Energy consumption data are influenced by the weather conditions. The
use of weather unadjusted energy data in analysis and forecasting coul
d give misleading and erroneous results. In this paper a new iterative
econometric technique is developed to correct for abnormal weather co
nditions in published energy consumption data. The technique is applie
d to sectoral natural gas sales data for the province of Ontario, Cana
da. The empirical results reveal that following the methods available
in the literature results in weather-corrected data that still contain
a significant portion of the abnormal weather impact. In addition, th
e study suggests that residential and commercial natural gas data requ
ire more weather correction than the data for the industrial sector.