Probabilistic and preferential sampling approaches offer integrated perspectives of Italian forest diversity

Aim: Assessing the performances of different sampling approaches for documenting community diversity may help to identify optimal sampling efforts and strategies, and to enhance conservation and monitoring planning. Here, we used two data sets based on probabilistic and preferential sampling schemes of Italian forest vegetation to analyze the multifaceted performances of the two approaches across three major forest types at a large scale. Location: Italy. Methods: We pooled 804 probabilistic and 16,259 preferential forest plotsassamplesofvascularplantdiversity across the country. We balanced the two data sets in terms of sizes, plot size, geographical position, and vegetation types. For each of the two data sets, 1000 subsets of 201 random plots were compared by calculating the shared and exclusive indicator species, their overlap in the multivariate space, and the areas encompassed by spatially-constrained rarefaction curves. We then calculated an index of performance using the ratio between the additional and total information collected by each sampling approach. The performances were tested and evaluated across the three major forest types. Results: The probabilistic approach performed better in estimating species richness and diversity of species assemblages, but did not detect other components of the regional diversity, such as azonal forests. The preferential approach outperformed the probabilistic approach in detecting forest-specialist species and plant diversity hotspots. Conclusions: Using a novel workflow based on vegetation-plot exclusivities and commonalities, our study suggests probabilistic and preferential sampling approaches are to be used in combination for better conservation and monitor planning purposes to detect multiple aspects of plant community diversity. Our findings can assist the implementation of national conservation planning and large-scale monitoring of biodiversity.