Food webs Thematic Assessment

Impacts of climate change on food webs

Climate change represents one of the main factors currently affecting marine ecosystems and their feeding interactions. Declines observed in the biomass and abundance of phytoplankton and zooplankton might have important implications for upper trophic levels due to cascading effects. Additionally, changes in the spatial distribution patterns and displacement of marine species, towards either northern areas or greater depths, represent an important driver that is affecting food web interactions. Besides triggering ocean warming, climate change has increased the terrestrial dissolved organic carbon load into coastal ecosystems, altering primary productivity, and has increased ocean acidification, with effects on shell-forming species. Synergies between these stressors may contribute to ecosystem food web stabilization and alter food web energy flow and material cycling (Chapman et al.,2020).

The warming of the Arctic Waters has resulted in substantial changes in the ecosystem (ICES ecoregion overviews for the sub-regions Greenland Sea, Icelandic Waters, Norwegian Sea, Barents Sea and the Arctic Ocean) (Berge et al., 2012; Edwards et al., 2016; Fossheim et al., 2015; González-Pola et al., 2019; IPCC 2019; Jørgensen et al., 2022; Kunisch et al., 2020; Stige et al., 2019). Some Arctic endemic and ice-associated species are more strongly impacted. For example, the polar cod population has declined, owing in part to the loss of sea ice, its spawning habitat, in the Region. In the meantime, large predators or generalist species such as the north-east Arctic cod have progressed northwards across the whole region. The record high cod stock is in need of wider feeding space. As the ice has contracted, the area available to the north and north-east of Svalbard has now become additional feeding grounds for cod. Both near eastern Greenland and in the Iceland Waters and Norwegian Sea, temperate fish species like mackerel and blue-finned tuna, and sea birds like eider ducks, are also extending their northern ranges. Minke whales are declining near Iceland, probably owing to less food availability. The thinning and reduction in ice cover has resulted in changes in ice biota communities. These include a reduction in the territory and ice conditions needed for seals to give birth and increased risk of killer whale attack. It is feared that the impacts on Arctic Sea mammal species, both seals and whales, will lead to major changes in food webs (ICES WGICA 2018; ICES WGIBAR 2021; ICES WGINOR 2022). 

Climate change is expected to affect primary production rates and trends but also the composition of phyto and zooplankton species, which may trigger potential changes in trophic interactions. Since the 2017 Intermediate Assessment, several new analyses of primary production sensitivity to environmental controlling variables have been published for the North Sea and/or the North Atlantic in the context of climate change and involving comparison of computational models. Temperature, light climate, mixing layer depth, nutrient availability, and grazing pressure of secondary herbivorous producers are the variables most frequently mentioned to explain the dynamics of primary productivity calculated in relation to climate change scenarios. Overall, this work shows a decreasing trend in primary productivity for the North Atlantic (Tagliabue et al., 2021) and the North Sea (Capuzzo et al., 2018). Concerning the latter ecosystem, there is uncertainty about the type of model to use in order to accurately reproduce primary production (Spence et al., 2022). The decreasing trend in primary productivity characterized for several OSPAR Regions is consistent with the results of the global model of Lotze et al., (2019), while contrasting with the results from the ecosystem modelling by Thorpe et al. (2022) for the North Sea (upward trend). As summarized in Sathyendranath et al., (2020), there are still uncertainties about the nature of the climate change-related primary production trend, both for ecosystem models in climate change studies and for models based on satellite data. This latest study highlights the accuracy of satellite-based calculations and discusses whether it could be improved by more intense measurements of key parameters and processes (such as underwater light field, chlorophyll-carbon ratio θ, photosynthesis-irradiance parameters and photoacclimation). Finally, the recent study by Spence et al., (2022) recommends, consistent with the conclusions of Sathyendranath et al., (2020) and the work of Capuzzo et al., (2018), developing regional models including processes occurring over short time scales (biological parameters cited above). Similarly, Thorpe et al., (2022) recommend that further work be done to improve the representation of bottom-up processes to ensure that ecosystem models can capture limitation by nitrogen and other elements, and not only food/energy uptake, particularly in higher trophic level models developed to understand the impact of fisheries.

