Anyone who tracks the weather closely soon becomes aware of a surprising fact: it’s not easy to get a forecast tailored to your local watershed – perhaps the most important natural terrain feature that determines water supply, water quality and flood risk.

In the United States, most weather forecasts pay no attention to watersheds. Instead, predictions are made within broad “climate divisions” that do not necessarily recognize the finer scale of watersheds. The National Weather Service (NWS) does have a network of river forecast centers, mainly to provide vital streamflow predictions. These are mostly aimed at a technical audience and do not break out information by watershed.

Now there is a new tool to fill this forecast gap. A partnership between the National Center for Atmospheric Research (NCAR) and the University of Colorado Boulder has produced a forecasting website that offers real-time weather prediction maps based on watershed boundaries across the nation.

Not only is the new system watershed-based but it also offers temperature and precipitation forecasts up to a month ahead. That is well beyond conventional forecasts offered by the NWS.

Known as S2S Climate Outlooks for Watersheds, the system is still considered a prototype. But it has already begun to fill an important need by applying weather and climate data to watersheds. Water Deeply recently spoke to Andy Wood, a project scientist at the NCAR and co-leader of the project, who explained how the system was developed and how it can be used.

Wood emphasized he is eager to hear feedback from users (via this form) about how to improve this new tool.

Water Deeply: Why did you want to bring forecasting to the watershed level?

Andy Wood: This is something I’ve been interested in doing for a while. I used to work for NOAA [the National Oceanic and Atmospheric Administration] in two different river forecast centers: in Salt Lake City for the Colorado Basin, the other in Portland for the Northwest region. As somebody who looked at climate forecasts in that context, it always struck me that it would be nice to have it communicated on a spatial basis, but connected a little more strongly with the kind of modeling and analysis that we would do for river forecasting. And one major basis for that is watersheds and watershed scale.

For instance, if you manage water in the Yakima Basin [in Washington state], you would be able to find information that’s kind of tailored to that watershed. So it’s a pretty simple concept, in a way.

Water Deeply: It sounds like a simple concept. But I’m guessing it wasn’t simple to carry out.

Wood: It’s not that bad. I think there were some tough choices over how to display things. It took a bit of engineering, but nothing serious.

What I think is challenging is thinking about the various directions you can go with it. Right now, for instance, we have two products we’re putting on the website. One is really just a raw forecast that kind of spatially remaps into these watersheds. Another is a set of products where we don’t just take the raw forecasts, but correct them for systematic biases. One of the major advances in the last decade has been a move toward not only making forecasts in these national centers but also making watershed-scale hindcasts, where you go back and take a forecast model and run it many times to figure out what the biases are and correct for them. That’s the start for forecasts that are little bit more enhanced and tailored. There are a lot of ways to add skill and value to that. That’s a little bit more where the science comes in.

Water Deeply: How do these watershed forecasts help with snowmelt forecasting?

Wood: Water supply is dependent on the runoff forecasts. Not so much with the precipitation forecast, because that still is very challenging. But at least where snowmelt is controlling part of the signal in the runoff forecast, we can start to employ temperature forecasts to help with that. You know to expect the spring is not probably going to be like the normal spring you might have seen back in the 1980s and 1990s, but probably a little warmer than that.

There are a myriad uses that are helped by having that forecast info. This is probably more important for the western U.S. than in the East, because in the West you have this dry-summer, [wet]-winter pattern. It’s important in the western U.S. even beyond water supply allocations that impact cities like Los Angeles, Phoenix, Las Vegas. You also have water allocation for environmental reasons, for fish, for hydrologic operations, for navigation that ends up depending on long-lead forecasts.

In California, you have the spring water supply forecasting that goes on for all those reservoirs up and down the Sierras that help control what allocations the state and federal government will make to different users. In the Colorado Basin, one interesting example of forecast use is the prediction of end-of-water-year reservoir levels in lakes Powell and Mead, which put the basin into different operating tiers. Depending on where they project flows to be, that will directly impact releases on Powell down to Mead, allocations from Mead down to Southern California and other lower-basin states. There’s still this fundamental reliance on predictions of where these storage levels will be that’s based on seasonal runoff forecasts.

Water Deeply: Is this a reliable tool now for watershed forecasting and water management?

Wood: I think so. These are based on real data being run by NOAA and the National Centers for Environmental Prediction. At the Climate Prediction Center, the way those forecast products get put together makes use of the same information in these maps but also other tools we have. In that regard, they are a real part of what our nation uses for climate prediction.

My understanding of the way people make decisions on water and hydrologic forecasting is, typically it doesn’t depend on one single product. I probably wouldn’t advise this be used solely to make decisions. But I see no reason why somebody in a water management context wouldn’t be able to look at this and add it into their understanding of their situation in their watershed. There’s certainly very good information in these forecasts. But also I would never take it alone versus looking at other forecasts like, for instance, those put out by the Climate Prediction Center.

Water Deeply: These new forecasts made me think of John Wesley Powell, who urged us to think and plan at the watershed level way back in the 1800s. Was he an inspiration to you in this project?

Wood: As a long-time hydrologist, I really enjoy seeing these watersheds that I’m familiar with as kind of the basis for analysis. I really get a lot out of seeing that, even on a finer scale.

I would like to see more products that are tailored and made accessible on a watershed basis. I think it makes a lot of sense. Sometimes it’s hard to interpret a (forecast) product that spans two sides of the Continental Divide, where you have very divergent climate and hydrology.

Water really is a critical, critical determinant of life in the western U.S., where you have this tendency to experience these big droughts that are very devastating. I do think we can do more to connect the science we have in geography, weather and water prediction to users who have to manage and make decisions on a watershed basis.