Metro Madrid
Metro Madrid
High resolution forecasting to improve a public service
Wind gusts that tear down trees, blocking the traffic out; heavy rains that delay bus frequencies; or temperature drops that create black ice on the highways. Weather conditions disrupt transport in many ways, and the underground is no exception.
That’s why, since 2016, we have worked with Metro Madrid (Madrid’s underground system) to provide a daily forecasting service. Here’s what it entails:
- High resolution forecasts: we use the WRF model to cover the whole Metro network, providing forecasts on the next 72 hours.
- Multiple ouputs: not only temperature and precipitation levels, but relative humidity as well. To cover all possibilities, like frost forming on the railway surface when relative humidity is high and temperature drops.
- Ensemble forecasting: instead of providing a single forecast for each variable, we provide a range of possibilities provided by our models, to have a grasp of the uncertainty of the prediction.
- Daily feeds: in the case of Metro Madrid, their preferred channel of communication is mailing, so we provide a daily update on the forecast, that is delivered to nearly 1,000 Metro workers. In total, we have delivered nearly 1.5 million e-mails.
- Alert notification system: we send out punctual notifications to alert the users of the service when certain values that trespass pre-defined thresholds, tailored to Metro Madrid’s demands.
This service enables Metro Madrid to:
- Cut down on operational costs: having a prediction on the weather enables to keep metro stations operational and comfortable, estimating the energy required by the grid and allowing to bid earlier and better on the energy trading market.
- Have an early alert on weather that can cause disruptions: from intense precipitations in specific points of the grid, that could interrupt the service, to potential frost on the catenary, that could become a safety issue.
Key features
Daily forecast
Alert notifications
High resolution forecasting
Coverage of the underground network