This repository hosts the validated TETIS_Runtime_Predictor a Machine Learning (ML) predictive tool developed to estimate the computational performance (runtime) of the TETIS v9.1 hydrological model. The package contains the trained Random Forest (RF) regression models and the associated Python scripts required for execution. This tool enables researchers and end-users to predict the runtime of two critical processes: Topolco.sds generation (Parallel process) Hydrological Simulation execution (Serial process) This resource is vital for planning large ensemble experiments, optimizing resource allocation, and improving the operational reliability of TETIS software. The figure Runtime TETIS experiments shows runtime performance of hydrological simulation for experimental design. Columns correspond to hardware configurations, while rows represent the main variables of interest: number of basin cells, number of time steps, input gauge density, and output gauge density. Each marker represents an individual simulation; marker colors indicate the combined input–output gauge density, while marker shapes represent the number of time steps. This figure presents the complete runtime distribution underlying the summary results shown in Figures 5 and 6. Cite the associated article when using this predictive tool: Scalability and Computational Performance of an Ecohydrological Model Using Machine Learning-Based Prediction DOI: https://doi.org/10.3390/w18040466