An ecotone pattern is revealed only after controlling for spatial blurring effects

Background/Question/Methods

Changes in community structure along biogeographic transition zones may reflect changes in species occupancy and habitat preference due to climate change at the margins of their distribution. We examined patterns of spider community similarity between patches along the sharp and fragmented climatic gradient of Israel. Our main aim was to test predictions regarding whether the transition zone in our study relates to an ecocline pattern (with a stable intermediate state that is a mixture of the Mediterranean and arid faunas), or an ecotone pattern (with no such stable intermediate state). We sampled a total of 14,854 individuals from 199 species, in samples stratified according to 9 different microhabitats. The samples were distributed in 35 patches, located within 3 landscapes along the climatic gradient. We used one-way permanova to explore patterns of similarities between patches, according to six distance categories – 3 categories of two patches from the same landscape, 2 categories of two patches from adjacent landscapes and 1 category of two patches from distant landscapes. We repeated the analysis under 3 stratification schemes – all samples from all microhabitats, exclusion of samples from structurally simple microhabitats, and taking only pairs of samples from the same complex microhabitats.

Results/Conclusions

Spider abundance and richness per-sample, as well as occurrence data, reveled that the microhabitats may be divided into two groups -- simple structure and complex structure -- with the first being a random subset of the later. The transition zone forms a mosaic ecotone pattern, with no stable intermediate state. The ecotone pattern agrees with other published results from this area and with the results of hierarchical cluster analysis according to the relative cover of the different microhabitats. However, the ecotone pattern was evident only after controlling for two blurring effects – 1) reducing within-object variance that is irrelevant to the question by excluding the simple microhabitats from the analysis, and 2) conserving between-object variance that is relevant to the question by taking only pairs of similar complex microhabitats. As these two blurring effects are an inherent property of most large-scale studies, they must be carefully considered in such studies before any conclusion is declared.