Water gushing over a dam that supplies hydroelectric power

Keeping the Lights on When Drought Threatens Power Generation

By Jim Chido

MITRE's new forecasting model can help water authorities and power plant operators and planners analyze drought scenarios 10 years in the future.

Less water—in the case of drought, for instance—doesn't just mean less to drink or shower with. It can mean power shortages, rolling brownouts, enforced reduction in use, and even blackouts.

In 2007 for instance, a serious drought affected the southeast. Water levels dropped, and in some areas, power production had to be reduced or halted.

This isn't likely to remain an isolated event. Power plants depend on a steady supply of water. Rising demand due to population and economic growth, coupled with recent climate change scenarios, could easily lead to more droughts—and more power outages. 

MITRE researchers saw an opportunity to help power plant operators, water authorities, and regulators understand the long-term potential for water scarcity and its possible impacts. Their work resulted in a model, the "Interdependent Critical Infrastructure Model," or ICIM.

ICIM analyzes how power and water interact when it comes to a region's ability to produce power and still meet demand for water for drinking and industrial uses. The decision-makers' choices will ultimately affect the safety, health, and prosperity of millions. 

Studying a System that Mirrors the Nation

To provide a relevant example for demonstrating the model's effectiveness, the MITRE team chose the Lower Colorado River Basin in Texas, which provides drinking water for one in 10 Americans.

"The Lower Colorado offered a balance of the different types and techniques of power generation and provides power across Texas," explains Kris Rosfjord, who launched the work. "Texas is one of the nation's three main power grids [along with the East and West Coasts], so this offered us an opportunity to study a more closed system that mirrors much of the country."

Rosfjord assembled a core team able to attack the problem from multiple angles—a MITRE strength. "We were different pieces of a puzzle that came together to look at the Lower Colorado River water and power systems as a whole, but in a different way than perhaps others would have." 

Forecasting Future Challenges

ICIM can be used either to model current operations of a system or for long-term forecasting, an unusual combined capability as models go.

Using a technique called agent-based modeling, ICIM offers plant operators and regulators various scenarios for power demand, water availability, and temperatures up to 10 years out. Every input or entity in the model is called an "agent." Power plant agents include steamelectric plants powered by nuclear or fossil fuels. Other agents are wind, solar, and hydro.

"You can run variations, add a plant, take a plant away, add a drought, and compare the results," says James Thompson, an expert modeler who led ICIM's development. "You can identify all the trade-offs. By considering all the variables, the model can help inform policy with an eye toward the future."

For instance, much of the nation's power generation utilizes a thermal steam cycle. For a plant to generate at peak capacity, water must flow into the plant at a rate necessary to condense the steam and recirculate it back through the boiler and steam turbine.

"But if the in-take cooling water is too hot, it will reduce the efficiency of the plant," says Michael Cohen, who brings a strong background in critical infrastructures and on climate change scenarios. "And if the water is still too hot when you discharge it into the cooling pond or lake for the plant, it will cause massive fish kills."

To avoid this, the plant might have to be shut down. Which, in turn, could lead to widespread brownouts or even blackouts. The same can happen as a result of water shortages caused by drought.

"It could develop into a situation where you have to go to a brownout just to preserve the water supply for drinking," Cohen says.

"Our model showed that as we enter an era where demand for water for power, irrigation, and drinking goes up and temperatures increase, we have to start looking much more seriously at water availability in the future."

Balancing Economic and Environmental Factors

The team also estimated the economic impact of power shortages and compared that against what it might cost to mitigate it, such as by employing air-cooled rather than water-cooled condensers for plants as well as renewable energy sources such as solar photovoltaics that do not require a thermal steam cycle or water cooling.

In addition, the team used Environmental Protection Agency data on emissions and Department of Energy data on plant-cooling systems. "This is another strength of the model," says Kenneth Hoffman, an expert on power systems who built one of the first national energy models.

"ICIM deals not only with the effects of climate change on water limitations and power reductions, but also with the effects of power production on emissions of carbon dioxide and other greenhouse gases."

Informing and Arming Decision-Makers

There are many parties—plant operators, regional authorities, and local, state and federal regulators—involved in the regional decision making.

"We built ICIM to facilitate decisions," says Thompson. "Stakeholders can use it as a central tool to reach consensus." Decisions might include building a new non-water-utilizing power plant or changing regulatory constraints.

The team has briefed the model and their results to more than a dozen federal, state, and local agencies, as well as regional councils in Texas and the Washington, D.C., metropolitan area. The U.S. Department of Energy and the Electric Reliability Council of Texas stated that they will be taking the team's results on the dependency of power generation on water resources into account in their predictions of future power supply.

"We'd like to apply this modeling capability to other regions of the country, particularly to regions susceptible to future drought," Cohen says.

Hoffman notes, "You could even add variables so the model can look at, say, heat island effects and emissions in an urban area, coupled with water and power limitations." He adds that this is exactly the kind of problem where MITRE can make a difference with our systems perspective.

"These aren't just regional issues, but global ones."