CLICK HERE TO REGISTER
Note: all event times are in eastern (EST/EDT)
Join us for an insightful webinar on HydroSphereAI, where we explore a cutting-edge Machine Learning (ML)-driven approach to hydrological forecasting. Learn how our models improve streamflow predictions, even in ungauged basins, and how real-time data integration enhances forecasting accuracy. We’ll also demonstrate the HydroSphereAI platform and its practical applications for water resource management.
Abstract:
Hydrological forecasting remains a critical challenge for water resource management, especially in northern climates where seasonal snowpack and drought profoundly affect agriculture. This presentation outlines a promising Machine Learning (ML)-driven approach to operational forecasting that advances beyond traditional rainfall-runoff (RR) models.
We begin by outlining the core technology underlying our ML forecasting models, including key predictors and the training process; a critical component that will also be discussed in some detail is the training dataset.
Evaluation of historical simulations against observed streamflow demonstrates excellent skill, even in ungauged basins, where traditional RR models struggle. We also show how ML forecasts can be interpreted in physical terms and extended to other critical variables such as snowpack.
In the operational forecasting segment, we discuss real-time acquisition and processing of weather forecast data, as well as integration (assimilation) of real-time observational data into the ML models.
We close with a brief demonstration of our HydroSphereAI forecasting and decision support platform, showcasing practical use cases for stakeholders. Overall, this presentation aims to bridge the gap between AI/ML research and operational forecasting, and to offer a roadmap for deployment of ML-driven hydrological forecasts across Canada.
Presenter Bio:
Dr. Andre Erler is a Senior Climate Scientist at Aquanty Inc. His main expertise lies in regional climate modelling and applications to hydrology and hydro-climatic extremes. Originally from Germany, he received the equivalent of a M.Sc. in Meteorology from the University of Mainz in 2008 and a Ph.D. in Physics from the University of Toronto in 2015. Other interests also include open source software, sustainable development, and food security. For his dissertation he performed high resolution climate simulations for western Canada and studied hydrological impacts of climate change in the Athabasca and Fraser river basins, as well as changes in precipitation extremes due to climate change.
Andre joined Aquanty in 2016 in order to study the impact of climate change on water resources and agriculture in Canada and provide insights into climate change impacts to end-users.