Phytoplankton functional response to spatial and temporal differences before and after the regime shift of Caohai Lake, China Authors: Liangliang Dai, Yunchuan Long, Daibo Wang, Jingcai Lyu, Guo Yuan, Shinong Liu, Tianwen Hou Published: 2026 ABSTRACT This study aimed to explore the functional responses of phytoplankton to the ecosystem regime shift of Caohai Lake, a typical plateau karst lake located in Weining County, Guizhou Province, China. Since 2020, the lake has undergone a significant transition from a clear-water steady state dominated by submerged macrophytes (e.g., Potamogeton lucens L., Myriophyllum verticillatum L.) to a turbid-water steady state dominated by phytoplankton, accompanied by ecological degradation such as extensive macrophyte die-off and severe eutrophication. To address this, water and phytoplankton samples were collected in July 2018 (pre-regime shift) and July 2023 (post-regime shift). A suite of environmental variables was measured, including electrical conductivity (EC), total dissolved solids (TDS), total nitrogen (TN), total phosphorus (TP), ammonia nitrogen (NH3-N), turbidity (Tur), chlorophyll a (Chl a), pH, and dissolved oxygen (DO). Statistical analyses, including one-way analysis of variance (ANOVA), redundancy analysis (RDA), analysis of similarities (ANOSIM), and similarity percentages (SIMPER) analysis, were applied to examine differences in environmental conditions and phytoplankton communities between the two periods. Results showed significant post- regime shift changes in environmental conditions: EC, TDS, TN, TP, NH3-N, Tur, and Chl a increased significantly (P < 0.05), while pH and DO decreased. Phytoplankton cell density rose markedly from 15.48 × 106 cells/L (2018) to 58.86 × 106 cells/L (2023), Cyanobacteria remained the dominant class in both periods, the cell density were 13.58 × 106 cells/L and 48.42 × 106 cells/L respectively, indicating a dramatic expansion post-regime shift. The composition of phytoplankton functional groups (FGs) shifted from pre-regime shift dominant groups Lo/M/N/S1/X1 to post-regime shift C/F/J/Lo/M, with Lo (e.g., Merismopedia sp.) and M (e.g., Microcystis sp.) being the most abundant across both periods. RDA indicated that the main environmental parameters affect Cyanobacteria, Bacillariophyceae and Chlorophyceae were NH3-N, Sal, Tur, Cond, Chl a, TDS and TN. SIMPER analysis revealed an average dissimilarity of 53.30% in FGs between 2018 and 2023, with Lo (29.42%) and M (17.78%) contributing most to this difference. In conclusion, the aggravation of eutrophication following the regime shift was the key driver of changes in the phytoplankton community of Caohai Lake. Nutrient enrichment created a turbid, eutrophic habitat that favored the proliferation of FGs adapted to such conditions, highlighting the critical role of nutrient regulation in restoring the lake’s degraded ecosystem. Introduction The key indicator of shallow lake ecosystem degradation is the shift in water clarity-specifically, the transition from clear water conditions to a turbid state. When a shallow lake maintains a clear water steady state, its water remains transparent, and this stability is often linked to extensive coverage of submerged macrophytes. In contrast, a turbid water steady state is characterized by a high density of phytoplankton, which directly reduces water transparency and results in noticeably turbid water quality1. In lake ecosystems, the transition from clear water to turbid water is defined as a regime shift. Ecosystem degradation arises from a synergistic interplay of external drivers, which encompass the influx of exogenous nutrient loads, fluctuations in lake water levels (such as artificial elevation), damage to lakeside riparian zones, and biotic regulatory processes2. The rapid decline of submerged macrophytes leads to a gradual increase in phytoplankton biomass, which in turn triggers a sharp reduction in water transparency. Furthermore, the loss of these macrophytes not only strengthens the resuspension of sediments but also disrupts the lake’s food web3. This disruption forces omnivorous fish to forage in the substrate, a behavior that further worsens sediment resuspension. Under the combined effects of elevated phytoplankton cell density and intensified sediment resuspension, water transparency decreases even more significantly—ultimately driving a complete shift in the lake ecosystem from a clear-water steady state to a turbid-water steady state2,4. The environmental response of communities can be studied through a taxonomical perspective (species-specific) or via a functional one5,6. Several functional strategies have been formulated to explain and predict the distribution patterns of phytoplankton species across environmental gradients7–9. The functional group (FG) concept sensu Reynolds8,10 categorizes phytoplankton taxa into over 40 groups (known as codon), with grouping criteria based on the taxa’s occurrence in a specific habitat. This functional method has been widely adopted for the study of phytoplankton communities across the world11–14, functional classifications based on species traits have been developed as well15. Classification of species relies on their morphological, physiological, or phenological characteristics, which are associated with processes including growth, reproduction, nutrient acquisition, light harvesting, and defense against grazing8,10,16. Although certain species lack data on their physiological characteristics, the