Tracking Oil Demand From Space Using Real-Time Satellite Pollution Data

A new study finds that daily satellite measurements of nitrogen dioxide, a pollutant linked to fossil fuel use, can significantly improve real-time forecasts of oil demand across major economies. By offering timely, global and high-frequency data, satellites help close the gap left by delayed official energy statistics.

When oil demand collapsed during the COVID-19 lockdowns, the world felt the shock immediately. Planes were grounded, roads were empty and factories slowed to a halt. Yet official oil consumption data confirming the scale of the crash arrived months later. By then, markets had already reacted and policymakers were working with incomplete information.

A new study by researchers from the European Central Bank, Banque de France, Paris 1 Panthéon-Sorbonne University, the Laboratoire d’Économie d’Orléans, QuantCube Technology and Equancy suggests that the answer to this timing problem may be orbiting above us. Their research explores whether satellites can help predict oil demand in real time. The conclusion is clear: satellite pollution data can significantly improve oil demand forecasts.

Why Nitrogen Dioxide Matters

The key lies in nitrogen dioxide, or NO2. This gas is mainly produced when fossil fuels are burned in cars, trucks, airplanes, power plants and factories. Unlike greenhouse gases that stay in the atmosphere for years, NO2 disappears within hours or at most a day. That makes it a near-instant signal of fuel use.

If traffic rises or factories increase output, NO2 levels quickly rise. If economic activity slows, they fall just as quickly. In short, NO2 acts like a daily pulse check of combustion activity, which is closely linked to oil consumption.

Satellites such as the European Space Agency’s Sentinel-5P, equipped with the TROPOMI instrument, measure NO2 across the globe every day. These readings are available within hours. Compared to traditional oil demand statistics that are released with a two- to three-month delay, this is a major advantage.

Turning Raw Satellite Data Into Economic Signals

Satellite data, however, are not perfect. Clouds, snow and sunlight angles can distort readings. To make the data reliable, the researchers developed a detailed cleaning and processing system. They filtered out low-quality measurements, combined fast but less refined data with more accurate versions, and focused on large cities where fuel use is concentrated.

They then smoothed daily fluctuations using a 28-day moving average, which helps track monthly trends. The final result was a national NO2 index for each country, updated continuously.

The study examined ten major economies: the United States, China, India, Japan, South Korea, Australia, France, Italy, Spain and the United Kingdom. Together, these countries represent about 60 percent of global economic output and NO2 emissions.

Strong Results, Especially in Crisis Times

The researchers tested whether adding satellite-based NO2 data improves oil demand forecasts. They compared traditional models with and without the pollution index.

The results were striking. On average, forecast errors fell by around 25 percent when NO2 data were included. Even when traditional factors such as industrial production, prices, car registrations and weather conditions were already part of the model, adding NO2 improved accuracy by more than 20 percent.

The biggest improvements came during crisis periods. In 2020, when the pandemic sharply reduced fuel use, models that included satellite data performed more than 30 percent better than simple traditional models. In countries like China and Spain, the gains were even larger.

Importantly, the advantage did not disappear after the crisis. Between 2021 and 2024, satellite data continued to improve forecasts, showing that NO2 is useful not only during emergencies but also in more stable times.

A Powerful Tool for the Future

The researchers also compared satellite pollution data with other high-frequency indicators, such as Google Mobility data and online search trends. While those tools were helpful during lockdowns, their usefulness declined afterward. Satellite-based NO2 data remained reliable both during and after the pandemic.

To take the analysis further, the team used machine learning models, including neural networks. These advanced tools are designed to capture complex patterns in data. Even in these sophisticated models, adding NO2 data improved results. This suggests that the relationship between pollution and oil demand is not simple, but satellite data help reveal it more clearly.

The broader message is powerful. Satellite pollution data are daily, global and free. They do not depend on smartphone usage or online platforms. They capture emissions from households, businesses and public infrastructure alike.

For policymakers, investors and energy analysts, this could mark a turning point. Instead of waiting months for official statistics, they may soon rely on real-time signals from space to track the world’s energy use.

From hundreds of kilometres above Earth, satellites are quietly mapping the rhythm of the global economy. The smoke rising from cities is no longer just pollution. It is information.