Reply #30: one month's data isn't enough. Here's some results of taking 5 yrs of (hourly) measurements

[JohnWxy]

This article is about a study of using wind farms distributed along the East Coast - over a considerable distance.

Balancing the Power of Offshore Wind

~~
~~

Kempton believes that wind farms up and down the coast, under the power of local weather conditions, could perform the same load balancing function for one another that gas turbines provide onshore.

He and his co-authors Dana Veron and Felipe Pimenta from University of Delaware and Stony Brook University associate professor Brian Colle tracked coastal winds over five years using hour-by-hour data from 11 offshore weather buoys and meteorological stations.

Then they put together a computer simulation that substituted wind farms for weather stations and included the proposed interlinking cable. They generated data simulating five years’ worth of power feeds from the hypothetical wind farms.

While the researchers found that output from individual wind farms varied widely over time, the power output of the network as a whole held steady. Fluctuations, when they did occur, were gradual, but at no time during the study period did power output on the simulated grid drop to zero.

(more)

as stated in OP, the variability becomes an issue when you reach 20% of your power coming from wind. Also, in OP, there are ways of storing large amounts of power being developed. While this is in the nascent stage, we have a few years to get this 'down'. Vanadium redox batteries have been used http://en.wikipedia.org/wiki/Vanadium_redox_battery adn can be scaled up.