Caribbean Journal of Science, Vol. 42, No. 1, 53-66, 2006 Copyright 2006 College of Arts and Sciences University of Puerto Rico, Mayagu¨ez Phytosociology of Vascular Plants on an International Biosphere Reserve: Virgin Islands National Park, St. John, US Virgin Islands Sonja N. Oswalt1,*, Thomas J. Brandeis1, and Britta P. Dimick2 1U.S. Department of Agriculture Forest Service, Southern Research Station, Forest Inventory and Analysis, 4700 Old Kingston Pike, Knoxville, TN 37919, U.S.A 2University of Tennessee Department of Forestry, Wildlife & Fisheries, 274 Ellington Plant Sciences, Knoxville, TN 37916, U.S.A. *Corresponding author: [email protected] ABSTRACT.—We investigated the relationships of vegetation communities to environmental variables and compared the relative contribution of native and introduced species in extant forest communities on St. John, US Virgin Islands, using an island-wide forest vegetation inventory and monitoring network of permanent plots. We detected 2,415 individuals of 203 species, 5 percent of which were introduced. Cluster analysis, Indicator Species Analysis, and Nonmetric Multidimensional Scaling (NMS) ordination detected four broad species communities divided primarily by moisture and disturbance gradients. Group 1 was characterized by rocky, low-to-mid elevation dry scrub forest on soils with fairly low soil nutrient content; Group 2 was distinguished by low soil nutrient content, high-elevation moist basin forests on steep slopes; Group 3 was indicative of disturbed communities on a wide range of elevations with gentle or no slope across a range of soil types; and Group 4 represented mid-elevation moist forests across a range of steep slopes on nutrient- rich soils. Though introduced species are present and widespread on the island, they do not appear to be dominating most plant communities. Exceptions may be those communities with long-standing histories of human disturbance. Achieving an adequate sample of forest types of limited extent or linear spatial patterns such as mangroves and gallery moist forests is difficult with a systematic design. Future sampling should consider some form of stratification targeting these under-sampled forest types. KEYWORDS.—Caribbean forest vegetation, disturbance, introduced species, inventory and monitoring, phy- tosociology, Virgin Islands National Park INTRODUCTION lent itself to sugarcane, cotton, tobacco, and coffee cultivation (Woodbury and Weaver Extant Caribbean forests have been 1997). As a result, the vegetation has been shaped by natural disturbances, specifi- dramatically altered through time, and cally hurricanes, and a more recent history many of the current island landscapes are of human exploitation (Lugo et al. 2000). dominated by secondary forest (Weaver Land use history on St. John is similar to and Chinea-Rivera 1987; Ray and Brown that of other islands in the Lesser Antilles. 1995; Rogers and Reilly 1998). Some An- Forests on St. John were first utilized by tillean forest types, such as the once- pre-Columbian native Taino Indians for common subtropical dry forest, have been food, shelter, and raw materials for boat impacted to the point that they are now construction. Danish settlers arrived in the considered endangered ecosystems (Mur- early 1700’s and cleared large areas of for- est for agriculture, farming, and building phy and Lugo 1986; Ray et al. 1998; Linares- construction (Ray and Brown 1995; Wood- Palomino and Ponce Alvarez 2005). On bury and Weaver 1997). A favorable cli- many of the islands, forest cover consists of mate contributing to long growing seasons fragments scattered across the landscape and much of what remains are secondary forests containing various introduced spe- ms. received November 14, 2005; accepted Decem- cies whose long-term impacts on species ber 22, 2005 composition and ecosystem function are 53 54 SONJA N. OSWALT ET AL. yet to be understood (Chinea and Helmer ent Caribbean forest landscape, surpris- 2003; Grau et al. 2003; Lugo and Helmer ingly few studies have documented pat- 2004). terns and composition of island vegetation, Though St. John shares a history of dis- particularly in the U.S. Virgin Islands. turbance-dominated landscapes common Studies that have explored species- in the Lesser Antilles, it has avoided the environment relationships and forest pro- subsequent urban development that has oc- ductivity in the Caribbean have focused curred across the Caribbean basin. In 1956 primarily on woody species (Weaver and approximately 65 percent of St. John was Chinea-Rivera 1987; Ray et al. 1995; Ray et donated to the U.S. National Park Service al. 1998; Rivera et al. 1998; Lugo et al. 2004; for the creation of the Virgin Islands Na- Linares-Palomino et al. 2005). Aside from a