This authors' personal copy may not be publicly or systematically copied or distributed, or posted on the Open Web,

except with written permission of the copyright holder(s). It may be distributed to interested individuals on request.

Vol. 744: 53–67, 2024 MARINE ECOLOGY PROGRESS SERIES

Published September 5
https://doi.org/10.3354/meps14664 Mar Ecol Prog Ser

Variation in top-down control of red algae epibiosis

in the White Sea
Alexandra Chava1,2,*, Anna Artemieva2, Eugeniy Yakovis2
1
P.P. Shirshov Institute of Oceanology, RAS, Laboratory of Ecology of Coastal Benthic Communities, Moscow 117997, Russia
2
St.-Petersburg State University, Department of Invertebrate Zoology, Saint-Petersburg 199034, Russia

ABSTRACT: Epibiosis is shaped by a complex interplay of biotic interactions involving hosts, epi-
bionts, and mobile consumers. In temperate waters, consumer control by mesograzers prevents
complete overgrowth of seaweeds. In polar waters, the mechanisms determining the abundances of
sessile organisms associated with seaweeds are unknown. We empirically assessed the strength of
the consumer control effect on the colonization of the sub-arctic red seaweed Phycodrys rubens by
caging individual plants in the field in the shallow subtidal of the White Sea (65° N). We compared
epibiosis on plants in consumer exclosure cages, in cages with the mesopredatory shrimp Spironto-
caris phippsii, in semi-enclosed cages, and on unmanipulated plants in a cold year (2014) and a
warm year (2015). Despite the dramatic interannual variation in consumer control, the mean total
cover of epibionts in the absence of consumers never exceeded 15 %. While consumers had a sub-
stantial effect on the total epibiont cover in the warm year and a nearly negligible effect in the cold
year, the total cover of unmanipulated algae was similar in 2014 and 2015. Bryozoans, which were
selectively impacted by consumers — particularly shrimp — dominated in both years. However,
bryozoan abundance was much lower in 2015, when the abundance of hydroids, sponges, and
bivalves — less affected by consumers — increased. Consumer control is not a key factor prevent-
ing most Phycodrys plants from being heavily overgrown. Yet, smaller plants, which have a higher
epibiont cover, may indirectly benefit from consumers. Future climate changes are likely to make
the Phycodrys epibiosis increasingly top-down regulated.

KEY WORDS: Top-down control · Interannual variation · Predation · Epibiosis · Mesopredator ·

Shrimp · Red algae · White Sea
Resale or republication not permitted without written consent of the publisher

1. INTRODUCTION (Connolly & Roughgarden 1999). Though it is widely

accepted that the intensity of biotic interactions
Interspecific biotic interactions are key processes decreases with latitude (Schemske et al. 2009, Baskett
determining the structure and dynamics of natural & Schemske 2018), acute competition for stable hard
communities. Negative interactions such as competi- substrate space has been observed from tropical
tion and consumer control have long been considered (Chadwick & Morrow 2011) and temperate (Nanda-
critical for understanding and predicting patterns of kumar 1996) to polar waters (Konar & Iken 2005).
abundance and diversity in living organisms (Menge Selective predation and grazing have profound direct
& Sutherland 1976). Since most marine benthic com- and indirect effects on species composition and diver-
munities are constituted predominantly by organisms sity by controlling competitively superior species
relying on larval or spore dispersal, the models ap- (Paine 1974) and subordinate consumers (Estes &
plied to these systems have usually incorporated an Palmisano 1974) which, in turn, affect multiple infe-
interplay of larval supply and negative interactions rior competitors and prey species, respectively. Sub-

*Corresponding author: [email protected] © Inter-Research 2024 · www.int-res.com

 54 Mar Ecol Prog Ser 744: 53–67, 2024

stantial consumer (‘top-down’) control has also been temperate shallow subtidal habitats, another common
Author copy

