Retailers are increasingly exposed to climate variability, especially when it affects customer behaviour, seasonal product demand, or warehouse logistics. In a new study published in the journal Climate Services, researchers from the Earth Sciences Department at the Barcelona Supercomputing Center-Centro Nacional de Supercomputación (BSC-CNS) co-developed a climate service tailored for Decathlon Spain, combining sub-seasonal and seasonal forecasts to support operational decisions.

The collaboration began after a mismatch in product supply during an unusually late summer heatwave in 2018. Over 3.5 years, BSC scientists and Decathlon teams worked together to build a real-time forecast system that predicts key climate variables, such as maximum temperature and rainy days, which directly influence product demand.

"A key part of our project was user engagement, both in co-developing and co-evaluating the service," said Dr Albert Soret, leader of the Earth Science Services group at BSC and first author of the study. "Some users had access to the climate information, while others didn’t, allowing us to analyse its practical impact over a year in real-world conditions."

The service leverages state-of-the-art climate forecasting models and a multi-model ensemble approach to improve prediction reliability. It integrates global datasets, applies regional downscaling and bias correction, and delivers forecasts specific to the company’s geographical areas of interest.

This climate service has already shown positive results. Tested in selected Decathlon stores, the forecasts helped improve planning, optimise stock distribution, and increase sales compared to control locations. Its user-centric design includes interactive visualisation tools, enabling decision-makers to integrate climate insights into strategic planning.

This is one of the first examples of a climate service co-developed specifically for the retail sector. The approach is now being explored in other regions and industries, including fashion and food, where seasonal demand and climate sensitivity play a crucial role.

"A frequent topic was the need to simplify climate information; users often prefer deterministic forecasts over probabilistic ones," Soret explained. "But probabilistic forecasts are crucial!"

This work highlights BSC’s commitment to applying climate science in practical contexts, fostering industry partnerships and building business resilience in the face of an increasingly variable climate.

“This kind of collaboration shows how climate science can drive practical, low-carbon innovation in industries far beyond the lab,” added Soret.