Summary: | A major impact of climate change is expected to materialize on energy demand for space heating and cooling needs in the residential sector. To quantify this impact, a set of regression models were tested to study the relation between residential energy demand for space heating in Iceland and explanatory variables such as Heating Degree Days and GDP per capita. Considering the nonstationarity of the time-series, three methods were studied to cope with this condition: Cointegration, differencing and detrending.The evaluation statistics of the three models for the validation period showed that the modified detrending approach is the most reliable method. It became obvious that including the seasonal dummy variables and AR component significantly improve the power of the model to predict monthly energy demand for residential space heating in Iceland. The developed model can be used to project climate related changes in demand for low-geothermal heat. Cointegration, climate change, heating degree day, space heating
|