A changing Mediterranean Sea climate: do we know where we are heading?

The Mediterranean Sea climate is undoubtedly changing at unprecedented rates, affecting its surface, shelf regions, intermediate waters, and even its deepest layers and dense water formation sites. Beyond physical changes, the impacts on biogeochemical processes and living organisms remain poorly understood and—despite existing regional climate projections—are largely unknown in terms of how they will respond to ongoing physical transformations. Here, we present several cases that reveal recent physical changes occurring in the Mediterranean Sea. These are based on observations from coastal and deep-sea observatories, opportunistic sampling, long-term ocean stations with nearly centennial time series, and modern observing platforms such as Argo profiling floats. At the surface, we document a widespread occurrence of salinity maxima, not only in the Levantine Basin but across all Mediterranean basins. The second example describes a paradigmatic shift in the properties of North Adriatic Dense Water, with much warmer and saltier waters now occupying the bottom of the Adriatic Sea. Dense-water formation has become predominantly haline-driven, while wintertime cooling now plays a reduced role in dense water cascading. A lack of precipitation and changes in precipitation regimes in the Alpine region have led to higher-than-average salt accumulation in the shallow northern Adriatic. This affects not only dense-water formation but also the summertime spreading of freshened waters—primarily of Po River origin—off the western Adriatic coastline, accelerated in the situations of stronger stratification. In case of warmer surface ocean during late spring and summer as in 2024, such rapid spreading might result in occurrence of extensive mucilage events. Salinity positive shifts have also occurred in the Eastern Mediterranean following the winter of 2022, superimposed on the steadily increasing salinity observed during the Argo era. Many of these physical changes are not adequately captured by climate models, that is, they are not projected to occur at the rates observed. This raises the question of whether current projections remain valid and how they might be improved. In this context, we discuss our findings and outline possible pathways for future research.