The ongoing global biodiversity loss caused by multi-stressors has raised concerns about the potential consequences of species extinctions on the functionality and stability of ecosystems. It has been widely recognized that biodiversity could stabilize the ecosystem by enhancing the asynchronous dynamics among species within a single trophic level. However, in natural multi-trophic ecosystems, asynchronous changes between trophic levels (trophic asynchrony) are expected to increase trophic mismatch and alter trophic interactions, which may consequently alter ecosystem stability. Under changeable environments, it is currently unclear how biodiversity in different trophic levels affects the stability of food web by changing the trophic interactions (represented by resource use efficiency, RUE) and trophic asynchrony, which is key to providing further insights into the biodiversity effects on ecosystem stability.

Using a 6-month mesocosm experiment, we tested how multi-stressors—microplastics, eutrophication, dissolved organic carbon, and invasive fish—affected species richness and RUE of phytoplankton (producer) and zooplankton (consumer), subsequently influencing temporal stability of food web via altering the trophic asynchrony.

Our results demonstrated that multi-stressors could shape ecosystem temporal stability through species richness-RUE-trophic asynchrony pathways, emphasizing the cascading impacts of multi-stressors on ecosystem functions and stability. Increasing species diversity of phytoplankton and zooplankton enhanced the RUE within that trophic level. However, greater producer diversity decreased RUE of consumers, probably due to an increase in inedible producer species. This suggests that benefits of biodiversity of producer could diminish functions at higher trophic levels. In addition, zooplankton RUE increased trophic asynchrony, whereas phytoplankton RUE decreased it. Crucially, trophic asynchrony stabilized phytoplankton and food web but destabilized zooplankton community, highlighting contrasting impacts of trophic asynchrony on stability across trophic levels.

Our study showed that producer diversity may indirectly decrease the functions and stability of consumers under multiple stressors, which challenged the simplistic expectation that more diversity universally stabilized ecosystems. This study provided a new insight into the biodiversity-ecosystem stability relationships by highlighting the mediating roles of RUE and trophic asynchrony under realistic environmental change scenarios.

1.1. Experimental design

Our research was conducted in the Mesocosm Control Experiment System at Chenggong Campus of Yunnan University (25.054058 N, 102.702144 E) from July to November 2023. Forty plastic buckets (diameter 1.58m, height 1.45m) were randomly distributed in the experimental site. Each bottom of the plastic bucket was covered with 100 mm thick sediment from Dianchi Lake (24.800556 N, 102.671389 E). Each tank was filled with 10% Dianchi Lake water and 90% tap water. Each bucket was cultivated with an equal number of Potamogeton maackianus, Vallisneria natans, Myriophyllum spicatum, and Hydrilla verticillata. Following the planting, the mesocosms were maintained for a 2-month period to allow the plants to stabilize.

We set up a control group (C) and seven treatment groups, including NP (nitrogen and phosphorus, by adding NaNO₃ and K₂HPO₄ 2 mg L^-1 and 0.15 mg L^-1 respectively every two weeks), MP (microplastics, by adding high-density polyethylene 5 g), F (fish, by adding 20 Pseudorasbora parva, fishes were replenished daily based on the observed mortality), DOC (dissolved organic carbon, by adding mixture of 4:1 C molar ratio sodium humate (C₉H₈Na₂O₄) and fulvic acid (C₁₄H₁₂O₈) 15 mg L^-1 every two weeks) and MP related treatments (MP+NP, MP+F, MP+DOC), with a total of 3 combinations. There were five replicates for each control or treatment groups, which were randomly assigned to forty water tanks.

We collected phytoplankton and zooplankton once a month. Phytoplankton was collected from 1 liter of surface water and then 1.5 ml of concentrated Lugo’s solution was added to each bottle. We then concentrated the sample to 30~60 ml by siphoning after standing for 48 h. We employed 13# plankton nets (aperture size is 112 µm) to gather zooplankton from a 20-liter volume of surface water. Subsequently, the collected samples were transferred into 80 ml square bottles, wherein 2 ml of formaldehyde (Shanghai Wukai Biotechnology Co., LTD. in China) was added to fix the zooplankton.

1.2. Laboratory analysis and microscopic examination

Total nitrogen (TN) and total phosphorus (TP) in water were measured according to standard methods. Specifically, the samples were first digested with potassium persulfate (Sigma Aldrich, Germany) in a sterilized pot (GR85DA, Zealway Instrument Inc., America) at 120 ℃ for 30 minutes.

For phytoplankton, 0.1 ml of concentrated solution was placed in the phytoplankton counting frame under microscopy at 200-400 magnification and each piece was counted 100 visual fields. Each sample was counted twice, and then the average value was calculated. The difference between each count should be within 15%. Finally, the calculation of density was based on the following formula:

N=(A/A_c)*(V_w/V)**n

where, A and A_c are counting frame area and counting area (mm²); V_w and V are the volume (ml) of concentrated solution and counting frame, respectively; n is the number of individuals. We then used N to calculate the phytoplankton biomass by standard methods.

Zooplankton was counted using 5 ml counting boxes and 40 magnification (Olympus CX41) and the body length (mm) was measured using the mesh micrometer (scale of the mesh micrometer was 0.025 mm) for subsequent biomass calculations. The specific process was as follows: the biomass of crustacean zooplankton was calculated according to the empirical regression equation (W = qL^b) of body length (L) and weight (W).

1.3. Calculation of species richness, RUE, trophic asynchrony and stability

Zooplankton and phytoplankton species richness were calculated by the “specnumber” function of the “vegan” package. The RUE of phytoplankton for nitrogen or phosphorus (Phytoplankton RUE_TN and RUE_TP) was calculated as the ratio of phytoplankton biomass (mg L^-1) to TN or TP (mg L^-1), respectively, which was commonly used in previous studies. Zooplankton RUE was calculated as the ratio of zooplankton to phytoplankton biomass (mg L^-1).

We calculated the statistical metric of trophic asynchrony according to Eq (1). Stability of zooplankton, phytoplankton and food web were calculated according to Eq (2).

   Trophic asynchrony =  1-(σ_total²)/(σ_zooplankton²+σ_{phytoplankton}²)                       Eq (1)

where, σ_total is the standard deviation of total biomass of zooplankton and phytoplankton. σ_zooplankton and σ_zooplankton are the standard deviation of zooplankton and phytoplankton biomass, respectively.

          Stability =   u/σ                                                                                   Eq (2)

where u and σ represent the mean and standard deviation of the total biomass of zooplankton, phytoplankton or the sum of zooplankton and phytoplankton, respectively. Earlier work highlighted that producer diversity was a key driver of whole-food-web stability and that its importance increased with trophic complexity. Consequently, it is reasonable to assign producers a relatively large weight rather than manually setting equal weights to zooplankton when assessing food web temporal stability, and using the total biomass to quantify temporal stability was already a common practice.