[1] The purpose of this study is to determine if concepts in terrestrial channel network analysis provide insight on intertidal creek network development and to present new metrics for their analysis. We delineated creek network geometry using high-resolution digital images of intertidal marsh near Georgetown, South Carolina. Analyses reveal that intertidal creek networks may be topologically random. Length-area relationships suggest that salt marsh and terrestrial networks have similar scaling properties, although the marsh networks are more elongate than terrestrial networks. To account for recurrent water exchange between creek basins at high tide, we propose that the landscape unit of geomorphic analyses should be the salt marsh island as opposed to salt marsh creek drainage basin area. Using this approach, the relationship between maximum channel length per island and island area is well described by a power function. A similar power relationship exists for cumulative channel length versus island area, giving a nearly unit slope; this implies that marsh islands have a spatially uniform drainage density. Since the island boundaries are easily identified in remote sensing, taking the island as the unit of geomorphic analysis will eliminate discrepancies in delineating basin boundaries and preclude the need for defining basin area in intertidal landscapes. Analyses and results presented here may be used to quantify salt marsh reference condition and provide indicator variables to assess salt marsh disturbance.