Linking biodiversity to ecosystem functioning in forest ecosystems: biotic drivers of wood decomposition
Final Report Abstract
Deadwood represents a globally important carbon pool. Quantifying wood decomposition and identifying its drivers is of central importance to understand carbon cycles and to predict how they might be affected by global change. Fungi and insects are the major agents of wood decomposition, but their relative contribution remains unknown. Moreover, it remains unclear whether decomposition rates increase with the number of decomposer species or functional groups or whether the presence of certain species, e.g. large-bodies species, is more important. Following a classic biodiversity-ecosystem functioning approach with mesocosms, we conducted an experiment manipulating beetle communities colonizing fresh spruce deadwood to create independent gradients of species number and functional diversity. Wood decomposition rates were measured after one and three years using both the traditional approach estimating dry mass loss and a novel approach using computer tomography (CT) scanning which was developed as part of this project. However, the processing of CT-imagery data created larger challenges than expected leading to significant delay of the work progress. While after one year, beetle galleries could be identified automatically based on grey-scale value and orientation with rather high accuracy, the high number of cracks in the wood due to shrinkage and fungal activity required development of a new approach. Here, machine learning reveal high potential being able to identify areas altered by fungal decay. Thus, the combination of CT scanning and machine learning has further potential than initially assumed. Decomposition measures based on dry mass loss and CT scanning were positively but only weakly correlated (R = 0.28). CT-based measures of decomposition generally correlated more strongly with beetle community metrics than mass loss data indicating that CT-based measures capture decomposition by beetles more accurately. Correlation coefficients between decomposition and beetle community metrics decreased from species richness to biomass to group richness. Variance partitioning revealed that the largest portion of variation in decomposition was explained by the combination of species richness and biomass. This suggests that biodiversity effects of beetles on wood decomposition are caused by a selection effect, i.e. at high levels of species richness, chance are higher that a large-bodied species of high functional importance, such as Monochamus sutor is present. Detection and measuring of single beetle galleries indicates unexpected interactions between bark and wood-feeding beetles and highlights 3D spatial analysis as another field of application of CT scanning in deadwood ecology. Fungal DNA was extracted from beetles used to stock mesocosms revealing high diversity of fungal OTUs carried and potentially vectored by beetles. As extraction and sequencing of fungal DNA from wood required more time than planned, the obtained OTU data and thus the effects of beetles on fungal communities inside wood could not be analyzed yet. Methodological challenges delayed the progress of the project considerably. The three main publications will be finished within the next months.