VBET & GNAT → RME (riverscapes metric engine) - New DGO Metric

[joewheaton]

Can we take the contributing drainage area raster from TauDEM and have a metric for every DGO, which is nothing more than the max drainage area in the DGO (in square kilometers I think). This should simply produce something similar to the channel network drainage area that we get from NHD+HR, but would have the benefit that we could calculate it even for custom Riverscape Contexts not based on NHD (coming very soon), and it would provide a nice diagnostic check if we are screwing up in the handoff from upstream HUCs on the drainage area outside the raster grid.

@jtgilbert, @lauren-herbine and @shelbysawyer FYI

[philipbaileynar]

Are you asking for direct, hillslope drainage (i.e. drainage wing area) or total upstream drainage? The former is doable. The latter is not.

At best, the latter will only work for first and maybe second order streams. Anything bigger will have a drainage area extends beyond the HUC10 extent at which we intend to run models. There is no, and has never been, spatial handoff between model runs. We only have in situ hand off between models in the waterfall.

[joewheaton]

Are you asking for direct, hillslope drainage (i.e. drainage wing area) or total upstream drainage? The former is doable. The latter is not.

Former... it is more then just drainage wings though. It is just the within input DEM bounds.

At best, the latter will only work for first and maybe second order streams. Anything bigger will have a drainage area extends beyond the HUC10 extent at which we intend to run models. There is no, and has never been, spatial handoff between model runs. We only have in situ hand off between models in the waterfall.

@jtgilbert what did we do in BRAT to handle this and does TauDEM do anything for this? I'm just trying to understand how this problem will play out as we move to other countries and other custom extents.

[philipbaileynar]

There are lots of questions to be answered.

Can we inject a pre-existing reach network into the tauDEM workflow by rasterizing it and then using it to determine the stream catchments? This pre-existing reach network won't perfectly match the topography. Is that a problem?

Alternatively, we could run the full tauDEM workflow to derive the reach network first and then the catchments. This method should work, but there's no crosswalk between these new tauDEM reaches and the official network that has the global identifiers (NHDID, COMID or whatever system is used in other countries).

I'm not saying this can't be done. But topology is hard, brittle and squirrelly. Prepare for this to take considerable effort and to fall apart in lots of places (typically flat places where it is most interesting).

[joewheaton]

Understood @philipbaileynar. There are things I just want us to start thinking about going forward... not expecting everything to be easy. In meantime, this is still a useful DGO metric and it flags differences between what NHD might give us and what we are seeing.