The phytoplankton associations in HRB and Dishui Lake
Twelve phytoplankton associations in HRB were identified by combinations of two to three taxa (Table 5) based on support>=50%, and confidence values of >=0.8. All associations can be divided into three types: (1) The flagellate algae association (R01-1, R01-2, R01-9): pollution-tolerant species (e.g., Cryptomonas ovata andCryptomonas erosa) co-existed; mixotrophic chrysophytes (e.g.,Chrumulina sp.) was observed to occur with Cryptomonas erosa , Chroomonas acuta , Chlamydomonas globosa from different taxon groups. (2) The diatoms with the flagellate algae association (R03-2, R05-2, R06-2, R07-2, R08-2, R10-2, R11-2 and R12-2): benthic diatoms with motility (e.g., Nitzschia palea ,Navicula cryptocephala) frequently co-occurred with the cryptophytes. It should be noted that a small centric planktonic diatom (Cyclotella meneghiniana ) also occurred with the flagellate algae. (3) The diatoms association (R04-3: Nitzschia palea andCyclotella meneghiniana ).
Fifteen phytoplankton associations in Dishui Lake were identified by combinations of two to three taxa (Table 5). These associations can be divided into three types: (1) The flagellate algae association (L11-1): mixotrophic chrysophyte Chrumulina pygmaea was observed to occur with Chlamydomonas globosa . (2) The mixed association-diatoms or colonial green algae with the flagellate algae association: this association can be further divided into three smaller associations including mixotrophic chrysophyte Chrumulina pygmaeawith diatoms or green algae (L03-2, L07-2, L09-2, and L12-2),Chlamydomonas globosa with diatoms or green algae (L01-2, L02-2, L04-2, L05-2, L08-2, L10-2 and L13-2), and Chromulina pygmaea andChlamydomonas globosa with diatoms or green algae (L14-2 and L15-2). (3) The diatoms with colonial green algae association (L06-3).
Relationships between phytoplankton associations and environmental variables
We analyzed the phytoplankton assemblages characterized with 12 phytoplankton associations in HRB using detrended correspondence analysis (DCA). DCA results showed that the maximum gradient length of the four axes was 2.63. Subsequently, we selected a redundancy analysis (RDA) to detect the relationship between phytoplankton associations and environmental factors (Fig. 2). Approximately 12% of the variance in phytoplankton associations can be explained by environmental factors (axis 1: 8%, axis 2: 3%). Forward selection in RDA identified six significant environmental factors (Fig. 2). Turbidity was positively correlated with axis 1, TN/TP ratio was negatively correlated with axis 1; Conductivity and TN positively correlated with axis 2, stream order negatively correlated with axis 2. Most of the mixed associations had a positive relationship with TN and turbidity except that association 3 displayed a positive relationship with conductivity and negative correlation with turbidity (Fig. 2).The flagellate algae association had a positive relationship with turbidity, DO and stream size while the diatom association was positively associated with TN(Fig. 2).
We used the 15 phytoplankton associations in Dishui Lake for detrended correspondence analysis (DCA) with the maximum gradient length of the four axes as 1.55. RDA showed that 32% variance in phytoplankton associations can be explained by environmental factors: axis 1: 27%, axis 2: 4.8%. Forward selection in RDA identified three significant environmental factors (Fig. 3). Salinity and transparency were negatively correlated with axis 1; pH was negatively correlated with axis 2. The mixed associations including Chromulina had a positive relationship with pH and transparency while the mixed associations including Chlamydomonas had a negative relationship with pH (Fig. 3). The flagellate associations positively correlated with salinity (Fig. 3).