Ábrahám Levente (szerk.): Válogatott tanulmányok VI. - Natura Somogyiensis 19. (Kaposvár, 2010)

SALAMON-ALBERT É., HORVÁTH F., & ORTMANN-AJKAI A.: Climatic conditions and habitats in Belső-Somogy, Külső-Somogy and Zselic as vegetation-based landscape regions II. Temperature and precipitation sensitivity of woodlands

54 NATURA SOMOGYIENSIS Table 1: Basic statistics of the climate surface and the envelope in the landscape regions Climate SURFACE ranee std BIOCLIM1 Annual mean temperature 1.5 9.8 10.8 11.3 ±0.2 BlOCLIM-4 Temperature seasonality 49 738 764 787 ± 12 BIOCLIM-8 Mean temperature of wettest quarter 2.7 17.9 194 20.6 ±0.6 BIOCL1M-9 Mean temperature of driest quarter 1.8 1.1 2.1 2.9 ±0.3 BIOCLIM12 Annual precipitation 191 562 674 753 ±44 BIOCLIM17 Precipitation of driest quarter 34 100 121 134 ±8 BIOCLIM-1

8 Precipitation of warmest quarter 57 188 223 245 ±11 BIOCLIM-1

9 Precipitation of coldest quarter 34 109 130 143 ±8 Climate ENVELOPE ran ire min mean max std BIOCLIM-1 Annual mean temperature 1.5 9.8 10.7 11.3 ±0.2 BIOCLIM-4 Temperature seasonality 47 739 762 786 ± 12 BIOCLIM-8 Mean temperature of wettest quarter 2.7 17.9 19.4 20.6 ±0.7 BIOCLIM-9 Mean temperature of driest quarter 1.7 1.2 2.1 2.9 ±0.3 BIOCLIM12 Annual precipitation 180 570 681 750 ±42 BIOCLIM17 Precipitation of driest quarter 34 100 122 134 ±8 BIOCLIM18 Precipitation of warmest quarter 55 190 225 245 ± 1

1 BIOCLIM19 Precipitation of coldest quarter 33 110 131 143 ±8 envelope of woody habitats. In the first step, scatterplots were constructed from the relative distribution (%) on total area of the region as the regional climate surface, on total area covered by any semi-natural vegetation as the regional climate envelope and on woodland types as the habitat envelope according to bioclimatic variables. Data originated from the linked dataset of habitat occurrence and climatic variables, were sorted for the analyses representing all of the sampling points (MÉTA hexagons) in the three vegetation based landscape regions for the regional climate surface (n= 16300) and the envelope (n=9187). In second step, area version of Gaussian function as a nonlinear single or multipeak analysis was executed on each scatterplot, computing LevenbergMarquardt algorithm as an iterative procedure by Origin 6.1 (OriginLab Inc). Gaussian model describes a bell-shaped curve like a normal probability distribution function. It is characterized by the center of the peak that matematically represents the mean of designated variable and the width of the peak at half height, that is the standard deviation. Representative parameters of distributions were calculated for describing position of the peaks: average (center) and minimum-maximum x-value at half hight (width) of the Gaussian curve. Resulted these parameters, they were statistically compared by a oneway analysis of variance (ANOVA). Pairwise significant differences were counted if p<0.05. Climatic sensitivity of a habitat was interpreted as the distribution variability in number and position of the peaks, that were adjusted by ascendent order from lower to higher values of a given bioclimatic variable (e.g. ENV1 to ENV5, see in Table 2). Significant differences to each other and the regional climate envelope were presented by their representative x-value parameters among multipeaks within the envelopes. Peaks are the function of the climate parameter from habitat occurrence point of view. Overlapping peaks without any significant difference are interpreted as a climate gradient, peaks that have significant difference to the others are defined as a regional climate or habitat functional group. Based on data of above-mentioned climate GIS databases, distribution maps were prepared in order to connect regional spatial and habitat occurrence data, using ArcMap 9.2 ESRI (Figs 4 to 6). Number and limits of ranges in bioclimatic data displayed on maps, corresponded to those of Gaussian multipeaks. In the case of BIOCLIM-1 that has