How Weather Models Move Natural Gas Prices
Weather models influence natural gas prices because heating and cooling account for a substantial share of natural gas demand in North America, and forecasted temperature changes alter expectations of near-term consumption. Short-term price moves in natural gas futures often reflect new weather data releases that shift the probability distribution of heating or cooling degree days. The magnitude and direction of these price moves depend on the forecast horizon, the degree of confidence in the model output, how far the predicted temperature deviates from seasonal norms, and the state of inventory and competing fuel prices. This mechanism creates a predictable pattern in which unexpected forecast revisions drive volatility, but the relationship is neither deterministic nor stable across market regimes.
Core Concept
Natural gas in North America is primarily consumed for space heating, water heating, and power generation. Collectively, these uses account for approximately 40% of U.S. Natural gas demand, with seasonal variation concentrated in winter months [1]. The remaining demand comes from industrial processes and other uses that are less temperature-sensitive. This demand concentration in temperature-dependent categories means that any change in the probability of future temperatures alters the expected quantity of natural gas needed in the coming days or weeks.
Weather forecasts do not predict demand directly. Instead, they predict temperature. Meteorological agencies, chiefly the National Weather Service (part of NOAA), produce deterministic point forecasts and ensemble forecasts that quantify uncertainty in future temperatures. The energy market interprets these forecasts by comparing predicted temperatures to historical seasonal norms to estimate heating degree days (HDD) or cooling degree days (CDD), which serve as proxies for demand. A 5-day forecast calling for temperatures 15 degrees Fahrenheit below normal in Chicago would imply higher heating demand than the seasonal average; traders incorporate this expectation into the price of natural gas futures, which typically reflect the market's estimate of supply-demand balance over the next month [2].
The lag between forecast release and the actual weather outcome creates a trading opportunity. Weather models have significant forecast skill out to about 10 days; beyond two weeks, predictability declines sharply [3]. Within this window, model forecasts are regularly updated (typically twice daily in the United States), and each update may carry new information about upcoming temperatures. If a new forecast significantly raises the probability of cold weather, natural gas prices typically rise within minutes to hours of the model release, because traders repricing the contract expect higher near-term demand. Conversely, a surprise forecast for warmer-than-expected weather tends to lower prices.
Mechanical Operation
The transmission of weather model information into natural gas prices operates through four sequential steps: model run, forecast interpretation, inventory response, and futures pricing.
Model Run and Distribution. The National Weather Service produces forecasts every 6 hours (00Z, 06Z, 12Z, 18Z UTC). Major forecast models include the GFS (Global Forecast System) and higher-resolution deterministic models, as well as the GEFS ensemble, which provides probability forecasts of extreme temperatures [1]. These are disseminated immediately to traders, meteorologists, utilities, and financial data vendors. Commodity traders monitor specific locations: population centers and heating-demand hubs such as Chicago, Boston, and Denver.
Interpretation into Demand Proxies. Traders do not directly trade temperature; they trade natural gas contracts. To convert a temperature forecast into a demand estimate, traders compute heating degree days (HDD = max(0, 65, T), summed over the forecast period, where T is the daily mean temperature). A 7-day cumulative HDD forecast is then compared to the historical median for that calendar week, yielding an anomaly. This anomaly is combined with recent actual HDD data and current inventory levels to form an expectation of near-term supply-demand imbalance.
Inventory Response Timing. Natural gas prices do not respond instantly and uniformly to all weather information. The first mover advantage belongs to traders who own or operate storage facilities. If a forecast predicts unexpectedly warm weather, storage operators may shift from injection to withdrawal plans. This operational decision occurs within hours of the forecast release and affects the physical gas market. Financial speculators in futures markets react to the same information, but often with a slight lag as they wait for storage index data or infer the physical response indirectly.
Futures Price Discovery. The Henry Hub natural gas futures contract, traded on the NYMEX (part of CME Group), reflects the market's consensus forecast of average daily natural gas prices over the contract month [2]. For the current month or front-month contract, price moves typically occur within hours of a material forecast revision. For contracts further out (e.g., next winter), weather forecasts have little predictive power and prices move mainly on supply news (production, LNG exports) or demand shocks (economic data, policy).
Worked Example: Winter 2021-2022
In late November 2021, the National Weather Service issued a forecast for an unusually cold December across the central and eastern United States. The GFS and GEFS models, issued repeatedly over the preceding 10-14 days, showed increasing agreement on a -1 to -2 standard deviation temperature anomaly for the first two weeks of December. The December Henry Hub futures contract traded near $3.00/MMBtu in mid-November. As the forecast confidence rose, traders revised upward their estimate of heating demand, and the December contract rose to approximately $3.80/MMBtu by early December [3]. In this case, the price move preceded the actual cold spell and reflected traders' conviction in the weather model forecast.
When December arrived and temperatures did indeed fall below normal, realized heating demand matched the forecast, and prices stabilized. This outcome represented a successful case in which the weather model was accurate and prices moved appropriately. However, the relationship is not always so clean: if the actual weather had been warmer than forecast, prices would have retreated, crystallizing losses for traders who had positioned for cold.
Limitations
Several structural and empirical limitations constrain the explanatory power of weather models on natural gas prices.
Forecast Error and Uncertainty. Weather forecasts are probabilistic and carry irreducible uncertainty. The GFS model, for instance, has typical root-mean-square temperature errors of 3-5 degrees Fahrenheit at the 10-day horizon [3]. This means that even a consensus forecast carries a wide confidence interval. Traders do account for this uncertainty by using ensemble forecasts and probability weighting, but traders often disagree on how to interpret model uncertainty or may overweight recent model performance. Large forecast errors are not rare and can cause prices to reverse sharply when model guidance changes or when actual weather exceeds the previously forecast range.
