Edited by: Peter J. Lammers, Arizona State University, United States
Reviewed by: C.-Elisa Schaum, University of Hamburg, Germany; Maria Stockenreiter, Ludwig Maximilian University of Munich, Germany
†These authors have contributed equally to this work
This article was submitted to Marine and Freshwater Plants, a section of the journal Frontiers in Plant Science
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Periphyton plays an important role in lake ecosystems processes, especially at low and intermediate nutrient levels where periphyton contribution to primary production can be similar to or exceed that of phytoplankton. Knowledge of how periphyton responds to key drivers such as climate change and nutrient enrichment is, therefore, crucial. We conducted a series of mesocosm experiments over four seasons to elucidate the responses of periphyton communities to nutrient (low and high, TN-0.33 mg L–1 TP-7.1 μg L–1 and TN-2.40 mg L–1 TP-165 μg L–1, respectively), temperature (ambient, IPCC A2 scenario and A2 + 50%) and plant type (two submerged macrophytes with different morphological structural complexity:
Periphyton plays an important functional role in lake nutrient cycles and food webs, especially at low and intermediate nutrient levels (
Nutrient and temperature are among the most important factors that drive biological processes and limit primary production of shallow lakes at a global scale (
Habitat structure complexity, here expressed as the architecture and morphological characteristics of the host that provides attachment sites for periphyton, can influence the establishment and development of periphyton (
We studied variations in the biomass development and composition of periphyton during four seasons in an existing mesocosm facility which has been running since 2003 (
The mesocosm system we used is situated in Central Jutland, Denmark (56°140N, 9°310E) and includes 24 experimental tanks, which each is 1 m deep and has a diameter of 1.9 m (
A factorial-design consisting of three temperature scenarios and two nutrient levels (six blocks in total) is applied to the mesocosms, each with four replicates. Among the 24 mesocosms, 8 are unheated (Ambient, AMB), while 8 are heated according to IPCC climate scenario A2 (warming, W) and 8 according to A2 + 50% (enhanced warming, EW). The temperature increase in the A2 scenario was calculated as the mean air temperature in one particular month with reference to period 1961 to 1990 and the temperature in the same month in 2017 to 2100 according to the IPCC climate model A2, the values being 2.74°C in May, 3.84°C in August, 3.76°C in November and 2.76°C in February (
Each temperature level is crossed with two nutrient levels. In this experiment, the mean concentrations of nutrient were TN-0.33 mg L–1 and TP-7.1 μg L–1 in the clear-state and TN-2.40 mg L–1 and TP-165 μg L–1 in the turbid-state. No additional nutrients are added to the clear-state mesocosms with low-nutrient levels, while nutrients are added to the turbid-state mesocosms to maintain a constant loading of 27.1 mg N m–2 day–1 and 7 mg P m–2 day–1. Nutrient addition was initiated in May 2003 when freshwater communities were established in the mesocosms, while heating was started in August 2003 when submerged macrophyte beds dominated in the non-enriched mesocosms and phytoplankton or filamentous algae dominated in the enriched mesocosms. Submerged macrophytes, mainly
Our study was conducted in the mesocosms described above which had been running for 15 years when our experiment was conducted. Microorganisms in the mesocosms might have evolved and adapted to the current condition during this time period, which makes the experiment more realistic than if study was conducted in mesocosm that had just been established as is often the case. Our experiment was designed to test the responses of periphyton to nutrient, temperature and substrate type. Four types of substrate were selected for periphyton cultivation: the submerged macrophytes
Four identical experiments were conducted in May 2017, August 2017, November 2017, and February 2018, representing the four seasons (spring, summer, autumn, and winter) in Denmark. A 3-week experiment of periphyton cultivation was carried out independently in each season. In the beginning of each experiment, four types of plant substrate — natural
Schematic representations of the detailed experimental design. White-triangles represent natural
After 3 weeks’ cultivation, the four types of plant substrate were harvested from each tank. All above-ground natural/artificial plants were cut off and taken back to the laboratory. Here, the periphyton and zooplankton (Cladocera, Rotifera, and Copepoda) attached to each substrate were rinsed off carefully and extracted in 1000 ml water. Once rinsed, the plant substrates were placed in a plastic bag and digital photographs were taken. Photographs were scanned in ImageJ software and their surface area were estimated to twice that of the scanned area (
Environmental variables were measured
Three-way ANOVA and
All the periphyton density data were square-root transformed and zooplankton data (attached to substrates) were 0–1 transformed to reduce the influence of a few dominant taxa, and a Bray–Curtis matrix of similarity among treatments was constructed. Two-way nested Analysis of similarities (ANOSIM, with treatment, plant type nested within treatment as the fixed factors) was conducted to compare the responses of periphyton and epiphytic zooplankton composition to treatment and plant type. To visually present possible differences, non-metric multidimensional scaling (NMDS) was performed using the R (version 3.6.3) vegan package. Besides, in order to expediently describe the seasonal variability in periphyton community structure, periphyton genera found in each season were classified into functional groups according to
