Estimates of mercury emissions from individual sources and source categories are needed to understand relationships between the emissions and resulting deposition and to evaluate possible approaches to reducing those emissions. We have developed geographically-resolved estimates of annual average mercury emission rates from current anthropogenic sources in the 48 contiguous United States. These estimates were made by applying emission factors to individual facility operating data and to county-wide source activity levels. We apportioned the emissions to an Eulerian modeling grid system using point source coordinates and the fractions of county areas in each grid cell. Point sources account for about 89% of the 48-state total mercury emissions of 146.4 Mg/yr. Most of the emissions in the inventory are from combustion of mercury-containing fossil fuels and municipal waste, located primarily in the mid-Atlantic and Great Lakes states as well as in the Southeast. The major uncertainties in the emission estimates are caused by uncertainties in the emission factors used to develop the estimates. This uncertainty is likely a result of variability in the mercury content of the combusted materials and in the removal of mercury by air pollution control devices. The greatest research need to reduce uncertainties in mercury emission estimates is additional measurements to improve emission factors.

In the practice of soil remediation, organoleptic observations such as the smell or the colour of contaminated soil play an important role when determining well-defined volumes of contaminated soil. A GIS is then used to combine quantitative measurements with such soft data. In this study general procedures concerning how to deal with this type of observations are presented. The procedures were applied to a former gas works site, which was contaminated with cyanide and polycyclic aromatic hydrocarbons in the Netherlands. The volume of contaminated soil was determined. Use of soft data reduced the uncertainty in the volume of contaminated soil with 4 to 16%.

The relationship between acidity and benthic invertebrate communities was investigated in two small streams in coniferous forests in central Jutland, Denmark, during 1992 and 1993. Stream pH was acid (mean pH 4.8 and 5.4) and alkalinity fluctuated greatly with mean values around zero. During rain events, water from springfed reaches upstream were mixed downstream with brown humic acid water with high aluminium content, and pH declined from circumneutral near the springs down to pH below 4 over a very short distance (0.5–1.5 km). The benthic invertebrate communities in the most acidic downstream reaches of both streams were dominated by filipalpian stoneflies, mainly Leuctra nigra (Olv.). Community composition remained very similar throughout the sampling period despite variations in pH. In the upper (neutral) reaches invertebrates intolerant of low pH such as Gammarus pulex L. were found. The concentrations of chloride and sulphate in the streams were more than double the concentrations in a nearby moorland stream indicating a strong impact of acidifying atmospheric deposition on the two forest streams. Results of this study indicate that afforestation with conifers on sandy Danish soils with low buffering capacity (about 25% of the total Danish land area) may lead to an increased acidification of surface waters and thereby also to an impoverished aquatic fauna.