Analysis of forest naturalness and tree mortality patterns in Estonia

New methods for evaluating structural properties of stands and individual tree mortality within forests are needed to enhance biodiversity assessment in forest inventories. One approach is to assess the degree of naturalness in a forest. We assessed forest naturalness by examining patterns and causes of mortality and deadwood amount and spatial distribution as indicators of naturalness, or degree of anthropogenic disturbance. This study is based on 5-year interval measurement using 294 permanent samples plots from a forest growth network in Estonia. The average annual mortality was 1.3% from stem number counting 29% of Scots pine, 27% of silver and downy birch and 20% of Norway spruce. Most common reasons for the individual tree death were growth-dependent reasons (45%), fungi (23%) and wind damage (16%). Modelling showed that relative diameter of a tree in a stand is significantly related to mortality probability. Modelling the reasons of tree death showed that with increasing relative diameter there was a greater probability that mortality was caused by wind or damage from game (mostly moose (Alces alces L.)), insect or fungi and a lower probability that mortality was due to competition between trees. Use of structural variable such as deadwood mingling, which was based on the neighbouring trees, improved the assessment of forest naturalness and helped to distinguish recent disturbances. A comparison of deadwood mingling and nature value scores in managed and semi-natural forests showed that dead trees were more dispersed and the naturalness score was higher in semi-natural forest stands. The nature score was significantly correlated with the diversity of mortality causes indicating that mortality causes are more diverse in semi-natural stands. Mean values and distribution of the deadwood mingling index in managed and semi-natural forests were not significantly different. In middle-aged semi-natural forests, mortality is spatially more random than in managed forests, thus there is no evidence of gap formation yet. (C) 2009 Elsevier B.V. All rights reserved.