This dissertation examines the community ecology of trees and fungal endophytes in the lowland rainforests of New Guinea. Forest inventory and soil nutrient data from a large permanent plot in addition to a chronosequence of regenerating forest were used to quantify carbon stocks and forest dynamics. These data were coupled with intensive sampling of foliar fungal endophytes at three sites to investigate local and regional variation in community composition of endosymbionts associated with a diverse sample of rainforest trees. The Wanang 50 ha forest dynamics plot is located in mature lowland wet rainforest in the Middle Ramu area of the Madang province in Northern Papua New Guinea. Established in 2009, the Wanang plot is the first spatially explicit, large-scale, long-term forest plot in Oceania aimed at studying forest dynamics and biodiversity through time. Chapter 1 investigated carbon storage in the Wanang forest dynamics plot and spatial and demographic variation in biomass. During the first census, every stem >=1 cm in diameter at breast height (1.3 m above ground) was tagged, measured, and mapped to the nearest 10 cm. A subsequent species identification survey was carried out to associate every stem with a species or morphospecies concept. The first census at Wanang recorded a total 253,350 individual trees comprising 581 taxa, including 531 species and 50 morphospecies in 253 genera and 85 families. Our estimate of biomass averaged 222.3 Mg per hectare (95% CI: 211.3-232.7). This finding agrees closely with two previous estimates derived from a distributed network of small forest plots, suggesting this figure is a reasonable approximation of aboveground biomass in lowland forests in New Guinea. We found that there was significantly more carbon held in small trees than is typically assumed for lowland rainforest and that there is substantial fine-scale spatial variation in forest carbon (range: 161.5-324.43 Mg/ha). New Guinea forests have substantially lower biomass than tropical rainforests on average (global average: 373.7 Mg/ha), possibly driven by frequent natural disturbance. Chapter 2 explored the extent to which species composition and basal area in Wanang forest is associated with topography and soil nutrients. Additionally, the reputation that New Guinea rainforests are unusually dynamic was tested for the first time by comparing mortality rates and indicators of forest regeneration stage to other tropical rainforests. Soil nutrients and topography account for 29% of variation in tree species composition but only 4% of variation in basal area among 20 m by 20 m quadrats at Wanang. Basal area and gap phase character was significantly different than that observed at a comparable forest dynamics plot at Barro Colorado Island, Panama. Comparison of basal area in the mature forest of the Wanang plot to an adjacent regenerating forest chronosequence produced an estimate that 6% of the Wanang plot was naturally disturbed in the past decade. Disturbance at Wanang, in addition to tree falls, includes frequent small land slumps that result in catastrophic mortality of trees in portions of forest up to approximately one hectare. Mortality rate was also higher in Wanang than has been reported for any other lowland rainforest. These results lend quantitative support to the notion that New Guinea lowland rainforests are exceptionally dynamic. This dynamic character likely plays a role in the lower than average biomass and low density of large trees at Wanang described in chapter 1. A large forest plot, such as Wanang, offers the opportunity for biodiversity discovery and studies of species interactions otherwise not possible. Chapter 3 focused on the relative importance of host specificity and dispersal limitation in structuring community composition of fungal endophytes in rainforest trees. Fungi were isolated from a diverse sample of eleven tree species representing five genera in the lowland rainforests of Papua New Guinea. Collections resulted in the culturing and sequencing of 2,079 fungal endophytes from three sites. Sequences of the internal transcribed spacer (ITS) region, a common fungal ribosomal barcode, were clustered into molecular operational taxonomics units (MOTUs) at 95% similarity. A total of 670 endophytes were collected from a single site in 2010, comprising 61 MOTUs. A slightly modified collection method carried out at three sites in 2011 resulted in isolation of 1,409 endophytes belonging to 191 MOTUs. Composition of endophytes varied by host species and genus both within individual sites and in aggregate. Spatial proximity of host trees did not correlate with similarity of endophyte composition within species, genera, or trees (i.e. regardless of taxonomy). The endophyte community both within sites and regionally was comprised of few abundant host generalist and many rare taxa. Regional turnover of fungal endophyte taxa was low, illustrated by sites separated by hundreds of kilometers having high similarity in endophyte community composition. These findings are consistent with the hypothesis that host specificity plays a larger role than dispersal limitation in structuring fungal endophyte communities. The findings of these three chapters provide a novel perspective of forest ecology in New Guinea that capitalized on the strengths of the long-term, large-scale plot setup of the Wanang forest dynamics plot. Additional plot censuses, focused studies of seed dispersal, and more intensive sampling of foliar endophytes will further elucidate the patterns described in this dissertation.
University of Minnesota Ph.D. dissertation. August 2015. Major: Plant Biological Sciences. Advisor: George Weiblen. 1 computer file (PDF); xi, 253 pages.
Community ecology of New Guinea rainforest trees: Carbon storage, dynamics, and fungal endosymbionts.
Retrieved from the University of Minnesota Digital Conservancy,
Content distributed via the University of Minnesota's Digital Conservancy may be subject to additional license and use restrictions applied by the depositor.