Thousands of years of climate data refined to better predict future environmental changes

Thousands of years of climate data refined to better predict future environmental changes

A collaborative effort between the universities of Cordoba and Granada improved the spatial resolution of monthly climate data from 22,000 years ago to the year 2100, enabling more localized and detailed studies.

Knowing the climate of the future is one of the current challenges and an essential issue when studying biodiversity and making long-term predictions regarding species' behavior or making decisions today. To know what a Spanish fir forest will be like 200 years from now, climate models that forecast future changes are needed.

“When developing climate models, you have to make sure that they work well. As we do not know what the climate of the future will be like, and are unable to verify it, we validate it with past data that is real, projecting the model backwards to verify that it works”— explains Diego Nieto, a researcher with the Basic and Applied Plant Biology group at the University of Córdoba — “If these backward projections work well, those of us who are dedicated to the natural environment use them to determine how vegetation or diversity has changed depending on the climate.”

One of the main challenges they face when it comes to understanding the climate and vegetation of the past, and being able to create predictions for the future, is that climate models work with very great resolutions; that is, they present data on very large areas (between 150 and 250 km), which makes it difficult to carry out studies and projections for specific areas, such as the Mediterranean.

Hence, the latest work by Diego Niego and Daniel Romera, a researcher in the same group, together with researchers from the University of Granada, has been the rescaling of a set of monthly climate data from 22,000 years ago, to the year 2100, obtaining a much more detailed resolution of 11 x 11 km, making possible better analysis of the phenomena of more specific ecosystems.

“The idea was to generate a set of past climate data at a level of detail that would make sense for biodiversity studies.” The team used one of the climate change models (TraCE-21ka) and rescaled it with a number of novel tools that had not previously been used to do this kind of rescaling at this resolution.

With this work, published in the journal Scientific Data, a ‘toolbox’ is made available to the scientific community to apply advanced rescaling techniques to large-resolution climate datasets using the R0 programming language, an open and free code that facilitates its easy use. The set of tools it contains allows for the rescaling of data corresponding to 7 climatic variables, including average temperature, solar radiation, precipitation and wind speed. In addition to the tools that enable each research team to rescale data to a resolution of 11x11 km for the area of greatest interest, a set of data that has already been rescaled is shared for a region recognized as a global biodiversity hotspot: the Western Mediterranean.

Thanks to this advance, studies can be carried out that allow us to know how biodiversity has responded to changes in the climate and make predictions regarding how it will change in the face of the effects of current climate change. In this way more reliable tools will be generated to, for example, manage species in natural parks, or long-term conservation plans.

Reference:

Romera-Romera, D., Alba-Sánchez, F., Abel-Schaad, D. et al. Replicable Fine-Spatio-Temporal Climate Data for Long-Term Ecology in the Western Mediterranean. Sci Data 12, 747 (2025). https://doi.org/10.1038/s41597-025-05067-9