Here's a quote from a blog on a website hosted by SNL interaactive, written by Bill Burson:
The main problem is the wind doesn't blow all the time, and the backup generation necessary when the wind stops blowing can generate more pollution than what was saved originally. Public Service Co. of Colorado uses more wind power in its average hourly electric generation mix than any utility in the country, but Denver's air pollution problem has not improved.
Here is our response from AWEA's Michael Goggin:
It is unfortunate to see another recitation of the myth that "the backup generation necessary when the wind stops blowing can generate more pollution than what was saved originally." This myth is so persistent that we've taken the time to create a fact sheet to debunk it:
The reality is that wind energy typically adds only a small increment of variability to the power system, beyond the large degree of variability in electricity supply and demand that already exists on the grid (from factors like millions of people turning air conditioners on and off, steel mills turning electric furnaces on and off, and power plants experiencing forced outages). Since wind's variability tends to be uncorrelated with these other sources of variability over the short time periods relevant for system balancing, much of wind's variability is canceled out by other sources of variability. The incremental variability introduced by wind can also be greatly reduced through a number of cost-effective changes to the power grid, like building a more robustly interconnected grid, creating larger balancing areas, implementing faster generator dispatch intervals, making greater use of dynamic scheduling or energy imbalance markets, and expanding use of energy and ancillary services markets.
The incremental variability that wind energy does add to the system tends to occur very gradually, over the course of an hour or more. This type of variability can be cheaply and efficiently accommodated by non-spinning reserves, provided by hydroelectric plants or natural gas plants, including the quick-start plants mentioned by Mr. Burson.
Mr. Burson is correct that some coal power plants, particularly old, inflexible, inefficient ones, may have difficulty responding to dispatch instructions to change their output. However, the likely outcome is that these inflexible coal plants will turn off and remain off during periods of the year when wind output is likely to be high, since doing so is more economic than staying online and essentially paying wind plants to use their superior flexibility to curtail their output. As the results of NREL's recent Eastern Wind Integration and Transmission Study (EWITS) indicated, much of that coal energy and capacity will likely be replaced by natural gas generation, thereby reducing emissions by even more than the amount directly displaced by wind energy.
For example, EWITS found that CO2 emissions decrease by more than 25% in the 20% wind energy scenarios and 37% in the 30% wind energy scenario, compared to a scenario in which our current generation mix was used to meet increasing electricity demand. This is because coal generation declined by around 23% from the business-as-usual case to the 20% wind cases, and by 35% in the 30% wind case.
Thus, Mr. Burson's fear that emissions will somehow increase with a larger amount of wind, based on an anecdotal claim that air quality in the Denver area has not improved as more wind has been deployed (which, if true, I'm guessing is caused by an increase in fossil fuel use by the transportation, heating, and/or industrial sectors), is directly refuted by the most detailed model ever built of how wind energy actually interacts with the power system. And common sense.