Abstract
The complex dependence relationship between peak district heating demand and outdoor temperature is examined by using the copula function.The aim is to derive a copula-based conditional probability distribution of heat demand given weather conditions and provide a probability lawsensitive to extreme events. We analyse data concerning the district heating system of the city of Bozen-Bolzano. Daily maxima heating demandand the corresponding outdoor temperature have been observed from January 2014 to November 2017. We perform a three-step analysis. First, theserial correlation of the two time series of interest, taken separately, is analysed through a seasonal autoregressive integrated moving average model.Second, the dependence relationship between the two series of uncorrelated residuals is investigated using copula models. Third, the conditionalprobability distribution of peak heat demand given outdoor temperature is derived. The selected copula model indicates that the two investigatedphenomena tend to co-move closely together for the whole observed period and, when taking into account the conditional behaviour of interest,in the case of extreme climatic events, we find a high probability of thermal energy demand reaching its peak. From this finding we derive usefulimplications for the management and production of thermal energy.