It should be noted that the work of Capuzzo et al., (2018) as well as the Candidate Indicator Pilot Assessment of Primary Productivity, which use some empirical models based on chlorophyll-a (Chla) and light (with local parametrization), show downward trends that can actually vary across hydrodynamic regions of the North Sea. Negative correlations are also obtained between this process and temperature, while positive correlations are highlighted with the smallest copepods (herbivores, from CPR records) as well as an index of recruitment from seven commercially valuable fish species (sandeel, sprat, herring, Norway pout, cod, haddock and whiting from the ICES stock assessment report, 2016). This work has therefore concluded in favour of bottom-up control (from physics to plankton, to planktivorous fish) of the food web, as did the previous work of Lynam et al., (2017). In conjunction with declining food web base biomass, increased energy demand from organisms, related to increased temperature, may also explain the decrease in fish biomass (Bryndum-Buchholz et al., 2019; Carozza  et al., 2019; Lotze et al., 2019). In addition, it can be caused by predator-prey decoupling between primary and secondary production. This will cause the phytoplankton to sink to the sea bottom and then be lost to secondary production (Morrison et al., 2019).

Changes in the spatial distribution patterns of marine species are being affected by global warming and increasing temperatures at an unprecedented rate in marine environments (Cheung et al., 2009; Burrows et al., 2011; Punzón et al., 2020). The distribution ranges of marine species are shifting towards higher latitudes in processes of tropicalization (Horta e Costa  et al., 2014), meridionalization (Punzón et al., 2016) and borealization (Fossheim et al., 2015). These regions are becoming increasingly dominated by species with warmer affinities (Lenoir et al., 2011; Simpson et al., 2011). 

The displacement of species toward greater depths is an additional change detected in the distribution of those species that seek a better niche condition (Dulvy et al., 2008; Hofstede et al., 2010). With temperatures displaying marked spatial gradients in latitude, longitude and depth, the thermal niche arises as a fundamental factor in the distribution of fish species (Righton et al., 2010; Bruge et al., 2016; Kleisner et al., 2017). 

Changes in the distribution of species, whether in latitude, longitude or depth, may cause changes in predator-prey interactions. Food webs should be seen as networks where all compartments are interconnected. Therefore, the disappearance of specific prey for some predators and/or changes in the distribution of top predators could lead to changes in trophic interactions and thus in ecosystem structure (in the trophic level of the species) and, ultimately, functioning. As the species in an ecosystem have different environmental preferences, climate change does not impact them homogeneously. This means that the trophic network will be impacted in different ways by climate change depending on species composition, but also on temporal and spatial dynamics (Nogues et al., 2022). The potentially strong structuring effect of climate change on both the functioning and the spatial and temporal organization of ecosystems thus calls for climate variables (e.g. seawater temperature, salinity, current velocity, precipitation, change in winds, depth of thermocline) to be implemented in the models, which will shed light on the impact that climate change may have on trophic levels at the population, community and ecosystem levels.

Global warming has been shown to be a vector that enhances invasions of a marine environment by new species, since most of them originate from warmer oceanic and coastal regions. Food web models using Ecological Network Analysis indices show that the arrival of non-indigenous species change the fate of organic matter within the ecosystem, with higher cycling, relative ascendency and a chain-like food web. It causes new trophic interactions in the food web that could lead to competition and thus modify food-web structure and functioning, with lower omnivory and higher detritivory (Le Marchand et al., 2022). Results from ENA indices show a decline in trophic efficiency which can be linked to a possible ecosystem shift caused by invasive species (Baird et al., 2012). Studies by Jung et al., (2020) used ENA to establish that increasing biomass of the invasive Atlantic jackknife clam (Ensis leei) coincided with a 70% increase of trophic carbon transfer from primary to secondary producers and an 80% increase from secondary producers to detritus in a western Wadden Sea food web. Carbon flows from secondary producers to higher trophic levels were reduced by more than 60% (Jung et al., 2020). Baird et al., (2019) used ENA to evaluate networks and their quantitative trophic interactions between living and non-living components at different temperature scenarios. Results from the trophic analysis revealed that detritivores clearly showed persistent increases in higher temperature scenarios, as did the Detritivory: Herbivory ratio. A similar trend was observed in the amount of detrital material returned to the detrital pool from the integer trophic levels, suggesting a shift in the system’s trophic activity to lower trophic levels. The main reason for these trophic shifts was the increased metabolic and katabolic physiological processes which are expected to occur at increased ambient temperatures (Baird et al., 2019).

