MIT: Risk of Weather Data Sabotage on the Rise

MIT Technology Review reports that as prediction markets and AI forecasting become widespread, the incentive to manipulate weather data is growing, threatening forecast accuracy. Industries worldwide—including airlines, power companies, and farmers—rely on morning weather forecasts to make critical decisions regarding flight operations, electricity pricing, agriculture, and life-saving rescues. Recently, as prediction markets that bet on real-world events, including the weather, have emerged as a booming industry, there have been confirmed instances of attempts to manipulate weather data for profit.

In one incident, evidence was captured of someone using a hairdryer or a lighter to heat a thermometer. Although the actual average temperature at the time was around 18 degrees Celsius, the observation equipment recorded a spike to 22 degrees Celsius. A gambler who targeted this discrepancy bet on the result and won a large sum of money. Members of a French climate non-profit organization discovered the anomaly, exposing the full scope of the incident.

Accurate weather forecasting relies on observational data collected from weather stations at airports, utility companies, and transportation services. Traditional operational systems, such as the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System or the Weather Research and Forecasting (WRF) model, estimate future weather patterns by combining observational data with numerical approximations. They possess a safety mechanism known as data assimilation to prevent errors caused by equipment malfunctions or upgrades, and all measurements are cross-referenced with the expected values of physical models and readings from nearby observation stations.

Experts currently view these risks as manageable, but they warn that we must prepare for the possibility of these issues expanding into system-wide problems in the future.