Introducing situation variables into subnational commercial sector energy demand models

The purpose of this thesis was to determine the value of using situation variables in subnational commercial sector energy demand models. Traditionally, such models have consisted of the extension of models built for other states or regions or the reformulation of national models to the regional or state level. The importance of scale and the place-related characteristics of the subnational area were largely overlooked.

A test was established comparing a typical subnational model to a model including situation variables for one geographic area. Choice of the area was limited by the availability of energy demand data for a subnational areal unit. The state of Wisconsin was chosen for study due to the existence of a detailed, accurate, appropriate data set.

The control model included general business activity, commercial sector economic activity, fuel price, and climate variables. The experimental model added metropolitan character, suburban influence and topographic location variables. Input of real data and comparison of the results showed the situation variables to account for a significant increase in the predictive power of the model.

Several sources of bias and possible implications of the results were examined.