Upstream Tech releases next generation of best-in-class AI water forecasts: HydroForecast ST-3

Boston, July 9, 2024 - Upstream Tech, an innovator of scalable AI-based climate software solutions, has announced today the release of HydroForecast Short-term 3 (ST-3), the latest generation of its short-term water forecasting product that predicts streamflow 0-10 days ahead with industry-leading accuracy. HydroForecast is a preferred solution for hydropower operators and water resource managers owing to its accuracy, reliability, and interpretability, all three of which are substantially bolstered in HydroForecast ST-3 due to AI model breakthroughs.

With increased frequency of extreme weather events, climatic unpredictability is causing significant challenges for sectors dependent on water resources – especially hydropower, which is subject to reduced power generation under drought conditions, and, conversely, damage to hydropower infrastructure and downstream areas owing to storm surges. Consequently, these sectors have become reliant on short- and medium-term water forecasts to make decisions, where accuracy, reliability, and interpretability are paramount.

However, water forecasts are only as reliable as their inputs, like meteorological forecasts, which historically face disruptions from power outages, technical issues, and other unexpected events. In the context of extreme weather, these kinds of disruptions jeopardize the safe and efficient operation of infrastructure.

By leveraging a distributed machine learning model, HydroForecast ST-3 is robust to missing or delayed inputs, giving users peace of mind that an accurate forecast will always be available during crucial operational events. For hydropower operators,  this results in safer operations and base load resilience into an increasingly renewable grid.

By reinforcing its commitment to resilient, accurate and timely forecasting through HydroForecast ST-3, Upstream Tech now provides hydropower and water resource industries the following advantages:

• Enhanced reliability: HydroForecast ST-3 is designed to withstand individual input failures, ensuring uninterrupted operational reliability even during weather forecast outages.
• Improved accuracy: The AI breakthroughs employed in ST-3 result in an accurate distributed model and additionally facilitates seamless integration of specialized regional weather forecasts, such as NOAA's High Resolution Rapid Refresh.
• Increased frequency: HydroForecast ST-3 produces forecasts more frequently, helping users better refine their operational planning, especially during extreme weather events.
• Improved forecast interpretability: HydroForecast ST-3 provides detailed per-weather source forecast traces, in addition to its primary forecast, enhancing interpretability and usability.

"We are trying to change the way organizations make water decisions by arming them with the best possible forecasts. ST-3 is the culmination of years of R&D conducted by our team and represents a step change in the kind of accuracy and resiliency we can offer to our customers." said Marshall Moutenot, Chief Executive Officer, Upstream Tech.

"We are passionate about the energy transition and the role hydropower is playing in increasingly renewable energy grids. Arming our hydropower partners with information that will lead to safer operations in the face of extreme weather, flexible operations as our grid evolves, and optimal generation to improve revenue - those are huge wins that my team is exceedingly proud of."

To learn more about HydroForecast ST-3, get in touch with our team today.