COMPARISON OF 3 METHODS OF ESTIMATING ATMOSPHERIC MERCURY DEPOSITION

Three methods far estimating atmospheric Hg deposition were compared at a site in north central Wisconsin. Mean volume-weighted Hg concentrations varied 2-fold among methods: 8.6 ng/L for manual, event sampling (wet only); 12.3 ng/L for automated, wet-only collections; and 15.5 ng/L for triplicate bulk collections. Mean Hg loading estimates ranged from 0.24 to 0.44 mu g m(-2) week(-1). Agreement among the three bulk collectors was good: 0.44 +/- 0.03 mu g m(-2) week(-1) (mean +/- SD). Potassium dichromate preservative produced high blanks in the automated, wet-only sampler (2-55% of the measured Hg). Data from two of the collectors showed a strong negative correlation between Hg concentration and precipitation volume. This correlation became weaker when the data were lumped into multi-week intervals and disappeared altogether at a collection interval of 1 month. A negative power function fit the weekly data well, and simple modeling demonstrated that missing early storm stages and small rain events can strongly bias estimates of Hg concentration and loading.