Statistical analyses were performed with IBM SPSS 19.0. In the case of TN removal and TP uptake by floating algae, data were distributed normally and variances were equal, so an ANOVA was performed. In cases of significant differences between treatments, the ANOVA was followed by post hoc testing using Tukey's HSD test. In all other cases, a non-parametric Kruskal–Wallis test for multiple comparisons was applied to test for differences between treatments because the data were not distributed normally and/or homogeneity of variance was lacking. When significant differences between treatments were present, Mann–Whitney U tests were performed. A Bonferroni correction was applied to overcome inflated Type I errors, reducing the level of significance to 0.05/ 3 = 0.017. During all other statistical tests, the level of significance was held at 0.05.