NERC-CEH/ghana_hydro_training_2026

Repository for scripts and data for the Ghana training workshop planned in early 2026 # UKCEH West Africa Hydrological Forecasting Training Workshop 2026 Welcome to the repository for the UKCEH West Africa Hydrological Forecasting Training Workshop 27-29 January 2026. This repository supports a three-day training workshop focused on building end-to-end capability in hydrological forecasting, from data access and exploration to modelling and operational forecast applications. The workshop progresses from working with hydroclimate datasets, through machine-learning-based water resources modelling, to the generation and interpretation of sub-seasonal to seasonal hydrological forecasts for decision making. The materials are organised by day and theme, with each section containing follow-along notebooks, scripts, example datasets, and guidance to enable participants to reproduce and adapt the methods introduced during the sessions. --- ## Training Objectives The high-level objectives of this training workshop are: - Demonstrate how to access, retrieve, and work with remote hydrological and hydroclimate datasets using Python and common data access protocols. - Introduce key hydrological and hydroclimate datasets and explore practical techniques for data processing, visualisation, and exploratory analysis. - Provide hands-on experience with water resources modelling using machine-learning approaches, including feature engineering, model training, hyperparameter tuning, and model evaluation. - Introduce conceptual hydrological modelling for catchment-scale applications and guide participants through model calibration and simulation. - Demonstrate the generation and interpretation of sub-seasonal to seasonal hydrological forecasts, including ensemble streamflow prediction. - Explore approaches for visualising, categorising, and communicating forecast products to support decision-making in operational contexts. ## Running the Training Notebooks on Google Colab To run the training notebooks using *Google Colab*, follow these steps: 1. **Log in to your …