Analysis of landscape patterns and their relationship with oak (Quercus Humboldtii Bonpl.) regeneration in the municipality of Popayan, Cauca

Landscape patterns were determined for three different areas having oak populations in the Popayan municipality (Clarete, Rejoya and Pisoje). Two Landsat images from different years and polygons with areas equal or greater than 1.5 hectares were used for land use classification. Patch Analysis software was used in order to determine quantitative variables. Structure description included: number of patches, mean patch size, mean patch index, mean patch fractal dimension and mean perimeter-area ratio. Dispersion and fragmentation were evaluated with the three indexes: Mean Nearest Neighbor Distance, Mean Proximity Index and Interspersion Juxtaposition Index. Community variables included: basal area, terrain slope, light percentage and regeneration density, and were measured in an area of 3600 m2 for each landscape. Landscape and community information were analyzed using principal component analysis (PCA). The first two components explained 91.4% of data variability; they were determined mostly by landscape variables than community factors. Correlation analysis and the Kruskal-Wallis test showed that the variable of major importance regarding oak tree regeneration were the Neighbor Distance in secondary forest patches, the Mean Proximity Index in oak tree forest patches and the Juxtaposition Index in patches of planted forests.