Seagrass is protected from ragworm pressure by a newly discovered grazer–ragworm interaction; implications for restoration

Coastal ecosystems (e.g. coral reefs, seagrass beds, mangrove forests, salt marshes, oyster reefs) are often formed by foundation species, supporting biodiversity, coastal protection, carbon sequestration, nutrient cycling and recreation (Angelini et al., 2016; Whitfield, 2017). They are, however, in rapid global decline, associated with eutrophication, climate change, development overfishing (Waycott 2009). These anthropogenic disturbances generally also affect ecosystem structure interspecific interactions (Östman 2016). More specifically, they can alter trophic interactions, bottom-up (resource-driven) top-down (consumer-driven) non-trophic which together regulate the stability of these (Garbary 2014; Moksnes 2008; Östman Valdemarsen 2011; Whalen 2013). Changes key can, turn, cascade into further destabilisation degradation eventually loss functioning. Clear examples changing their effects, be found many shaped habitat-forming species. In kelp overexploitation sea otters relieved urchins from predator pressure resulted elimination macroalgae (Estes 1978). marches, blue crabs released marsh periwinkles predation, resulting higher grazing on plants led to barren mudflats (Silliman & Bertness, 2002). predation mesopredators mesograzers allow for algae blooms occur compete light nutrients Positive mainly established via habitat-modification capabilities species (Borst 2018; van der Heide 2011), affected factors as well. For example, removal reefs inhibit long-distance facilitating effect making them more prone wave energy (Gribben 2019; Meyer 1997). destabilised systems, negative become dominant well; seed burial lugworms Arenicola marina inhibits recruitment recently degraded systems (Valdemarsen 2011) introduced invasive shore Carcinus maenas clip shoots 2014). Although highly important, much about interaction between among ecosystem-associated remains unknown. A discovered eelgrass Zostera L. (hereinafter referred Zostera) ragworm Platynereis dumerilii Platynereis) showed that efforts restore Lake Grevelingen were severely inhibited ‘glueing’ behaviour (Cronau 2022). is globally distributed meadows (Daly, 1973; Gambi 2000; Jacobs Pierson, 1979; Riera 2020). known glue algal thalli construction dwelling tubes, used a refuge against predators 2000). At our restoration site, we glueing leaves significantly hampered planting unit survival increased epiphyte growth The latter likely explained called gardening, where attaches epiphytes its tube later consumption structural consolidation (Bedford Moore, 1985; It speculated this has because caused drastic shifts composition Grevelingen. That is, large predatory fish decreased while polychaeta benthic (Tangelder 2019). Moreover, gastropods Hydrobia ulvae Littorina littorea Littorina), both important grazers, have almost disappeared (Hoeksema, similar those other (Moksnes As increasingly being recognised crucial conservation (Gagnon 2020; Sievers 2022), mechanisms responsible should clarified. recognising an early stage could prove timely management or success. To mechanistically understand how why was affects under enriched non-enriched conditions, conducted field experiment mesocosm laboratory experiment. We tested following hypotheses: (1) leaf-glueing negatively performance, (2) enrichment (bottom-up forcing) increases growth, thereby impact (3) (top-down decreases cover, counters enrichment. experiment, studied naturally occurring densities addition system all but devoid gastropod grazers. examined interacting effects Platynereis, plant using controlled, full factorial design. Our at Grevelingen, Netherlands (Figure 1a). Covering 10,800 ha, lake former estuary 1971 consists sand banks deep channels. Nowadays, water-level amplitude only 5 cm salinity varies 27.7 31.6 psu throughout due small opening Brouwers dike connecting sea, well surface runoff surrounding land (van Donk 2021). Historically, occurred low abundance, started dominating simultaneously drop shortly after dammed, extending maximum coverage 4600 ha 1978. again late 1990s, most human driven forming bottleneck present population had adapted less saline environment Katwijk 2023). Currently, no seagrasses remained natural recolonisation, ephemeral taken over role primary producers within lake. Whilst numbers declined values below 1 m−2 one numeric reported local average 490 (Sistermans 2006). trends around multiple (Hughes 2004; Monroy-Velázquez Natural variable vary 0 up 2000 (Table S3) 200 (Saier, 2000) subtidal areas. Both inhabit littoral areas, relying hard substrate, depths ranging 3 m deep, respectively, observed 80 (Gambi Their geographic distribution focused Northern Atlantic Ocean, been show cosmopolitan occurrences continents except Antarctica (GBIF.org, accessed