Characterizing and mapping coastal vegetation on a sea island in the southeastern USA

Vegetation Classification and Survey 6: 279–299 doi: 10.3897/VCS.157787 International Association for Vegetation Science (IAVS) RESEARCH PAPER SPECIAL COLLECTION ISLANDS AND ARCHIPELAGOS Characterizing and mapping coastal vegetation on a sea island in the southeastern USA Cody Goodson1, John Holloway2, Zak H. Bartholomew1, Anne C. Axel1, Pamela Puppo1, Peyton Debowsky1, Kyle A. Palmquist1 1 Department of Biological Sciences, Marshall University, Huntington, West Virginia, USA 2 Environmental Division – Natural Resources Section, The Marine Corps Recruit Depot Parris Island, Parris Island, South Carolina, USA Corresponding author: Kyle A. Palmquist () Academic editor: Jorge Capelo Received 12 May 2025 ♦ Accepted 25 October 2025 ♦ Published 23 December 2025 Abstract Questions: Our objectives were to classify and map coastal vegetation types on a sea island that is undergoing rapid change from storm surge intensification and sea-level rise (SLR). Study area: The Marine Corps Recruit Depot Parris Island (MCRDPI) is a sea island in South Carolina, USA, dominated by tidal salt marshes with large, inland forested areas and developed areas supporting Marine Corps training activities. The climate is humid subtropical. Methods: We established 57 permanent vegetation plots within all the major coastal vegetation types present on the island, made detailed plant community measurements, and characterized the main environmental variables associated with each vegetation type. We employed cluster analysis and non-metric multidimensional scaling ordination to delineate these types based on vegetation plot data and then cross-walked those types onto existing associations within the United States National Vegetation Classification (USNVC). Finally, supervised deep learning classification was employed on 0.5 m high-resolution RBG imagery to map the distribution of vegetation types and other land cover types. Results: We identified four marsh associations and eight forest/woodland associations. The distributions of these vegetation types were primarily related to salinity and elevation. In addition to these 12 vegetation types, we mapped 11 additional land cover types for a total of 23 distinct classes. The marsh types dominated the landscape, covering ~57% of the classified area, while forest types covered ~13%. Accuracy assessment yielded an overall accuracy of 62% and a Kappa statistic of 0.6. Conclusions: We identified 11 existing USNVC vegetation types that are widely distributed in the southeastern USA and mapped their spatial distribution. The permanent vegetation plots and vegetation map provide fundamental baseline information that can be used to monitor the effects of climate change on coastal vegetation, evaluate wildlife habitat use by species of conservation concern, and facilitate decision-making in a region that has been identified as a global hotspot of salt marsh loss. Taxonomic reference: Weakley (2022). Syntaxonomic reference: USNVC (2025). Abbreviations: DBH = diameter at breast height; MCRDPI = Marine Corps Recruit Depot Parris Island; ROI = Regions of Interest; SLR = Sea-level rise; TEC = total exchange capacity; USNVC = United States National Vegetation Classification. Keywords Climate change, coastal vegetation, deep learning, sea-level rise, supervised classification, vegetation classification, vegetation mapping Copyright: This is an open access article distributed under the terms of the CC0 Public Domain Dedication. 280 Cody Goodson et al.: Characterizing and mapping coastal vegetation Introduction Coastal ecosystems are among the most functionally important habitats and are among the most sensitive to climate change. Upland forests and wetlands along the coast significantly buffer storm surges and the longer-term effects of erosion (Watson and Byrne 2009; Sousa et al. 2010; Feagin et al. 2010; Alizad et al. 2018; White et al. 2022). Coastal wetlands (i.e., salt marshes) have significant filtration capabilities that can assimilate pollutants and prevent eutrophication and sequester large amounts of carbon above- and belowground, thus acting as carbon sinks (Sousa et al. 2010; Humphreys et al. 2021; White et al. 2022). Coastal landscapes also support a diverse array of wildlife that rely on these habitats including migratory birds, reptiles, amphibians, rodents, and invertebrate species (Simas et al. 2001; Rosencranz et al. 2019; Marcot et al. 2020; Anderson et al. 2022; White et al. 2022). Native coastal vegetation is well adapted to the cyclic rise and fall of the tides. These plants trap sediment, mediating accretion of salt marsh substrates and potentially succession to mature, coastal forests. However, climate change is amplifying the intensity and frequency of disturbance, and rapidly altering conditions to which the dominant species are adapted (Feagin et al. 2010; Naumann et al. 2009; Kearney et al. 2019). Eustatic sea-level rise (SLR) resulting from rising ocean temperatures and melting land ice (Simas et al. 2001; He and Silliman 2019) is imposing significant stress on coastal vegetation (Warren and Niering 1993; Lucas and Carter 2013; Alizad et al. 2018; Cahoon et al. 2021) and the intensity of storm surge events associated with tropical storms are increasing (Donnelly et al. 2001; Webster et al. 2005; He and Silliman 2019; White et al. 2022). Coastal habitats can respond in very different ways to these disturbances. Some marshes experience submergence and conversion to open water (Reed 1995; Simas et al. 2001), while other areas accrete sediment and keep pace with SLR (Field et al. 2016; Kirwan and Gedan 2019; Rosencranz et al. 2019). If saltwater intrusion begins to compromise upland forest habitats, then marshes may migrate upland, driving habitat conversion (Murray et al. 2022; White et al. 2022). These differential responses are modulated by site-specific conditions, so a critical need is to characterize the composition and spatial distribution of coastal ecosystems in relation to environmental variables to improve our understanding of the impacts SLR may have on coastal habitats (Raposa et al. 2017). Accurate mapping of vegetation types is vital for tracking shifts in species distributions and to understand how plant communities are responding to changing conditions. Coastal vegetation is highly zonal, with high compositional turnover at small spatial scales. The most influential environmental variables affecting its distribution are salinity (Reed 1995; Watson and Byrne 2009; Cunha-Lignon et al. 2011; Du and Hesp 2020; Humphreys et al. 2021) and inundation (Warren and Niering 1993; Tiner 2013; Alizad et al. 2018). Coastal plant communities assemble along a salinity gradient which is influenced by hydroperiod, the amount of time a given area is inundated (Watson and Byrne 2009; Tiner 2013; Woods et al. 2020) and relatedly, elevation (Costa et al. 2003). Low marsh species are adapted to high frequency tidal dynamics and high salinity. (...truncated)