sæsonvariation

Sæsonvariation, or seasonal variation, refers to regular and predictable fluctuations in a variable that occur at the same time each year. These patterns result from seasonal factors such as weather, daylight, holidays, and cultural practices, and they are distinct from long-term trends and random variation. Seasonal effects typically repeat with a fixed calendar period (for example monthly or quarterly).

In time series analysis, seasonal variation is often separated from the trend and irregular components. In additive models, the observation is seen as Trend plus Seasonal plus Irregular; in multiplicative models, it is Trend multiplied by Seasonal and Irregular. Seasonal components can be estimated using methods such as classical decomposition, STL (Seasonal and Trend Decomposition using Loess), or formal seasonal adjustment tools (X-12-ARIMA, SEATS). Seasonal indices quantify the average impact of each period, such as a specific month or quarter.

Common examples include higher electricity demand in winter for heating or in summer for cooling, increased retail sales around holidays, and tourism peaks in summer in many regions. Agriculture and education calendars also show seasonal effects tied to growing periods or school terms.

Modeling approaches frequently used to handle seasonality include SARIMA (seasonal ARIMA) and ETS (Exponential Smoothing) models, which incorporate seasonal terms with a defined seasonal period s (for example s = 12 for monthly data). Properly identifying and adjusting for seasonality improves forecasting accuracy and supports planning in areas such as inventory, staffing, energy use, and marketing.

Limitations include changing seasonality over time due to climate change, policy shifts, or behavioral changes, as well as irregular or evolving seasonal patterns in shorter datasets.