Does Phragmites expansion alter the structure and function of marsh landscapes?: Patterns and processes revisited

We assess the probability and importance of different spatial distributions of Phragmites australis (Trin Ex Steud) within brackish tidal marshes of the mid-Atlantic United States coast. The comparative impact of Phragmites expansion on the larger coupled marsh-estuary system may partially be a function of the landscape area dominated by Phragmites, the landscape position occupied by Phragmites, the landscape pattern created by Phragmites expansions, and the resulting impact on tidal drainage networks. We find evidence that Phragmites establishment can occur at many landscape positions, and that Phragmites spread within a marsh can occur via colonization (new patches), linear clonal growth (along a preferred axis), or circular clonal growth (non-directional, random spread). Early intervals of Phragmites spread were dominated by colonization for all sites except for Piermont Marsh (which appeared to be dominated by linear clonal growth) and Lang Tract (which appeared to be dominated by circular clonal growth). Although 46-100% of new patches of Phragmites occurred within 5 m of drainages, at only one site (piermont Marsh, New York) did Phragmites populations remain concentrated along creek banks. Except for Iona Island, New York, which appears to be in an early stage of Phragmites invasion, patch dynamics at all sites showed an increase followed by a decrease in patch number, as independent patches became established, expanded, and coalesced. We also found some evidence for a loss of first order streams at later stages of Phragmites invasions in several sites (Hog Island, Lang Tract, Silver Run).