New tool to help decision-makers consider possible futures for the Colorado River

The Colorado River is a vital water source for the western United States, providing drinking water to homes and irrigating farms in seven states. But the basin is under increasing pressure from climate change and drought. A new computational tool developed by a research team led by Penn State scientists could help the region adapt to a complex and uncertain future.

Their tool, the Framework for Narrative Storylines and Impact Classification (FRNSIC), can help decision-makers explore many plausible futures and identify consequential scenarios—or descriptions of what critical futures might look like—to help planners better manage the uncertainties and impacts presented by climate change. They presented their findings Sept. 19 in the journal The future of the Earth.

“One way states like Colorado prepare for the future is to plan for how things will evolve based on available science and input from different stakeholders,” said Antonia Hadjimichael, assistant professor in the Department of Geosciences at Penn State University and lead author of the study. “This scenario planning process recognizes that planning for the future comes with a lot of uncertainty about climate and water needs, so planners need to consider different possibilities, such as a high-warming scenario or a low-warming scenario.”

Hadjimichael said the scientific community and policymakers around the world often turn to scenarios to describe what future conditions might look like, but that approach can consider only a few possibilities and discount other alternatives.

These scenario planning approaches often involve a relatively small number of scenarios (e.g., what drought conditions might look like under different levels of warming) and may fail to capture the complexity of all the factors involved.

Scientists also use a technique called exploratory modeling, in which models simulate thousands or even millions of possible futures to discover which ones are important. But this approach is often not practical for policymakers, the scientists said.

“We wanted to come up with something that was in between,” Hadjimichael said. “We wanted to create something that bridged the gap, that took into account the complexities, but also boiled down to something that was a little more actionable and a little less intimidating.”

Their tool, FRNSIC, first uses exploratory modeling to study a large number of hypothetical plausible future conditions. It then uses this data to rank and identify locally relevant and meaningful scenarios, the scientists explained.

“Our approach is essentially to explore plausible future impacts and then say, ‘For this stakeholder, this is the scenario that matters most, and for this other stakeholder, there’s a different scenario that they should be concerned about,’” Hadjimichael said. “That adds a little more pluralism and a little more nuance to how planning scenarios are done.”

In the Colorado River Basin, policymakers face a complex set of factors, including how to provide enough water to growing populations and farmers while ensuring their state doesn’t use more than its allotted share of the river’s flow, Hadjimichael said.

“The problem is there’s no single criterion that takes into account all the stakeholders and their concerns,” she said. “You may have a very large farm, and I may have a very small one. And we may grow different things. It’s hard to use a single factor to come up with scenarios that would make us all happy or unhappy.”

The scenarios produced by FRNSIC can be used in future work in the Colorado River Basin, for example on the impact of drought events as populations adapt and make changes.

“This allows policymakers to look at different states of the world and consider how different interventions could affect the basin under each scenario,” Hadjimichael said. “These drought scenarios can be used to inform potential outcomes and therefore be used in negotiations or to solicit stakeholder input.”

Cornell University professor Patrick Reed, University of Virginia assistant professor Julianne Quinn, geospatial scientist Chris Vernon, and software engineer Travis Thurber at Pacific Northwest National Laboratory also contributed to the project.