Identical aliquots (split samples) from lakes in the southeastern United States were analyzed in laboratories in the U.S. and Norway. A second set of split samples from lakes in the northeastern U.S. was analyzed in laboratories in the U.S. and Canada. Methods used by the laboratories were identical for most analytes. For 8 of 11 analytes, there was a statistically significant difference between the results from Norway and the U.S. For 15 of 17 analytes, there was a statistically significant difference between the results from Canada and the U.S. Linear equations describing the relationship between results from the U.S. and Norway (or Canada) explained over 90 of the variance for most analytes. Notable discrepancies occurred for labile inorganic aluminum and acid-neutralizing capacity, two analytes critical to acid deposition studies. A linear model explained about 35 of the variance for labile inorganic aluminum. Considerable laboratory bias was evident for acid-neutralizing capacity in both sets of split sample measurements. Difference in methodologies and holding times may account for these discrepancies. For any of the analytes, the practical significance of the differences is dependent on the objectives and constraints of each application. The effects of laboratory and methodology were statistically confounded. Standardizing analytical methods would assist the chemist in interpreting data from studies, notably those of lake acidification, in other nations.
展开▼
