Ábrahám Levente: Biomonitoring a Dráva folyó magyarországi szakasza mentén 2000-2004 - Natura Somogyiensis 7. (Kaposvár, 2005)

Horváth, Győző, Molnár D., Németh Tamás - Csete Sándor: Landscape ecological analysis of barn owl pellet data from the Drava lowlands, Hungary - A Dráva menti gyöngybagoly köpetekből nyert adatok tájökológiai elemzése

184 NATURA SOMOGYIENSIS prey taxa (M arvalis, Apodemus spp. and S. araneus) and their most characteristic habitat patches (ploughland, broad-leaved forests). As suggested by G-tests, the relative frequency of the typically forest-dwelling small mammal C. glareolus, found at lower proportions in the samples, is also determined by the ratios of closed, dry deciduous forests. Thus, in these species the relative frequency calculated for a particular Drava section differed from values obtained for certain localities, whic is because of the fact that the habitat patches typical of these species differed significantly too when locaities and largerscale areas were compared (Table 4). The relative proportions of C. glareolus and S. araneus in the Drava sections were noted as characteristic for forest patches; and relative proportions of M. spicilegus es a С suaveolens as characteristic for natural grasslands. When patch ratios and the relative frequencies of these species were charted together, it suggested that relative prey species abundance in the food of the owls depends on the extent of the characteristic small mammal habitats.(Figure 1-2). The correlations obtained by homogeneity tests were then tested with regression analysis too, to reveal significant correlation between relative frequencies of habitat patches and that of the particular species, in the case of M arvalis, Apodemus spp., and С glareolus (Figure 3-5). As patch proportions grew, relative frequencies of wood mice {Apodemus spp.) and common vole increased with an exponential curve, whereas for the bank vole significant linear correlation was found, showing that as the proportions of forested patches grew, the abundance of this species in the owl pellets also increased. When relationships between landscape ecological parameters and small mammal community parameters derived from pellet data were analysed in the case of localities, significant correlation was found between patch diversity and prey species number, meaning that the number of species indicated by pellets is in linear correlation with patch diversity (Figure 6). Conclusions Spatial heterogeneity is one of the most essential factors determining processes in populations and communities. The effect of heterogeneity is seen mostly in landscapes and habitats transformed by humans (KOZAKIEWICZ 1983). In respect of the Drava reaches subject to barn owl pellet collecting and analysis as part of small mammal indirect monitoring, landscape ecological analyses showed that among the three studied areas the one in the lower section was considerably different from the two areas in the upper reach. When interpreting the results, it has to be noted that the proportions of wet forest patches were considerably higher and those of ploughlands were lower in the upper Drava section. An important finding was the significant difference in the frequency of marshy areas in the lower (Baranya county) section. These results clearly indicated landscape ecological differences between the various Drava reaches, and, accordingly, differences occurred also in the compositions of their small mammal communities as shown by owl pellets. The results obtained are remarkable because due to seasonal differences in the hunting strategy of the barn owl, and due to the presence of prey preferences, i.e. to density-dependent hunting, the over-representation or under-representation of certain species can considerably influence small mammal abundance values obtained from pellets. Selective prédation is studied in the barn owl (COLVIN and MCLEAN 1986, DICKMAN et al. 1991, TAYLOR 1994), but there are no data showing how much significance prey selection has in comparing data of various temporal and spatial scales, in relation to small mammal availability expected on the basis of mosaic habitat patches.