“Dad, is Halloween going to be cancelled again this year because of the weather?” It’s not a pleasant thought for kids dreaming of bags full of candy. Well, we are approaching the end of October, and for those of us who have been in Connecticut for the past few years, it seems appropriate to discuss big storm events. With the Halloween snowstorm of 2011, and then Sandy in 2012, we have some catastrophic events in recent memory. But how often can we expect a large event, such as a “100-year storm”, to occur?

The answer to this question is based on historical data. Rainfall data for a particular site, and for a particular duration (such as 24 hours) are sorted, from highest to lowest. Then a calculation is performed to determine the chances of an event of any given size occurring. The recurrence interval is an estimate of the likelihood of an event, and is related to the probability. So as you can see from Table 1, there is only a 1% chance that a 100 year storm will occur in any given year.

Table 1. Chances of different events and recurrence intervals.


So what happens if our rainfall totals are increasing as a result of climate change? Will we be seeing these large events more often? Well, I’m glad you asked. Here in southern New England, our yearly precipitation totals have been increasing, and we have also seen a rise in the intensity of individual events. Other parts of New England have not seen large changes. Figure 1 shows how the annual rainfall total in Connecticut has increased over the last 110 years. The past 40 years have had higher totals, with almost all years greater than the average (solid black line in figure).

Figure 1. Annual rainfall totals, state of Connecticut (from NOAA at http://www.ncdc.noaa.gov/cag/time-series/us).


Historically, engineers used values from pre-1961 as a basis for sizing stormwater pipes and other structures. Researchers at Cornell have used recent rainfall totals to adjust the values for different recurrence intervals. Updated values for Storrs, CT showed that for smaller intervals, there is not much difference, but for 50-year and 100-year events, there has been quite a change (see Table 2). The old 100-year storm (7.0 inches) is about equivalent to a 50-year storm (6.8 inches) now! So the chances of us getting a storm of this size in any given year have doubled.

Table 2. Older and updated storm sizes for different recurrence intervals.


So, should I tell my daughter to be prepared to miss Halloween again this year? Not a chance!

Dr. Michael Dietz