ALBERT

Notes: Lake ecosystems throughout the world are threatened by numerous chemicals. Most people have heard about the major chemical threats to aquatic systems, such as mercury, radionuclides, sulphur and acid rain, and nutrients causing eutrophication effects. How are these threats manifested in ecosystems? What is threatened and why? What can be done for remediation? The complexity of ecosystems has often led to the predictive failure of ecosystems models, and a decade ago these models were rightly regarded with suspicion. At that time, many lake models predicted target variables, such as mean concentrations of toxins in water, sediments and fish, within a factor of 10. There has been very rapid development in predictive modelling for water pollutants during recent years. Today, lake models for water pollutants predict as well as can be measured (i.e. within a factor of 0.25–0.5). One major reason for this rapid development is the Chernobyl accident. Large quantities of radiocaesium were released in a pulse during April–May 1986. In order to follow the radiocaesium along ecosystem pathways, the identification of fundamental transport processes and the structuring of models has been developed. To obtain a high predictive power in ecosystem models, it is crucial to include only the key processes (the optimal size problem). Effect-load- sensitivity (ELS) models play a paramount role in the environmental sciences and lake management. The aim of ELS models is to provide quantitative predictions relating operationally defined ecological effect variables to load and environmental sensitivity variables. Validated ELS models provide a tool to simulate practically feasible remedial measures, and important concepts like critical load, natural load and guideline limits/values can be defined from ELS models. The first objective of this paper is to discuss some basic structures, principles and components of ELS models. Empirical (static) and dynamical (time-dependent) ELS models exist for mercury, radiocaesium and phosphorus in lakes. These models have high predictive powers for a wide range of validation lakes. The second objective of this paper is to demonstrate what can actually be done in water management in terms of remediation of the effects asociated with acidification (lake liming), eutrophication and contamination by mercury and radiocaesium.