Atmospheric rivers (ARs)–also commonly referred to as a river in the sky, Pineapple Express, moisture plume, water vapor surge, cloud band–are long, narrow bands of concentrated moisture that result in intense precipitation. As a vital component of the hydrological cycle, ARs pose both challenges and opportunities for water resource management.
HydroForecast empowers water managers and dam operators to predict ARs with greater accuracy by leveraging AI to reduce uncertainties. Accurate forecasts provide critical insights to enhance dam safety, reduce flood risk, minimize asset damage, and optimize water usage.
Increasing in intensity and frequency globally, ARs present many uncertainties that are difficult to forecast.
ARs are extremely thin, measuring several hundred miles across and thousands of miles long, and they hold a huge volume of water, on average carrying the mean flow of the Mississippi River (NOAA). When they make landfall, they unleash long and narrow downpours that can overwhelm water systems. Minor fluctuations in AR trajectories can result in meaningful differences in precipitation for adjacent watersheds, creating a wide range of possibilities for how water managers prepare for and respond to these extreme events.
This occurred in spectacular fashion in 2017 at the Oroville reservoir in California where higher than predicted precipitation threatened dam collapse, causing roughly 200,000 people downstream to preemptively evacuate.
Water managers must weigh uncertainties in weather forecasts while also factoring in how precipitation flows into rivers. Key questions that impact their decision making include:
The complex nature of the timing and magnitude of ARs introduces further uncertainty. Even in the days leading up to the peak, AR forecasts often show a wide distribution of possibilities to prepare for, creating difficult decisions around when and how much water to hold or release.
Despite these challenges, ARs also offer potential benefits that have a major impact on water supply. For example, ARs are a critical source of water in California–half of the state's water originates from ARs–but they are also responsible for 84% of flood damage throughout Western US.
Dam operators must constantly hedge risks to ensure that dams effectively mitigate negative impacts of intense precipitation events like flooding while capturing the water for energy generation and drinking water needs during drier periods.
All of these uncertainties make accurate, reliable forecasting for AR all the more important.
By learning from errors and reducing uncertainties, our AI-driven forecasts allow water resource managers and dam operators to integrate advanced technologies into their operations for more actionable insights.
Using multiple leading meteorological forecasts (GEFS, ECMWF, HRRR), satellite data sources (SNODAS, VIIRS), and a physics informed machine learning model, HydroForecast provides more accurate streamflow forecasts compared to other leading industry models.
Here’s a look at how HydroForecast provided early warnings in a recent intense precipitation event in California from February 2025.
.gif)
Although ARs are commonly associated with the West Coast, they occur across the world - posing similar risks and rewards to countries in western Europe, southern South America, and in Oceania. The learnings from the US examples are applicable globally to all regions that experience ARs and any intense precipitation events, such as hurricanes, monsoons and bomb cyclones.
HydroForecast enables more informed decisions before and during intense precipitation events with more confidence.
Talk to our team about how HydroForecast can help your team maximize benefits while minimizing the associated risks of intense precipitation events.