Machine Learning for Rainfall–Runoff Modelling: Fundamentals, Case Study & Model Comparison

In this live session, we take you from the fundamentals of ML for rainfall–runoff modelling to developing models for a real catchment in NSW, and then present a practical comparison of different modelling approaches. This practical session connects ML concepts, hydrological modelling, and a real-world case study. The outline & timestamps of this live webinar are as follows: Timestamp 00:00 Introduction: 02:41 Introducing Water Edu Hub 03:22 Introducing the speakers 05:09 Scope 07:23 Housekeeping & Engagement 07:51 Why Machine Learning for Rainfall–Runoff Modelling? 08:07 Challenges in Simulating Rainfall-Runoff Process 11:59 Opportunities in Machine Learning 13:58How Does an ML Rainfall–Runoff Model Work? An End-to-End Workflow for Building an ML-based Rainfall-Runoff Model 14:16 Data Requirement 16:35 Data preprocessing 22:00 Input selection 30:50 Model Training & Validation 33:38 Model Testing & Performance Metrics 34:49 Meet the Three ML Algorithms 35:11 Why these three? 37:03 Artificial Neural Network (ANN) 39:56 Random Forest (RF) 42:40 XGBoost 47:02 The Benchmark: Conceptual Modelling with GR4J 47:37 GR4J Structure & Calibration Parameters 49:36 Why use it as a benchmark? 50:04 Case Study: Clarence Catchment, NSW, Australia 50:37 Catchment details and available data 56:57 Input selection 1:05:35 Model training & validation 1:07:34 Model Results: Comparison, Benchmarking & Interpretability 1:07:39 Performance metrics 1:09:18 Visual Performance Assessment: Hydrographs & Scatterplots 1:10:19 Peak Flow Estimation Performance 1:12:08 Overall Model Comparison 1:12:59 Peeking Inside the Black Box (Interpretability) 1:18:34 Live Demo of Model Calibration (handling the overfitting problem) 1:37:02 Practical Key Takeaways 1:37:08 What worked & what didn't? 1:40:03 When should we use ML? 1:41:42 What should we be cautious about? 1:43:52 What’s next? (Overview of upcoming training workshops) 1:45:14 Q&A Also, watch a related video on different types of machine learning techniques for rainfall-runoff modelling:    • Machine Learning in Rainfall Runoff Modell...   #machinelearning #hydrologicalmodelling #hydrology #hydrologytutorial #rainfallrunoffmodel #flowsimulation #ann #randomforest #xgboost