This scenario isn’t hard to imagine: It’s the middle of summer. The entire year has been unusually dry, drought conditions took over months ago, and they’ve only gotten worse. Your phone dings, and you see a notification telling you something that’s been in the back of your mind for a while. Officials are now asking people to use less water, and if things don’t improve, tougher water-conserving measures might have to happen.
But what if water resource managers had more useful weather information that helps them make plans to hold back water early on and potentially avoid such situations, even though they can’t stop the drought itself?
That’s where subseasonal-to-seasonal forecasts, or S2S forecasts, come in. It might not seem apparent at first, but weather at the S2S time scale — two weeks to two years in the future — has a big effect on our daily lives and the economy.
“Forecast-Informed Reservoir Operations” are a growing strategy among water resource managers to ensure adequate water supplies. Power utilities can anticipate demand with greater lead time, making proactive preparations for grid resiliency weeks ahead of extreme heat or cold. Longer outlooks can help farmers decide whether to adjust their fieldwork plans or expected yields. But impacts go beyond the farm: If the Mississippi River is low because of drought, barges can’t load up as much grain, increasing prices at the grocery store. And in public health, extreme heat and cold contribute to thousands of deaths across the U.S. each year; municipalities and organizations could have more time to prepare and allocate resources such as warming or cooling centers.
These are just a few examples that illustrate the power and value of S2S information for decision-making. However, developing these forecasts presents a unique set of challenges.
Short-term forecasts (“what’s tomorrow’s high temperature?”) depend largely on what’s happening in the atmosphere right now. Meanwhile, long-term forecasts (“how might regional rainfall patterns change over the next 10 years?”) come more from the slower-moving parts of the land-ocean system, such as sea surface temperatures. But S2S forecasts are basically a blend of all of that, making them an especially tricky problem to solve. It’s a whole “Earth system” puzzle — atmosphere, land, ocean, sea ice, and so on. The interactions among them are complex and difficult to understand.
Communicating S2S forecasts is also difficult. If a stakeholder is told “the chance of drought conditions developing in the next two months is 70%,” what do they do with that information? Is there something they should do to prepare? What things can they do that they haven’t yet considered? These are tough questions, and because users’ needs vary so widely, there’s no one-size-fits-all answer. This is where S2S efforts bridge into the social sciences — the forecasts can be much more effective when we understand how people use probabilistic information and tailor it to help them make appropriate decisions for their situation.
Despite these challenges, scientists are making progress. At the center of those efforts is making better models.
First, there’s data assimilation. In essence, this involves figuring out how to take all the right observations of the Earth system and feed them into a model so that it starts with a satisfactory representation of what’s really happening. One such example is the Joint Effort for Data assimilation Integration (JEDI). It’s a community-based initiative to provide a unified data assimilation framework to model the Earth system.
Those data assimilation efforts tie into another area of progress: coupling of the various Earth systems within a model. This includes connections like land-atmosphere and ocean-atmosphere. Energy transfers, moisture, and even aerosols like dust and smoke all play a role.
It’s not enough to just have a better understanding of how to do those things; they have to be put into practice. Emerging technologies like artificial intelligence and machine learning are beginning to help scientists in both the public and private sector. While it’s too early to say how much these techniques are improving S2S forecasts just yet, they’ve already shown promise in other weather forecasting applications.
The Weather Program Office supports this critical S2S research, championing a range of projects, including its dedicated S2S Program. By accelerating the development of these unique forecasts, WPO helps ensure that communities and decision-makers receive the vital information they need to better prepare for challenging weather weeks and months ahead.
This story was published on March 16, 2026.
