[feedback] missing forcing or attributes // area discrepancies // PET estimates // poss.forcing bias

[Wouter Knoben]

Hi everyone,

I figured I'd kick of a feedback discussion with a few points that I encountered during my use of CAMELS-US. To start with, CAMELS has been a great resource and I'm only mentioning these  points in the hopes of making the data set even better.

1) I made a note at some point that there seem to be 7 more catchments in the meteo data than there are in the CAMELS attributes. These are: 01150900, 02081113, 03448942, 34489420, 06775500, 06846500 and 09535100.

2) As far as I can tell, there are several catchments in CAMELS that are not part of the HCDN-2009 data used in Newman et al (2015). These are 2096846, 2427250, 3338780, 5592575, 6440200, 6623800, 7263295, 7362587 and 10205030.

3) I cross-referenced the area estimates from Geospatial fabric and Gages II with data from the USGS website and found a few discrepancies (see attached figure 1; data in txt file).

4)  found some differences between the mean PET using my estimated time series (using Priestley-Taylor) and the provided mean PET attribute in CAMELS (see attached figure 2). I couldn't quite figure out why that happened and have put it down to different assumptions being used in the PET estimation and/or the way mean values are estimated (using hydro-years might lead to a small part of the time series being discarded?). I can share my PET code if you'd like to dig deeper.

5) I removed the catchments that fall outside the Budyko curve. There's some geographical patterns to these catchment (see attached figure 3). Perhaps there is some bias in the forcing data in the northwest?

6) There are some differences between the locations of catchments on a map in the Newman et al (2015) paper and the CAMELS data. Catchments are roughly in the same place but not exactly. I'm pretty sure that's due to the plots in Newman15 using catchment centroid, whereas CAMELS gives gauge lat/lon. Not sure if this is very relevant, but I figured I'd mention it.

Best,

Wouter