Sampling for the copper concentration in water was performed three times per week using new 60 mL polypropylene syringes in each mesocosm. Immediately after sampling the water was pushed through an UM 729 washed 0.45 μm pore size cellulose acetate membrane filter (25 mm Ø, Whatman). The filtrate was collected in 50 mL LDPE acid washed polythene bottles and acidified to pH < 2 (HCl, Suprapur, Merck). Samples were transported within 1 h to the laboratory. Here, the samples were diluted ten times with a 0.1 M nitric acid (HNO3 Suprapur, Merck) solution in demineralised (Milli-Q) water. The diluted solutions were than analysed for copper content using inductively coupled plasma atomic mass spectrometry (ICP-MS) at a mass of 63 m/z. The measurements were also performed at an alternative mass of 65 m/z. The difference between the results at grasslands biome two masses was not considered to be significant (< 10%). Peak intensity of rhodium was used as internal standard at the mass of 103 m/z, for calculation of the amount of copper. For the quality control (QC) of the analyses a sample of mesocosm water spiked at 10 μg Cu/L at the beginning of the study was included in each of the 12 batches of copper analysis. The average concentration determined in these QC-samples was 9.73 μg Cu/L with a standard deviation of 0.28 μg/L confirming the correctness and reproducibility of the analyses.