Hurricanes are important drivers of periodic disturbances on tropical forests of the Caribbean region. The immediate impact of this kind of disturbance is on the redistribution of the biomass from canopy compartments to the detrital pool of the forest floor which is affected by microclimate change when the canopy is opened.  Earlier studies on naturally occurring hurricanes in the Luquillo Mountains of Puerto Rico showed that the interaction of those two primary effects propagate through the system in complex ways determining long-term behavior of ecosystem responses.  To assess the independent effect of the deposition of canopy debris on the forest floor, and the changes in microclimate on community structure and ecosystem processes we initiated during 2004-2005 a large scale manipulation experiment involving canopy trimming, at selected sites of the Luquillo mountains near the El Verde Field Station (18^o 20'N, 65^o 49'W). This experiment was also designed to simulate increased hurricane frequency by repeating the treatments every 6 to 10 years. The objective of the present analysis is to assess whether the design effectively captures short- and medium-term responses to the treatments, given variability in baseline measurements taken before treatments.

Results/Conclusions

A completely randomized block design with three replicates was used for establishment of treatments (each 30x30 m) and involved: 1) canopy trimmed, cut biomass distributed on forest floor, 2) canopy trimmed, cut biomass removed from the plot, 3) canopy not trimmed, canopy biomass from a trimmed plot distributed on the forest floor, and 4) canopy not trimmed, no canopy biomass added to forest floor. For the trimming, professional arborists cut all branches <10 cm diameter in the trimming plots but no trees were felled. We selected a series of measurements representing key biotic characteristics and ecosystems processes for continuous monitoring. We also monitored the changes in activity of various sets of decomposer organisms. The results suggest that in spite of the spatial variability associated with the geomorphological features of the landscape, the experimental design was effective in decoupling the effects of microclimate and detrital pulses for some short- and medium-term aspects of community and ecosystem processes influencing forest dynamics.  Many responses closely resembled those seen after natural hurricanes. However, for a few measured variables, mainly changes in faunal density, the treatments did not show a response, probably due to the scaling of the experimental design and the capability of organisms to move.