Damage and responsiveness of Jamaican montane tree species after disturbance by a hurricane

Ecology, Dec, 1995 by P.J. Bellingham, E.V.J. Tanner, J.R. Healey

INTRODUCTION

In long-lived plant communities, many studies of the effects of disturbance on community stability have focussed on two aspects: resistance (the extent to which communities resist change by disturbance) and resilience (the extent to which communities recover to their original condition after disturbance) (Leps et al. 1982, Halpern 1988). Disturbance caused by severe storms and hurricanes usually results in differential resistance of species (e.g., Wadsworth and Englerth 1959, Whitmore 1974, Foster 1988a, Webb 1988, Walker 1991) and differential resilience of species (e.g., Merrens and Peart 1992). We used permanently labelled trees in the mountains of Jamaica to measure the damage and response of all the commoner tree species in order to discover whether or not damage and response were correlated with species characteristics. Such correlations would allow us to predict the effects of hurricanes. We also used the results to predict the likely changes in community composition and structure over the next few decades.

Previous studies of Hurricane Gilbert, which passed over Jamaica on 12 September 1988, showed that levels of mortality and damage caused to the montane rain forests during the hurricane were low compared with other forests affected by hurricanes (summarized in Brokaw and Walker 1991). These earlier studies (Bellingham 1991, Bellingham et al. 1992) examined patterns of mortality and damage caused by the hurricane and subsequent recovery (sprouting) at a community level; in this paper, we explore variation in damage and response within the community. For all tree species in the community, we assessed damage by measuring mortality, crown loss, defoliation, uprooting, and partial breakage. We looked for correlations between mortality and stem size, wood density, and leaf characteristics. Response to the hurricane was assessed by recruitment of each species to tree size class ([greater than or equal to]3 cm dbh), growth rates of each species, and sprouting. We used measures of damage and response to produce a two-dimensional depiction of species, which can then be loosely classified into resistant, susceptible, resilient, and usurper. This classification is probably applicable to a wide range of plant communities.

METHODS

Study site

The study area was the crest and the area 250 m south and north of the crest (1300-1920 m above sea level) along the 5 km of ridge between John Crow Peak and High Peak in the western Blue Mountains of Jamaica (18 [degrees] 05[minutes] N, 76 [degrees] 38[minutes] to 76 [degrees] 40[minutes] W). Here, upper montane rain forests vary in composition according to exposure to the prevailing northeast trade winds, to physiography, and to differences in soils (Shreve 1914, Grubb and Tanner 1976, Tanner 1977). The geology of the study area is highly complex, with at least eight rock types in highly faulted strata, and includes metamorphic blue schists, green schists, and amphibolites, marine shales and reef limestones, and igneous rocks. Soils of most of the study area are immature lithosols on the steep slopes of the range. On less steep areas of the slopes and on ridge tops, soils are more stable, less stony and more mature, yellow brown or pale brown loams (Grubb and Tanner 1976).

Jamaica is at the edge of the western Caribbean zone of frequent hurricanes (Alaka 1976). Between 1903 and 1988, nine hurricanes passed within 60 km of eastern Jamaica (including the Blue Mountains), five between 1903 and 1916, and one each in 1944, 1951, 1980, and 1988 (Neumann et al. 1978, Thompson 1983, Lawrence and Gross 1989). Hurricane Allen, in 1980, caused substantial damage to plantations of Pinus caribaea but relatively little damage to the montane rain forests in the Blue Mountains (Thompson 1983). Hurricane Gilbert was the strongest hurricane yet recorded in the Caribbean (Lawrence and Gross 1989); it caused substantial damage to forests throughout Jamaica, including the Blue Mountains (Bellingham et al. 1992).

We collected data from three sample areas totalling 1.10 ha:

1) The Tanner and Healey (T/H) plots recorded February-April 1989 (5-7 mo after the hurricane), four sites in close proximity on or near the Grand Ridge of the Blue Mountains, with contiguous permanent 10 x 10 m plots sampling 0.46 ha in total, (0.35 ha set up in 1974; Tanner 1977, Tanner et al. 1990), and three discrete 10 x 10 m plots nearby (control plots of Healey 1990, set up in 1986), for a combined total of 0.49 ha. The T/H plots were enumerated in January-February 1974 and May-August 1984 (i.e., a 10.4-yr prehurricane period), and February-May 1989 (i.e., a 4.6-yr hurricane period). A 0.35-ha subset of these plots was recorded after the hurricane in January-May 1991 (i.e., a 2-yr post-hurricane period).

2) The Mabess River transect (MRT), recorded May 1989, 26 nonbounded plots at 40-m intervals along a transect parallel to the Grand Ridge of the Blue Mountains in the Mabess River Valley, containing 20 individuals per plot (after the method of Hall 1991), and sampling a total area of 0.29 ha.