Abstract:
Mortality rates of 205 tree and shrub species were estimated during two intervals, 1982-1985 and 1985-1990, in two size classes, 1-10 and -10 cm in diameter, in a 50-ha census plot in tropical moist forest on Barro Colorado Island in Panama. The severe dry season of 1983 was the focus of the study, since prior observations had demonstrated that it caused mortality in the forest. Here we document that forest-wide mortality was -3%/yr during the drought interval but only 2%/yr during the period afterwards, and that excess mortality during the first interval amounted to 2% of stems in the larger size class and 1% in the smaller. Overall, just under 70% of all species had higher mortality during the first census interval, but not all species were equally affected. Canopy trees had significantly higher mean mortality rates during 1982-1985 than during 1985-1990, but treelets and shrubs showed no or slight differences. This was counter to our prediction that species with short root systems would suffer more from a long drought. Shrubs did, however, have higher mortality rates than trees and treelets during both census intervals. We also evaluated mortality rates for subgroups of species that specialized on different microhabitats in the forest. As we predicted, colonist species (those associated with light gaps) had higher mortality rates than generalist species, 7-10%/yr compared to 2-4%/yr, but only in the smaller size class. Unexpectedly, colonizers had similar mortality rates as non-colonizers in the larger size class. Gap colonizers and generalist species were similarly affected by the drought-both had elevated mortality during 1982-1985. Species whose distributions were associated with moister soils (on the slopes around the island's plateau or in a swamp in the midst of the 50-ha plot) also had elevated mortality during the drought period, but no more so than generalist species. This was counter to our prediction that species from moist microhabitats would suffer more during an extended drought than generalists. Understory treelets that were slope specialists had higher mortality than generalists during both census intervals, but not large trees that were slope specialists. Our conclusions emphasize diversity as well as pattern. Every trend we illustrated had well-documented exceptions: large trees with lower mortality during the drought period, for example. Clearly, accurate predictions about how tropical forests will respond to climatic perturbations will require much detailed information from many species.