In Nogues et al., (2022), ENA was applied to a spatialized food web model (Ecospace model) of the Seine Bay to assess the impact of changing species distribution. This showed a significant impact by climate change for many ENA indices like trophic efficiency, system omnivory, system recycling and the relative redundancy of trophic pathways. The effect of climate change in the Seine Bay differed depending on the studied time frame, but the results tended to indicate that in the long run (2100), climate change will have a negative effect on ecosystem functioning by reducing its ability to resist external perturbations (Nogues et al., 2022). Moreover, climate change seems to have a structuring effect on the ecosystem of the Seine Bay, in terms of both network topology and spatial organization. Changes in ecosystem functioning may also impact the way in which other subsequent ecosystem drivers impact trophic network. This could lead to unexpected cumulative effects on ecosystems, where climate change is combined with other drivers. In this connection, numerous non-additive effects were observed on ENA indices when climate change was combined with the reef and reserve effect of an operating offshore wind farm in the Seine Bay (Nogues et al., 2021).

The impacts of climate change on the marine ecosystem are expected to reach the deepest parts of the ocean. Projections point to significant changes in deep-water mass properties worldwide (Sweetman et al., 2017). In the North Atlantic, such changes could include a decrease in seawater temperature, a loss of dissolved oxygen, a decrease in the flux of particulate organic matter to the seafloor, a decrease in pH and saturation of the minerals involved in the carbon cycle (e.g. calcite and aragonite) (Gehlen et al., 2014; Sweetman et al., 2017; Perez et al., 2018). The multitude of impacts caused by changes in these environmental properties are difficult to predict. Nevertheless, several studies have demonstrated climate-driven effects on the productivity, biodiversity, metabolism and distribution of various deep-sea organisms (Levin & Le Bris, 2015; Levin et al., 2019; Xavier et al., 2019; Morato et al., 2020; Puerta et al., 2020). Climate-driven changes in the spatial distribution of deep-sea fauna raise important questions about associated ecosystem-wide impacts. For example, climate change could reduce suitable habitat for cold-water corals and deep-sea fishes, forcing them to shift to northern latitudes (Morato et al., 2020). As this shift occurs, new interactions could emerge as immigrant and resident species from northern habitats interact (Kortsch et al., 2015). These interactions could result in new food web configurations that might affect ecosystem functioning, as already documented (Beaugrand et al., 2015; Blanchard et al., 2015). Thus, it is critical to channel research efforts into seeking an understanding (among other questions) (a) of how poleward shifts in deep-sea fauna might affect food web structure and (b) of the extent to which such changes might affect the ability of the deep-sea ecosystem to provide essential services.

An important step towards answering such questions is to improve current capabilities for modelling deep-sea ecosystems through network analysis or trophic models. To date, the development of complex, holistic modelling tools specific to deep-sea systems has been limited (Morato et al., 2016; Woodstock et al., 2022). This is primarily because these models are very data-intensive, which naturally limits their application to traditionally data-poor systems such as deep-sea ecosystems. New information on the life histories, feeding preferences and habitat use of different deep-sea species/functional groups is therefore of paramount importance to operationalize the development of models suitable for examining the ecosystem-wide impacts of climate change on deep-sea ecosystems. This is one of the priorities of the modelling efforts currently underway in the Azores.

Climate warming has significantly altered the phenology of a wide range of taxa across ecosystems, but responses frequently vary among species, potentially disrupting the synchronisation of key trophic interactions in the food web. In particular, failure of a predator to overlap the period of peak resource demand (typically breeding) with peak prey availability may lead to ‘trophic mismatch,’ and such decoupling may alter food web structure and ecosystems. One example is mismatch between the timing of seabird prey occurrence and periods of peak energy-demand (e.g. chick-rearing) (Burthe et al., 2012). 

Besides ocean warming, ocean acidification can have an impact on food web interactions, with consumers expected to be affected by changes to the nutritional quality of their prey (Rossoll et al., 2012). Anthropogenic CO₂ can function as a resource that boosts productivity throughout food webs, while warming can reverse this effect by acting as a stressor to trophic interactions (Goldenberg et al., 2017).