CHRYSOPHYCEAN CYST ASSEMBLAGES AND THEIR RELATIONSHIP TO WATER CHEMISTRY IN 71 ADIRONDACK-PARK (NEW-YORK, USA) LAKES

Chrysophycean cysts were identified and enumerated from the surface se diments of 71 lakes in Adirondack Park, New York, and their relationsh ips with measured limnological variables were examined with canonical ordination (redundancy analysis). Gradients in the cyst distributions were most strongly related to conductivity, dissolved organic carbon, Secchi depth, and aluminum. Lesser amounts of variance in the cyst dat a were explained by pH, acid neutralizing capacity (ANC), maximum dept h, and altitude. We derived predictive models for 13 environmental var iables based on the cyst assemblages using partial least squares (PLS) regression. The best models, in terms of regression coefficients (r(2 )) and root mean square errors of prediction (RMSEP), were obtained fo r pH and ANC (r(2) = 0.64; RMSEP = 0.59 and 0.46 ln (mu eq/L+41)), res pectively, followed by [Mg] (r(2) = 0.53, RMSEP = 0.36 ln (mu eq/L+1)) , total aluminum (r(2) = 0.48, RMSEP = 0.86 ln (mu g/L+1)), conductivi ty (r(2) = 0.46, RMSEP = 0.19 ln (mu S/cm+1)), and monomeric aluminum (r(2) = 0.35, RMSEP = 0.66 ln (mu M+1)). These models can be used to i nfer historical levels of these variables in paleolimnological studies .