Articles | Volume 16, issue 14
Model evaluation paper
20 Jul 2023
Model evaluation paper |  | 20 Jul 2023

Modeling river water temperature with limiting forcing data: Air2stream v1.0.0, machine learning and multiple regression

Manuel C. Almeida and Pedro S. Coelho

Model code and software

mcvta/WaterPythonTemp: Release 0.2.0 M. C. Almeida and P. S. Coelho

Short summary
Water temperature (WT) datasets of low-order rivers are scarce. In this study, five different models are used to predict the WT of 83 rivers. Generally, the results show that the models' hyperparameter optimization is essential and that to minimize the prediction error it is relevant to apply all the models considered in this study. Results also show that there is a logarithmic correlation among the error of the predicted river WT and the watershed time of concentration.