functional trait (FT) method remains valuable for forecasting which strategies are favored under specific environmental conditions17. Once a species successfully surpasses dispersal limitations, the suitability of its growth in a specific habitat is determined by local environmental factors, such as the physical structure of the water column, nutrient accessibility, and grazing intensity18. Compared to taxonomic analysis, Reynolds’ FG system8 integrates ecological traits (nutrient requirements, light tolerance, buoyancy) to better link community changes with environmental drivers10, which is critical for understanding regime shift mechanisms. Despite extensive research on phytoplankton responses to lake regime shifts, few studies have focused on spatial and temporal differences in phytoplankton functional group dynamics in karst shallow lakes. Moreover, the key environmental drivers regulating these dynamics before and after regime shifts remain unclear, limiting our understanding of the ecological mechanisms underlying regime shift processes in such unique aquatic ecosystems. Here, we studied the phytoplankton community of Caohai Lake from a functional perspective. Caohai Lake (Guizhou Province, China) is a typical plateau karst lake and is facing serious environmental problems, including enhanced sedimentation, pollution of recharge water sources, and deterioration of water quality19. Since 2020, a large area of submerged macrophytes in Caohai Lake has died out, and with obvious turbid water bodies and serious water eutrophication, the aquatic ecosystem has been degraded, Caohai Lake has transitioned from a clear-water state to a turbid-water state19. Earlier studies have focused on the response of phytoplankton communities to environmental factors, and these researches rely on hydrochemical parameters and sediment records—data that fails to offer direct evidence of algal community succession for identifying regime shifts20,21. We studied phytoplankton and environmental data in July 2018 and July 2023, encompassing the possible regime shift in Caohai Lake, to achieve two objectives: (1) to investigate the occurrence of a regime shift in Caohai Lake; (2) to identify the main factors driving phytoplankton community sucession before and after the regime shift, if one did occur. Materials and methods Study area Caohai Lake (26°47′35" N-26°52′10" N, 104°9′23" E-104°20′10"E; altitude: 2170 m) is located in Weining County, Guizhou Province (Fig. 1). It is the largest natural shallow freshwater lake in Guizhou Province and the largest natural karst wetland in China. The protected area is 120 km2 and the average depth of the lake is 2 m. In 1992, the area around Caohai Lake was protected as a natural reserve. The lake area is characterized by plateau subtropical climate; and on yearly average, temperature is 10.6 °C and rainfall 980 mm20. The lake is rich in aquatic plants, primarily Potamogeton lucens, Myriophyllum verticillatum, Potamogeton wrightii, Potamogeton maackianus, and Najas graminea. Great importance is attached by the government to the protection and treatment of Caohai Lake, with external pollution effectively controlled. However, due to ecological disequilibrium inside the lake, submerged macrophytes have vanished over extensive areas, and the lake’s regime has shifted: it has changed from clear water (dominated by submerged macrophytes) to turbid water (dominated by phytoplankton)22. Water level elevation may be a potential contributing factor, as it can alter hydrological connectivity and sediment resuspension in shallow lakes19. Meanwhile, the fish community has experienced a significant expansion in both its composition and quantity, rendering the aquatic ecosystem more vulnerable20, fish community effects remain a hypothetical secondary factor, requiring future research to verify causal links with regime shift of Caohai Lake.Fig. 1Sampling site locations in Caohai Lake, Guizhou Province, China. Sampling site locations in Caohai Lake, Guizhou Province, China. Field sampling and laboratory analyses Water samples were collected prior to (in July 2018) and after (in July 2023) ecosystem degradation. Sampling included vertical profiles at 0.5 m intervals for temperature (T), pH, electrical conductivity (EC), salinity, total dissolved solids (TDS) and dissolved oxygen (DO) taken with a multiparametric probe (YSI professional Plus). Water turbidity (Tur) was measured using a turbidimeter (HACH 2100Q). A total of 12 surface (at a 0.5 m depth) water samples were taken using 500-mL polyethylene bottles; these bottles were then stored in a refrigerator until analysis. Total nitrogen (TN) was obtained from samples digested to nitrates, and their absorbance was measured at 220 nm and 275 nm. Total phosphorus (TP) content was measured through spectrophotometric methods. The determination of ammonia nitrogen (NH3-N) adopts the Nessler’s reagent spectrophotometric method23, chlorophyll a (Chl a) was analyzed via spectrophotometry after being extracted with 90% ethanol. A total of 12 phytoplankton samples were also collected, following the method outlined by Hu and Wei24. The 12 sampling sites were identical in 2018 and 2023, with coordinates recorded using GPS for precise replication. Sampling sites were selected using a grid sampling method, covering different hydrological zones (inflow, central lake, outflow) and habitat types (shallow littoral, deep pelagic) of Caohai Lake. These samples were taken from the surface water layer (0–50 cm depth) using 1-L polyethylene bottles, then fixed with 15 mL of Lugol’s solution and left to stand at room temperature for 48 h. The siphon method was employed to extract the algae-free supernatant from the upper part of the solution via a rubber hose. The remaining 50 mL of sample sediment was transferred into an 80 mL polyethylene bottle for storage, after which it was sent to the Guizhou Institute of Biology (Guizhou Academy of Sciences) to measure algal density, and species composition. Statistical analyses One-way analysis of variance (ANOVA) was employed to analyze the variation in environmental variables, with a significance level of 5%, and the analysis was conducted using SPSS software (Statistical Package for Social Sciences, Version 20.0). Pearson’s correlation in SPSS 20.0 was used to determine the relationships between environmental variables and FGs. To evaluate the relationship between environmental variables and phytoplankton classes, redundancy analysis (RDA) was also performed. RDA was performed using the vegan package in R 4.4.1, The significance of RDA axes and environmental variables was tested via 999 permutation tests. Phytoplankton cell density data were log-transformed (log10(x + 1)) to meet normality assumptions, with consistent units (cells/L) reported for all analyses. One-way analysis of similarities (ANOSIM) was conducted to assess significant temporal variations, and the results were visualized through metric multi-dimensional scaling (MDS) analysis. Finally, similarity percentages (SIMPER)25 analyses were executed to determine the percentage contribution and average dissimilarity before and after the regime shift. The Bray–Curtis similarity index-used to measure differences in species composition—was applied to examine similarity matrices. This index ranges from 0 to 1, a value of 0 indicates complete overlap in species between samples, and a value of 1 indicates no shared species. SIMPER analyses were analyzed using PRIMER 6 software, while MDS was performed using the vegan package in R software (Version 4.4.1). Results Environmental conditions Environmental variables varied significantly in July 2018 and July 2023. EC (Fig. 2A), TDS (Fig. 2B), Salinity (Fig. 2C), TN (Fig. 2D), TP (Fig. 2E), NH3-N (Fig. 2F), Turbidity (Fig. 2H), Chl a (Fig. 2I) showed evident variabilities at each study site except DO (Fig. 2G) (Table 1). Furthermore, significantly lower (P < 0.05) values of most water quality parameters were observed prior to ecosystem degradation. Conversely, significantly higher (P < 0.05) nutrients, i.e., TN (Fig. 2D) and TP (Fig. 2E) as well as EC (Fig. 2A), were observed after the ecosystem degradation.Fig. 2Physicochemical water quality parameters on the site prior to (in July 2018) and after (in July 2023) ecosystem degradation in Caohai Lake.Table 1Comparisons of environmental variables prior to (in July 2018)and after (in July 2023) ecosystem degradation in Caohai Lake.ParameterJuly-2018July-2023Mean ± SDMean ± SDTemperature (°C)22.61 ± 0.48a21.43 ± 0.26bpH8.62 ± 0.34a8.06 ± 0.24aElectrical conductivity (µS/cm)255.44 ± 20.86b430.21 ± 27.57aTotal dissolved solids (mg/L)174.36 ± 14.92b310.79 ± 24.86aSalinity(‰)0.13 ± 0.01b0.22 ± 0.02aTN (mg/L)1.01 ± 0.11b1.99 ± 0.20aTP (mg/L)0.05 ± 0.02b0.10 ± 0.02aNH3-N (mg/L)0.35 ± 0.04b0.88 ± 0.17aDO (mg/L)6.49 ± 0.48a5.78 ± 0.27aTurbidity (NTU)3.85 ± 0.56b9.08 ± 1.31aChla (μg/L)4.63 ± 0.98b13.86 ± 1.41aThe superscripts of different lowercase letters indicated that the index had significant differences in different years (P < 0.05). Physicochemical water quality parameters on the site prior to (in July 2018) and after (in July 2023) ecosystem degradation in Caohai Lake. Table 1Comparisons of environmental variables prior to (in July 2018)and after (in July 2023) ecosystem degradation in Caohai Lake.ParameterJuly-2018July-2023Mean ± SDMean ± SDTemperature (°C)22.61 ± 0.48a21.43 ± 0.26bpH8.62 ± 0.34a8.06 ± 0.24aElectrical conductivity (µS/cm)255.44 ± 20.86b430.21 ± 27.57aTotal dissolved solids (mg/L)174.36 ± 14.92b310.79 ± 24.86aSalinity(‰)0.13 ± 0.01b0.22 ± 0.02aTN (mg/L)1.01 ± 0.11b1.99 ± 0.20aTP (mg/L)0.05 ± 0.02b0.10 ± 0.02aNH3-N (mg/L)0.35 ± 0.04b0.88 ± 0.17aDO (mg/L)6.49 ± 0.48a5.78 ± 0.27aTurbidity (NTU)3.85 ± 0.56b9.08 ± 1.31aChla (μg/L)4.63 ± 0.98b13.86 ± 1.41aThe superscripts of different lowercase letters indicated that the index had significant differences in different years (P < 0.05). Comparisons of environmental variables prior to (in July 2018)and after (in July 2023) ecosystem degradation in Caohai Lake. The superscripts of different lowercase letters indicated that the index had significant differences in different years (P < 0.05). Phytoplankton composition and assemblage structure In July 2018, seventy-eight species of phytoplankton belonging to 7 classes were identified. Chlorophyceae, Bacillariophyceae, Cyanobacteria, Euglenophyceae, Dinophyceae, Cryptophyceae , and Chrysophyceae constituted 43, 14, 10, 5, 3, 2, 1 species with the percentage contribution of approximately 55.13%, 17.95%, 12.82%, 6.41%, 3.85%, 2.56% and 1.28%, respectively. The total cell density of phytoplankton varied between the sites, the highest total cell density (50.53 × 106 cells/L) was observed in sampling site S6 and the lowest(2.92 × 106 cells/L) in sampling site S11 (Fig. 3A). Among the seven phytoplankton groups in the all sample sites, Cyanobacteria (1.36 × 107 cells/L) was the most abundant, while Chrysophyceae (1.40 × 103 cells/L) was the