Weather-load regression
L_t = beta_0 + beta_H HDD_t + beta_C CDD_t + gamma' Z_t + epsilon_tSeparate heating and cooling terms allow load to rise in both cold and hot regimes.
Time Series
Temperature–Load Relationship
Temperature drives heating and cooling demand; a simple weather–load model underpins many power forecasts.
Mental model
Weather is an explanatory driver that links temperature conditions to power demand.
Heating and cooling degree days translate temperature into expected demand for electricity and gas.
Load forecasting models often regress demand on temperature, calendar effects, and trend terms.
Explore the mechanism
How does load respond differently to colder and hotter temperatures after calendar and trend effects are separated?
This curve shows a stylised temperature–load relationship. Heating and cooling slopes translate degrees below or above a comfort temperature into extra demand.
Load at comfort (18°C) ≈ 30.0 MW
Numbers
Load at cold extreme ≈ 86.0 MW
Load at comfort ≈ 30.0 MW
Load at hot extreme ≈ 38.0 MW
Interpretation
A simple weather–load model says: demand is roughly linear in degree days around a comfort band. Heating and cooling slopes encode how sensitive the system is to cold and heat.
For forecasting and risk you care less about any single temperature realisation, and more about the gradient of load with respect to temperature. That gradient propagates weather uncertainty into load, then into prices, then into portfolio P&L.
UNDERSTAND
What to notice
Weather risk enters portfolios through the sensitivity of load, and hence prices, to temperature deviations.
The quantitative structure
Notation
L_t- load at time t
T_t- temperature
HDD_t, CDD_t- heating and cooling degree measures
Z_t- calendar and trend controls
This model assumes
- The selected weather station or aggregate represents the served region.
- Calendar, trend, and structural changes are controlled well enough for the intended forecast.
Where it appears
- forecasting electricity and gas demand
- constructing weather-normalized load
- stress-testing weather-sensitive portfolios
Where the intuition breaks
What this view leaves out
- Temperature-response estimates change with region, season, time of day, technology adoption, and behavioral adaptation.
Common mistakes
“Load changes linearly with temperature everywhere.”
Not quite.Heating and cooling responses are often nonlinear, asymmetric, and conditional on calendar and location.
Where this concept sits
Follow the strongest prerequisites, applications, extensions, and contrasts.
Connections for Temperature–Load Relationship
Temperature–Load RelationshipYOU ARE HERE
Foundation ofLoad Profile and Volume RiskTemperature is a major driver of demand shape and therefore of load-volume uncertainty.
UsesLinear Regression (OLS) in Factor ModelsA regression is a simple baseline for quantifying how load changes with temperature and other explanatory variables.
Supported byKalman FilterState-space filtering can update latent demand states or time-varying relationships as new temperature and load observations arrive.
SupportsPower Forward Seasonality and ShapeWeather-driven demand patterns contribute to seasonal structure in expected power prices and load.
Foundation of · Load Profile and Volume Risk
Temperature is a major driver of demand shape and therefore of load-volume uncertainty.
Explore concept →