Coastal seas are highly involved in fundamental biogeochemical processes controlling nutrient and organic matter cycling (Middelburg and Herman 2007; Bouwman et al. 2013; Carstensen et al. 2020). They receive nutrients and organic matter from both marine (e.g. primary production) and terrestrial (e.g. riverine inputs of terrestrial material, anthropogenic outfalls) origins and act as a filter between both realms (Asmala et al. 2017). In addition to climate change, human activities influence nutrient cycling through eutrophication, resulting in changes in both absolute values and ratios between nitrogen (N) and phosphorus (P). Major changes in nutrient supply are likely to affect primary producers’ requirements with regard to elemental building blocks, and have repercussions on processes regulating consumers elemental homeostasis (ecological stoichiometry sensu Sterner and Elser 2002) with consequences for food web functioning and biogeochemical cycling.
Archived biological samples from environmental monitoring programs can be retrospectively analysed for elemental composition (C, N, P) and stable isotope ratios of carbon and nitrogen (δ13C, δ15N) to study nutrient cycling and reconstruct food webs in relation to a changing environment. Carbon isotopes provide information about ultimate carbon sources for primary production (Fry and Sherr 1984), while nitrogen isotopes can be used to trace specific nitrogen sources, such as anthropogenic (e.g. Connolly et al. 2013) or diazotrophic sources (Rolff 2000; Karlson et al. 2015), or to quantify trophic position in consumers (Vander Zanden and Rasmussen 1999). Organisms at the base of the food web such as filter-feeding bivalves or grazing snails, with low motility and long-life span, can be used as proxies for isotope baselines (i.e. the ultimate C and N sources) since they integrate intra-annual variability of nutrients (Vander Zanden and Rasmussen 1999; Post 2002). Perennial macrophytes typically reflect nutrient sources during the growth period (e.g. Savage and Elmgren 2004), hence the relatively fast turnover rates of macrophyte tissue may allow for the detection of subtle variations in N sources compared consumers such as bivalves, which provide an integrated measure of N signal over longer time spans, several months or even year(s) (Post 2002). Whether primary producers and primary consumers reflect large-scale and decadal changes in nutrient conditions in a similar manner is largely unknown.
In the Baltic Sea, the suspension-feeding blue mussel Mytilus edulis trossulus species complex (Kijewski et al. 2006; Stuckas et al. 2009; hereafter referred to as Mytilus or blue mussel) and the ephemeral filamentous green algae from the genus Cladophora are highly abundant. Mytilus occur in densities of up to ~ 100 000 individuals m− 2 (Westerbom et al. 2008) and constitute, through their efficient suspension-feeding, an important link between the pelagic and benthic ecosystem, promoting nutrient cycling (Kautsky and Wallentinus 1980; Kautsky and Evans 1987; Attard et al. 2020). Cladophora is widely distributed in the Baltic Sea (Zulkifly et al. 2013), and mainly occurs from the surface down to 1 or 2 m depths. It is perennial, but overwinters as a small tuft attached to shallow rocky substrates. During summer it reaches its full growth and benefits from nutrient enrichment (Thybo-christesen et al. 1993). Isotope composition in bivalves, including Mytilus, is commonly used to study pelagic organic matter origin (Magni et al. 2013) and its variability over time and space (Briant et al. 2018; Corman et al. 2018). Bivalves are considered suitable baselines for food web studies (e.g. Abrantes and Sheaves 2009; Willis et al. 2017) and in contaminant monitoring (Karlson and Faxneld 2021). Cladophora is mainly used as a proxy of nutrient levels in coastal waters, in the Baltic as well as elsewhere (Mäkinen and Aulio 1986; Planas et al. 1996).
The pronounced latitudinal gradients of temperature, salinity and nutrients combined with the historical excessive anthropogenic nutrient inputs in the Baltic Sea provide an ideal study system to link environmental and nutrient conditions with those measured in archived mussels and algae. Riverine inputs of organic carbon and nutrients have increased in the recent time period, especially in the Bothnian Sea (Wikner and Andersson 2012; Asmala et al. 2019), and this ‘brownification’ increase is expected to continue (Andersson et al. 2015a). In the Bothnian Sea, the N:P ratio of the dissolved inorganic pool is similar to the Redfield molar ratio of 16 (ca. 13, although the slight N limitation has increased in recent years; Rolff and Elfwing 2015), while it is considerably lower in the Baltic Proper (ca. 4), indicating strong N limitation (Savchuk 2018). Diazotrophic primary producers, such as some species of bloom forming cyanobacteria, can bypass this N-limitation by directly fixing N2. Satellite images of surface accumulations indicate that these blooms have increased since the 1980s (Kahru and Elmgren 2014), and this internal N loading now exceeds external loadings from rivers and atmospheric deposition in the Baltic proper (Olofsson et al. 2020). Cyanobacterial blooms benefit from denitrification and phosphate release from hypoxic sediments, which exacerbate the N:P imbalance in a ‘vicious cycle’ (Vahtera et al. 2007). The most recent decade has also seen regular occurrence of cyanobacterial blooms in the Bothnian Sea (Olofsson et al. 2020).
