engleski

Grouping of pedunculate oak provenances on the basis of biomass by applying cluster analysis

During the pedunculate oak provenance trial the tree biomass was analysed on the basis of 5 variables (stem weight, live branches, dry branches, leaves). Through cluster analysis (hierarchy and k-means) pedunculate oak provenances were grouped and the differences in monitored features between resulting clusters were analysed using the univariate and multivariate variance analysis. Results show that the inclusion of individual provenances into clusters corresponds in both the hierarchy and non-hierarchy procedure (in total 4). Clusters don’ t follow geographic links, therefore we assume genetic connectedness. Cluster 1 (Vrbanja) shows the highest values for all analysed features except the dry branches, which indicates that the process of branch dying and stem cleaning in this provenance has not yet started. Cluster 3 (Motovun, Dubica, Skakavac, Guševac, Morović, V.Gorica) shows the lowest values for all analysed features except the dry branches. Clusters 2 (Orlovac, Spačva, Zdenački Gaj, Ključevi, Novska, Gunja, Lipovljani) and 4 (Okučani and Đurđenovac) show very similar results for every feature except the dry branches where cluster 4 has statistically a significantly higher value than cluster 2 although they are geographically relatively close.

pedunculate oak provenances; biomass; cluster analysis

nije evidentirano