This post summarizes a new study investigating whether extreme high-temperature weather can help forecast crude oil prices.
The research, led by Donglan Zha, Shuo Zhang, and Yang Cao at Nanjing University of Aeronautics and Astronautics, introduces a novel High-Temperature Index (HTI) derived from meteorological observations at oil production and storage sites connected to China’s International Energy Exchange (INE).
The study evaluates HTI alongside traditional predictors using several forecasting models and finds meaningful improvements in prediction accuracy when heat extremes are included.
What the researchers did and why it matters
The team compared the predictive value of HTI against 15 conventional oil-price predictors across five classical forecasting frameworks.
Their goal was to test whether a climate-oriented index—focused specifically on extreme high temperatures—adds explanatory power for INE oil prices beyond standard financial and macroeconomic indicators.
High-Temperature Index (HTI) and data sources
HTI was constructed from meteorological data at oil production and storage locations linked to the INE.
By focusing on sites with direct relevance to China’s crude market, the index captures localized heat stress that can affect production, storage integrity, and transport logistics.
The researchers then incorporated HTI as an additional predictor in forecasting models to evaluate its incremental value.
Forecasting experiments and headline results
The study tested HTI in out-of-sample forecasts across five classical models and benchmarked performance against 15 traditional predictors, including macroeconomic and financial variables.
Across the board, adding HTI improved forecast accuracy, signaling that weather extremes contain information not captured by standard indicators.
Model performance and the stand-out RNN
The best-performing model was a recurrent neural network (RNN) with an input sequence length of one.
This RNN achieved a mean absolute error (MAE) of 14.379, a root mean square error (RMSE) of 19.624, and a directional symmetry rate of 66.67%, demonstrating superior ability to capture short-term price movements when HTI is included.
HTI frequently ranked as the third most important predictor in the optimal RNN model, outperforming some traditional variables such as stock market data.
Accumulated local effects (ALE) analysis further revealed a clear positive correlation between extreme heat and INE oil prices—strong evidence that heat extremes are a non-negligible driver of price dynamics.
Why extreme heat affects crude oil markets
Extreme high temperatures can stress infrastructure, reduce workforce productivity, force production throttling, and increase energy demand for cooling.
All of these factors feed into oil supply, storage, and transport dynamics.
Practical implications for investors and regulators
From a market-practice perspective, the study’s findings suggest several concrete steps:
Can extremely high-temperature weather forecast oil prices?—the paper published in Frontiers of Engineering Management (2025, Vol. 12, Issue 3)—adds a timely climate dimension to oil price modeling.
For traders, analysts, and policy makers, climatic extremes are an emerging source of market risk and a useful signal for forecasting when properly quantified.
Here is the source article for this story: Can extremely high-temperature weather forecast oil prices?