Classification of fens near southern James Bay, Ontario, using vegetational physiognomy

Abstract

On the basis of vegetational physiognomy, 47 sites within 1 to 43 km of the southern James Bay coast were classified in the field into four fen types: graminoid, low shrub, graminoid-rich treed, and sphagnum-rich treed. The four types are directly related to differences in vegetational cover and in soil and water parameters, specifically depth to water level, peat thickness, selected groundwater nutrients, and distance from the coast. Detrended correspondence analysis was used to ordinate the vegetational cover of the fen sites. No one or two of the 16 soil and water parameters obtained in this study could be used to discriminate conclusively among fen types. Linear discriminant function (LDF) analysis, however, indicated that the better discriminators were pH, peat thickness, SO4−, K+, and depth to water level. When all water and soil parameters were used, regrouping by LDF analysis into the four a priori groups was achieved with 78% accuracy. Canonical analysis also showed separations when soil and water parameters for sites were plotted in two dimensions. Because of isostatic rebound, distance from the coast represents a temporal as well as a spatial gradient. Peat depth in the fens increases with distance from the coast, at a mean rate of 4.7 cm for each kilometre inland. Na+ plus Cl− in the groundwater of the fens decreases asymptotically with increasing distance from the coast.