Geomorphology may be an important predictor of vegetation pattern in system
s where susceptibility to disturbance is unevenly distributed across the la
ndscape. Salt marsh communities exhibit spatial pattern in vegetation at a
variety of spatial scales. In coastal Georgia, the low marsh is a virtual m
onoculture of Spartina alterniflora interspersed with patches of species th
at are more typical of the high marsh. These localized disturbances are mos
t likely created by wrack mats, mats of dead vegetation which can compact a
nd smother underlying vegetation creating bare patches for colonization by
high marsh species. We investigated the spatial pattern of disturbed patche
s along a 2 km section of Dean Creek, a tidal creek at the southwestern end
of Sapelo Island, Georgia, U.S. We used a discriminant model to explore th
e relationship between tidal creek morphology (e.g., the presence of draina
ge channels and creek bends) and the spatial distribution of disturbed patc
hes. The model predicted vegetation pattern along the creek with relatively
high accuracy (> 70%). Areas where water movement is slowed or multidirect
ional (e.g., along creek bends and near drainage channels) were most suscep
tible to disturbance. Our findings suggest an important functional linkage
between geomorphology and vegetation pattern in salt marsh communities.