least abundant group. The average cell density of phytoplankton in the all sampling sites was 15.48 × 106 cells/L (Fig. 3C). In July 2023, sixty-five species of phytoplankton belonging to 7 classes were identified. Chlorophyceae, Bacillariophyceae, Cyanobacteria, Euglenophyceae, Dinophyceae, Cryptophyceae, and Chrysophyceae constituted 43, 9, 10, 2, 1, 2, 1 species with the percentage contribution of approximately 61.54%, 13.85%, 15.38%, 3.08%, 1.54%, 3.08% and 1.54%, respectively. The total cell density of phytoplankton varied between the sites, the highest total cell density (69.68 × 106cells/L) was observed in sampling site S12 and the lowest (30.78 × 106 cells/L) in sampling site S3 (Fig. 3B). Among the seven phytoplankton groups in the all sample sites, Cyanobacteria (48.42 × 106 cells/L) was the most abundant, while Cryptophyceae (5.16 × 103 cells/L) was the least abundant group. The Average cell density of phytoplankton in the all sampling sites was 58.86 × 106 cells/L (Fig. 3C), the cell density in July 2023 was significantly higher than the cell density in July 2018 (t = 7.022, df = 11, P = 0.000). In the riparian area near the county, pre-regime shift conditions were characterized by uneven phytoplankton density: relatively high concentrations were observed at sites 3, 6, 9, and 12, in contrast to the relatively low densities at the other sites. Post-regime shift, this scenario altered substantially—phytoplankton density increased universally across all sites, and the previously distinct density disparities between sites narrowed, leading to a more homogeneous distribution.Fig. 3Phytoplankton cell density observed in each sampling site prior to and after ecosystem degradation in Caohai Lake (A: in July 2018, B: in July 2023, C: Mean and standard deviation). Phytoplankton cell density observed in each sampling site prior to and after ecosystem degradation in Caohai Lake (A: in July 2018, B: in July 2023, C: Mean and standard deviation). In July 2018, the most abundant phytoplankton species recorded with a dominance Y > 0.01 was Microcystis sp. (Y = 0.77), followed by Merismopedia sp. (Y = 0.07), Pseudanabaena sp. (Y = 0.01), Ankistrodesmus sp. (Y = 0.01), Cosmarium sp. (Y = 0.01) as shown in Table 2. In July 2023, The most abundant phytoplankton species was Merismopedia sp. (Y = 0.43), followed by Microcystis sp. (Y = 0.41), Crucigenia sp. (Y = 0.01), Nephrocytium sp. (Y = 0.01).Table 2Abundance (D, × 106cells/L) and dominance (Y) of dominant phytoplankton species in Caohai Lake.SpeciesJuly 2018July 2023DYDYCyanobacteriaMerismopedia sp.1.250.0724.330.43Microcystis sp.11.930.7723.150.41Pseudanabaena sp.0.220.01––ChlorophyceaeAnkistrodesmus sp.0.190.01––Cosmarium sp0.240.01––Crucigenia sp.––0.670.01Nephrocytium sp.––0.710.01“–”represent those species were not dominant species in this year. Table 2Abundance (D, × 106cells/L) and dominance (Y) of dominant phytoplankton species in Caohai Lake.SpeciesJuly 2018July 2023DYDYCyanobacteriaMerismopedia sp.1.250.0724.330.43Microcystis sp.11.930.7723.150.41Pseudanabaena sp.0.220.01––ChlorophyceaeAnkistrodesmus sp.0.190.01––Cosmarium sp0.240.01––Crucigenia sp.––0.670.01Nephrocytium sp.––0.710.01“–”represent those species were not dominant species in this year. Abundance (D, × 106cells/L) and dominance (Y) of dominant phytoplankton species in Caohai Lake. “–”represent those species were not dominant species in this year. The phytoplankton assemblage significantly differed in July 2018 and July 2023. A two-dimensional MDS based on the Bray–Curtis similarity index divided the phytoplankton abundance into two distinct groups prior to (in July 2018) and after (in July 2023) ecosystem degradation (Fig. 4).Fig. 4Classification of phytoplankton assemblages with the help of MDS plot in July 2018 and July 2023 in the Caohai Lake. Classification of phytoplankton assemblages with the help of MDS plot in July 2018 and July 2023 in the Caohai Lake. SIMPER analysis exhibited an overall average dissimilarity of 53.30% in July 2018 and July 2023. The five functional groups (FGs) most significantly responsible for this difference in July 2018 and July 2023 are Lo (29.42%), M (17.78%), C (11.34%), J (9.04%), and F (5.16%). The average similarity in July 2018 is 63.47%, the most significant contributory FGs potentially responsible for this similarity were M (37.68%), Lo (14.02%), and J (10.04%). The average similarity in July 2023 is 85.51%, the most significant contributory FGs potentially responsible for this similarity were Lo (30.14%), M (29.31%), and C (11.45%). After considering segmentation and stations, the RDAs were performed to understand the complex relations of phytoplankton groups with environmental parameters (Fig. 5). The distance point of phytoplankton classes from environmental parameter directional tendencies indicated the perceptual ties between them. For instance, Cyanobacteria showed a closer association with EC, Sal, and TDS in July 2018, Bacillariophyceaeshowed a closer association with EC, Sal, and TDS as well, Chlorophyceae showed a closer association with Chl a, Tur, pH (Fig. 5A). Cyanobacteria showed a closer association with TN, TP, TDS and pH in July 2023, Chlorophyceae showed a closer association with DO, EC, T, Sal, Tur, NH3-N (Fig. 5B). Considering average RDA, Cyanobacteria, Bacillariophyceae and Chlorophyceae are affected closely by the associated environmental parameters at the sampling area (Fig. 5C), the main environmental parameters affect Cyanobacteria, Bacillariophyceae