Measured physico-chemical variables were compared with analysis
of variance (ANOVA) [30]. Tukey's honestly significant difference
(HSD)multiple range testswere also carried out for comparing environmental
changes. Spearman's rank correlation analysis was used to
evaluate relationships among the environmental variables. Detrended
correspondence analysis (DCA) was used to determine the gradient
length of phytoplankton composition, which was a unimodal response
before performing CCA [31]. Canonical correspondence analysis (CCA),
a direct gradient analysis technique, was used to elucidate the relationship
between predictor variables (environmental factors) and response
variables (phytoplankton species) in the pond. Environmental variables
were transformed (ln (x + 1), except the pH) to reduce skewness [32].
Obtained data were analyzed with forward selection of Monte Carlo permutation
test (499 unrestricted permutations). Net-phytoplankton
composition with binary data was used in this analysis. The program
CANOCO was performed for the ordination analyses [31,32]. Weighted
averaging (WA) regression of the CALIBRATE program [33] was carried
out to estimate species optima (uk) and tolerance (tk) levels for environmental
variables.