Analyzing selective harvest events in three large forest observational studies in North Eastern China

• The study is based on three 10-ha observational studies with mapped trees in multi-species forests.
• New approaches for assessing the effects of a selective harvest are explored.
• Species are grouped into cohorts to evaluate the change of forest composition.
• New spatially explicit harvest selection preferences within tree neighborhoods are employed.
• The modification of tree size distributions and spatial patterns is assessed.

This study presents an analysis of selective harvest events in three 10-ha field plots with mapped trees, representing different forest successional stages in the temperate multi-species forests of North-Eastern China. Non-spatial methods of analyzing the harvest events include species and size selection preference and an assessment of harvest intensities for four species groups (identified using bivariate mixed dbh/height distributions). Spatial aggregation increased very slightly after the harvest event, but in most cases, the change was hardly noticeable. In addition, tree selection preferences involving nearest neighbor structure units are presented, using the attributes “species mingling” and “dominance”. The removals occurred within a broad array of neighborhood constellations, involving suppressed as well as dominant individuals. Previous approaches involving harvest event analysis in multi-species forests were limited to assessing size and species selection preferences. This study uses more advanced methods and presents more detailed interpretations, due to the large and detailed observational datasets and improved analytical tools that have become available recently. Models of tree growth and survival, which represent the overwhelming result of traditional observational studies, only describe a part of forest dynamics. Equally important are the modifications caused by regular human disturbance. There is thus increasing motivation for analyzing selective harvesting activities as presented in this contribution.