METER SCALE VARIABILITY OF SOIL PROPERTIES IN A LAKE ONTARIO COASTAL WETLAND

Little work has been published on natural variability of wetland soils. This data becomes important in evaluating changes in these soils over time. Sampling at multiple scales in the Ellison Park wetland was conducted to evaluate spatial differences. The wetland is a delta-type coastal wetland covering over 423 acres at the southern end of Irondequoit Bay in an urban area adjacent to Rochester, NY. The wetland extends for about 3 km with a main channel and secondary channels through the wetland.

Soil samples were collected in two paired locations, along five 60 m transects and at two additional locations within the wetland. The paired locations represent levee and depressed areas, and were randomly sampled within a 2m by 2m square. The 60 m long transects proceed from a levee at the stream bank through the adjacent depressed area. The two remaining sites are an elevated area adjacent to a levee, and a depressed area bordering steep slopes bounding the wetland. Soil samples were analyzed for particle size distribution by laser diffraction and for soil chemical properties including major and trace metal concentrations by ICP, organic matter content, pH, and total nitrogen by combustion.

Variability within individual sampling sites was determined by descriptive statistics. Levee samples typically have a smaller range of values and lower standard deviations than depressed area samples for most parameters. Organic matter ranges from 1.7% to 6.9% for all levee samples while depressed area samples ranged from 0.7% to 13.4%. Levee samples tend to be symmetrically distributed while depressed area samples are slightly skewed.

Levee samples from different areas of the wetland are not comparable. While the descriptive statistics are similar, showing a narrow range of values, the mean values for most parameters vary with location. Box plots show the IQR for levee samples do not always overlap, however, whiskers do overlap for all levee sites. Depressed area samples show an IQR that overlaps for most parameters, which have a wider range of values.

In the middle wetland area, 60 m transect and 2 m by 2m square sampling sites were collocated. Comparing these observations shows similar results for most parameters. A statistical analysis using a t-test supports the hypothesis that the two data sets are not significantly different for each parameter.