The Canopy Trimming Experiment (CTE), located in subtropical wet forest in eastern Puerto Rico, simulates the principal effects of a severe hurricane through manual disturbance of the forest canopy and forest floor.  Although these two aspects of a natural hurricane occur simultaneously, the CTE separates the relative effects of canopy disturbance and forest floor disturbance on vegetation, nutrient cycling, and trophic interactions. Hurricane damage was simulated in 30 x 30 m plots by trimming large (>10 cm diameter) branches from all trees.  A factorial design crossed canopy trimming and addition of detritus, each replicated three times, to separate the confounded effects of canopy opening and detrital inputs on forest functioning.  The detritus produced by trimming was cut into small (≤1 m length) portions and distributed evenly over the plots.  The experimental treatments began in the fall of 2004, and trees > 1 cm DBH were sampled in 2003, 2004, and 2007.  

Results/Conclusions Basal area did not change significantly from 2003 to 2007, but density increased significantly as a result of canopy trimming (P<0.001). Tree mortality was significantly greater in the untrimmed plots (P=0.05) and recruitment was substantially greater in the trimmed plots, particularly in the trim plus detritus addition plots (significant interaction, P<0.001). Analysis of individual species showed that two pioneer tree species (Psychotria berteriana and Cecropia schreberiana) accounted for the greatest increase in density in the trimmed plots, a response to the effects of canopy opening noted in previous hurricanes at our site.  Psychotria berteriana and the shrub Miconia prasina accounted for the most mortality in the untrimmed/control plots and appeared to reflect ongoing dynamics set in motion by Hurricane Georges in 1998. Species richness increased significantly from 2004-2007 in the trimmed plots (P<0.001). Likewise, in the untrimmed plots there was a decline in species richness from 2003-2007 a pattern, we note, that is consistent with the Intermediate Disturbance Hypothesis. Simpson’s Index (D), an index of community dominance, showed a small but significant rise (P=0.04) in the treatments with detritus addition. We have analyzed tree responses to determine if our method of simulated hurricane damage can produce a vegetation response similar to naturally occurring severe hurricanes, which it did.  These results give us confidence that the detailed analysis of biogeochemical and trophic responses will allow us to disentangle the effects of increased canopy openness and detrital inputs on these critical ecosystem functions.