This will be based on the global DEM, but with processing possibly done at up to 90m resolution before aggregating to produce the globally consistent 1km product. Ming reports that a 90m flow accumulation (catchment area) layer has been completed for the Oregon case study, but with "some artefacts still present". Whatever procedure was used to produce this layer should as least serve as a good starting point, but we also need clarification on what are the remaining issues.