Weathering power's demand

Using temperature forecasts to predict power demand has some major pitfalls. Here Martin Fischer and Michael Grossman suggest ways to glean more from forecasted temperature data

Temperature is the single most important predictor of electricity demand variability - and the variability of much else, for that matter. It is estimated that 25% of global GDP is directly influenced by weather and climate variability.1 In absolute terms, calendar factors, such as the day of the week or public holidays, may be responsible for a greater cyclicality in demand. However, these are easily incorporated in to existing models, and there is no forecasting science involved, as such. The