Optimal plot size for carbon-diversity sampling in tropical vegetation

Abstract

Obtaining high-quality information on vegetation attributes directly linked to ecosystem services in tropical forests, such as carbon stock and biodiversity, is essential to develop ecological information to be used in planning conservation strategies. Defining the optimum vegetation sample plots size is an important part of the process of obtaining information, since is crucial for its quality, and also for its financial costs and time required. In this context, we evaluated whether: (i) the optimum plot size selected in two approaches (basal area/carbon stock vs biodiversity) are different since they are associated with different ecological drivers; (ii) different tropical forests vegetation types present different optimum sample plot size due to their ecological specific attributes, such as species composition, structure and functioning. For this we evaluated the vegetation sample plots optimum size of different vegetation types of tropical savanna and forest of broad occurrence in the tropics (Atlantic Semideciduous Forest – ASF; Cerrado Woodland – CW; Cerrado Forest – CF; and Seasonally Dry Tropical Forest - SDTF), in two approaches: forest basal area and carbon stock, since they are related variables; and biodiversity variables by species composition similarity and heterogeneity in communities. We used 4 forest fragments in Brazil that were sampled in a way that allowed the arrange of subplots to form different plot sizes. In order to evaluate the optimum sample plot size in the basal area/carbon stock approach we used four methodologies, in which the obtained variability was compared to a reference size (1000 m2). For the biodiversity approach we used the floristic similarity and composition heterogeneity between units as metric, evaluating differences in the similarity patterns and possible reductions or increases in variability compared to the reference size. We found there are no significant differences for biodiversity variables between sample plot sizes for all vegetation types, thus the optimum size may be selected using just the basal area/carbon stock approach. In the basal area/carbon stock approach, most of the evaluated methods generally presented lower plot size values compared to the reference size (1000 m2) in all vegetation types, except for the ASF. There are differences between vegetation types that may be grouped in two groups related to the optimum size selected, that are related to their ecological restriction patterns: CW, CF and SDTF with optimum size between 600 and 1000 m2; and the second with ASF, with optimum sample plot size between 1000 m2 and 1400 m2.