Food web rewiring drives long-term compositional differences and late-stressor interactions at the community level

In this blogpost, Francesco Polazzo talks about how pesticides can cause a drastic change in the organisation and strength of interactions of aquatic food webs. Particularly, after pesticide exposure, changes in the strength of trophic interactions seem to be driving long-lasting effects in community structure. When multiple pesticides act together, they modify both food web structure and interaction’s strength, causing late-stage non-additive effects between stressors. 

Natural communities in a stressed world

Ecosystems and communities are increasingly stressed by human activities. Such human-induced pressures can not only alter the number of species living in an ecosystem and their total biomass, but also their identity and relative abundances. Studies have shown that, after a short-term disturbance, the number of species composing a community, as well as their total biomass can be regained rapidly, even after an initially significant decline. However, the same does not hold true for their identity and relative abundance, which might take substantially more time to be recovered or might not be recovered at all. Yet, scientists have to explain the mechanisms driving different recovery patterns at the community level.

Often, anthropogenic activities result in multiple stressors impacting natural communities and ecosystems simultaneously. When these stressors act together, they can result in additive (simple sum of their individual effects), antagonistic (less than the sum), or synergistic (more than the sum) effects. Yet, whether stressors will combine in a non-additive fashion depends on when the effects are measured. Thus, non-additive interactions between stressors are temporal scale dependent, meaning that they may appear only at a later stage, even after cessation of the short-term disturbance. Late synergistic or antagonistc effects have been reported often in natural systems. However, we still lack suitable methods to understand what processes may be driving the (late) non-additive effects of multiple stressors at the community level. Here, we apply for the first time quantitative ecological network analyses to try to elucidate the mechanisms driving long-term community composition dissimilarity and late-stage stressor interactions. The food web analyses used here are based on the measurement of the strength of the interactions linking the species by quantifying  the energy fluxes between them.

Our study design

We used a freshwater mesocosm experiment with a full factorial design to study the effects of multiple stressors on an aquatic food web (Fig. 1). We applied a long-term perturbation of nutrients (P and N, lasting for the whole experimental time) along with short-term perturbations of an insecticide (chlorpyrifos 1 µg/L) and an herbicide (diuron 18 µg/L) and focus on three time points related to the timing of pesticide application: before stress, maximum effects phase, and post-exposure recovery phase. We analysed the response of species richness, total biomass, and the multivariate community composition (considering species identity and abundance simultaneously). Furthermore, we measured how the structure of the food web responded to the stressors, as well as how the interaction’s strength within the food web is affected by the stressors.

Figure 1 Mesocosm facility at the IMDEA Water Institute (CC Francesco Polazzo, IMDEA)

Study highlights

We found that at the moment of maximum effects, the number of species, total biomass, and food web structure were significantly impacted by the stressors in isolation. In the recovery phase, the number of species, total biomass, and food web structure were all recovered. Yet, the multivariate community composition of the communities treated with single pulse application of either of the pesticides were still significantly dissimilar from the control. This means that the relative abundances of the different species were still modified compared to the pre-disturbance conditions. Mirroring this multivariate dissimilarity, also the interaction’s strength between species was significantly modified by the single pesticides in the recovery phase. We identified the reorganisation of interactions strength as the mechanism driving the long-term dissimilarity in community composition. Supporting this conclusion, we found a significant correlation between the relative change in the strength of the species interactions and the relative change in multivariate community composition (Fig. 2).

Figure 2 Spearman’s rank correlation between the Log response Ratio (LRR) of the energy fluxes and the LRR of the Bray – Curtis dissimilarity between the treatments and the control over the whole experimental time (n = 63). The line represents the linear trend between the two variables and the coloured area represents the confidence interval.

Multiple stressors (insecticide  x herbicide) interacted non-additively only in the recovery phase, reducing the total number of species and changing their relative abundances. Analysing the changes in the interaction’s strength between the species in the recovery phase, we found that they were significantly modified by the mixture of the pesticides. We found that the outgoing energy fluxes (our way to measure interaction  strength) in this treatment was dominated (> 80%) by the basal species, whereas top predators strongly declined in both biomass and in the interaction’s strength they exerted, which resulted in a complete reorganisation of the interaction strength in the food web (Fig. 3). Moreover, the reduction of the total number of species changed drastically the structure of the food web.

Here, we show for the first time that the reorganization of the interactions’ strength in a food web, which led to a loss of strong interactions with predators, is sufficient to trigger a complete rewiring (i.e., re-organisation of interaction strength as well as change in structure) of the food web under multiple stress sources. The methods presented here provide opportunities  to unveil the mechanisms in which pesticides and other stressors affect complex freshwater communities in the long-term.

Figure 3 Link-weighted representation of the food web in the recovery phase. The networks, weighted links, and node distribution show how the energy fluxes within the community change in the presence of different stressors. N: nutrients, I: insecticide, H: herbicide.

The paper titled ‘Food web rewiring drives long-term compositional differences and late-disturbance interactions at the community levelwas authored by Francesco Polazzo, Tomás I. Marina, Melina Crettaz-Minaglia, Andreu Rico and has been published in Proceedings of the National Academy of Sciences 119, no. 17 (April 26, 2022): e2117364119.