100 Year Events
By Paul Homewood
h/t John F Hultquist
http://md.water.usgs.gov/publications/wsp-2375/ma/index.html
John sent me the above report by the USGS, concerning floods and droughts in Massachusetts between 1927 and 1988.
The report goes into some detail on certain major floods, but what struck me was the table above, which listed the recurrence intervals. Just note how many are ranged up to 100 years or more.
Given that there are 14 floods listed (and this only includes major events), spread over 61 years, this works out as one every 4 years. What we need to bear in mind, of course, is that there are maybe dozens of locations that are vulnerable to such flooding, just in this one state. Whilst the likelihood of a major flood at any single location may be once every fifty years or so, if there are, say, ten potential locations, one of these would flood on average every five years.
Note that on the list above, each flooding event is a separate event, on a separate date.
Remember this when you so often read about claims of “100 Year Events”.
Comments are closed.
NOT A LOT OF PEOPLE KNOW THAT 
Two points:
1. Note line two in the table — Drought 1929-32; this was relevant to a recent post about rainfall in England at the same time. So, one place has rain while another doesn’t. Just weather, don’t panic.
2. About two years ago a researcher from a technical school in Sydney posted on the problems of trying to get a grip on the analysis of such things as “100 year events.” He believed that what has and is still done is very wrong and thus misleading. I’ve lost that reference. Maybe someone with better search skills can find a link to that report.
“Whilst the likelihood of a major flood at any single location may be once every fifty years or so, if there are, say, ten potential locations, one of these would flood on average every five years.”
Except that the factors which trigger major floods tend to be widespread over large areas, so that given one major flood event in any particular year, the chances of other major flood events elsewhere in that year are much higher than one in fifty. It probably makes more sense to try and calculate the probability of a ‘flood year’, without worrying too much about the specific locations involved.
In which case we seem to flood years once in every four!
This site offers a good run down on flood types: http://water.usgs.gov/edu/100yearflood.html