In lowland tropical forests, high biodiversity makes it difficult to isolate the influence of tree species and functional traits on local biogeochemistry in situ. We used high-fidelity remote sensing data from the Carnegie Airborne Observatory to identify ¼ ha forest plots with either relatively high or low canopy nitrogen (N) on the Osa Peninsula in Costa Rica (n =5). Plots were located within 1 km2 on a minimally-eroded surface with similar parent material. We hypothesized that tree communities would differ between high and low canopy N plots. We also hypothesized that litter N and soil inorganic N availability and cycling would be correlated with canopy N. We identified emergent trees, measured diameter at breast height, and compared communities using a Monte Carlo Permutations Test. We installed litter traps in each plot and collected and weighed litterfall every two weeks for 18 months. We measured % carbon (C) and N in a subset of these litter samples. We extracted soluble C from four composite litter samples, each representing three months of litterfall. We calculated litter decay rates from a reciprocal litterbag experiment. Finally, we measured KCl-extractable nitrate (NO3--N) and ammonium (NH4+-N) as well as net nitrification and N mineralization quarterly over one year.
Results/Conclusions
We found that the high and low canopy N plots host significantly different communities of emergent trees (P=0.036). Few putative N fixers were identified in any of the plots. Annual litterfall mass was similar among plots as were rates of litter decay (both P>0.005). However, litter C:N in the high canopy N plots was lower (32±0.89) than in the low canopy N plots (42±1.0; P<0.001). Additionally, the C:N of litter leachate from the high canopy N plots was roughly half that of the low canopy N plots (30±1.4 vs. 61±1.4; P<0.001). Over the course of a year, soil NO3--N was 10-22 times higher beneath high N canopies (mean: 2.19±0.50μg/g) than beneath low N canopies (mean: 0.13±0.05μg/g; P<0.001). Net nitrification was also higher under high canopy N plots (1.7±0.30 μg/g/d) relative to low (0.48±0.11 μg/g/d; P<0.001). The strong correlations between canopy N and litter and soil N support our second hypothesis. These associations suggest that emergent tree assemblages may drive local variability in soil N via differences in their litter chemistry. Our findings suggest that remote sensing of foliar characteristics may offer an effective way to study spatial patterns in soil biogeochemistry in diverse tropical forests.