and Chlorophyceae are NH3-N, Sal, Tur, EC, Chl a, TDS and TN. In RDA biplots, the length of environmental vectors indicates the strength of correlation with phytoplankton community structure—longer vectors (e.g., TN, TP in 2023) have stronger explanatory power. The direction of vectors shows the positive/negative correlation with axes: vectors pointing in the same direction are positively correlated, while opposite directions indicate negative correlations. The shift in cyanobacterial association from EC/Sal/TDS (2018) to TN/TP (2023) implies that nutrient availability has become the dominant limiting factor post-regime shift. In 2018, cyanobacteria were more constrained by ionic strength (EC/Sal/TDS), but eutrophication in 2023 has elevated TN/TP to levels where they now drive cyanobacterial proliferation, consistent with global eutrophication trends.Fig. 5Redundancy Analysis (RDA) biplots of phytoplankton classes and environmental variables. (A) 2018 pre-regime shift; (B) 2023 post-regime shift; (C) Combined RDA of 2018 and 2023 data, showing temporal shifts in community-environment relationships. T, temperature; EC, electrical conductivity; Sal, salinity; TDS, total dissolved solids; DO, dissolved oxygen; Tur, Water turbidity; TN, Total nitrogen; TP, Total phosphorus; NH3-N, ammonia nitrogen; Chl a, chlorophyll a. Redundancy Analysis (RDA) biplots of phytoplankton classes and environmental variables. (A) 2018 pre-regime shift; (B) 2023 post-regime shift; (C) Combined RDA of 2018 and 2023 data, showing temporal shifts in community-environment relationships. T, temperature; EC, electrical conductivity; Sal, salinity; TDS, total dissolved solids; DO, dissolved oxygen; Tur, Water turbidity; TN, Total nitrogen; TP, Total phosphorus; NH3-N, ammonia nitrogen; Chl a, chlorophyll a. Phytoplankton functional group The phytoplankton species recorded during this investigation were classified into 21 functional groups (FGs), as presented in Table 3., The total biomass proportion of Chodatella sp., Nostoc sp., Phacotus sp., Quadrigula sp., Raphidonema sp. and Westella sp. in the two-year samples was all less than 0.1% (calculated based on cell density), and their low abundance had a minimal impact on community-level analysis. In addition, there is no unified functional group classification standard for these genera in mainstream classification frameworks, so these genera were excluded from functional groups classification. The dynamics observed by the relative abundance of the functional groups displayed a relative abundance of > 1% prior to ecosystem degradation (in July 2018), while the representative FGs were Lo/M/N/S1/X1. However, after the ecosystem degradation in July 2023, the dominant FGs were characterized as C/F/J/Lo/M. Therefore, Lo/M could be interpreted as the most abundant functional group in this study (Table 3).Table 3Abundance (D, × 105cells/L) and Percent relative abundance of phytoplankton FGs observed in the Caohai Lake.Functional groupsHabitatSpeciesJuly 2018July 2023DRelative abundanceDRelative abundanceCMixed, eutrophic small-medium lakesCyclotella sp.0.920.59%31.274.49%DShallow, enriched turbid waters, including riversFragilaria sp., Nitzschia sp., Synedra sp.1.150.73%3.240.47%EUsually small, oligotrophic, base poor lakes or heterotrophic pondsDinobryon sp.0.010.01%0.310.04%FClear epilimniaMicractinium pusillum, Nephrocytium sp., Oocystis sp., Sphaerocystis sp., Elakatothrix sp.0.830.53%8.541.23%GShort, nutrient-rich water columnsEudorina sp.0.570.37%0.180.03%H1Dinitrogen-fixing NostocaleansAnabaena sp., Aphanizomenon sp.0.250.16%0.560.08%JShallow, enriched lakes ponds and riversCoelastrum sp., Crucigenia sp.4.612.95%34.174.90%KShort, nutrient-rich columnsPediastrum sp., Scenedesmus sp., Tetraedron sp., Tetrastrum sp.––1.550.22%LoSummer epilimnia in mesotrophic lakesChroococcus limneticus, Coelosphaerium kutzingianum, Merismopedia sp., Ceratium hirundinella, Peridinium sp., Gymnodinium sp.13.388.57%243.5434.95%MDielly mixed layers of small eutrophic, low latitude lakesMicrocystis sp., Oscillatoria sp.119.5676.56%231.4633.22%MPFrequently stirred up, inorganically shallow turbid lakesAchnanthes sp., Cocconeis sp.,Cymbella sp., Cymbella sp., Gomphonema sp., Navicula sp.1.440.92%4.250.61%NMesotrophic epilimniaCosmarium sp.3.642.33%0.190.03%PEutrophic epilimniaClosterium sp., Staurastrum sp., Staurastrum sp., Melosira sp.0.100.06%3.470.50%S1Turbid mixed layersPseudanabaena sp. Limnothrix redekei2.181.40%1.210.17%S2Shallow, turbid mixed layersRaphidiopsis sp.––1.090.16%TcEutrophic standing waters, or slowflowing rivers with emergent macrophytesGloeocapsa punctata, Phormidium sp.0.760.48%––W1Small organic pondsEuglena sp., Phacus sp.0.290.19%0.150.02%W2Shallow mesotrophic lakesTrachelomonas sp.0.030.02%0.080.01%X1Shallow mixed layers in enriched conditionsAnkistrodesmus sp., Schroederia sp., Chlorella vulgaris1.891.21%0.690.10%X2Shallow, clear mixed layers in meso-eutrophic lakesChlamydomonas sp.0.010.01%0.030.00%YUsually, small, enriched lakesCryptomonas sp.0.640.41%0.050.01%“–”represent those species were not observed in this year. Functional group codes follow Reynolds et al.8. Table 3Abundance (D, × 105cells/L) and Percent relative abundance of phytoplankton FGs observed in the Caohai Lake.Functional groupsHabitatSpeciesJuly 2018July 2023DRelative abundanceDRelative abundanceCMixed, eutrophic small-medium lakesCyclotella sp.0.920.59%31.274.49%DShallow, enriched turbid waters, including riversFragilaria sp., Nitzschia sp., Synedra sp.1.150.73%3.240.47%EUsually small, oligotrophic, base poor lakes or heterotrophic pondsDinobryon sp.0.010.01%0.310.04%FClear epilimniaMicractinium pusillum, Nephrocytium sp., Oocystis sp., Sphaerocystis sp., Elakatothrix sp.0.830.53%8.541.23%GShort, nutrient-rich water columnsEudorina sp.0.570.37%0.180.03%H1Dinitrogen-fixing NostocaleansAnabaena sp., Aphanizomenon sp.0.250.16%0.560.08%JShallow, enriched lakes ponds and riversCoelastrum sp., Crucigenia sp.4.612.95%34.174.90%KShort, nutrient-rich columnsPediastrum sp., Scenedesmus sp., Tetraedron sp., Tetrastrum sp.