|
publications > paper > woody debris in the mangrove forests of south florida > methods
METHODSStudy Sites.—Downed woody debris was sampled from 23 mangrove sites spanning from Rookery Bay National Estuarine Research Reserve in southwestern Florida, to the Taylor River and Joe Bay areas in southeastern Florida (Fig. 1). All locations were initially selected based upon proximity to the path of Hurricane Andrew, which impacted the study region in August of 1992. The reported maximum sustained windspeed from Hurricane Andrew was 232 km/h (Armentano et al. 1995), and the eyewall of the storm was over the Everglades landscape for at least 4 h (Platt et al. 2000), with peak wind gusts, probably associated with tornado activity, ranging from 253 to 333 km/h. (Wakimoto and Black 1994). The strongest sustained winds from a hurricane generally occur around the eyewall (Jordan et al. 1960), yet tornadoes or unpredictable turbulent eddies spawning in areas to the immediate right of a storm's path (cf., Novlan and Gray 1974) often register the greatest amount of damage (Shea and Gray 1973, Wakimoto and Black 1994). In order to understand potential differences in woody debris stores relative to the hurricane's path, we stratified our sampling locations into five regions. These regions included two that were of low hurricane impact, with RB-Right actually having lower sustained wind speeds than the other regions and TAY-Left escaping much of the wind damage by having a reduced forest stature. ENP-Left, ENP-Eye, and ENP-Right correspond to the immediate left (0-35 km), eyewall, and immediate right (0-35 km) of Hurricane Andrew's path, respectively (Table 1, Fig. 1).
All sites were interior forest locations associated with either a nearby river, creek, or embayment; overwash mangrove islands were not sampled. Forests were composed of three principal tree species: Rhizophora mangle L., Avicennia germinans L., and Laguncularia racemosa Gaertn. f. These species make up the majority of the surveyed downed woody debris. Trees of the mangrove associate, Conocarpus erectus L., were sparse and present only in TAY-Left and one site (SRU) along the Shark River. Rainfall in the region is between 1010 and 1650 mm/yr depending upon location, occurrence of periodic droughts, and aperiodic tropical storms (McPherson et al. 2000). Prior to Hurricane Andrew, no major storm impacted our survey sites since 1960 (Hurricane Donna) and 1965 (Hurricane Betsy) (Craighead & Gilbert 1962, Doyle et al. 1995). Likwise, no major storm has impacted survey sites since 1992. Woody Debris Measurements.—Downed wood was measured from May 2001 to October 2002 using a line-intercept technique originally described by Van Wagner (1968) and Brown (1974), and later applied to mangroves (Allen et al. 2000). On most sites, four 20-m-long transects were demarcated from two random azimuth offsets at 15 m from an established plot center. For Shark and Taylor River sites, between 6 and 12 nonoverlapping transects were established from fixed points 10 m apart along a systematic grid. Survey sites were separated into five regions as described above; between 16 and 31 transects were analyzed for each region. Coarse woody debris >7.5 cm in diameter, intersecting the line at any location, was measured to 0.1 cm with metal calipers and categorized as sound, intermediate, or rotten. Rotten and intermediate classes were indicative of trees and debris downed by Hurricane Andrew, with classes being assigned based upon penetration ease of the metal calipers similar to Robertson and Daniel (1989). Volume adjustments were made for downed, rotten logs containing hollow cores by measuring the width of the remaining outside shell and subtracting the inner decomposed core volume. Fine woody debris between 1 and 7.5 cm was tallied as numeric counts over the first 4 m of the transect and separated into two diameter categories (1-2.5 cm; 2.5-7.5 cm). Fine debris with a diameter <1 cm was subsampled as numeric counts over the first 2 m of each transect. Volume for fine woody debris from numeric count data and for coarse woody debris >7.5 cm in diameter was calculated as follows (Van Wagner 1968, Allen et al. 2000):
with v representing woody debris volume (m3/ha), di as the diameter of an individual piece of woody debris (m), L as the sample line length (m), and k as the per hectare conversion constant (10,000 m2/ha). Approximate conversions from woody debris volume to mass were made to compare woody debris estimates among forests. We used density data for mangroves from Robertson and Daniel (1989), and from Polit and Brown (1996) we used relative scaling relationships of 0.5 t/m3 for woody debris <7.5 cm and of 0.5, 0.35, and 0.2 t/m3 for large woody debris classified as sound, intermediate, and rotten, respectively. Components of woody debris could not be identified conclusively to species and prevented our use of species-specific densities. Average canopy height was measured on at least three codominant individuals per species on each site using a laser height device (Impulse 200, Laser Technology, Inc., Englewood, CO), and overstory canopy density was determined from four readings per plot at cardinal directions with a spherical densiometer (Model A, Forest Densiometers, Bartlesville, OK). Statistical Analysis.—The analysis was conducted as a split-plot design, since volume of downed wood within a sample region was not independent. An analysis of variance (ANOVA) with a nested error structure (i.e., site within region) was used to determine if differences existed in combined downed woody debris volume (coarse + fine debris) among regions and for coarse woody debris volume and fine woody debris volume, separately, among regions. All data were log-transformed (+0.5) to improve normality and homogeneity of residual variances. Statistical groupings were determined with a Tukey's Studentized range test ( |
U.S. Department of the Interior, U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/papers/woody_debris/methods.html
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 22 September, 2005 @ 10:58 AM(TJE)
= 0.05). Relationships between volume of downed wood and canopy height for a given range of predicted windspeeds were determined through linear regression. Windspeed estimates were derived for each site based on a hurricane simulation model (HURASIM) designed to simulate wind trajectory, magnitude, and circulation of Hurricane Andrew (Doyle 1998).