Land use determines Mediterranean ecosystems' multifunctionality more than plant richness or habitat composition

1. Local and landscape diversity loss, habitat fragmentation and land-use changes commonly co-occur and impair ecosystem multifunctionality, yet they are often studied in isolation. Therefore, we ignore the relative importance of these drivers of ecosystem change or whether or not they interact to determine ecosystem functioning. 2. We measured how changes in local (plant richness) and landscape (different land uses) diversity, land use (orchards, shrubland, pine, oak and mixed forests) and habitat sise, and their interactions, affected the functioning of Mediterranean ecosystems. At 49 plots, we measured 17 above- and below-ground functions, related to nutrient cycling, carbon sequestration and biotic interactions. Additionally, we generated different combinations of plots to compose artificial landscapes where multifunctionality and biodiversity conservation can be maximised, in order to aid land managers to preserve or design functional and diverse areas. 3. The relative importance of local and landscape attributes varied slightly depending on the target function, yet land use was by far the strongest predictor of most functions and multifunctionality above local plant biodiversity, habitat sise or landscape diversity. Oak and mixed forest were the most multifunctional land uses but were functionally wcomplemented by other land uses that maximised some individual functions. 4. The relative proportion of each land use within optimal landscapes varied if the target was biodiversity conservation (with an even dominance of land uses), carbon stocks or multifunctionality (highly dominated by native oak and mixed forests). 5. Synthesis and applications: Our results highlight the importance of remnant native forest to provide multiple ecosystem functions and the potential to restore them and complement these uses with sustainable agriculture. By merging landscape and ecosystem approaches we provide specific numbers regarding the proportion that each land use should have in order to maximise biodiversity conservation and/or functioning in these charismatic environments.

This research is funded by the Spanish Ministry of Science, Innovation and Universities (project FOBIASS; RTI2018-098895-A-100). Miguel Berdugo acknowledges support by a Ramón y Cajal grant (RYC2021-031797-I) from Spanish Ministry of Science.