Features of the spatial-typological variability assemblage study of certain bird species groups and their distribution (on the example of Corvids)
Yu. S. Ravkin1, O. A. Odintsev2, I. N. Bogomolova1
1Institute of Systematics and Ecology of Animals SB RAS, Novosibirsk, Russia
2Omsk State Pedagogical University, Omsk, Russia
Keywords: Corvidae, Northern Eurasia, distribution, territorial heterogeneity of the assemblage, quasi-fractality, cluster and gradient analysis, qualitative approximation, environmental factors
Abstract
Results of bird counts from 1880 to 2019 (intermittently, and mainly since 1960), averaged over the first half of the summer (16.05-15.07) were analyzed to identify the spatial-typological structure and organization of the corvid assemblage of Northern Eurasia. 352 researchers have been involved in collecting of this material over 110 years. Data processing was carried out using multivariate statistics methods, including cluster analysis and linear qualitative approximation of connection matrices, as well as gradient and expert approaches. As a result, the main territorial trends in corvid assemblage and the natural and anthropogenic regimes correlating with them have been identified, and the relationship between the spatial variability of the assemblage of these birds and the heterogeneity of habitats has been assessed. Quasifractal spatial differentiation of assemblage was shown. The analysis of corvid groups identified using factor classification gave an unsatisfactory result of approximating the spatial and typological variability of their assemblage in terms of abundance at the level of landscape tract type and a rather high information content of species classification by abundance in all 8144 habitats. A satisfactory explanation for the heterogeneity of corvid assemblage was achieved only after averaging the abundance indices over landscape groups. It was not possible to get an idea about them using cluster analysis, and only the use of gradient and expert (speculative) approaches, followed by an assessment of the information content of the representations, can be considered acceptable for obtaining such generalizations. The use of these techniques increased the assessment of the information content of representations by averaged abundance indicators by seven times in terms of explained variance and three times in terms of the multiple correlation coefficient. At the same time, structural classifications by species, and especially by territorial communities, acquire a two-dimensional reticular (network) form. The formalized classification of corvid species according to the similarity of distribution is seven, and the assemblage is almost twice as informative as expert-speculative representations.
|
