Major Inconsistency: Outlier in flood storage per square mile

[ghost]

08079000 has a problem. It is responsible for three and certainly two massive outliers in unit flood regulation. It has CDA = 114 square miles but tot_basin_area=104060.1 km^2 * 0.3861 ---> 40,200 square miles. I think we have a flow accumulation BUST. Somehow NHD+ or other techniques are grabbing vastly more reservoirs than reality. I bet the entire Brazos River basin in Texas as leaked into the record for this streamgage.

decade tot_nid_storage tot_norm_storage 1950 1117648 711769.1 1960 5650500 2008682.1 1970 10949171 3366243.5

Example computation: (10949171 - 3366243.5)/114 = 66,500 acre-feet(?) per square mile.

When I check the record for 08116650 Brazos River Rosharon and 1970, I get 10948911 for tot_nid_storage, which is awfully close to that for 08079000. This supports my concern not well stated in Nashville that the HUC-NHD+ discussion that we had on the whiteboard does not in fact adequately capture full real upstream properties for a given streamgage.

Circa 1999, I did similar aggregations for 327 streamgages in Texas using just a base map of 1:100k streamlines and click,click,clicked vertices of a quick bounding polygon for the basin upstream of the streamgage from which I then aggregated the cumulative storage and cumulative normal storage using ArcInfo dumps and a FORTRAN "script" that I had written.

[scworland]

Here is the the total NHD+ basin boundary for 08079000:

and for 08116650:

So yeah, they are basically the same thing. Is that unexpected? Maybe a distance metric needs to be used for the dam storage?

[ghost]

I do not disagree with an idea of a distance metric. I hypothesis (informed though) that proximity of bulk storage to a watershed outlet is important. I live in the far distal reaches of the delineation and reservoir storage in my part of the world has no bearing on flow into the Gulf---particularly given a lot of storage in the middle part of the basin towards the Gulf.

The problem is one that needs recification. Presumably the screen capture below has about 114 square miles associated with it. I hack at about 50 miles crow-fly-length and say watershed width about 3 to 4 miles (Google Earth), then I hack into about 150 to 200 square miles, which is of the same order as the recognized CDA by the USGS. (Morphologically, I suspect Oyster Creek is ancestral Brazos River, but short of hurricane type rainfall, the flows are disconnected.) Google Earth viewing suggests just a little reservoir storage associated with suburban development.

What we should have our hands on for RESTORE is the ASCII dump of the NID itself. I don't have a copy anymore (mine was circa 1999). If we had it, then I or my student could integrate the NID and we could do some field replacement in your feather file.