The relationship between environmental variables and periphyton genus composition in each season was analyzed in R (version 3.6.3) vegan package. For all periphyton data, the largest gradient length of detrended correspondence analysis was less than 3 standard deviation units. Redundancy analysis (RDA) was therefore considered most appropriate. RDA was carried out seasonally in the clear-low-nutrient state and the turbid-high-nutrient state using periphyton density data as response variable and environmental data as explanatory variables. Periphyton density data were Hellinger-transformed and all the environmental data were lg(
During the experiment, the water temperature differed significantly among the three temperature levels in each season (
In May and August, periphyton Chl
Results of three-way ANOVA comparing the effects of nutrient (two levels: low and high), temperature (three levels: ambient, warming, and enhanced warming), nutrient*temperature and plant type (four levels: natural
d.f. | May |
August |
November |
February |
|||||
Nutrient (N) | 1 | 20.588 | 7.582 | 0.116 | 0.734 | 0.208 | 0.650 | ||
Temperature (T) | 2 | 0.299 | 0.742 | 5.970 | 1.963 | 0.148 | 0.797 | 0.455 | |
Nutrient*temperature | 2 | 2.124 | 0.127 | 0.748 | 0.477 | 1.358 | 0.264 | 2.131 | 0.126 |
Nutrient*temperature (plant type) | 18 | 0.664 | 0.834 | 0.418 | 0.980 | 2.110 | 1.855 | ||
Nutrient (N) | 1 | 7.917 | 15.474 | 0.100 | 0.753 | 1.760 | 0.189 | ||
Temperature (T) | 2 | 0.813 | 0.448 | 11.234 | 0.245 | 0.784 | 0.598 | 0.553 | |
Nutrient*temperature | 2 | 1.431 | 0.246 | 3.569 | 0.161 | 0.852 | 1.478 | 0.235 | |
Nutrient*temperature (plant type) | 18 | 0.889 | 0.593 | 1.510 | 0.112 | 1.843 | 4.015 |
Periphyton Chl
Periphyton density in May was strongly influenced by nutrient, being significantly higher in the turbid-high-nutrient state than in the clear-low-nutrient state (
Periphyton density on four types of plant substrate in each treatment divided into season.
In total, 78 periphyton genera (20 Bacillariophyta, 39 Chlorophyta, 8 Cyanophyta, and 11 others) were observed. Bacillariophyta were the dominant phyla and accounted for more than 60% of the abundance in each season (
The genus composition of periphyton differed significantly among six treatments (
Results of two-way nested ANOSIM comparing the effects of treatment and plant type (nested in treatment) on the genus composition of periphyton.
d.f. | May |
August |
November |
February |
|||||
Global R | Global R | Global R | Global R | ||||||
Treatment | 5 | 0.169 | 0.488 | 0.534 | 0.126 | ||||
Treatment (plant type) | 3 | −0.077 | 0.958 | −0.073 | 0.953 | −0.044 | 0.811 | 0.296 |
Non-metric multidimensional scaling (NMDS) plots of periphyton communities in six treatments based on Bray–Curtis similarities calculated from genus density (square-root transformed).
The genus richness of periphyton was significantly higher at ambient temperature (mean value 18.53) than in the two warming scenarios (mean values 16.34 in warming scenario and 15.97 in the enhanced warming scenario, respectively) in August, while being significantly lower at warming temperature (mean value 12.28) than at enhanced warming temperature (mean value 15.03) and ambient temperature (mean value 15.19) in November. The Shannon–Wiener index of periphyton in May was notably lower in the turbid-high-nutrient state (mean value 2.35) than in the clear-low-nutrient state (mean value 2.48), while in August it was notably higher in the turbid-high-nutrient state (mean value 2.54) than in the clear-low-nutrient state (mean value 2.39). Significant plant type effects on the Shannon–Wiener index were found in May and February (
Redundancy analysis revealed that the seasonal periphyton composition was significantly affected by different environmental variables in the clear-low-nutrient state and the turbid-high-nutrient state (
Tri-plots of RDA results showing the differences in periphyton genera, treatments and environmental variables between the clear-low-nutrient state and the turbid-high-nutrient state in
During the four seasons, a total of 23 Reynolds functional groups were identified, differing in importance among seasons (
Apart from periphyton, 60 taxa of epiphytic zooplankton (12 Cladocera, 27 Rotifera, and 21 Copepoda) were also found in our study. The density of Cladocera in May was significantly higher in the turbid-high-nutrient state than in the clear-low-nutrient state, while the density of rotifers in August was significantly higher at ambient than at warming temperatures (A2 and A2 + 50%) (two-way ANOVA,
Our study showed that periphyton abundance (Chl
A significant effect of temperature on periphyton abundance was only found in summer where Chl
Plant-type had a significant effect on periphyton abundances in autumn and winter, periphyton Chl
The ANOSIM results confirmed that periphyton composition was sensitive to the interaction between nutrient and temperature. Among seasons, the genus composition differed significantly between any two treatments except between W and EW, W&NP, and EW&NP. This suggests that the temperature difference between scenario A2 and A2 + 50% (mean value 1.60°C) was too low to affect the periphyton composition in both the clear-low-nutrient state and the turbid-high-nutrient state. Similar results have been reported by
