A simple method for evaluating the relative importance of environmental conditions within habitats and connectivity among habitats in odonate metapopulations

The biodiversity of freshwater fauna is now suffering from an impending extinction crisis. The loss of habitats and their consequent fragmentation is the primary threat, though degradation within the habitats is also a serious issue. We developed a new method that can evaluate the relative importance of the environmental conditions within habitats and the connectivity among habitats from one-time field observations. In this method, three kinds of linear regression models were constructed for the species abundance. One is a multi-regression model involving the environmental factors. Another is spatial trend-filtering using the principal coordinates of neighbor matrices (PCNM). The third is a uni-regression model involving the connectivity indices assuming the Hanski type of metapopulation. The relative explanatory power of these models was determined in terms of their own pure elements and their mutually overlapping elements by the method of variation partitioning. We applied the method to the actual ‘snapshot’ observations of odonata metapopulations in rural ponds. Copera anulata, which is a residential damselfly, was affected by both the connectivity among ponds and the environmental conditions within ponds. On the other hand, Pseudothemis zonata, which is an abundant and ubiquitous species, showed small effects of the connectivity and the environment. Sinictinogomphus clavatus, which is a large dragonfly and a long flyer, had a strong environmental preference while it was slightly influenced by spatial and connectivity effects. These results corresponded with each species behavior and biology reported in preceding works.