Light model and GPP estimates for 173 U.S. rivers

This data release contains products from Savoy & Harvey (2021) which modeled photosynthetically active radiation and gross primary productivity (GPP) for 173 streams and rivers across the continental United States. Estimates of photosynthetically active radiation (PAR) were generated using a model that considers the contribution of light attenuation from riparian zones as well as within the water column. PAR estimates at the stream surface and within the water are both provided and each of these two estimates were converted to GPP estimates using a fixed light use efficiency. This data release includes the site parameters, model inputs, model outputs, and model source code use to generate these estimates.

Items in this data release consist of:
Site information and parameters : Information on site identifiers, location information, and model parameters. The format is a tab-delimted file with 1 row per site.
Model inputs and outputs: Timeseries of the necessary input data used to estimate photosynthetically active radiation as well as the resulting modeled estimates. The format is a tab-delimited file for each site that contains all model inputs and outputs at hourly timesteps.
Validation dataset: A common validation set for estimates of gross primary productivity generated from estimated light and a fixed light use efficiency. This dataset represents a subset of all predictions where all necessary data was available and independent estimates of gross primary productivity from Appling et al. (2018) (https://www.sciencebase.gov/catalog/item/59bff507e4b091459a5e0982) were also available. The format is a single compiled tab-delimited file with 1 row per site-day.
Model source code and documentation: Model source code for StreamLight v0.2.0 built as a binary package, an R environment with all site parameters, model driver files, and model source code from this data release loaded. An example markdown file is included to give a brief summary of the model and a demonstration of running the model.
This work was supported by the USGS Water Availability and Use Science Program.