I agree with your figures (I previously misstated this year's 1Q at IAD as 31.6 degrees). Bear in mind, though that these figures are unweighted averages of January, February, and March average monthly temperatures, which gives too much weight to February, as it has only 28 days (29 in leap years), compared to 31 days each in January and March. Weighting each of the first 90 days of each year equally results in 1Q 1970 being colder than 1Q 1978 at IAD (31.4 vs. 31.7). 2015 would still be third coldest, but at 32.0 degrees, rather than 31.8. Perhaps some day weighted averages will be used to calculate quarterly and seasonal averages.