Vegetation structure and diversity on the CaribbeanCaribbean island of StSt.. John, U.S. Virgin Islands Structure de la vvégétationégétation et diversitdiversitéé sur l l’île’île CaribCaribéenneéenne de Saint John, dans les Iles Vierges US. Diversidad y estructura de la vegetacivegetaciónón en la isla Cariben ̃ã̃ de St. John, Islas Virgenes Virgenes,, USA. Sonja N. Oswalt1,, BrittaBritta P.P. DimickDimick2,, andand ThomasThomas J.J. BrandeisBrandeis3 1Forest Resource Analyst, USDA Forest Service, Knoxville, TN USA; 2Research Associate, University of Tennessee, Knoxville, TN USA; 3Research Forester, USDA Forest Service,Service, Knoxville,Knoxville, TNTN USAUSA

Distance (Objective Function) INTRODUCTION RESULTS 5.8E-03 2.8E+00 5.6E+00 8.5E+00 1.1E+01 The vegetation of the Caribbean island of St. John has been dramatically impacted by anthropogenic •We recorded 2415 individuals of 203 species Information Remaining (%) disturbance, beginning with Taino culture and extending to modern times. Following acquisition by the (63 families) on St. John forestland (21 percent 100 75 50 25 0 United States, approximately 65 percent of the island was donated to the National Park Service for of all vascular species described on the island) Vegetation Groups 1 2 3 4 conservation as the Virgin Islands National Park. Building on an island-wide forest inventory framework, we (Acevedo-Rodriquez, P. 1996). investigated the structure and diversity of the forest vegetation on St. John. •First-order jackknife estimates indicated the true species richness was 255.3 (± 12.15 SD).

•5% of recorded species were introduced

•79% of subplots contained at least 1 introduced species; 39% contained 2 or more. OBJECTIVES Measure the composition, relative abundance, and vertical position of all vascular plants on •Guapira fragrans was the most common forested plots vascular plant, occurring on 91% of subplots.

Quantify populations of native and introduced vegetation occurring on forestland •Cluster analysis yielded maximum results at 4 groups.

Identify environmental variables important in structuring the distribution of vascular plants on Group Scientific Name Indicator the island Number Value 1 Plumeria alba L. 75.7 Acacia muricata (L.) Willd. 63.9 METHODS Eugenia cordata (Sw.) DC. 63.8 •The major division occurred between Group 3 and Groups The study was conducted on the island of St. John, 2 Morisonia americana L. 61.3 Cordia alliodora (Ruiz & Pavon) Oken 46.4 1, 2, and 4 along a disturbance gradient, with Group 3 U.S. Virgin Islands, located at 13°22’N latitude and Chionanthus compactus Sw. 45.9 indicative of past and present severe disturbance and 64°40’W longitude in the Caribbean Sea, 3 Opuntia repens Bello 69.7 colonization by “weedy” and introduced species. approximately 30 km east of the island of Puerto Leucaena leucocephala (Lam.) de Wit 65.0 Rico. Schaefferia frutescens Jacq. 60.7 •Groups 1, 2, and 4 were divided along a moisture gradient, with species in Group 1 typical of dry forests, and species Data collection methods followed protocols 4 Sabinea florida (Vahl) DC. 61.2 in Groups 2 and 4 typical of moist forests. outlined by the USDA Forest Service Forest Ardisia obovata Desv. Ex Hamilton 60.7 Guettarda scabra (L.) Vent. 60.6 Inventory and Analysis program. Detailed Guettarda scabra (L.) Vent. 60.6 Vegetation Groups 1 descriptions of methods may be found at descriptions of methods may be found at •Group 1 = Rocky, low to mid-elevation dry scrub forest on 2 http://fia.fs.fed.us/. 3 Data source: Conservation Data Center, University of the Virgin Islands, 2001. soils with low productivity 4

•26 plots containing 4 subplots each were located •Group 2 = Low-productivity, high elevation moist basin

on an unbiased, systematic grid across the island. forests on steep slopes Distance to Roads Axis 3 Axis Bulk density •Only 100% forested subplots were sampled •Group 3 = Disturbed communities on a wide range of pH elevations with gentle or no slope, range of soil types Total Clay LEP SAR AWC Gypsum Elevation •Subplots were analyzed as independent samples to CEC satisfy statistical requirements, particularly species- •Group 4 = Mid-elevation moist forests across a range of area relationship assumptions, though some degree steep slopes on highly productive soils of spatial autocorrelation was expected. CONCLUSIONS Axis 2 •Vascular plant composition and structure, soil Introduced species, though widespread, do not dominate plant communities. Past and present characteristics, and environmental variables were collected on each subplot. site disturbance plays an important role in defining the species present and dominant in the communities on the island. Vascular plant communities are structured primarily on the basis •Clustering was performed using Sorenson distance of moisture, soil structure and productivity, elevation, slope, and aspect. measures with flexible beta set at 0.25. Nonmetric multidimensional scaling (NMS) and Multi- ACKNOWLEDGEMENTS response permutation procedure (MRPP) were used The authors gratefully acknowledge the USDA Forest Service Southern Research Station and the International Institute of Tropical Forestry for to relate vegetation to environmental axes. funding. We appreciate the efforts of Humfredo Marcano, Luis Ortiz, and Hana Blumenfeld during data collection, Gary Ray for his help in identifying voucher specimens, and the National Park Service for use of facilities.