River and stream discharge and Salmon River flux modeling using Daymet gridded precipitation, watershed boundaries, and gaged rivers in northern Alaska, USA.

The difference between gains of Fe and loss by settling determines Fe flux. To calculate flux requires stream discharge, but because discharge is ungaged within the Salmon River basin we relied on modeled estimates of discharge to calculate Fe flux. We constructed models by regressing log-transformed discharge at USGS gage stations against the log-transformation of a gridded precipitation product, Daymet, Thornton et al. 2022) integrated over watershed boundaries. We then applied the models to Salmon basin catchments above each Fe sample location by integrating precipitation across sampling period and catchment area. Because the Kobuk River sample was near its gage, we adjusted gage discharge on the sample date with Daymet to estimate Fe flux in that river on that day. For the high discharge period, we estimated fluxes of Fetot, Fedis, and Fesus = Fetot - Fedis; for median discharge we estimated only Fedis flux.Estimating iron flux To estimate mean iron fluxes in dissolved, total, and suspended particulate forms of iron over each of three 27-day periods (27 Jul-22 Aug 2022, 30 Jun-26 Jul and 19 Jul-14 Aug 2023) we multiplied mean discharge, Q, averaged over each period by the point sample value of [Fe] in its dissolved, total, and suspended particulate forms. Because none of the Salmon River’s reaches or tributaries are gaged, we modeled discharge using a gridded (1km) daily precipitation (mm/day) product known as Daymet (Thornton et al. 2022). We doubly integrated Daymet over the watershed area A and sampling period D then used that value for each period at n = 10 Arctic Alaska rivers as gaged by the USGS to construct a linear model of log10-transformed gage discharge regressed on log10-transformed values of the integrated Daymet rasters. Daymet rasters were extracted from netcdf data downloaded as Daily Surface Weather Data on a 1-km Grid for North America, Version 4 R1 from https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=2129. We downloaded files for Jul and Aug 2022 and Jun-Aug 2023 as daymet_v4_daily_na_prcp_2022.nc and daymet_v4_daily_na_prcp_2023.nc.We calculated A from a 5m resolution interferometric synthetic aperture radar (IFSAR) digital elevation model (DEM) downloaded from the Elevation Portal of the Alaska Division of Geological and Geophysical Surveys (https://elevation.alaska.gov/). Applying hydrological tools in the R package whitebox to the DEM with gage station location as the lowest point on a watershed’s boundary where water flows out of a catchment (pour point) allowed us to delineate the catchment upstream of each pour point as a spatial polygon. Next, we extracted all Daymet pixels within the polygon for each day during a given sampling period, D, and summed them over the period and the watershed boundary, A. We calculated the double integral as “discharge from Daymet”asmeanQ = [1/(d2 - d1)] x [1 day/(24 x 60 x 60) sec] x (m/103 mm) x (106 m2/km2) x sum(pixela,t)where the sum is over all those pixel values at position a in A for the days between d1 and d2 during D and the rest of the numbers scale the days and mm of Daymet precipitation in 1-square-kilometer pixels as cubic meters per second.We calculated observed mean discharge from the gagesfor each of the n = 10 gaged watersheds and each of the three time periods and regressed log-transformedagainst log-transformedusing simple linear regression. This provided three models that we then applied to calculate Q for Fe sample locations on the Salmon River and its tributaries that we treated as hydrological pour points. The .kml files are polygons representing watershed boundaries. Some are gaged and have river names in their path names. Others are tributaries of the Salmon River and have river locations as km markers that also correspond to Salmon River watersheds.Reference: Thornton, M.M., R. Shrestha, Y. Wei, P.E. Thornton, and S-C. Kao. 2022. Daymet: Daily Surface Weather Data on a 1-km Grid for North America, Version 4 R1. ORNL DAAC, Oak Ridge, Tennessee, USA. https://doi.org/10.3334/ORNLDAAC/2129