Non-Linear Demand Response. Heating and cooling demand does not respond linearly to temperature. Demand is inelastic within a range: one degree of additional cold does not necessarily increase demand by a fixed amount, because buildings are insulated and heating systems operate near maximum capacity during extreme cold. Also, price elasticity of demand matters: if natural gas prices rise sharply due to a cold forecast, some industrial users or power plants may switch to alternative fuels or reduce consumption, dampening the demand spike that the weather model predicted. This nonlinearity means that large price moves driven by forecast extremes often overstate the actual demand change.
Competing Information. Weather is only one of several drivers of short-term natural gas prices. Storage inventory levels, LNG export volumes, production disruptions, power generation fuel mix, and geopolitical events (pipeline outages, wars affecting supplies) all influence prices simultaneously. During periods of supply stress or extreme inventory draws, price moves can overwhelm or reverse weather-driven signals. For example, in the Texas winter storm of February 2021, exceptionally cold weather did drive prices higher initially, but the price spike was amplified by supply disruptions (frozen wells and pipelines) that had little to do with the forecast model itself. Conversely, during a warm winter with high storage inventory, even an accurate cold forecast may produce only muted price moves.
Forecast Persistence and Herding. Traders often move in the same direction in response to the same model run, creating momentum that can exceed the economically justified price adjustment. A surprise downward revision to a temperature forecast can trigger a cascade of liquidations in long positions, driving prices lower faster than physical supply-demand changes would warrant. This herding behavior is itself a form of noise, not directly tied to the accuracy or skill of the weather model.
Limited Horizon. Weather model skill degrades rapidly beyond 10 days. The Henry Hub contract is monthly, and prices further in the future are priced by supply and demand fundamentals, not weather. Thus, weather models exert measurable influence only on the near-term contract (1-2 weeks forward), while longer-dated contracts are anchored to different factors. This limits the practical use of weather models for medium- or long-term hedging.
Summary
Weather models move natural gas prices by influencing traders' expectations of near-term heating and cooling demand. The mechanism is straightforward: a model forecast of cold weather raises expected demand, demand forecasts feed into futures pricing, and prices rise. However, the link is noisy and subject to multiple confounds. Forecast errors, nonlinear demand, competing supply shocks, and herd behavior all weaken or invert the weather-price relationship. The strongest weather-driven price moves occur when model forecasts are consistent and extreme relative to seasonal norms, inventory is near seasonal averages (neither abundant nor depleted), and no major supply disruptions are underway. Outside these conditions, weather models remain one input among many, and their predictive power for prices is modest.
Key Definitions
Heating Degree Days (HDD): A measure of cumulative daily temperature shortfall below 65°F, used to estimate space heating demand; computed as the sum of (65 − T) for each day when the mean temperature T is below 65°F.
Deterministic Forecast: A single-valued prediction of a future weather variable (e.g., "tomorrow's high is 42°F"), as opposed to a probabilistic ensemble.
Ensemble Forecast: A set of multiple weather simulations, each with slight perturbations to initial conditions, used to estimate forecast uncertainty and the range of probable outcomes.
Henry Hub: A natural gas trading hub in Louisiana, the reference point for U.S. Natural gas futures contracts traded on CME Group's NYMEX.
Root-Mean-Square Error (RMSE): A measure of forecast accuracy, calculated as the square root of the average squared difference between predicted and observed values.
Forecast Skill: The accuracy of a forecast relative to a baseline (e.g., climatological average), often expressed as a percentage improvement.
Inventory Injection/Withdrawal: The flow of natural gas into or out of underground storage facilities, driven by seasonal demand and price expectations.
References
-
U.S. Energy Information Administration, "Natural Gas Explained: Use of Natural Gas," EIA (accessed 2025). Https://www.eia.gov/energyexplained/natural-gas/use-of-natural-gas.php
-
CME Group, "Henry Hub Natural Gas Futures Contract Specifications," CME (accessed 2025). Https://www.cmegroup.com/markets/energy/natural-gas.contractSpecs.html
-
National Weather Service, "National Forecast Accuracy," NOAA (accessed 2025). Https://www.weather.gov/media/epz/wxcalc/rhTdFromWetBulb.pdf
-
Investopedia, "Heating Degree Day (HDD): Definition and Why It Matters," Investopedia (accessed 2025).
-
U.S. Commodity Futures Trading Commission, "Commitments of Traders Report," CFTC (accessed 2025). Https://www.cftc.gov/MarketReports/CommitmentsofTraders/index.htm
Educational research on historical data only. Not investment advice, not a signal, and never a performance promise. Past results do not predict future performance. Every reference is link-verified before publication and every paper is re-audited weekly against the library's editorial standard.
Last reviewed by the PropLedger research pipeline: 2026-11-08. Educational research on historical data, not financial advice.
Keep reading
The Crack Spread: Refining Margins as a Tradeable Relationship
The crack spread represents the price difference between crude oil and refined petroleum products, most commonly gasoline and heating oil, and it reflects the economic margin of the
EIA and USDA Reports: How Scheduled Releases Move Oil, Gas and Grains
The U.S. Energy Information Administration (EIA) and Department of Agriculture (USDA) publish inventory and supply forecasts on fixed schedules that are followed closely by energy and agricultural traders.
OPEC Announcements and Oil: How Supply News Gets Priced
OPEC production announcements contain forward-looking information about global crude oil supply that the futures market reprices in minutes to hours following release. This price discovery mechanism reflects both