This section provides code for reproducing the figures in Rahmani et al. (2023b).

This section provides model code described by Rahmani et al. (2023b). This code accepts basin attributes and forcings and predicts stream temperatures using a differentiable model with neural network and process-based equation components. Code files are contained within code.zip. A description of each code file is given in the 01_code.xml metadata file and also in code_file_dictionary.csv. Instructions on how to run the code are given in code_readme.md.

This section provides model code described by Rahmani et al. (2023b). This code accepts basin attributes and forcings and predicts stream temperatures using a differentiable model with neural network and process-based equation components. Code files are contained within code.zip. A description of each code file is given in the 01_code.xml metadata file and also in code_file_dictionary.csv. Instructions on how to run the code are given in code_readme.md.

This section provides model simulation outputs from the models described by Rahmani et al. (2023b), as well as a subset of model outputs produced by Rahmani et al. (2021) that were used for comparison within Rahmani et al. (2023b).

This model archive (Rahmani et al. 2023a) provides all data, code, and model outputs used in Rahmani et al. (2023b) to improve model representations toward improved prediction of stream temperature and groundwater/subsurface flow contributions to stream temperature. Briefly, we modeled stream temperature at sites across the continental United States using a hybrid differentiable model that combines neural network components with differentiable implementations of several structural priors, i.e., process-based equations. The differentiable framework permits estimation of parameters and comparison of structural priors as well as prediction of stream temperature.

The data are organized into these child items:

This model archive (Rahmani et al. 2023a) provides all data, code, and model outputs used in Rahmani et al. (2023b) to improve model representations toward improved prediction of stream temperature and groundwater/subsurface flow contributions to stream temperature. Briefly, we modeled stream temperature at sites across the continental United States using a hybrid differentiable model that combines neural network components with differentiable implementations of several structural priors, i.e., process-based equations. The differentiable framework permits estimation of parameters and comparison of structural priors as well as prediction of stream temperature.

The data are organized into these child items:

This data release component contains shapefiles of river basin polygons and monitoring site locations coincident with the outlets of those basins. Three file formats describing basin attributes, and three file formats describing forcing and observational data, are also included. These data were used to train and test the stream temperature prediction models of Rahmani et al. (2023b).

This data release component contains shapefiles of river basin polygons and monitoring site locations coincident with the outlets of those basins. Three file formats describing basin attributes, and three file formats describing forcing and observational data, are also included. These data were used to train and test the stream temperature prediction models of Rahmani et al. (2023b).