Salinity and temperature are both lower in the Bothnian Sea compared to the Baltic Proper and are the primary factors affecting species distribution, including that of Cladophora and Mytilus. Predicted increase in temperature and decrease in surface salinity of the Baltic Sea (Räisänen 2017 and references therein) are hence expected to affect organisms, food webs and ecosystems (Andersson et al. 2015a; Vuorinen et al. 2015). Recent studies have shown a decrease in mussel populations over recent decades in the southern Baltic Proper, linked to a changing environment (Franz et al. 2019; Westerbom et al. 2019; Liénart et al. 2020). A shift in dominance from the canopy-forming perennial macrophyte Fucus towards opportunistic ephemeral Cladophora has been reported since the 1980s in different areas of the Baltic Sea (Kraufvelin and Salovius 2004), likely linked to eutrophication (Kautsky et al. 1986; Råberg 2004; Torn et al. 2006). However, Fucus recovery has been observed recently in some areas (Rinne and Salovius-laurén 2020). Nonetheless, higher temperature and declining salinity promote filamentous green algae (Takolander et al. 2017), suggesting climate change will enforce the shift towards ephemeral macrophytes in the Baltic.
Tracing C and N origin in the Baltic Sea is complex, due to multiple interacting sources, especially in the coastal area. For instance, in the Baltic Proper, eutrophication is associated with elevated δ15N values in sediment (Voss et al. 2000) and sewage waters enriched in 15N, which is traceable in macrophytes (Savage and Elmgren 2004). However, the depleted 15N signal of synthetic N fertilizers used in agriculture (Bateman and Kelly 2007) can be confounded with the similar signal of diazotrophic cyanobacteria (Rolff 2000), the latter considered an indirect effect of eutrophication (Vahtera et al. 2007). In the Bothnian Sea, the naturally low δ15N-NO3 from pristine rivers (Voss et al. 2005) should equally be reflected in relatively depleted 15N baselines. The typically low δ13C of terrestrial carbon from the extensive riverine input in the north (Rolff and Elmgren 2000) is similarly expected to be reflected in a low δ13C baseline in the Bothnian Sea. However, low temperatures and light availability (the latter from brownification) can also result in lower δ13C values for macrophytes (Wiencke and Fischer 1990; Hemminga and Mateo 1996). In the Baltic proper, higher δ13C may reflect eutrophication due to increasing plankton biomass (Oczkowski et al. 2018) with the exception of cyanobacteria (generally low δ13C; Rolff 2000). In addition, an organism’s physiology (e.g. rapid growth, nutritional stress, reproductive stages) can lead to substantial isotope variability in consumers (Doi et al. 2017; Gorokhova 2018). Osmoregulation is an especially N-demanding process for Mytilus experiencing low-saline conditions in the Baltic Sea (Tedengren and Kautsky 1987). This likely influences its 15N fractionation and hence confounds the dietary origin of the δ15N signal in the mussels (Liénart et al. 2020). To better trace ultimate N sources, the end-member δ15N signal can be measured in source amino acids (e.g. phenylalanine), which show very little 15N fractionation during assimilation, physiological processes or trophic transfer (McClelland and Montoya 2002) compared to the bulk δ15N signal. Finally, N and P elemental composition is expected to be driven by nutrient background and taxa elemental requirements. The fast-growing algae, which are under little homeostatic control compared to slow growing consumers like mussels (Smaal and Vonck 1997), are expected to reflect basin-specific nutrient conditions, with higher nutrient concentrations in the N limited but nutrient rich Baltic Proper than in the Bothnian Sea. N and P concentrations in mussels might also reflect this difference, but in the low-saline Bothnian Sea this might be confounded by high N requirements during osmotic stress.
In this study, we take advantage of the high temporal resolution of pelagic monitoring data and the archived macroalgae and mussel samples from phytobenthic monitoring in the Baltic Sea to explore whether eutrophication and climate-related changes are mirrored in these key taxa. We retrospectively analyse elemental (C, N and P) and stable isotope (δ13C, δ15N) composition of archived samples of the summer growth of the filamentous ephemeral algae Cladophora and the slow-growing sessile blue mussel Mytilus from three contrasting regions (coastal Bothnian Sea, coastal Baltic Proper and open sea Baltic Proper), spanning 8 to 24 years data depending on region. We first document potential differences and temporal changes in region-specific elemental and isotope baselines, and compare between baselines for algae and mussels. Further, we compare bulk δ15N composition with a smaller data set of δ15N in source amino acids. Then, we test whether is it possible to explain the observed year-to-year variability and long-term trends with environmental and oceanographic data using various statistical approaches. Finally, we link biomass data for both taxa with the stable isotope and environmental data.
We expect elemental and isotope composition of both taxa to reflect the latitudinal gradient in nutrients, with higher values of δ13C, δ15N, N% and P% in the Baltic proper compared to the Bothnian Sea. We expect Cladophora to better reflect nutrient background in its elemental composition than Mytilus, due to more homeostatic control of its elemental composition. Finally, we expect Cladophora and Mytilus to respond differently to environmental and oceanographic variables, and reflect different processes in their ecology and physiology.