Species composition and wood density variation: Responses to logging history in Kibale National Park, Uganda

Background/Question/Methods

Due to impacts of logging on forest canopy and soils, logged forests undergo changes in resource availability that lead to shifts in forest composition, structure and function. Depending on time since disturbance, logged forests are expected to be dominated by species that maximize resource acquisition over those that maximize resource conservation. Therefore, recovery of tropical forests with a logging history can be assessed using wood density, a key functional trait with tight associations to the growth strategy and successional status of a tree. To assess recovery of forests logged 38 years ago in Kibale National Park, Uganda, we compared light levels, species composition, and wood density for most individuals in five transects (200 m x 10 m) in logged forest to five transects in unlogged forest. Further, we explored temporal dynamics (1989, 1999, and 2006) in species composition and ecological similarity among 26 transects differing in logging history. We predicted that logged forests will be characterized by higher light levels, favoring a distinct species composition dominated by species that maximize resource acquisition and have lower average wood density. Furthermore, we predicted that the species composition of logged forest will become more similar to that of unlogged forest over time.

Results/Conclusions

Logged transects had higher mean light intensity, and greater range in wood density, but average wood density did not vary between logged and unlogged transects. Nonmetric Multidimensional Scaling showed that species composition differed between logged and unlogged transects, except for one unlogged transect that had species composition and mean light intensity more similar to logged transects. Variation in species composition between transects was related to differences in mean light intensity, with only a few species significantly associated to each forest type. Cluster analysis of ecological relationships among transects showed clustering that only in part corresponds to logging history. Additionally, few transects shifted from one cluster to another with a different logging history over time. Differences in mean light levels for logged and unlogged forest is consistent with the arrested succession typical of logged areas in Kibale forest. The lack of significant variation in average wood density between logged and unlogged transects may be related to the ecological similarity of some transects despite different logging histories. Rather than indicating forest recovery to pre-logging conditions, ecological similarity between transects differing in logging history may be due to similarities in landscape matrix, topography (e.g. slopes), and other recent disturbances such as elephant activity.