Reproduction varies, widespread distribution, winter spring Europe (Barroso 2007; Bedford 1985). test during circumstances, enclosure containing grazer addition, control no-enclosure treatment two different sites historic presence Zostera; Veermansplaat (VP; 51°44′55.6578″N, 4°0′10.548″W) Sirjansland (SJ; 51°41′19.2228″N, 4°1′6.2358″W) selected 90 110 cm, depth highest historical (Nienhuis 1996). relatively sheltered southwest winds. Permission conducting fieldwork provided Rijkswaterstaat Staatsbosbeheer, managers No ethical approval required experiments presented paper. ramets harvested Limfjord, Denmark (57°0′57.69736″N, 9°29′23.0172″E) approximately 12 June 2019. Ramets carefully collected rinsed sediment. cleaned stored crates covered wet cloths keep moist transport. After transport, aerated basins filled artificial water. next day, units created bundling three iron wire, attaching 15-cm nail. All least living rhizome intact roots. 14–15 2019, when 720 (2160 shoots) evenly 30 plots (15 per site) 24 (72 m−2. Planting planted × pattern first slot replaced long rebar marker. Each plot enclosed cage (1.13 m2, high), constructed perforated (5 mm holes) aluminium sheet. cages pushed 10 sediment anchored 1.5 rebars. Next, five site given adding 250 specimens 1b), based Limfjord meadow improve productivity (Jaschinski Sommer, 2011). Plots visited monthly snorkelling minimal disruption. replenished counts dropped 100 plot. density, units, glued 1c), total leaf count quantified. Littorina, eutrophication combinations Zostera, manipulated (P), (N) (L). This eight treatments: (C), P, N, L, PN, PL, NL PNL four aquaria (35 20 50 cm) replicates. treatments received 18 aquarium 1d), density 257 m−2, ranges Nutrient concentration μmol L−1 PO4, 45 NH4 15 NO3 NaH2PO4, NH4CL NaNO3, water levels available data S1; waterinfo.rws.nl). 1e), Seagrass transported October 2019 same method location Within donor bed, handpicked closed container transportation, placed reservoirs seawater (Tropic Marin Inc.; 28 psu) acclimatisation (18°C; 14 h light, dark). Artificial 160–205 s−1 (photosynthetic photon flux density; mid-depths aquaria) reach requirements optimal (Olesen Sand-Jensen, 1993). 23 October, handpicking free-floating macroalgae, Sargassum muticum. removed 12°C shells serving substrate food source. Aquarium consisted 90% 10% (28 introduce epiphytes. Salinity maintained stable demineralized counter evaporation. applied continuously through flowthrough Masterflex pumps refreshed column weekly. equipped CO2-pump pH (7.5–7.9 pH) balance CO2/HCO3− potential photosynthetic capacity. Constant circulation guaranteed pump circulating passed UVC-lamp prevent phytoplankton S1). randomised block design 4 blocks. single apical shoot minimum leaves, including young (<10 rhizome. Wet weight ramet measured groups six gently layer sieved (1 mesh) Subsequently, capped 1-cm silver reduce leakage. Finally, according size, added reservoirs. Every ~10 days, monitored number bite marks, percentage cover ramet. checked daily adjusted necessary. stable, dead stock. On day 70, terminated roots washed, razor blade. Plants dissected below-ground above-ground parts, counted. Biomass dried 70°C 48 dry measurement. first, analysed generalised linear mixed model (GLMM) time explanatory variables random account repeated observations time. Second, likewise, replacing omitting non-cage model. Third, last monitoring round (GLM) variable. averaged GLMMs. terms included variables, incorporated run bites aquarium. Only Responsive (i.e. ramet, weight) LMs variables. Laboratory analyses followed posteriori multiple-comparison (Tukey HSD) Benjamini–Hochberg corrections glht function multcomp package (Hothorn 2008). Distributions GL(M)Ms family. log +1 transformed (upper bound − x) case upper-bound left-skewed necessary meet assumptions corresponding statistical test. Poisson models overdispersion. P-values retrieved ANOVA Anova command (car package; Fox Weisberg, 2018) L(M)M GL(M)M, Type III analysis terms. tests considered significant alpha level p ≤ 0.05. Means text figures standard error (SE). Statistical r v4.2.0 (R Core Team, contained epiphyte-covered tubes. found. However, time, slightly moment second (F(1, 35) = 5.45; 0.025). When considering cage-containing plots, high sites, mean (± SE) 61.14 ± 4.28% (n 19), 61.04 20) SJ VP, constant period strongly reduced 22.8) 15.43, 0.007; Figure 2a) 21.3 6.6%. end demonstrated 58.3% (Chi 9.91, 0.002) compared controls 2b). Additionally, 10.2 15.2 1.1 plots. Epiphyte (F(6, 144) 67.41; < 0.001) virtually clean 27.00 1.57% (range: 0%–90%) 3a; Table S2). 24) 10.51, 0.003), 167.97, able fully halt increase until 3.20, 0.006; F(6, 8.77, 0.001). Nutrients especially weeks it plateaued, suppressed slight few weeks. immediate leaves. 