Environmental determinants of Phragmites australis invasion in a New Jersey salt marsh: Interactions among human activities, disturbance, and edaphic conditions

Causal connections among human activities, disturbance, and biological invasions are elusive in many cases, in part because these entities are often described on too high a level of aggregation to make testable causal hypotheses. I examined human causation of Phragmites australis invasion in a New Jersey salt marsh by disaggregating the invasion into historically verifiable units, human activities into case specific combinations of events, and disturbance into property-specific changes. Historical evidence suggested that invasion could be disaggregated into establishment and subsequent postestablishment spread. Establishment was associated with episodes of mosquito ditching that buried large rhizomes in well-drained areas during periods of lower salinity. Field experiments suggested that large rhizome burial was needed to avoid propagule desiccation, and rhizomes needed to he hurled near well-drained features. Greenhouse experiments suggested that establishment depended on the fluctuating hydrology created by ditches rather than changes to sulfides or salinity. These experiments also suggested that while lower salinity improved survival, short periods of lower salinity did not consistently improve performance. These data suggest that ditching caused invasion at the site because it buried large rhizomes in well-drained areas.; Post-establishment spread was not confined to ditched areas, and occurred in anoxic areas with high salinity and sulfides. Experimental limitations of translocation and venturi-induced ventilation suggested that initial spread into these areas was supported by ditch-side stands. Once established, Phragmites oxygenates the rhizosphere sufficiently to relieve anoxic stress and lower sulfide concentrations. This improves nitrogen uptake and leads to further expansion across the marsh surface. Thus, postestablishment spread is self supporting, requiring no human intervention.; These experiments support the assertion that site-scale Phragmites invasion needs to be disaggregated into its components to find evidence of human causation. Comparisons with other sites and other episodes of ditching also suggest that the activity "ditching" causes invasion only when it buries large rhizomes in well-drained areas. Therefore, whether ditching in fact causes an invasion depends on how it was practiced in a particular case.