Extreme weather generally involves intersecting weather systems. For example, 'The Perfect Storm' of movie fame was the intersection of the remnants of a hurricane coming up from Gulf of Mexico with a polar cold front coming across Canada and as I recall a low pressure system out around Nova Scotia. The timing of these events was basically random; if the events had been staggered a few days apart instead of hitting at the same time, it wouldn't have been a perfect storm.
So, where models predict an increase in such events comes down to increasing probabilities of the individual events, kind of like throwing the dice twice as often as before, then you're going to get double sixes (i.e. extreme events) more frequently.
As far as the attribution of specific events, well, you can look back in history and say, historically we've had these events every 100 years or so, so we shouldn't expect three such events over a twenty-year period. This is similar to how a casino might suspect a player is using a loaded die.
However, for one specific event, is it possible? Plausibly you can examine the details of a single event and ask, okay, are there physical characteristics here directly related to climate change, in the same way one could examine a die to see if it was really symmetrical or not? One could perhaps use expectation results, i.e. "we expected this kind of extreme weather event to occur with greater frequency outside the regions where they normally take place, according to frequency results in models forced with extra CO2" etc.
Really though, it's when you get 500-year floods every 5-10 years in Europe, then you know you're 'testing the historical limits'.
So, where models predict an increase in such events comes down to increasing probabilities of the individual events, kind of like throwing the dice twice as often as before, then you're going to get double sixes (i.e. extreme events) more frequently.
As far as the attribution of specific events, well, you can look back in history and say, historically we've had these events every 100 years or so, so we shouldn't expect three such events over a twenty-year period. This is similar to how a casino might suspect a player is using a loaded die.
However, for one specific event, is it possible? Plausibly you can examine the details of a single event and ask, okay, are there physical characteristics here directly related to climate change, in the same way one could examine a die to see if it was really symmetrical or not? One could perhaps use expectation results, i.e. "we expected this kind of extreme weather event to occur with greater frequency outside the regions where they normally take place, according to frequency results in models forced with extra CO2" etc.
Really though, it's when you get 500-year floods every 5-10 years in Europe, then you know you're 'testing the historical limits'.