Development of ecosystems to climate change and the interaction with pollution—Unpredictable changes in community structures

• Soil invertebrate species composition was affected by temperature and Cu exposure.
• Temperature effects were species-dependently related to population growth curves.
• Cu caused large changes in the community reducing species abundance.
• The combined effect of Cu and temperature was not predictable.
• Cu-tolerant species seemed to benefit from reduced competition.

Climate change has serious impacts on ecosystems, e.g. species diversity and abundance. It is well known that changes in temperature may have a pronounced influence on the reproductive output, growth and survival of various terrestrial species. However, much less is known on to how changes in temperature combined with exposure to pollution will influence biodiversity, the interaction between species, and the resulting change in species composition. In order to understand the effects of changes in temperature and copper pollution (individually and in combination) on soil communities and processes, a factorial multispecies experiment was performed. Six animal species (representing different functional groups) were exposed in control (30 mg Cu/kg) and copper-contaminated soil (1000 mg Cu/kg) to four temperatures (10, 14, 19, and 23 °C) representing the “summer” range (low to high) for Denmark, and three exposure periods (28, 61, and 84 days). The species composition, feeding activity and OM turnover were assessed throughout. Multivariate analysis displayed significant changes in the food-web both with different copper levels and temperatures, resulting in different species composition for each exposure scenario. The most important species were Enchytraeus crypticus (most sensitive to copper and temperature) and Folsomia candida (most abundant). Major changes in abundance due to temperature occurred in the first 28 days of exposure, where population growth was higher. A temperature dependent population growth rate could be modeled for an exposure period of 28 days, whereas after 61 and 84 days of exposure the data did not fit the model. Especially for treatments that also included Cu, modeling of the population growth was no longer possible. The results of our study indicate that when climate change occurs in polluted areas, the consequences on populations cannot be predicted based on data from non-polluted areas. The risk may be synergistic for certain species, as indicated in the present study, and the final balance may depend on the particular species composition of that ecosystem.