Crab-driven vegetation losses
Since 2006, Dr. Mark Bertness, Christine Holdredge, and other Brown University students have been researching the hypothesis that overgrazing by a native, herbivorous crab, Sesarma reticulatum (purple marsh crab), is responsible for dieback in the low marsh. The research team used crab-exclusion cages to show how plants that are protected from herbivory grow up tall and healthy, whereas those that are not proected are frequently grazed down to stubble and eventually die.
Crab density studies revealed that Sesarma populations are very high and that there is very little predation pressure in Cape marshes. The latter is what is what has presumably allowed them to proliferate in this area.
Perhaps the most convincing evidence that Sesarma is the primary cause of dieback comes from the spatial distribution of dieback. In marshes where there are no Sesarma crabs, or they are so rare as to be undetected, there have been no such vegetation losses whatsoever. So far, we have not found a single exception to this. In contrast, where Sesarma are plentiful, dieback is extensive.
It has been suggested that perhaps Searama colonizes dieback areas after the plants were already gone. However, there is no reasonable model for this. Sesarma do not benefit from decreased vegetation cover since the plants themselves are their main source of food, hide them from predators, and provide support for their elaborate burrows. Population explosions in nature typically result from a reduction in predation or a significant increase in resources (food). The former has been tested experimentally now and found to be true. The latter is in direct contrast to what is actually occurring on the ground (i.e., food resources are decreasing).
Finally, there is no good explanation other than continued herbivory for the lack of recovery in most dieback areas. From Bertness's caging experiments, we know that plants can otherwise thrive in these areas. Thus, the re-estabishment of plant cover through rhizomatous growth from the surrounding vegetation should have been rapid if dieback were due to some episodic event. However, this is clearly not happening - the reason being that continued herbivory has either maintained or significantly expanded the extent of denuded marsh.