Elevation-based post-stratification in Atlantic forest sampling
Cielo-Filho, Roque; Martins, Fernando Roberto
Question: Management and restoration of biodiversity hotspots worldwide might benefit from an intensification of sampling effort through the use of larger field plots than usually employed. However, besides being more expensive, small-range elevation variation and the related floristic composition gradient may compromise the desired internal floristic homogeneity in larger field plots. We present a method to remedy these problems, which consists in sampling the field plot and post-stratify the sample if a significant relationship between floristic composition and elevation is present. Location: A 6.5 ha Atlantic forest field plot with an elevational range of 40 m, located in Southeast Brazil (22º50'S, 46º55'W). Methods: We sampled the field plot with one hundred 10 m × 10 m sampling units and used Non-metric Multidimensional Scaling to extract significant floristic composition gradients (stems≥ 5 cm DBH). The median score of the axis that presented a statistically significant correlation with elevation was used to stratify the sample into two subsamples. Floristic separation between subsamples was evaluated through Multi-Response Permutation Procedure and the Indicator Value Method. Results: We recorded a total of 1080 living individuals of 119 arboreal species. The NMDS ordination resulted in a final solution with three significant axes of which only one presented a statistically significant correlation with elevation. We found a significant floristic separation between subsamples obtained on the basis of that axis and each subsample presented its own indicator species. Conclusions: Each subsample represents a particular community type and should be considered as an individual plot in the data bank to be used in the investigation of floristic complementarity in the landscape the study area belongs to. The approach proposed here may aid conservation efforts in other biodiversity hotspots, besides the Atlantic forest biome, favoring the definition of conservation programs at the plant community level.