Mercury (Hg) is regarded as a major environmental concern in many regions, traditionally because of high concentrations in freshwater fish, and now also because of potential toxic effects on soil microflora. The predominant source of Hg in most watersheds is atmospheric deposition, which has increased 2- to >20-fold over the past centuries. A promising approach for supporting current European efforts to limit transboundary air pollution is the development of emission-exposure-effect relationships, with the aim of determining the critical level of atmospheric pollution (CLAP, cf. critical load) causing harm or concern in sensitive elements of the environment. This requires a quantification of slow ecosystem dynamics from short-term collections of data. Aiming at an operational tool for assessing the past and future metal contamination of terrestrial and aquatic ecosystems, we present a simple and flexible modelling concept, including ways of minimizing requirements for computation and data collection, focusing on the exposure of biota in forest soils and lakes to Hg. Issues related to the complexity of Hg biogeochemistry are addressed by (1) a model design that allows independent validation of each model unit with readily available data, (2) a process- and scale-independent model formulation based on concentration ratios and transfer factors without requiring loads and mass balance, and (3) an equilibration concept that accounts for relevant dynamics in ecosystems without long-term data collection or advanced calculations. Based on data accumulated in Sweden over the past decades, we present a model to determine the CLAP-Hg from standardized values of region- or site-specific synoptic concentrations in four key matrices of boreal watersheds: precipitation (atmospheric source), large lacustrine fish (aquatic receptor and vector), organic soil layers (terrestrial receptor proxy and temporary reservoir), as well as new and old lake sediments (archives of response dynamics). Key dynamics in watersheds are accounted for by quantifying current states of equilibration in both soils and lakes based on comparison of contamination factors in sediment cores. Future steady-state concentrations in soils and fish in single watersheds or entire regions are then determined by corresponding projection of survey data. A regional-scale application to southern Sweden suggests that the response of environmental Hg levels to changes in atmospheric Hg pollution is delayed by centuries and initially not proportional among receptors (atmosphere > soils not equal sediments>fish; clearwater lakes > humic lakes). This has implications for the interpretation of common survey data as well as for the implementation of pollution control strategies. Near Hg emission sources, the pollution of organic soils and clearwater lakes deserves attention. Critical receptors, however, even in remote areas, are humic waters, in which biotic Hg levels are naturally high, most likely to increase further, and at high long-term risk of exceeding the current levels of concern: </=0.5 mg (kg fw)(-1) in freshwater fish, and 0.5 mg (kg dw)(-1) in soil organic matter. If environmental Hg concentrations are to be reduced and kept below these critical limits, virtually no man-made atmospheric Hg emissions can be permitted.