of a headwater catchment, with a patchwork of smaller HRUs below, delineated based on e.g., vegetation, riparian, soils, classes.
I've spoken to Martyn about this. His suggestion was to 1) change the input structure to a ragged array to accept a vector rather than a scalar of lateral HRU connections and associated contourLength values and 2) partition the calculated fluxes based on the proportion of the total (summed) contour lengths.
I wanted to start an issue to raise awareness and open discussion.