Depends on how radical the change is; if we found weather patterns traveled through mini-wormholes then it would dismiss the current theory entirely; of course, this is highly unlikely as the theory is well established by data, the most likely means of falsifying the theory would be to prove the data incorrect.
You are right, of course, however, it is more difficult to falsify a theory than someone might think at first. Take for example, the science of weather forecasting. A while back, I was walking in San Francisco at 10:30 in the morning and the weather announcer said that the weather for today will be sunny and clear all day. However, at the very time that he made that announcement, it was pouring rain, and even so, that did not falsify the science of weather forecasting. Similarly with the science of Global Warming.
There have been snowstorms in China. India has had three meters of snow, in Afghanistan 200 people and thousands of sheep have died from an extraordinary cold snap. In the Journal of Geophysical Research, Petr Chylek argues that the Greenland melt is now actually slower than it was in the 1900s, 1930s, 1940s, 1950s, and 1960s. For the first time in more than 60 years, there was snow in Baghdad, according to satellite data prepared by remss.com, in Summit County Colorado, at the Dillon weather station, the average snow depth for January is eight inches, but this year, Jan 2008, there were 18 inches, and the average daily maximum temperature was 24.1 degrees, well below the historic average of 31.3 degrees. There was a record low temperature of 26 degrees below zero in Yankton, SD on January 24, 2008. In Oregon, records were broken or tied for the early morning low temperature, and it was the coldest February 1 in 48 years in Santa Cruz, India http://timesofindia.indiatimes.com/Mumbai/Coldest_Feb_1_in_48_years/articleshow/2749848.cms
In La Ronge, Sask., Canada, temperatures of -43 degrees on January 30, 2008 broke all records.
But still all this has not falsified the science of global warming.