(map will go here pending publication)
Sites
All sites were accessed by canoe via channels and then walking along designated trails. Care was taken to not disturb soils or vegetation in the vicinity of study sites. Because of the marshÕs fragility and the need for systematic accessibility to study sites during the ~4-hour low tide period, neither a completely random nor square grid sampling scheme could be used. Since the area covered was far too large to construct boardwalks, 23 of the sites were placed along randomly located transects established in a 1991-1992 comprehensive vegetation mapping study (Hilgartner, 1995; Pasternack et al., 2000). One of these sites (5-P7) was lost due to distributary bank erosion during a storm. The other 7 sites were located in predetermined plant associations at randomly selected sites to generate more samples in those plant associations. Detailed procedures for measuring each variable and parameter described below are provided in Pasternack (1998), Pasternack and Brush (1998), and Pasternack et al. (2000).
Methods
Pasternack and Brush (1998) described an "anchored tile" method for monitoring sedimentation and erosion which is well suited for deltas experiencing significant spatial and temporal variability in transport processes. According to this approach, lightweight 1.22 m x 2.5 cm dia. (4' x 1") aluminum rods are sunk into the ground and capped with a detachable 20 cm x 20 cm (8"x8") ceramic tile flush with the marsh surface, as identified during low tide when the surface is exposed and there is no ponded water or fluidized mud layer present. Anchor rods are so firmly imbedded into the soil that they are difficult to move or adjust by hand once installed. As OPC vegetation and sediment is not subjected to ice ÒgrazingÓ or heaving, a force imposed by ice cannot lift the anchors. The detachment mechanism involves gluing a 5 cm long acrylic tube with a 2.5 cm inner diameter to the bottom of each tile. The ceramic tile/acrylic tube assembly caps the anchor rod and is not susceptible to motion unless subjected to extreme hydraulic lift forces.
The anchored tile at each of the 30 sites was visited once every 2 weeks during low tide and all accumulated materials on a tile were collected into pre-washed, pre-weighed glass jars. During the winter months, sampling was prevented because tiles were frozen into the marsh. Sedimentation rates were averaged over the last collection date in autumn and the first collection date after thawing. Tiles in areas of rapid accretion were raised by filling in their underlying acrylic tubes to maintain a position at the marsh surface. Surface samples adjacent to each tile were collected and analyzed for bulk density using the method of Pasternack and Brush (1998). While in the field, biweekly erosion was determined by measuring the height of each tile edge and the anchor rod above the marsh surface, averaging the measurements, and multiplying by bulk density. Local scour induced by tiles was observed to be negligible for all sites. Other potential sources of elevation change within the 1.22 m span of the anchor rod which might mimic erosion over long time scales, such as compaction due to respiration of organics or sediment consolidation, may be neglected due to their insignificance at the biweekly time scale. Respiration rates must be extremely low from November through March due to the cold and freezing climate. No sediment consolidation is possible when the marsh is frozen. Potential effects of compaction during the other seasons were considered on a site by site basis where erosion was significant, and those results are reported below. Sediment samples were returned to the laboratory and processed to obtain wet weight, dry weight, water content, organic content, and deposition rate. When both erosion and deposition were evident, the two methods were combined to obtain a net sedimentation rate. Biweekly net sedimentation rates (g cm-2 2weeks-1) were converted into annual rates (g cm-2 yr-1) by simple multiplication by a constant (365/14) to facilitate comparison with other studies where different methods and different sampling intervals are used. However, biweekly values are not necessarily representative of the average annual deposition at a site, so care should be used in interpreting individual data points. Organic content is reported as percent weight loss-on-ignition using the method of Pasternack and Brush (1998).
Results
The observed mean net sedimentation rate was 1.00 g cm-2 yr-1, with a range of Ð74.15 to 145.2 g cm-2 yr-1. No relations between delta sedimentation rate and total precipitation, peak precipitation intensity, or watershed discharge were found over time. Instead, three distinct temporal regimes in the data predominantly reflected seasonal patterns in vegetation life cycle. With regard to spatial patterns, nonparametric statistical tests demonstrated that each habitat had a unique cycle of sedimentation and erosion. When sedimentation rates were multiplied by habitat area, the floating leaf habitat was found to have sequestered 6,370 t yr-1. In contrast, the high marsh lost 624 t yr-1. These data indicate that the greater diversity of plant species in tidal freshwater marshes generates a wider variation in geomorphic processes than is possible for salt marshes.
