**3.3 Diatom analysis**

The sediment samples (1g) were dried in an oven and cleaned with potassium permanganate and hydrochloric acid according to the Simonsen technique [55]. For light microscopy (LM) analyzed, a Zeiss Axioplan Microscope (Carl Zeiss, Oberkochen, Germany) was used. The relative abundance of the taxa was carried out in slides seeking the minimum sample efficiency of 80% [56]. Species richness was estimated by the number of taxa present in the samples. The specific diversity was assessed using Shannon index (H′) [57] and Evenness equitability (E). The analysis of variance (ANOVA) was applied to test the significance among the community attributes, since the data presented a normal distribution. The PAST® software was used for these analyzes. The relationship between biotic and abiotic variables with canonical correspondence analysis (CCA), PC-ORD® version 6.08 was used. For the construction of the biotic matrix, only species with a frequency equal to or >5% were considered in at least one sample unit and for the abiotic matrix, 10 environmental variables were included (**Table 1**). The data were transformed into log10(x + 1) in order to normalize the variances [58]. The Monte Carlo permutation test was carried out to verify the significance of the ordination axes.


#### **Table 1.**

*Physical and chemical variables analyzed in Peixe Lagoon in the four seasons, from June 2011 to February 2012, in the North (N), Center (C), and South (S) sampling stations. Depth (cm); Secchi = Secchi transparency (cm); temp = temperature (°C); ORP = oxide-reduction potential (mV); cond = conductivity (mS.cm<sup>−</sup><sup>1</sup> ); DO = dissolved oxygen, salin = salinity (ppt); sil = silica (mg.L<sup>−</sup><sup>1</sup> ); PT = total phosphorus (mg.L<sup>−</sup><sup>1</sup> ); NT = total nitrogen (mg.L<sup>−</sup><sup>1</sup> ).*
