What Is Regulating Chironomid Populations? The Influence of Food Supply and Interference Competition on Development and Mortality in Chironomus riparius

Abstract

Density-dependent processes are important for a fundamental understanding of population regulation, as well as for understanding responses to and recovery from stressors. While exploitative competition is well-studied, interference competition is rather difficult to investigate, but it has been regularly observed to occur in many aquatic insect populations. We conducted laboratory experiments with the non-biting midge Chironomus riparius (Diptera: Chironomidae) to investigate the impact of different combinations of food supply and larval densities on development and mortality at a constant temperature of 20°C. The chosen two-factorial experimental design allowed a separate evaluation of exploitative (food) and interference (mortality) competition across a gradient of larval densities. The use of different vessel sizes between 50 cm² and 600 cm² made it possible to quantify the functional response at different food densities. To test mechanistic explanations for the statistically significant empirical relationships found in this study and to predict density-dependent processes, we used a dynamic process-oriented modeling approach. We extended a recently developed DEB-IBM full life cycle model for C. riparius and successfully applied it under variable food conditions at the population level under laboratory conditions. Our study showed that chironomid development and reproduction are primarily dependent on food supply, whereas larval density drives the density-dependent mortality rate. The interaction of food availability and interference competition determined the effective mortality over time. Killing by conspecifics was the most likely mechanism responsible for the intraspecific mortality of the larval stages. Combining data generated using a tailor-made experimental design with a mechanistic model provided insights into and quantified regulation mechanisms of chironomid populations, allowing future uses of this information in the context of population-level risk assessment from exposure to chemicals.

Keywords: Diptera; cannibalism; density dependence; dynamic energy budget model DEB; individual‐based model IBM; interference competition.

FIGURE 1

Graphical representation of the two‐factorial experimental design, which combines four food levels combined with five larval densities (10, 20, 40, 100, and 200 larvae per 50 cm⁻² beaker), as well as two larger aquaria (180 and 600 cm²) with 20 larvae each at a food level of 0.25 mg dw per larva and day. A no‐food scenario was also integrated as a control. All scenarios were replicated fourfold.

FIGURE 2

Cumulative emergence success in the beaker experiments (a) as a function of larval density (10, 20, 40, 100, and 200 larvae per 50 cm²) and four food levels (0.05, 0.125, 0.25, and 0.5 mg food larva⁻¹ d⁻¹), and (b) as a function of sediment area with 50, 180 and 600 cm² (20 initial larvae per aquarium 0.25 mg food larva⁻¹ d⁻¹). M ± SD of replicates. *Data not used for further analysis.

FIGURE 3

Regression between (1) EmT₅₀, (2) mortality rate [%/d], (3) mortality at test end [% of the initial number of larvae], (4) egg number per clutch, and food per larva (left, a) or larval density (right, b). Significance level of LM and GLM statistics in brackets. ns, not significant, *p < 0.05; **p < 0.01; ***p < 0.001. Trend model: y = ax^b (trend model coefficients are given in the Supporting Information S1).

FIGURE 4

Cumulative emergence success in measured data (mean values of replicates, symbols) and simulations (mean of 20 Monte Carlo simulations, simple lines). (a) Different larval densities (10, 20, 40, 100, and 200 larvae per 50 cm²) per food level [mg food larva⁻¹ d⁻¹]. (b) Enlarged sediment area with 180 and 600 cm² (20 initial larvae per aquarium at 0.25 mg food larva⁻¹ d⁻¹).

FIGURE 5

Correlation of model‐based predicted versus measured values: (a) EmT₅₀ (data corrected for mortality), (b) % mortality at test end, (c) % emerged adults at test end, (d) mean eggs per clutch (red triangles in brackets are considered as experimental outliers and were not included in the regressions).

FIGURE 6

Measured (left) and simulated (right) dependence of mortality [%] at the end of the experiments [%] on the initial larval density per beaker (50 cm²) and food level ([mg larva⁻¹ day⁻¹], different lines).

FIGURE 7

Comparison of experimental data (Hooper, Sibly, Hutchinson, and Maund 2003) and simulations. (a) Experiment 1 (long‐term study in aquaria). Survival to emergence [%] (black: measured; gray: simulated), separated for the first generation and the following generations from week 5 to 34 at two food levels (hf:high food; lf: low food). (b) Experiment 2 (56‐day study in beakers). Measured (black dots) and simulated (gray dots) percentage of cumulative emerged larvae within 56 days. Gray triangles: simulated % density‐dependent larval mortality. (c) Correlation between predicted and measured percentage of emerged larvae in experiment 1 (black dots) and experiment 2 (white dots).

FIGURE A1

Experimentally derived equation for the density‐dependent mortality rate [% day⁻¹] at a constant food level of 0.25 mg larva⁻¹ day⁻¹.