––1.550.22%LoSummer epilimnia in mesotrophic lakesChroococcus limneticus, Coelosphaerium kutzingianum, Merismopedia sp., Ceratium hirundinella, Peridinium sp., Gymnodinium sp.13.388.57%243.5434.95%MDielly mixed layers of small eutrophic, low latitude lakesMicrocystis sp., Oscillatoria sp.119.5676.56%231.4633.22%MPFrequently stirred up, inorganically shallow turbid lakesAchnanthes sp., Cocconeis sp.,Cymbella sp., Cymbella sp., Gomphonema sp., Navicula sp.1.440.92%4.250.61%NMesotrophic epilimniaCosmarium sp.3.642.33%0.190.03%PEutrophic epilimniaClosterium sp., Staurastrum sp., Staurastrum sp., Melosira sp.0.100.06%3.470.50%S1Turbid mixed layersPseudanabaena sp. Limnothrix redekei2.181.40%1.210.17%S2Shallow, turbid mixed layersRaphidiopsis sp.––1.090.16%TcEutrophic standing waters, or slowflowing rivers with emergent macrophytesGloeocapsa punctata, Phormidium sp.0.760.48%––W1Small organic pondsEuglena sp., Phacus sp.0.290.19%0.150.02%W2Shallow mesotrophic lakesTrachelomonas sp.0.030.02%0.080.01%X1Shallow mixed layers in enriched conditionsAnkistrodesmus sp., Schroederia sp., Chlorella vulgaris1.891.21%0.690.10%X2Shallow, clear mixed layers in meso-eutrophic lakesChlamydomonas sp.0.010.01%0.030.00%YUsually, small, enriched lakesCryptomonas sp.0.640.41%0.050.01%“–”represent those species were not observed in this year. Functional group codes follow Reynolds et al.8. Abundance (D, × 105cells/L) and Percent relative abundance of phytoplankton FGs observed in the Caohai Lake. “–”represent those species were not observed in this year. Functional group codes follow Reynolds et al.8. The correlation of water quality parameters and FGs showed a significantly strong relationship among and with the selected water quality parameters (Fig. 6). The most dominant FGs (Lo/M) were negatively influenced by pH and DO. At the same time, they showed a positive correlation with the EC, Sal, TDS, TN, NH3-N, TP, Tur, and such conditions prevailed as favorable at July 2023.Fig. 6Correlation analysis on the phytoplankton FGs and environmental variables observed in the Caohai Lake. The vertical axis represents phytoplankton functional groups, and the horizontal axis represents environmental factors. Correlation analysis on the phytoplankton FGs and environmental variables observed in the Caohai Lake. The vertical axis represents phytoplankton functional groups, and the horizontal axis represents environmental factors. Discussion Physiochemical variables It has been widely agreed that a natural lake has two distinct stable states, clean and turbid26. Many studies focus on these two different states and have given detailed descriptions of each26,27. The assessment of water quality of Caohai lake is critical for wholesomeness of the aquatic fauna and flora in transition between clean state and turbid state28. In this study, the water temperature, pH, and DO were stable in July 2018 and July2023, the mean value of water temperature were 22.61 °C and 21.43 °C, the mean value of pH were 8.62 and 8.06, the mean value of DO were 6.49 and 5.78 respectively. Although the DO decrease (from 6.49 to 5.78 mg/L) was not statistically significant (P > 0.05), it represents a 10.9% reduction. For shallow lakes, DO concentrations below 6 mg/L can affect the metabolic activity of aerobic organisms and promote the release of phosphorus from sediments, indirectly facilitating cyanobacterial proliferation—a key feature of the observed regime shift. This suggests the DO decrease, while subtle, may contribute to the lake’s transition to a turbid state29. EC, TDS, and Salinity are characterized by a remarkably variation which could be attributed to high evaporation rate of the surface water and dilution by precipitation30,31. Salinity is concluded could affect the diversity of phytoplankton30. In this investigation, a significant correlation relationship was observed among the different classes of phytoplankton with EC, TDS and salinity. As shown in the investigation, the concentrations of micronutrients were at the high level which indicates obvious turbid water bodies and serious water eutrophication to the system in July 2023. Total nitrogen, total phosphate, and ammonia nitrogen are widely used as indicators of eutrophication which positively corelate with the presence of phytoplankton32. Naderian et al.33 compiled a rich database, which includes primary water quality constituents (selected based on data availability), as well as morphometric characteristics and geographical information of Earth’s lakes. This database was built through extensive searches in accessible repositories, published articles, publicly available maps, and both private and governmental reports, and contains 866,390 water quality data points from 13,876 lakes located in 77 different countries, along with 55,629 morphometric characteristic data points. By comparing with global datasets, we found that the post-shift TN (1.99 