Use and misuse of 20-year average degree-days

In order to adjust fuel consumption figures to compensate for weather variations, it is common to consider annual consumption as a combination of fixed and weather-related portions, and to multiply the latter by a factor DS/DA where DA represents the actual degree-day value over the measured year and DS represents the number of degree days in a 'standard' weather year. The convention in the UK is to set DS = 2,463. In theory any figure could be used, but the use of this common basis yields the quantity that would have been used under fixed standard weather conditions. This facilitates comparisons not only between buildings in different regions, but also with published 'yardstick' values calculated on the same basis.

Some analysts who are only concerned with tracking trends in specific buildings have historically opted to use their regional 20-year average for DS. The logic behind this is that the adjusted quantity is what would have been used had the building experienced typical weather for the region in question (rather than some arbitrary national norm). Although this option precludes both inter-regional comparisons and comparisons with published yardsticks, it is safe provided that users choose a unique value for DS and stick to it. Unfortunately, some choose to use the rolling 20-year average, which changes through time. This gives the wrong answer, and in particular it flatters performance when the 20-year average is falling year on year, as at present, since the adjustment factor DS/DA is artificially deflated by the annual decline in DS.

Rolling 20-year average figures are really intended to help forecast consumption budgets. For more on the standard degree day values for both heating and cooling click here.

(V.V. 1/12/06)