COS 21-3
Changing interactions between soil properties, tree community composition, and soil fungal communities across succession in secondary tropical forests

Tuesday, August 11, 2015: 8:40 AM
319, Baltimore Convention Center
Andrew W Quebbeman, Ecology, Evolution, and Environmental Biology, Columbia University, New York, NY
Krista L. McGuire, Biology, Barnard College, Columbia University, New York, NY
Benedicte Bachelot, Department of Biology, Duke University, Durham, NC
Jess K. Zimmerman, Department of Environmental Science, University of Puerto Rico - Rio Piedras, San Juan, PR
Maria Uriarte, Ecology, Evolution, and Environmental Biology, Columbia University, New York, NY

The majority of tropical forests now exist as mosaics of old-growth and regenerating fragments as a result of human land-use. Understanding the distribution of soil fungal communities in secondary forests and their interactions with tree communities and nutrient cycling will provide important insight into forest recovery and biogeochemical cycles. We tested the hypothesis that fungal communities are correlated with tree community composition and soil conditions and that this relationship strengthens with forest stand age. To do so, we sampled soils in four forest stands of El Yunque National Forest in Puerto Rico which included an old-growth forest, two regenerating forests aged between 60-75 years, and one 35-60-year old forest. Soil cores were collected from 20 quadrats at each site and analyzed for pH, micro- and macro-nutrients, and fungal community composition using DNA sequencing on the Illumina platform. The tree community at each stand was obtained from a tree census conducted in 2013 where all trees ≥1cm dbh were identified, measured, and mapped. We conducted analyses at two scales: between forest stands and across the 20 quadrats at each site.


Both fungal community and tree community composition differed significantly between sites (ANOSIM; p<0.001). Using redundancy analyses across quadrats, we tested for correlation between fungal community composition and the parameters: soil Fe, P, K, C:N, pH, tree community richness, and total tree basal area. We found that fungal communities significantly associated with Fe and P at the youngest stand, Fe and pH at intermediate stand one, K at intermediate stand two, and soil C:N and pH at the oldest stand. Combined, soil Fe, K, P, C:N, and pH explained 33%, 32%, 44%, and 49% of variation in fungal community composition at the four plots, respectively. Tree community richness and tree basal area accounted for < 2% of variation in fungal communities between quadrats across sites. The increase in variation of soil fungal communities attributed to soil factors with stand age is consistent with the idea that historic land-use homogenizes soil conditions, thereby weakening the relationship between soil fungal composition and soil nutrients. The negligible variation in fungal community composition attributed to tree community properties suggests that other abiotic factors or land-use history parameters may be dominant drivers of microbial community composition in these stands.