Multiscale analysis of the species diversity of Maltese shrubland

Abstract

The aim of this study was to determine whether the species diversity and composition of a plant community depend on the scale at which the study is carried out, and to determine the effect of certain site characteristics and the presence of dominant species on species diversity at different spatial scales. The study was carried out using a hierarchically nested sampling design at four spatial scales (125m x 125m, 25m x 25m, 5m x 5m, and 1m x 1m) in six shrubland sites in the Maltese islands. The design was balanced across scales meaning that the change in scale from a smaller unit to a larger one was consistently incremental. Each sampling plot comprised five subplots of the next lower scale. Amalgamated abundance data for all perennial plant species was collected from the study plots between October 2015 and August 2016. Seven site characteristics (Aspect, Slope, Elevation, Rock type, Exposure, Soil pH, and Soil electrical conductivity) were measured for each site, but only the three that contributed most to the variation between sites (Elevation, Exposure, and Slope) were included in the analyses. The data was analysed using multivariate techniques (RDA, DCA, CCA, and PCA) and the vegetation distribution patterns were related to the site characteristics. The plant communities in each of the six sites had their own distinctive ‘vegetation signature’ that was recognizable at all scales, although less clearly at the 1m x 1m spatial scale. The observed species richness, estimated species richness (Chao-1 index), and N2 diversity all increased as the spatial scale increased in all six sites, while evenness decreased. Moreover, the observed and estimated species richness of a community at smaller scales was highly correlated with that at larger scales. The three site characteristics together explained 61.9% of the variability in community composition across the six sites at the largest spatial scale. When the abundance of Thymbra capitata was added as an explanatory factor, the explained variation rose to 86%. The abundance of Thymbra capitata explained more of the variation in community composition at the largest scale compared to the smallest. This effect of Thymbra capitata was unexpected, as previous studies had suggested that the effect of woody shrubs was greater at smaller scales. At all four spatial scales, higher species richness and diversity were associated with lower abundance of Thymbra capitata and vice versa. The results of this study showed that a multiscale approach towards species diversity would provide more relevant information than that obtained from a single scale. The results also showed that assessment of community composition should be carried out using study areas of at least 25m x 25m because at this scale, the vegetation ‘signatures’ of each community are still very distinct from each other.