Weighted averaging (WA) regression and calibrationbased optima and tolerances of lakewater pH andtemperature are presented for diatoms in ecologicallysensitive, subarctic Fennoscandian lakes. The studysites are mostly small, simple, oligotrophic,low-conductivity lakes with a pH range from 5.0 to7.7 and a temperature range (after data screening)from 9.3 to 15.0 °C. Experiments with inverse andclassical deshrinking, with or without tolerancedownweighting, were used to identify the bestcalibration functions. The model estimates wereadjusted by jackknifing procedures. WA by inversedeshrinking and with tolerance downweighting performedbest for pH prediction, whereas simple WA wasmarginally superior for predicting water temperature.The established pH model is accurate to within± 0.39 H units, and the temperature model towithin ± 0.88 degrees Celcius. Fifteen diatom taxawere identified as potential indicator species for pHand three for temperature.