Toward sustainable forest management indicators? A data mining approach to evaluate the impact of silvicultural practices on stand structure

ndicators are increasingly required to support a fine-tuning between sustainable forestry andmultiple environmental targets. A data mining strategy was implemented in this study toassess the overall impact of traditional and innovative silviculture on stand structure in asample of beech forests with varying dominant age, management history and stand structurein Italy. Harvesting intensity and stand sensitivity to treatment were investigated using aprincipal component analysis (PCA) run on a set of dendrometric and stand-structure vari-ables measured before and after practice implementation at the scale of forest compartment.The PCA decomposed the overall impact of silviculture on forest structure in two manipula-tive effects: (i) structural changes between control and treatments, and (ii) the net manip-ulative effect of innovative versus traditional treatment. Our approach informs the sustainablemanagement of forests, outlining between-site differences in stand structure and identifyinga diversity gradient shaped by silvicultural practices. Multivariate analysis of forest indicatorsfollowing practice’s implementation is a promising tool to design innovative silviculturecoherent with conservation of forests’structural diversity.