Environmental stressors, complex interactions and marine benthic communities' responses

Abstract

The increasing number and diversity of anthropogenic stressors in marine habitats have multiple negative impacts on biological systems, biodiversity and ecosystem functions. Methods to assess cumulative effects include experimental manipulations, which may identify non-linear responses (i.e. synergies, antagonisms). However, experiments designed to test these ideas are uncommon, generally focusing on single biological responses. We conducted a manipulative experiment to investigate the isolated and combined effects of warming (+ 6 °C), salinity variation (freshwater pulses or presses), and nutrient enrichment (natural or enriched) following one and three month's exposure, on responses measured at multiple levels of biological complexity in a simple bivalve assemblage. More specifically, we determined effects on bivalve mortality, growth, shell mineralization, and energy content, as well as microphytobenthos biomass. Salinity variation and nutrient enrichment, individually and combined, caused strong impacts on some of the measured variables and their effect varied through time. In contrast, warming had no effect. Our work highlights the prevalence of antagonistic interactions, the importance of examining effects of single and multiple stressors through time, and of considering multiple responses to understand the complexity behind stressor interactions.

Figure 1

Effect of salinity variation through time on microphytobenthic standing stock. Average chlorophyll a concentration (μg g⁻¹ dry mass) for salinity variation following one and three months of exposure. Pair-wise contrasts showed a significant difference between one and three-months of exposure within for freshwater presses (p < 0.05) as indicated by the asterisk. Histograms represent mean values, and error bars correspond to standard errors.

Figure 2

Effect of the interaction between salinity variation and nutrient enrichment on mortality levels in Mytilus sp. (a) Average mortality level (%) for salinity treatments (S) and nutrient concentrations (N) in Mytilus sp. Pair-wise contrasts showed a significant difference between pulses and freshwater presses with natural nutrient concentrations (p < 0.05) as identified by the asterisk. Histograms represent the mean value, and error bars correspond to standard errors. (b) Positive antagonistic interaction between freshwater presses and nutrient enrichment based on a null multiplicative model (yellow) ((S + N) − (S × N)). Single effects of stressors, compared to the least stressful conditions, are shown in light blue and gray and a combined effect is shown in dark gray. Error bars correspond to standard errors.

Figure 3

Effect of the interaction between salinity variation and nutrient enrichment through time on magnesium content in the shell of L. balthica. (a) Average magnesium content (mmol of Mg²⁺ kg⁻¹ dry mass) for salinity treatments (S) and nutrient concentrations (N) at one and three months of exposure in L. balthica. Pair-wise contrasts showed a significant difference between the two exposure times for different combinations of salinity variation and nutrient conditions (p < 0.05) as indicated by asterisks. Histograms represent mean values, and error bars correspond to standard errors. (b) Positive and negative antagonistic interaction between freshwater presses and nutrient enrichment through time based on a null additive model (yellow) (S + N). Single effects of stressors, compared to the least stressful conditions, are shown in light blue and gray and combined effect is shown in dark gray. Error bars correspond to standard errors.

Figure 4

Effect of nutrient enrichment through time on energy content in the tissues of L. balthica. Average energy content (kJ g⁻¹ dry mass) for nutrient conditions following one and three months exposure. Pair-wise contrasts showed a significant difference between one and three months of exposure under enriched nutrient conditions (p < 0.05) as indicated by the asterisk. Histograms represent mean values, and error bars correspond to standard errors.