Improving pH and alkalinity estimates for regional-scale acidification models: Incorporation of dissolved organic carbon

Dissolved organic carbon (DOG) can affect the accuracy of regional model projections of acid neutralizing capacity and pH. To improve predictive accuracy. we incorporated DOC into a modified titration curve and used a DOG-based correction factor to convert charge balance acid neutralizing capacity to an estimate of Gran alkalinity. Applying the new titration curve to 5612 lakes in Eastern Canada resulted in only a slight (0.08 unit) average improvement in the accuracy of pH predictions over a previous model without DOG. The DOG-based conversion of charge balance acid neutralizing capacity to Gran alkalinity had a greater impact on model projections: a median alkalinity decrease of 25.5 mu equiv./L over 3226 lakes, lowering average pH projections for each of 25 subregions of Eastern Canada by 0.1-0.4 pH units. Empirical relationships calculated separately for each secondary watershed had slightly improved predictive power over regressions computed from the entire data set. These modifications produce quantifiable improvements in the accuracy of regional model estimates of pH and acid neutralizing capacity, and shift pH distributions towards more toxic levels.