The work of Warren Katzenstein and Jay Apt, Peter Lang, and Peter Hawkins all seems to demonstrate that natural gas back up of wind generation imposes choices and inefficiencies, that almost or completely eliminates the CO2 emissions benefits of wind. It would appear from their work that stand alone natural gas systems using combined cycle gas turbines, are either nearly as efficient at lowering CO2 emissions or actually more efficient as a wind plus open cycle gas turbines. Supporters of wind generation have noted that that wind generators are well matched to hydro-electricity, and indeed that seems to be the case in a few parts of the world, for example Scandinavia where wind generators in Denmark appear to compliment hydroelectricity from Norway and Sweden. In the United States, hydro resources have been almost entirely utilized, and are currently inadequate for wind back up in most high wind areas. Pumped Storage has been suggested, the the high cost of past pumped storage facilities suggest that the cost of nuclear reactors is competitive with the cost of pumped storage facilities with similar rated capacities, while the nuclear facilities would be far more flexible, and produce as much as 2 times as much electricity on an annual basis as pumped storage would.
Compressed air energy storage (CAES) is a second form of backup proposed for wind generators. My investigation, however, revealed a surprising problem from CAES, radioactive radon gas would be brought to the surface with returning compressed air. The problem appears to be far more serious than the release of radioactive gases associated with nuclear power generation. My case study of proposed CAES project presented by the Ridge Energy Storage & Grid Services company to Texas State Energy Conservation Office in 2005 showed that 40% of the energy for the project would come from the burning of natural gas. CASE systems are huge geothermal heat pumps, and they return cold air. Humidity in the air freezes, and the ice can damage generator turbines. Heat lost in the CAES process represents lost energy from electricity used to compress the air. In evaluating the cost of wind generated electricity I stipulated
a cost for new West Texas wind of $2250 per name plate KW in 2009. Since the capacity factor of West Texas runs around .40, the adverage output West Texas wind producer can expect to pay $5625 produce KWs of electricity his windmill will average producing. Since only 70% of the electricity entering the CAES facility reaches the consumer, the wind producer must add 30% more capacity to compensate for the energy loss. Thus the price of the wind generated electry entering the CAES facility must compensate the wind producer for something like a $8000 capitol investment for every average kW sold to the CAES facility.
In addition the estimated cost of the Ridge Energy CAES facility was $765 per KW of electrical output, Thus we are looking at an investment of nearly $9000 per kW of electrical capacity and this does not count the cost of new electrical transmission lines from West Texas to energy hungry Dallas. In contrast
the 2008 cost of nuclear power is somewhere between $4000 and $5000 per kW (as opposed to an estimated $8000 to 12,000 figure during the middle of the next decade).
And nuclear plants can be located close to electricity markets. In addition, the nuclear plant would be far more flexible, and would produce more electricity on an annual basis than the wind + CAES combination. In addition I noted an alternative employment of the CAES system that no one seems to have thought of, the used of CAES in in nuclear cooling, that would produce a low cost nuclear CAES combined cycle:
It seems to have escaped the notice of most CASE advocates that CAES casn be teamed with nuclear power plants in innovative ways. Since it is more economical to keep reactors running at full power all night, suplus electricity produced at night could be used to store compressed air. During the day, compressed air can be used to expand the reactors daytime power output by as much as 40%. The air does not have to be heated with natural gas. Indeed the compressed air can be heated from the reactors waste heat, killing two birds with one stone, and conserving the water used for daytime reactor cooling, and the use of compressed air in cooling the reactor, would creat significant water use savings, allowing reactors to run even during drought conditions.
Just a thought, mind you.

I also looked at battery backup for wind, that was of course, way too expensive. In fact it was so expensive that I conducted a thought experiment,
Assume that the system operators chose to back up the 1 GW wind system with nuclear power rather than a redundant wind system plus batteries. The cost of the wind system would then drop to $2.7 billion plus $5 billion for nuclear backup or $7.7 billion. Quite obviously the nuclear backup would be cheaper, but now the wind is totally redundant, because the backup system can operate full time for just the added price of fuel. Thus the purely nuclear system would simply be a lower cost than wind a reliable wind system. The nuclear system would be more reliable, and could be counted on with a fairly high degree of certainty to produce at 100% of its rated capacity during peak electrical demand summer months.
Thus my conclusion was that Pumped Storage, CAES, and battery backups for wind were more expensive, less flexible, and would produce less electricity over time than electricity producing nuclear reactors.