6 diverge 12.82) 18.75, 0.001; 3b; Most importantly, leaf-gluing 47.21, mitigating even stronger Littorina–nutrient 4.91, 0.045). During took noticed t 27 days 4.12, 0.043), interacted 5.25, 0.022), cancelling out effect. differed 4a), 5.662, 0.026). 9.183, 0.006), countered 20.08, 4.31, 0.049) 16.36, countering 4b). parameters, Similarly previous 23.87, 19.71, adverse 4c). Contrasting response variables; 9.82, 0.005) conditions. results controlled newly Platynereis. mitigate presently largely line hypothesis confirmed lab Here, protects distinct pathways: regulated forcing hampering Platynereis–Zostera interaction. suggest potentially plays North wide disturbances. findings emphasise importance healthy mesograzer co-introduction fast-growing positively Extensive (Littler 2006; 2013), mesocosms alike 5). Top-down grazers relieve 2013) ecosystems, communities, maintaining Littler meadows, favours release competition 2004). (nutrient enrichment) (grazer absence) destabilise favour expense simulated study, entangled still attached 1c,d). previously 1985) study accordance Daly (1973), began matter minutes lab. treatments. not treatments, contrasting Cronau al. (2022). lack gardening diverse community did consume half-moon-shaped bites, typical 2000), treatment, lowest, suggesting partial diet shift becoming limiting. lab, weight. Glueing directly effective surface, mobility, spacing 1c,e). argue this, consequences acquisition (Sand-Jensen, 1977). Limited nutrition above- (Carroll 2008) shedding (McMahon field, decrease surviving showing entanglement harsher Species harm problematic already disrupted environmental conditions disruptors (Eriander visible absence allowing damaging activity. inclusion mitigates seagrass. experiments, reducing glueing. positive amplified nutrient-enriched offset biomass enrichment, allowed availability releases competitive 2004), loading reveals another, unknown, protective pathway mitigation signs exploitative interference consuming Hughes 2004) and, thus, compete. consumes assemblages unselectively scraping Consequently, paramount abandons every 2–3 relocate 1973), increasing worm. continuous grazing, negates releasing old investigated whether grazes new tubes effectively defends occupied Despite protecting plants, any parameters own Instead, appeared adversely isolation, although significant. An suggested unselective scraping, incidentally destroying outer layers Marsh scar cordgrass reductions Zieman, 2001). Based recorded behaviour, easily place seagrass-associated documented, along top such cod Gadus morhua documented mesopredators, turn populations, (Baden 2012; (co-) introduction would cause decline alleviated. achieved targeted action mremoval population, shown facilitation populations Another recent decades exotic alga S. muticum may enhanced modify epifaunal communities worldwide (Belattmania Hunn 2022; Therefore, options remove Sargassum, investigated. changes described above detected several Baltic Sea Pacific Ocean expect happening entire Atlantic, Interestingly, decrease, whereas Tangelder rise closely intervention concurrently, measures purposes. Recent studies facilitative success (Chin 2021; Temmink Valdez 2020), active rates (Derksen-Hooijberg Gagnon Examples direct include detoxification Lucinidae alleviation drought stress ribbed mussels 2012). studies, that, rather than trying absent area. environments structures, propose three-step strategy, evaluation, decision (2022): Transplant seagrasses, preferably web, co-introduce (historically) native system, (Littorina case) exclude inhibiting organisms manual removal; indirectly case, cascades; predators). regular co-introduced excluded animals exclusion methods implemented decide repetition Rens J. T. Cronau, Tjeerd Bouma, Jannes H. Heusinkveld, Dieuwke Hoeijmakers, Tjisse conceived designed experiment; Yvet Telgenkamp, Marieke M. Katwijk, Heide, Leon P. Lamers Hoeijmakers performed experiments; Jimmy de Fouw, data; wrote and/or edited manuscript. thank Staatsbosbeheer collaborating approving work chosen sites. financed Rijkswaterstaat. Fouw supported NWO Open Competition #ALWOP.203. authors conflict interest declare. Data DANS EASY Repository https://doi.org/10.17026/dans-x38-abzq S1: (a) Setup each individual experimental Grevelingen-sediment cm). CO2/HCO3 balance, connected circulated phytoplankton. reservoir delivered weekly refreshment rate. Excess left drainage hole (height (b) consisting ramets. exchange fauna. S2: Number represented medians quartile means moments since 27th non-nutrient ragworms, i.e. (P) (PL) Platynereis:Littorina, panel (PN) Platynereis:Nutrient (PNL) Platynereis:Nutrient:Littorina, right panel. Significant differences indicated letters Tukey post hoc comparisons. Mean concentrations September 3). Post (%) S3: across systems. Please note: publisher content functionality information supplied authors. Any queries (other missing content) directed author article.