I've been thinking a lot about temperatures in the Gulf of Maine. While recent temperatures have been very warm, and 2012 was extraordinary, we have only recently encountered conditions warmer than 1950:
Note that this time series (and the others below) have been smoothed using a two year running mean. This removes the high-frequency signal, including damping out big events like 2012. In fact, 2013 is now the peak year in the Gulf of Maine, and it is almost exactly the same temperature as 1950.
The causes of the 1950 event (actually, a warm period from 1945-1955) are interesting, but in some ways, they're not important. The Gulf of Maine, and the northwest Atlantic in general, is one of the most variable parts of the ocean:
This means that we can think of 1950 as giving us a glimpse of just how far we could get away from mean conditions. To do this, we need to know the mean conditions, in this case, the global mean sea surface temperature:
You'll notice that the warming is pretty steady when averaged over the globe. In 1950, the mean SST was 0.4° cooler than today. If you view the Gulf of Maine relative to the rest of the ocean, you see that 1950 was even more extraordinary:
This gives us one way to think about how warm the Gulf of Maine could get. If 1950 were to happen today, we would get an anomaly almost 0.5° above the 2013 average (line marked "max now"):
Climate models suggest that the mean temperature of the ocean is likely to rise by 3°C in latter part of the century ("mean future"). If 1950 were to happen in that climate, then we would have some very extraordinary temperatures ("max future").
By this same logic, we are just as likely to get a cold event of the same magnitude as 1950 as we are to get a warm event. If we were to get one today, it would be about the same as the minimum temperature in the mid 1960s. Every year, those temperatures become less and less likely, and in the future, a 1960s-like cold period would look a lot like our recent "warm" period.