When does bad weather become Extreme Weather?
In the UK, we're used to bad weather, we get it all the time. And we love to talk about it (or so the stereotype goes). The weather is indeed... variable, perhaps making it more interesting than say the predictable weather of some of our European cousins. But our weather is actually pretty tame compared with some other part of the world. In fact, we are blessed with a moderate climate that throws up relatively few extreme weather events. This is due to a few factors: being an island and our mid-latitudes location play a big role in keeping our weather stable, and the North Atlantic Current keep winter temperatures about 10 degrees warmer than the would be if it were to 'turn off'.
We do experience some extreme weather though, sometimes a hurricane gets a little lost in the Atlantic and heads our way:
Hurricane Gordon: 2006
As an island, we also experience heat waves, droughts, floods, cold snaps, snow storms and high winds. We typically suffer a few tornadoes a year, although we have a long way to go before we reach the levels seen in tornado alley. Some of these events are clearly extreme, although we still don't have a good definition for what 'extreme weather' is.
What is Extreme Weather?
Definition 1: outside of normal conditions
There are a few different answers depending on which definition of an extreme weather event you take. The first is that on average 5% of weather is extreme (according to the NOAA). That is, events that happen on average less than 5% of the time should be classed as extreme. Whilst this is a useful working definition, it suffers from one major drawback: the amount of extreme weather events will never increase above 5%. Clearly this is a problem if we want to look at whether these events are increasing or decreasing in frequency! So this definition works better when thinking about weather systems, as opposed to thinking about how weather events fit into the large climate systems. This definition of extreme weather does avoid seemingly nonsensical claims like "Extreme events will be the norm" by John Prescott though!
Definition 2: Outside of predefined limits
An alternative way of defining an extreme weather event, as taken from section 1.3.3 in this paper on extreme events, is to pick a value for a variable that you are monitoring, and define a maximum or minimum for that variable which is 'extreme' when averaged over a particular time period. According to the paper for example, a typical maximum value for temperature averaged over a month is 25°C. Any time that the temperature exceeds this is classed as an extreme weather event. Although the particular limit is somewhat arbitrary and will be location dependent, this has the advantage of allowing you to easily track the increase or decrease in the frequency of extreme weather events over time.
Definition 2: Outside of predefined limits
An alternative way of defining an extreme weather event, as taken from section 1.3.3 in this paper on extreme events, is to pick a value for a variable that you are monitoring, and define a maximum or minimum for that variable which is 'extreme' when averaged over a particular time period. According to the paper for example, a typical maximum value for temperature averaged over a month is 25°C. Any time that the temperature exceeds this is classed as an extreme weather event. Although the particular limit is somewhat arbitrary and will be location dependent, this has the advantage of allowing you to easily track the increase or decrease in the frequency of extreme weather events over time.
Closing remarks
All extreme weather events are interesting in their own right. However, some are not easily modellable by climate models. This is because the models are not run at a high enough resolution, either spatially or temporally, to capture the processes involved. A good example of this would be tornadoes; it is hard to capture an event that happens over ~100s of metres/half an hour if your model is running with a grid size of 10km and once per hour (although this may change!). So in the next few posts I'll be focussing on extreme weather events that are more amenable to being modelled over yearly timespans such as droughts, heat waves and cold snaps.
EDIT: Added a heading for each definition
EDIT: Added a heading for each definition
As with Donald Rumsfeld who was originally somewhat derided and then subsequently lauded for his famous saying of 'unknown unknowns' I can kind of see where Prescott is coming from with his saying. I tend to think of extreme weather as being storm related events (as opposed to extreme heat or cold) and as such appear to lend themselves more easily to an absolute categorization. ie we could all agree that what is currently categorized as a hurricane would be extreme and the level of this does not need to change with prevailing climatic conditions. So by counting these we could (for example) defined at a macro level a east-coast spring as being 'extreme' if more than 3 hurricanes were experienced. And if this then became the norm then I could see myself agreeing with Mr Prescott...
ReplyDeleteFor temperature it strikes me that almost unintuitively extreme weather woudl be characterized by periods of low volatility in the temperature (ie sustained periods of time when the temperature is either above or below the long term average for that month. So I guess in this case one of the important variables to use to measure extremeness is the time period used for the basic temperature observations - as opposed to a storm-type observation where it is the pretty mush only the instantaneous force that is important and not the longevity.
Whether or not Mr Prescott made sense really depends on which version of the definition for extreme weather you go for, and I'd agree that it's does make sense if you use the second one. This is pretty close to what you're saying about the number of hurricanes experienced by a section of coastline. On that note, you might be interested in this page, which looks at some evidence that the frequency/intensity of hurricanes has increased. I'll have a more detailed look at these issues in the future, but the evidence is split on frequency (although only 2 papers are taken into account), whereas intensity is increasing (again only one paper so take with a pinch of salt).
DeleteThere are actually more ways of defining extreme weather events, and the duration of being above a limit can be accommodated by either definition. In the case of the first one, if it's high enough for long enough then it can pass a statistical test that says it's outside of the 95% range and therefore counts as extreme. With the second definition, you can decide on your own criteria for judging whether a hot spell has lasted long enough to be classed as extreme. Section 1.2.2 in this paper (also linked in the article) goes into more detail about chronic vs acute extreme weather, which is exactly what you are talking about with the difference between heat waves vs storms. It's a good paper and well worth a read!