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

182 NATURA SOMOGYIENSIS open, dry or wet deciduous forests, as well as coniferous forests and forest plantations), natural and degraded grasslands, shrubby areas, unvegetated open areas, ploughlands, perennial cultivated plants (vineyards and orchards), agricultural areas with various types of cultivation, artificial surfaces (artificial, non-agricultural vegetated areas and urban areas). Two areas were selected in the upper Drava monitoring areas (between Zákány-Porrogszentkirály, and Berzence-Heresznye, respectively), whereas in the lower section only one larger area was chosen (Drávaiványi-Szaporca). The local landscape ecological analysis of the hunting areas was done on the basis of the patch composition in a 2 km radius around the nesting sites, whereas for the greater-scale analysis of the particular Drava section we used the patch composition of the area bordered by the artingents of 10 km radius circles around nesting sites and by the line of the Drava. As part of lanscpe ecological analysis, the patch diversity values of the entire river section areas were calculated, and the comparison of sampling localities and larger-scale areas was done by G-tests and patch overlap calculations (Schoener-index). In the case of all 17 places, the analysis of barn owl food composition was done using the cumulated data from pellet samples collected in 2002 (Table 1). For comparing small mammal proportions at various river sections homogeneity tests (G-test) were applied. Any correlation between habitat patches that were found to be occurring in considerably different rates within the various sampling areas and the abundance of their characteristic small mammals were tested using regression analysis. Results Regarding patch composition, the three Drava sections and the cumulation of their sampling localities were found by G-tests to be homogenous, with only one case of significant difference on the lower and middle section, and two patches with significant difference in the upper section. Patch overlap values also showed high degree of similarity in all three comparisons (lower section: 0.82; middle section: 0.74; upper section: 0.73). The statistical analysis of patch compositions of the three greater areas revealed higher patch overlap between the upper and middle sections (0.85), whereas these two areas were more different from the lower region in respect of their patch distributions, thus smaller overlap values were obtained (0.64 and 0.62). Based on G-tests the upper and Table 2: Homogeneity G-test of patch compositions in the various Drava reaches Patch type Pi (total reach) Pi (total reach) G-value s Upper reach Middle reach Mixed agricultural areas 9.49 Middle reach 2.23 Lower reach 4.84* Marshes Dry closed deciduous forest 0.18 46.87 4.36 19.63 4.77* 11.50*** Upper reach Lower reach Mixed agricultural areas Dry closed deciduous forest Forest plantation 9.49 39.08 3.58 2.66 19.63 11.51 4.06* 6.57* 4.38* Í.Ü5