QUESTIONING THE HGM PARADIGM: A GEOCHEMICAL INVESTIGATION OF WETLANDS IN THE CATSKILL/DELAWARE WATERSHEDS

We tested the hypothesis whether the Hydrogeomorphic Classification System, based on wetland water sources, also characterizes fundamental differences in wetland water chemistry in 22 wetlands in the Catskill-Delaware watersheds. Wetland sites were evenly classified as terrene and lotic wetlands.

We analyzed surface water from the wetland outlets for pH, specific conductance, and concentrations of dissolved oxygen (DO), Ca, Fe, Sr, Mg, K, Na, Si, Cl, DOC, NH4, SO4 and HCO3. We used univariate Shapiro-Wilk, Kolmogorov-Smirnov, and multivariate Mardia Skewness and Kurtosis tests to evaluate whether sample distributions were normal or log normal and used Hierarchical Cluster Analysis to determine if samples collected from lotic and terrene wetlands could be statistically grouped apart. Canonical Discriminant Analysis was then used to test whether the identified clusters were statistically different.

Mahalanobis distance and Wilks’ Lambda tests from these analyses show that different clusters are significantly separated, but they do not conform to the HGM classifications. Principal Component Analysis method further showed that >70% of all information in the data set can be explained by only 3 factors and 6 variables; Ca, Si, Sr, Mg, K, and SO4, many of which are related to water-rock interaction.

Fundamentally, there appears to be no statistical difference between lotic and terrene geochemistry. Piper plots independently agree that terrene and lotic sites do not occupy unique hydrochemical facies. The Hydromorphic Classification System may inappropriately characterize wetlands in the context of meaningful geochemical functions.