The RDA results revealed that temperature was only an important factor affecting the periphyton composition in autumn and winter (November and February) and that there was no significant difference between the clear-low-nutrient state and the turbid-high-nutrient state. Winter is normally associated with environmental minima, such as low temperatures associated with reduced solar radiation and low light intensity, the latter due to ice and snow cover, providing a harsh environment for primary producers (
The RDA results for spring and summer revealed that the environmental variables explaining periphyton composition differed notably between the clear-low-nutrient state and the turbid-high-nutrient state. The RDA showed that periphyton composition could best be explained by rotifers in the clear-low-nutrient state and by turbidity or TP in the turbid-high-nutrient state. These results concur with those of previous studies demonstrating that the grazing effects of herbivores on periphyton are weakened by high nutrient availability (
Multiple studies have reported that periphyton community composition is strongly influenced by seasonality (
We found a complex response of periphyton abundance to nutrient enrichment, temperature increase and the structure complexity of substrate, which varied with season. Periphyton abundance was strongly affected by nutrient addition and the temperature rise in spring and summer, whereas substrate structure was of great importance for periphyton abundance in autumn and winter when environmental conditions were harsh for periphyton growth. In contrast, the community composition of periphyton was significantly influenced by the interactions between nutrient and temperature and independent of seasonality. Our results suggest that the effect of warming on periphyton abundance and composition in the different seasons varied with nutrient state and host plant type in the shallow-lake mesocosms, and similar results are likely to occur in natural lakes.
The raw data supporting the conclusions of this article will be made available by the authors, to any qualified researcher with reasonable request.
This study did not involve listed endangered or protected species. No specific permits were required for this study.
BH designed the study, performed the research, analyzed the data, and wrote the manuscript. HW performed the research, collected samples, and analyzed the data. WZ and HJ collected samples. YC, EJ, and WL contributed to the data analysis and revisions. All authors have reviewed, discussed and agreed to the authorship and submission of the manuscript for peer review.
WZ was employed by the company Wuhan Planning & Design Co., Ltd., Wuhan, China. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank Anne Mette Poulsen for English edition.
The Supplementary Material for this article can be found online at:
Illustration and photo of the 24 flow-through experimental mesocosm setup studying how increased temperature will affect shallow lake systems (
Relative abundances of Bachillariophyta (Bac), Chlorophyta (Chl), Cyanophyta (Cya) and others (Oth) in periphyton on four types of plant substrate in each treatment on a seasonal basis. CON represents ambient temperature without nutrient addition; W represents warming without nutrient addition; EW represents enhanced warming without nutrient addition; NP represents ambient temperature with nutrient addition; W&NP represents warming with nutrient addition; EW&NP represents enhanced warming with nutrient addition.
Functional group composition and density of periphyton in May (spring), August (summer), November (autumn) and February (winter).
Results of two-way ANOVA comparing the effects of nutrient (two levels: low and high; d.f. = 1), temperature (three levels: ambient, warming and enhanced warming; d.f. = 2) and the interaction between nutrient and temperature (d.f. = 2) on environmental variables. All bolded values in the tables are significant at
Results of three-way ANOVA comparing the effects of nutrient (two levels: low and high), temperature (three levels: ambient, warming and enhanced warming), nutrient*temperature and plant type (four levels: natural
Results of Pair-wise ANOSIM tests for treatment effects (nutrient*temperature) in May, August, November and February. CON represents ambient temperature without nutrient addition; W represents warming without nutrient addition; EW represents enhanced warming without nutrient addition; NP represents ambient temperature with nutrient addition; W&NP represents warming with nutrient addition; EW&NP represents enhanced warming with nutrient addition. All bolded values in the tables are significant at
Results of three-way ANOVA comparing the effects of nutrient (two levels: low and high), temperature (three levels: ambient, warming and enhanced warming), nutrient*temperature and plant type (four levels: natural
RDA results of environmental variables selected by forward-selection procedures and permutation tests for the RDA models in the clear-low-nutrient state and the turbid-high-nutrient state in May, August, November and February. P values represent the results of permutation tests for the models, and R2 and R2 adjusted are the cumulative proportions explained by the models. All bolded values in the tables are significant at
Results of two-way PERMANOVA comparing the effects of nutrient (two levels: low and high), temperature (three levels: ambient, warming and enhanced warming) and their interactions on the functional group composition of periphyton. All bolded values in the tables are significant at