Assessing variability in total phosphorus measurements in Ontario lakes

In most Ontario lakes, phosphorus is present in trace quantities, making the precise measurement of concentrations difficult. The considerable variation that results in many datasets can be attributed to imprecise analysis. Even with precise analysis, substantial variation in ice-free whole-lake, mean mixed layer, and spring turnover total phosphorus (TP) concentrations often remain, both between years and within a given year. Lake managers have adopted many strategies to address these issues. Whole-lake TP concentrations expressed as ice-free season means are difficult to derive because they require the collection of numerous, volume-weighted samples often from multiple lake layers. As a result, spring turnover or ice-free-mean, mixed layer TP concentrations are often used to characterize the trophic status of a lake or to describe the nutrients available for primary production. These different phosphorus characterizations will vary between years, leading many lake managers to use long-term (multiple year) means to describe trophic status. The ability to interpret natural variation or trends in these data can be reduced by any bias introduced through sample collection and storage (container) methods or as a result of sample contamination by zooplankton, which can produce high sample TP concentrations. Much variability can be reduced through precise analysis, and most sources of error can be eliminated by sampling directly into the same borosilicate glass tubes used to digest samples prior to analysis and by coarse filtering water samples in situ to 80 mu m to eliminate large zooplankton. We demonstrated the importance of collecting precise TP data and presented the variation associated with 25 years of TP measurements in Ontario, Precambrian Shield lakes. We also demonstrated various sample collection strategies that can be used to characterize the nutrient status of lakes and explained how the interpretation of these can be affected by variability in the results.