mg/L) and TP (0.10 mg/L) in Caohai lake exceed the 85th percentile of global lakes, and its Chl a (13.86 μg/L) exceeds the 70th percentile, indicating that the nutrient enrichment of Caohai lake and algal bloom risks are consistent with global eutrophication trends. Post-shift conditions (high turbidity, elevated TN/TP) of Caohai lake are consistent with global findings that nutrient enrichment and reduced light availability promote cyanobacterial dominance, and that nutrient-to-chlorophyll conversion efficiency increases with turbidity34. In Caohai lake, increased turbidity post-regime shift reduces light penetration, favoring cyanobacteria over green algae and diatoms. Additionally, the lake’s shallow depth enhances mixing, which benefits FGs with tolerance to turbulent conditions (e.g., Lo, M), further reinforcing cyanobacterial dominance. The main phytoplankton FGs in July 2018 and July 2023 showed a positive correlation with the measured nutrients, example FGs Lo with reactive TN (r = 0.87*, p < 0.05), Lo with reactive TP (r = 0.73*, p < 0.05), Lo with reactive NH3-N (r = 0.78*, p < 0.05). FGs M with reactive TN (r = 0.56*, p < 0.05), M with reactive TP (r = 0.46*, p < 0.05), M with reactive NH3-N (r = 0.42*, p < 0.05) (Fig. 6). Turbidity is a parameter that is used in measuring water clarity. Turbidity is used to see the extent of the scattering of light in water. Thus, turbidity can reveal some suspended solids, algae, organic matter, and other very small particles that cause the liquid to become cloudy35. While, these parameters reflect the combined effect of the processes influencing turbidity—and thus affecting potential phytoplankton bloom development36. In this study, a significant difference in turbidity and Chl a were found in July 2018 and July 2023. The presence of Chl a is an indicator of phytoplankton biomass, Cyanobacteria, Bacillariophyceae, and Chlorophyceae show a positive correlation with Chl a (Fig. 5C). Patterns of phytoplankton assemblages The total phytoplankton abundance in the Caohai lake demonstrated remarkable temporal variations, with a significantly higher total cell density (58.86 × 107 cells/L) observed in July 2023, while the lowest (15.48 × 107 cells/L) in July 2018. The maximum total cell density recorded in July 2023, which reached the highest level, may be attributed to the regime shift in Caohai lake and the subsequent increase in EC, TDS, Sal, TN, TP, NH3-N, and Turbidity37. Nutrient enrichment was potentially accountable for the higher phytoplankton abundance in Caohai lake21. Since 2020, a large area of submerged macrophytes in Caohai lake has died out, and there have been obvious turbid water bodies and serious water eutrophication, as a result, the aquatic ecosystem has been degraded19. Once submerged macrophytes decline and die, the sediment is more likely to resuspend under the disturbance of wind and waves. This accelerates the nutrient cycling at the sediment–water interface, making it easier for the aquatic ecosystem to maintain a turbid water state38 .Therefore, wind and waves after the regime shift triggered the increase of the nutrient load in the water column. This could be one of the principal reasons for a higher enrichment of the Cyanobacteria group after the regime shift in July 2023 during this study. Before the regime shift, the phytoplankton density was relatively high at sites 3, 6, 9, and 12 in the riparian zone near county, while it was relatively low at the other sites. After the regime shift, the phytoplankton density increased at all sites and became more uniform. Phytoplankton assemblages and diversity Changes in the community structure, abundance, and biomass of phytoplankton can usually reflect the comprehensive environmental conditions and productivity level of an ecosystem. Therefore, phytoplankton plays an indicative role in evaluating environmental quality, estimating fishery productivity, assessing aquatic productivity, and monitoring water quality39. A two-dimensional MDS based on the Bray–Curtis similarity index divided the phytoplankton abundance into two distinct groups prior to (in July 2018) and after (in July 2023) ecosystem degradation. SIMPER analysis exhibited an overall average dissimilarity of 53.30% in July 2018 and July 2023, the five functional groups (FGs) that contributed most significantly to this difference are Lo (29.42%), M (17.78%), C (11.34%), J (9.04%), and F (5.16%). For July 2018, the average similarity stands at 63.47%, with the FGs that potentially act as the most significant contributors to this similarity being M (37.68%), Lo (14.03%), and J (10.05%). When it comes to July 2023, the average similarity reaches 85.51%, and the FGs that may be the most notable contributors to this similarity are Lo (30.14%), M (29.31%), and C (11.44%). Trends in the functional groups and water quality Categorizing phytoplankton communities into functional groups (FGs) and morphology-based functional groups (MBFGs) serves as a crucial tool for comprehending the behavioral patterns and species dynamics of these communities in response to environmental conditions40. Based on the habitat characteristics and morphological features of phytoplankton, Reynolds8 proposed a relatively comprehensive phytoplankton functional group classification method. The theory of this method is to classify algae that grow in the same habitat type and have similar sensitivity and tolerance into the same functional group. Padisák et al.10 further supplemented and improved