The dry grasslands of Abruzzo National Park, the oldest protected area in the Apennines (Central Italy): overview of vegetation composition, syntaxonomy, ecology and diversity

The Abruzzo, Lazio and Molise National Park was established in 1923 and is considered a flagship conservation area in Italy. It includes large extensions of semi-natural grasslands, maintained by traditional transhumant grazing for centuries. The patterns and drivers of grassland composition within the Park are still poorly investigated, and the scattered phytosociological data available were often based on relevés with varied and not precisely defined sizes. In order to provide for the first time a general overview of the Park’s dry grasslands, we analysed a dataset of 87 relevés with a fixed size of 2 × 2 m, precisely delimited in the field and GPS-located. Specific research aims were: (1) to classify the vegetation plots into floristic-ecological types, supported by an analysis of mean (Italy-specific) Ellenberg Indicator Values (EIVs); (2) to assign the types to up-to-date phytosociological syntaxa; (3) to identify the main environmental drivers for both composition and richness patterns; (4) to test the degree of correlation between (Italy-specific) EIVs and the measured environmental variables. Environmental predictors included high-resolution climatologies and remote-sensed standing biomass. Main vegetation types were identified using Hierarchical Cluster Analysis (HCA). Distancebased RDA was performed as a constrained ordination method to reveal correlations between floristic composition and environmental variables. Drivers of species richness were explored through partial correlation and Regression Trees. HCA and NMDS revealed four floristically and ecologically well-interpretable groups, in turn well corresponding to the level of phytosociological class (namely Molinio-Arrhenatheretea, Nardetea strictae, Festuco hystricis-Ononidetea striatae and Festuco-Brometea). Constrained ordination showed that most of the floristic variation was explained by biomass, annual precipitation (Pann) and mean annual temperature (Tm). Strong and significant positive correlations were found between biomass and EIV for Nitrogen (EIV-N), and between Tm and EIV for Temperature (EIV-T). Strong and significant negative correlations were found between Pann and EIV-T, EIV for Continentality (EIV-C) and EIV for soil Reaction (EIV-R). Species richness was positively correlated with slope inclination and negatively with elevation; richness was higher in sites with a high rock cover, and on limestone or clayey substrata than on siliceous ones. We conclude that in the study area: a) dry semi-natural grasslands are arranged at least into four distinguishable, high rank floristic-ecological groups; b) a mixture of climatic (especially precipitation) and edaphic (especially bedrock and soil reaction) gradients are the main drivers of such composition patterns; c) species richness is higher in sites more stressed by summer drought and/or nutrient scarcity; d) community-means of Italy’s specific EIVs are well correlated with environmental variables in grasslands, including a good correspondence between EIV-T and mean annual temperature.