Using climate models to forecast future energy supply and demand

How to best offset emissions to slow global warming is among the puzzles that face policymakers, scientists, engineers and the commercial sector. For instance, would carbon sequestration pose more advantages than disadvantages?

Energy producers would like to assess what climate change variables and policy options will most influence demand, supply, cost and reliability. In looking at climate change impacts, strategic decisions can be guided by tying together research in climate modeling and energy systems modeling.

That is a goal of a three-year Department of Energy project spearheaded in the National Security Engineering Division by Jeff Stewart, a project engineer and environmental economist, and colleagues. Partners include Princeton University, the University of Washington, UC Davis, and PG&E, which is one of the country’s largest public utilities.

The project kicked off in September with a workshop to discuss tools and technologies with representatives of companies that produce energy in 11 western states. The research addresses economic impacts of climate change and policies. Energy companies represented included PG&E, Seattle City Lights, Calpine Corp., Southern California Edison, and the solar energy company Recolte Energy.

Also represented were the quasi-governmental Bonneville Power Administration and U.S. Bureau of Reclamation, both of which produce hydroelectric power. Altogether, the groups participating in the workshop provide about 80 percent of the energy in the western region. That area spans Washington in the upper northwest, east to Montana, and south to New Mexico.

Some of the more eastern states of that region — Montana, Idaho, Wyoming, Utah, Colorado and New Mexico — were not represented at the initial meeting due to time constraints, but are welcome at next summer’s follow-up workshop. The workshop has been co-sponsored by the Engineering Department’s Corona Systems Center.

Future exploration

In a frank exchange at the first session, it was concluded that the most promising subjects to explore will be the impact of climate change on hydroelectric power generation and issues around carbon sequestration, a strategy meant to minimize greenhouse gas emissions, which is important in cases where coal is relied upon for energy generation, as in those more eastern parts of the region. Overall more than half of the electricity generated in the United States comes from coal.

"The biggest problem with coal is the emission of greenhouse gas," Stewart said. "In the area where it is produced, there also are concerns about mining and water pollution. Climate change is a global problem, but will require local solutions to mitigate impacts. There is a question of whether coal would be viable under current emissions regulations in the west if carbon sequestration works."

West coast climate change policies tend to ignore the role of burning coal to generate electricity, although energy providers anticipate that changes in resources and population may motivate their interest in putting coal on the table as a consideration in the energy supply mix.

For one thing, relatively steady water supplies from melting glaciers in the Pacific Northwest may become increasingly limited over time. As the glaciers diminish, water supply will become more seasonal, surging with spring snow melts. For another reason, in California, a massive population influx into hotter, drier regions is anticipated — putting a further strain on per capita energy demands.

Stewart’s post-doctoral fellow Noah Goldstein, who has a Ph.D. in geography, has been involved in forecasting future energy demands in the state. He put together a model that shows shifts in demographics and demand based on time and location.

"Usually energy demand models assume 2 percent to 3 percent annual growth," Stewart explained. "In California, the population is expected to grow from 35 million to 50–70 million in the next 40–50 years, almost double. If you’re doubling the population, you need twice as many power plants in the west."

However, incentives to buy energy-efficient appliances and the retirement of old inefficient ones could slow power demand so it might not rise as rapidly as the population, Stewart said.

Another tool is a code being developed by spatial statistician Gardar Johannesson to enable a closer look at how anticipated climate change may affect annual water runoff from watersheds. The current models typically average their predictions across 180–300-kilometer grids. The project is developing new climate models that rely on statistical downscaling, so that watershed analysis can be conducted on a focused grid of roughly 10 km.

These focused climate impact models will report the average condition in each grid cell, but also will extract the range of variation — such as daily lows and highs — not usually represented in the typical climate trend predictions derived from existing models, said Alan Lamont, a Lab engineer who is involved in the modeling effort.

Part of their vision is to create a publicly accessible tool in which researchers can enter latitude and longitude to forecast long-term changes over more than 50 years, out to the year 2100 — compared to typical projections of 2-3 years. Lamont explained that longer term forecasting should help utilities to plan electrical generation to accommodate fluctuating supply and demand. For instance, water supplies for hydroelectric plants would be affected by changes in watershed stream flow and runoff.

Such a tool also may be of interest in the developing world, he added. In India, for instance, glacial melt from the Himalayas will shift as those reserves shrink. He foresees a situation closer to the case in California, with massive snowmelt in the spring, followed by no flow in the summer.

For now, the model is being calibrated by looking back over historical records of precipitation, stream flow and daily temperatures. PG&E has kept such records stretching back into the 1880s.

Projections will initially focus on a few spots in the western region — one in the northwest, two in California and another in the Rocky Mountains.

Looking at variation that is normally averaged out of climate change predictions is enlightening when it comes to managing energy supply and demand, said Don Price , a senior consulting scientist at PG&E. For instance, PG&E scientists noticed energy demand didn’t drop when nightly low temperatures remained elevated over historical levels — the sort of thing that may play out in global warming.