the FG classification method in 2009. Compared with the traditional Linnaean homology classification method, the functional groups classification method makes it easier to assess the potential responses of phytoplankton to environmental changes, is more accurate in determining species evolution, and simplifies the complexity of the traditional biological classification system. The representative FGs in July 2018 were Lo/M/N/S1/X1. However, after the ecosystem degradation in July 2023, the dominant FGs were characterized as C/F/J/Lo/M. Lo/M could be interpreted as the most abundant functional group in this study. The habitat template of Lo was deep and shallow, oligo to eutrophic, medium to large lakes. The habitat template of M was eutrophic to hypertrophic, small- to medium-sized water bodies10. The codon Lo include Chroococcus limneticus, Coelosphaerium kutzingianum, Merismopedia sp., Ceratium hirundinella, Peridinium sp., Gymnodinium sp. The codon M include Microcystis sp., Oscillatoria sp. The relative abundance of codon Lo was 8.57%, the relative abundance of codon M was 76.56% in July 2018. While the relative abundance of codon Lo was 34.95%, the relative abundance of codon M was 33.22% in July 2023. The abundance of Lo and M increased after the ecosystem degradation. The relative abundance of codon Lo increased since the mass proliferation of Merismopedia sp. The dominance of FGs C, F, J, Lo, and M in the post-shift turbid state is driven by their adaptive traits. FG M (e.g., Microcystis sp.) possesses gas vacuoles that enable buoyancy regulation, allowing them to access light in turbid waters. FG Lo (e.g., Chroococcus limneticus) are small, non-motile, and have high nutrient uptake rates, thriving in nutrient-rich, mixed environments. FG C (e.g., Cyclotella sp.) and F (e.g., Micractinium pusillum) are shade-tolerant and can form colonies to compete for light and nutrients, also thriving in nutrient-rich, mixed environments. FG J (e.g., Coelastrum sp.) can settle into low light and thrive in shallow, enriched lakes. These traits align with the post-shift environment (high TN/TP, low light, frequent mixing), explaining their dominance8,41. Caohai lake is a hydrologically dynamic ecosystem. By using functional approaches to define the phytoplankton communities, we showed that before and after the regime shift, Caohai lake were two distinct habitats11. Before the regime shift, EC, TDS, Sal, TN, TP, NH3-N, Turbidity and Chl a were lower, while after the regime shift, EC, TDS, Sal, TN, TP, NH3-N, Turbidity, and Chl a have all increased. During the study period, the FGs (C/F/J/Lo/M) benefited from increased nutrient loads in the water column, these codons prefer eutrophic or turbid habitats8, while the FGs (N/S1/X1) appeared only before the regime shift and preferred habitats with high light levels and nutrient deficiency. The RDA plot of July 2018 and July 2023 also supported this conclusion. Since aquatic ecosystem had been degraded, the Caohai lake was enriched with nutrients after the regime shift, the lake turned into a turbid, nutrient-rich, and eutrophic waterbody, favoring an increased enrichment of FGs like C, F, J, Lo and M. Several researchers also reported similar observations10. Correlation analysis demonstrated a consistent relationship between functional groups (FGs) and water quality parameters, which aligns with the results previously reported by Calijuri et al.42, Donald et al.43. These researchers have noted that water temperature exerts a negative effect on codon D, while exhibiting a positive association with TN. Additionally, elevated TP were found to be beneficial for the growth of codon M, a result that is consistent with the observations reported by Okogwu and Ugwumba12. Hence, the assessment of phytoplankton functional groups (FGs) in Caohai lake revealed that the aggravation of eutrophication subsequent to the regime shift was the most significant factor influencing the phytoplankton community. The nutrient regime was crucial for shaping the phytoplankton community, which is consistent with the situation in other eutrophic water bodies around the world44. Conclusion This study was performed in a lake transformed from a clear water steady state to a turbid water steady state, which impacted by various factors, including EC, TDS, Sal, TN, TP, NH3-N, Turbidity. These various factors together give rise to an ecosystem suitable for certain phytoplankton communities and their corresponding functional groups. In conclusion, before the regime shift, EC, TDS, Sal, TN, TP, NH3-N, turbidity, and Chl a were at relatively low levels; in contrast, after the regime shift, EC, TDS, Sal, TN, TP, NH3-N, turbidity, and Chl a all showed an upward trend. Throughout the study period, the FGs (C/F/J/Lo/M) gained advantages from the elevated nutrient loads in the water column, and these functional groups tend to favor eutrophic or turbid habitat conditions. Results also showed a significant difference in the phytoplankton community before and after the regime shift. Again, assessment of FGs in the Caohai lake and its correlation with environmental variables exposed that the aggravation of eutrophication was the most critical factor influencing the phytoplankton community. This study is limited to summer (July) sampling in 2018 and 2023, so seasonal dynamics are not captured. Future research should include multi-seasonal sampling to better understand annual variations in phytoplankton functional groups post-regime shift.