Contrasting responses of habitat conditions and insect biodiversity to pest- or climate-induced dieback in coniferous mountain forests

International audienceNatural disturbances are major drivers of forest dynamics. However, in the current context of anthropogenic global warming, shifts in disturbance regimes are expected. Natural disturbances usually leave biological or structural legacies which are important for early-successional species. Nevertheless, these legacies are usually eliminated by forest managers through salvage logging. Here, we investigated the consequences of forest dieback and the following salvage logging on both forest habitat conditions and saproxylic beetle communities. We conducted our study in two types of conifer-dominated highland forests: Pyrenean silver fir (Abies alba) which has suffered drought-induced dieback and Bavarian Norway spruce (Picea abies) which has suffered bark beetle-induced (Ips typographus) dieback. In both of the forest contexts, dieback provided a biological legacy through an increase in deadwood resources; however, this increase was much greater in the spruce forests. Nonetheless, despite this increase in resources, neither type of forest gained in total abundance or species richness after disturbance, compared to healthy stands. Nevertheless, the species composition of saproxylic beetle composition was significantly affected by dieback in spruce stands, but not in the silver fir forests. In the spruce plots, saproxylic beetles responded positively to the large increase in deadwood in the declining stands, including a very strong positive response from red listed species. Saproxylic beetle assemblages in spruce forests were mainly drove by canopy openness and deadwood amount. In the silver fir plots, we did not observed responses from the saproxylic beetle communities to deadwood amount increase. This lack of response may be explained by the relatively low amount of deadwood generated by the drought-induced dieback. Concerning salvage logging, it caused stronger contrasts in spruce forests than in silver fir forests, where it generally had no significant impact. For example, in spruce forests, salvage logging reduced the density of large snags by 91% and large logs by 87% compared with unharvested declining plots. Most of the significant environmental effects on biodiversity associated with dieback were no longer significant after accounting for the salvaged plots in our study data. Then, forest dieback and salvage logging induced much sharper and stronger effects on environmental and community metrics in the spruce than in the silver fir forests. The contrast between Bavaria and the French Pyrenees seems partly related to dieback severity. Finally, we invite forest managers to conserve biological and structural legacies through patches of deadwood-rich areas