tional Park (VINP) (National Park Service limited number of field guides, taxonomic 2005). In recognition of its uniqueness, in floral descriptions, and broad scale com- 1976 the park was designated an interna- munity maps, the result has been a general tional biosphere reserve by the United Na- paucity of information regarding non- tions Educational, Scientific and Cultural woody vascular plant species, their rela- (UNESCO) Man and Biosphere (MAB) pro- tionships to woody species, and plant com- gram. With the addition of offshore areas in munity-environment relationships in this 1962 and the establishment of the Virgin region. Current VINP management objec- Islands Coral Reef National Monument in tives that could benefit from a comprehen- 2001, the VINP mission “to protect, man- sive phytosociological study and vegeta- age, interpret and preserve the park’s tion monitoring include: reduction of soil unique natural and scenic resources and erosion, landslides, and freshwater runoff values unimpaired for the education, en- that damage near-shore coral reefs and re- joyment and inspiration of present and fu- duce their value as an ecotourism resource ture generations” extends from the island’s (National Park Service 2005); elimination or highest point, Bordeaux Mountain at 389 reduction of feral white-tail deer (Odocoi- m, into the surrounding Caribbean tropical leus virginianus), donkey (Equus asinus), do- marine ecosystems (National Park Service mestic goat (Capra hircus), wild hog (Sus 2005). The result has been the protection scrofa), and domestic sheep (Ovis aries) from development of one of the larger con- populations that threaten forest regenera- tiguous forest areas in the Lesser Antilles, tion and soils (National Park Service 2004); and a unique opportunity for the study of an Avian Conservation Implementation island flora and fauna. Plan, directed toward maintaining and in- The species composition and vegetation creasing the amount of over wintering structure in St. John’s forests, considered a habitat for Neotropical migrants (Watson global biodiversity “hotspot” (Myers et al. 2003); and monitoring of land use change 2000, Helmer et al. 2002), are of substantial from construction and over development importance to understand current biodiver- (U.S. Department of Transportation 2004). sity in the Caribbean basin, and to evaluate The U.S. Department of Agriculture availability of habitat for multitudes of (USDA) Forest Service, Forest Inventory wildlife species. In addition to contributing and Analysis (FIA) program conducts sys- to global wildlife and vegetation biodiver- tematic surveys of forestland across the en- sity, the Greater and Lesser Antilles are an tire United States and its territories. In 2004, important archipelago for wintering neo- FIA initiated a special, intensified study on tropical migratory songbirds (Askins et al. St. John to investigate the structure and di- 1992; Wunderle and Waide 1993; Rappole versity of forest vegetation, and additional 1995). Studies have suggested that defores- forest attributes. This intensified inventory tation has led to losses of wintering habitat was not limited to within the VINP bound- for many warbler species in the Caribbean, aries, but encompassed the entire island be- and to population declines in these species cause conditions and management activi- (Rappole 1995). ties in the non-park portions of the island Despite the need to understand the pres- directly and indirectly impact the unique FOREST VEGETATION OF ST. JOHN 55 Caribbean ecosystems the park is meant to sign (USDA Forest Service 2003). In an ef- preserve. Specifically, our objectives were fort to minimize impacts to forest vegeta- to: 1) install a permanent forest vegetation tion, Forest Service FIA protocol does not inventory and monitoring network across allow for repeated visits to sample sites the island of St. John, 2) document the com- within a given year. To minimize any po- position, relative abundance, and vertical tential effects of seasonality we sampled at position of all vascular plants on forested the mid-point of the rainy season—the plots, 3) quantify and compare populations point at which any seasonal growth flushes of native and introduced vegetation occur- would have already occurred but drought ring on forestland, and 4) identify environ- would not have resulted in senescence. mental variables influencing the distribu- Twenty-six plots were arranged on an un- tion of vascular plants across the island. biased, systematic sample grid across St John. The grid on St. John was composed of hexagons covering approximately 200 ha MATERIALS AND METHODS each, with one sample plot located within each hexagon. Plots were located and Study site mapped using differentially corrected