gcb12139-sup-0001-TableS1.xlsxapplication/msexcel33KTable S1. Pearson's correlation coefficients of the phytoplankton and physico-chemical variables at the sampling sites. Asterisks indicate 5% statistical significance and thus similar dynamics at different sites. Multiple Friedman comparisons further showed that the dynamic patterns did not shift up or down in the different sampling sites for any of the variables studied (data not shown).
gcb12139-sup-0002-TableS2.xlsxapplication/msexcel36KTable S2. Weight of CCA environmental variables in the axis construction (expressed by the direction cosines) and correlation coefficient on all axes conjointly (r2). * = significant to 0.05; ** = significant to 0.01; *** = significant to ≤0.001 (Permutations test, n = 1000).
gcb12139-sup-0003-FigS1.jpgimage/jpg68KFigure S1. Location of the study area (Dampierre-en-Burly) on the river Loire.
gcb12139-sup-0004-FigS2.jpgimage/jpg259KFigure S2. Principal component analysis (PCA) (Bray-Curtis distances) of (a) the phytoplankton group abundances (cell L−1) and (b) the environmental variables for the period 1985–2008. The site locations upstream, discharge and downstream are identified as a, b, and c, respectively. P: orthophosphate; Si: silicate; SM: suspended matter; Cond: conductivity; T: temperature. The low dispersion of the sampling sites for each of the studied variables indicates the absence of any substantial difference between the sites in terms of environmental characteristics or phytoplankton community.
gcb12139-sup-0005-FigS3.jpgimage/jpg176KFigure S3. Temporal dynamics of the annual average river (a) discharge (m3 s−1) and (b) temperature (°C) estimated from the continuous records during the period 1977–2008. Least-square linear lines are shown.
gcb12139-sup-0006-FigS4.jpgimage/jpg237KFigure S4. Identification of time periods in the phytoplankton dynamic by a modeling approach. Ratio of residual sum of squares from models 1 (Abundances Abundances + Year) and 2 (Abundances Abundances + Period) (dashed line, see section 'Materials and methods'), superimposed over the annual maximum phytoplankton abundances during the monitoring period (May to October) (gray bar) and its fitted smooth spline function (bold solid line). Model 2 outperformed model 1 (ratio <1) when the period limit was around 1994, which demonstrated the existence of two phases in phytoplankton dynamics: prior to and after 1994.
gcb12139-sup-0007-FigS5.jpgimage/jpg311KFigure S5. Contribution (%) of the nanoplankton and microplankton cells to the total phytoplankton community during the monitoring period (from May to October).

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.