Effect of Sampling Protocol on Estimates of Phosphorus and Chlorophyll Concentrations in Lakes of Low to Moderate Trophic Status

Abstract

Methods of sampling lake phosphorus and chlorophyll concentrations were assessed by comparing mean values for 16 lake basins in southern Quebec. Integrating tube and discrete-depth samples were analyzed to yield six different subsurface, depth-integrated, and volume-weighted average values representing epilimnetic, euphotic, and trophogenic zones. The different protocols were found to be similarly precise. Epilimnetic samples yield low values, while trophogenic samples yield high ones; however, the differences in protocol accounted for less than 50%. Analysis of variance associated most of the uncertainty in these estimates with lake and sampling date and showed that sampling site rarely had a significant effect. Year had only a minor effect, but interannual variation is probably underestimated, since these data represent only two years. These patterns suggest that representative sampling for a given lake requires only a single sample per visit but several visits within a season. Since chlorophyll varied more spatially than phosphorus, sampling protocol affects its estimates to a greater extent. Phosphorus–chlorophyll regressions based on data collected in different ways were very similar, so differences in protocol and sampling depth seem unlikely to contribute to the high uncertainty in existing phosphorus response models.