# How Do Heating Degree Days Vary With Temperature?

Once again heating degree days (HDD). In two of our previous postings we investigated the correlation between HDDs and energy consumption. The first posting aimed at highlighting the influence of the number of HDDS on gross energy consumption, while the second focused on more specific data, namely the amount of energy devoted to heating purposes. The results were not very encouraging, since no strong correlation between the two parameters could be found. In fact, we might have expected otherwise.

In this posting we aim at a more fundamental approach. Is is clear from the definition of HDD that changes in temperature are reflected in the number of days where the heating needs to be switched on. This argument is straightforward on a daily basis. But does it also hold if we take monthly averages instead? Intuitively, the answer would be yes. But what we want to know is to what extent a montly average temperature may be considered a reliable measure for determining the value of HDDs.

In order to find the solution to this riddle we analysed the data from Sweden during the period 2003 – 2011. The baseline heating temperature for our investigation was taken to be 20°C, but HDDs for other baseline temperatures may easily be calculated. Our analysis lead to the conclusion that there is a very strong and reliable (negative) correlation between the average outside temperature and the number of HDDs over the years. As the annual data show consistently the same pattern we are not surprised to find that the same relationship holds for the multi-annual averages taken over a 30-year period as is shown in Fi.g 1.

Fig. 1 Average temperature and HDD in Sweden over a 30-year period

After these enouraging findings we might wonder if we could go one step further and look at the correlation between the annual data. Thus, we take the annual average temperature and relate it to the number of HDDs per year. As the resolution gets coarser we might expect a weakening of the relationship. However, the results are once again quite stimulating since at annual level the relationship between the two sets of parameters does not seem to loosen.  This is demonstrated in Fig. 2.

Fig. 2 Annual average temperature and HDD, 2003-2011

There is a nice negative correlation between the mean outside temperature and the number of HDDs, similar to the one we have seen for the monthly data.

A numerical analysis of our findings leads to the conclusion that, on a monthly basis, one degree of temperature difference (T,in – T,out) in centigrade corresponds to 30.5 HDD. Thus, if the monthly average temperature drops by 1 °C the number of HDD increases by 30.5. As a consequence, the number of HDDs may be directly calculated from the mean temperatures. Needless to say, that this is in perfect agreement with our own expectations.

# Heating Demand and HDD

Heating degree days (HDDs) are supposed to serve as an indicator for heating demand. At first glance, everything seems obvious. HDDs are a smart combination of the temperature difference between the interior of a dwelling and the outside temperature and the number of days where the difference is valid. (See one of our previous articles for technical details).

In an earlier posting we investigated the link between HDD and gross energy consumption. Then the findings were not as convincing as one might expect. An increase in HDD does not necessarily mean that consumption is rising in the same way. On the contrary, in some cases the two parameters might even go in opposite directions.

Surely, primary energy demand is certainly too crude a measure to be strongly correlated to simple variations in temperature. Too many other factors like energy demand for transport purposes and manufacturing come into play. At the end of the day, domestic needs and especially heating constitute just a fraction of the total energy cocktail. Thus it is tempting to go one step further and look into a potential, and hopefully more clearcut relationship between HDD on the one hand and consumption figures for heating purposes on the other.

We performed such an analysis for the case of Austria, covering the years from 2003 till 2010. The source data have been taken from Eurostat and Statistik Austria. The consumption figures refer to three different sources, namely heating, warm water and cooling.

Fig. 1 HDD and energy consumption for heating purposes in Austria, HDD* = HDD/30.

The correlation between the two sets of data is obviously better than in our earlier analysis which was based on gross consumption figures. Nevertheless, there are notable deviations from an ideal scenario which require some interpretation.

Before the economic crisis of 2008 the amount of TWh that went into heating was fairly stable. Up to 2008 and again in 2010 HDDs show much larger deviations from the mean value than the consumption figures. In some years, 2004 and 2008 to be precise, the deviations have a different sign, thus going in opposite directions as can be seen in Fig. 2.

Fig.2 Energy demand for heating and HDD, deviation from mean value.

Both, in 2006 and 2009, consumption figures differ more significantly from their mean value than their respective counterparts in HDD.  However, we may consider this as an exception. In general, energy consumption for heating, cooling and warm water seems to be more inert than the fluctuations caused by weather conditions.

This is good news because it shows that our heating systems are much less sensitive to outside conditions than what we might expect in the first place. On the other hand, it may also indicate that dwellings having a high degree of thermal insulation of which there are many in Austria are less exposed to temperature fluctuations.