Long$-$range forecasting

Long$-$range forecasting generally covers the period required to replace major resources. These forecasts are strategic in nature and cover applications such as facility and capacity planning, technology and design planning, research and development, and process planning.

Forecasting is aggregated into total groups of similar products and should indicate trends and resource imbalances. So a steel producer might forecast tonnes of output per year, an airline might forecast the number of passengers and tonnes of freight per route per month and an electricity producer might forecast megawatt hours $($MW h $={10}^6$ watt hours$)$ of electricity per month.

Forecast data for long$-$range forecasting should be bracketed in months, quarters or even years. If you were planning the electricity generation requirements for a significant region you might examine closely the peak seasons for the next $10$ or $20$ years. Daily requirements would not be important, but possible daily peak loads would be important. The actual day that the peak load will occur will not be considered, but the year that the new peak level is achieved would be important.

The forecasts at this level need to be in the "ballpark". In other words, absolute accuracy is not relevant as long as it is relatively close and clearly provides the trend data and resource requirements that will be required.