Conventional wisdom suggests that coal benefits from really hot weather in the summer. One will often see experts in the trade press during the summer explaining that the spot coal market is off because we did not have a hot summer or that the spot market is strong because weather was extraordinarily hot. The data on coal use and extreme temperatures simply does not support the view that extremely hot weather in the summer is good for coal.
Looking at the average daily coal burn for the United States by month for January 2000 through October 2003, the highest burn rates occur consistently during the summer peak months of July and August when coal-fired units typically burn around 3 million tons per day. Less than 200,000 tons separates the highest burn rate during these two months and the lowest burn rate. One coal buyer said that he ought to be shot if he missed his burn projection for July and August because it is always hot then and all his units run as hard as they can. While the coal burn rates for the warmer months of May, June, and September are lower, they also do not vary widely either (with the exception of September 2001 which is low as a result of a specific, non-weather event). But how do these burn rates and electricity generated from coal vary with extreme temperatures?
Coal burn is closely related to normal weather variations
Since every summer is warm, it is important to separate normal weather from unusually warm or cool conditions. The U.S. Weather Bureau tracks of cooling degree-days as a means of determining whether and by how much temperatures are warmer or cooler than normal. A cooling degree-day is calculated for each degree of temperatures that an area’s average temperature for a day exceeds 65ºF. For example, if a region averages 70ºF each day for a week, it will add 35 degree days to its monthly total (example: 70ºF average actual -65ºF standard X 7 days = 35 degree-days). The degree-day norms are based on the temperatures that have been observed over a 30-year period.
These normal or expected cooling degree-days are highly correlated with coal-fired power generation in the three census regions studied: South Atlantic, East South Central, and East North Central. These regions were selected because of their heavy dependence upon coal-fired generation and high summer peaking loads. These three regions accounted for 56.3% of the nation’s coal-fired generation in 2002 and 50.6% of coal used for electric power generation. Roughly half to two-thirds of the variability in coal-fired generation in these regions during the months of May through September for the years 1995 through 2001 is explained by the long-term average of normal cooling degree-days.
The unexplained changes in generation are not systematically related to whether a month was hotter or cooler than normal. The portion of generation that was not explained by normal weather was estimated by subtracting the portion of coalfired generation that is explained by normal cooling degree-days from actual generation. Whether a month’s weather was warmer or cooler than normal was estimated by subtracting normal cooling degree-days from actual degree-days. When the relationship between the unexplained coal-fired generation and cooling degree-day differences from normal was estimated, little or no correlation was found (i.e. less than 10% of the variation explained by actual cooling degree-day differences from normal). Coal-fired generation did not consistently increase when cooling degree-days were higher than normal and coal-fired generation did not consistently decrease when cooling degree-days were lower than normal.
Other factors such as the performance of non-coal generating units and non-weather changes in electricity demand probably explain the variation in coal-fired generation that is not explained by normal weather. However, it is clear that while coal-fired generation is strongly affected by normal weather, it is not systematically impacted by hotter or cooler weather.
Another way of looking at this is to consider how specific generating units performed during periods of very hot and very cool weather. For example, looking at the hourly dispatch for Miller Unit 4 during similar periods of August 1997 and August 1999, August 1997 was 9% cooler than normal in the East South Central Region and August 1999 was 16% higher than normal. Despite this extreme difference in temperatures, Miller Unit 4 generated the same amount of electricity (i.e. less than 1% difference). Hotter weather does not translate into significantly higher generation from baseloaded units such as Miller Unit 4 because these units are already operating at full capacity around the clock. There is no opportunity to increase generation from these units, even if temperatures reach record highs.
Although July 1999 was 41% hotter than normal in the East North Central Region and August was 25% cooler than normal, Gibson Unit 4 generated only 3.4% more electricity in July than in August, despite the extreme temperature differences. Gibson Unit 4 was fully loaded during the daytime peaks whether the weather was hotter or cooler than normal. The unit cycled down during off-peak nighttime hours. When the weather was extremely hot or other units were unavailable, the Gibson unit ran more during a limited number of nights. Cycling units have an opportunity to increase load but only during off-peak periods. Hot weather normally affects the peak daytime demand for electricity and generally does not lead to higher demand during off-peak overnight hours.
Seasonal peaking plants operate only during peak demand periods and may experience significant increases in generation during hot summer weather. However, the amount of coal-fired seasonal peaking capacity is small and growing smaller as generating companies retire or convert these units to natural gas to avoid retrofitting them with costly environmental cleanup technologies.
A hot summer may lead to less coal use
Now comes the zinger for coal. As temperatures climb higher cost sources of electricity generation are brought on line to generate power. Gas combined cycle (GCC) units will be dispatched to meet high peak demands. Peak electricity prices will climb to justify the operation of high cost generators. Since operators will want to make sure that they have adequate capacity to meet peak demand, they are likely to keep their GCC units running during the off-peak hours. Since GCC units do not cycle well, they generally will decrease their output by no more than 30% during off-peak hours. Some operators have had problems with start-up of their combined cycle units. Since they would not want to risk not having their gas units online during periods of high electricity prices, they also are likely to keep their GCC units running overnight.
Since coal-fired units are generally are more flexible and can take deep cycles, often operating at 20% of their peak load, they may be forced to shed load to make room for the gas combined cycle units. GCC units have maintenance and other costs associated with each start-up. While it may sound strange that it is more cost effective to operate a high fuel cost gas plant instead of a low fuel cost coal unit, this can be the optimal operating strategy for the generating units involved. Strangely, hotter than normal weather may lead to reduced coal use.
The conventional wisdom concerning the impact of weather on coal burn probably arose during a time when the average capacity factor was considerably lower and more plants operated in a seasonal peaking mode. This analysis does not imply that coal burn does not increase during summer months. It does, and generating plant operators plan for it to increase. Normal weather conditions are such that fuel planners know that more coal will be burned during normal summer temperatures. Hot temperatures, or unusual weather patterns, may affect the timing of spot purchases and give the illusion of increased demand because several utilities hit the market at the same time. However, even significant changes in cooling degree-days should not impact aggregate coal burn for the summer significantly.
With the addition of nearly 135 gigawatts of GCC capacity since 1999 and through the end of this year, the impact of hot weather on coal use may be changing further. Hot weather is likely to require GCC units to be brought on line during peak periods to meet demand. These units are likely to be kept on line during off-peak hours to ensure that they will be available for the following day’s peak requirements. This increase in gas generation during off-peak hours may lead to a decrease in coal burn as coal-fired units take deeper cycles to make room for the GCC generation.
Be careful what you wish for.