detected beyond tropical and temperate zones (Qui- foliose red alga, Chondrus crispus, is about 50 % cov-
jón & Snelgrove 2005). ered, primarily by ascidians and bryozoans, at the
The growing recognition of positive interspecific sites with a low abundance of mesograzers (snails),
interactions (i.e. facilitation) as an equally powerful and only 5–10 % covered at sites with a high abun-
driver of community structure (Bruno et al. 2003) dance of mesograzers (our approximations based on
prompted recent research on foundation species (FS). epibiont/host weight ratios reported by Stachowicz &
FS such as trees, corals, mussels, seagrasses, and sea- Whitlatch 2005). Importantly, experimental removal
weeds are strong facilitators ameliorating the effects of mesograzers results in a total overgrowth of algae
of environmental stress and negative biotic interac- by sessile epibionts, leading to severe growth deple-
tions for the dependent taxa (Ellison 2019). FS de- tion (Stachowicz & Whitlatch 2005).
crease consumer pressure by providing refuges (Ware In polar subtidal waters, the total cover of epibionts
et al. 2019) and reduce competition by generating hab- on red algae is relatively low, and the mechanisms that
itat space (Stachowicz 2001). In particular, common control fouling intensity and structure are mostly un-
seaweeds develop several square meters of blade sur- known (Chava et al. 2019). In the shallow subtidal of
face per one square meter of the bottom (Teagle et al. the sub-arctic White Sea (65° N), the foliose red alga
2017). Their growing blades are the least space-limited Phycodrys rubens provides substrate for an epiphytic
microhabitat in kelp forests, able to host weaker com- community composed predominantly of bryozoans,
petitors for space (Seed & Harris 1980) and thus poten- serpulid polychaetes, hydroids, and sponges (Chava
tially increasing species diversity. Although some FS et al. 2019). P. rubens has a lifespan of several years
indirectly benefit from their epibionts (Wahl & Hay (Schoschina 1996), and most substrate area provided
1995), algal hosts, when heavily fouled, generally for epibionts is represented by newly grown blades
demonstrate growth depletion and suffer from in- (Chava et al. 2019). These large and distinctly recog-
creased herbivory (D’Antonio 1985, Honkanen & Jor- nizable blades (hereafter referred to as ‘young blades’)
malainen 2005). Therefore, consumer control of epi- emerge in spring, are colonized by recruits of sessile
bionts by predators and grazers usually has a positive organisms in summer and fall, and partly degrade in
effect on algal hosts (Duffy 1990, Stachowicz & Whit- winter, adding to the smaller ‘old’ part of the plant
latch 2005). Overall, biotic interactions behind an ob- comprising 1–3 yr old blades. In fall, when the cover
served epibiont assemblage structure can be fairly of epibionts is highest, it averages 7–10 % on young
complex and difficult to disentangle. and 24–36 % on old blades (Chava et al. 2019). Con-
While seaweeds and seagrasses act as FS in coastal sequently, competition for space on young blades ap-
ecosystems across the globe (Amsler et al. 2014, Olafs- pears negligible.
son 2016), the knowledge of the functioning of their It is unclear whether the low cover of P. rubens
epibioses is mostly limited to tropical and temperate (especially on its newly colonized young blades) is
regions (Duffy 2006, Miller et al. 2015, Moore & Duffy mostly due to consumer control or alternative mech-
2016, Freestone et al. 2020, Lamy et al. 2020; but see anisms, e.g. recruitment limitation. To investigate
Amsler et al. 2014). Manipulating the abundance of this, we manipulated the presence of mobile con-
seagrass- and seaweed-associated mesopredators and sumers on P. rubens blades in a series of field caging
mesograzers in the field is technically complicated experiments. The mesopredatory shrimp Spironto-
and thus infrequently attempted (Whalen et al. 2013). caris phippsii is one of the most common generalist
Experimental research on red algae as hosts of depen- consumers in subtidal White Sea habitats (Grishankov
dent sessile assemblages is especially scarce despite et al. 1997), feeding on various mobile and sessile epi-
their wide distribution (Díaz-Tapia et al. 2018) and benthic organisms such as polychaetes, amphipods,
growing commercial potential (Cabrera et al. 2022). mollusks, and bryozoans (Yakovis & Artemieva 2019,
Red algae can become heavily overgrown, with con- 2021, Chava et al. 2024). We used exclosure cages iso-
sumer control being a key process regulating the few lating individual plants from any mobile organisms
systems studied in this respect. In tropical shallow unable to pass 2 mm mesh to assess the effect of con-
subtidal areas, the foliose red seaweed Cryptonemia sumer control on the epibiosis and enclosures with
seminervis supports a sessile community dominated S. phippsii to evaluate its possible role in top-down re-
by sponges and bryozoans and covering 50–90 % of gulation. Preliminary observations showed a notable
its surface area. Laboratory experiments show that interannual variation in S. phippsii abundance at our
the red algae covered by bryozoans are preferentially research sites. Considering the strong interannual
consumed by herbivores (da Gama et al. 2008). In variation in recruitment rates of sessile organisms in
 Chava et al.: Top-down control in sub-arctic algal epibiosis 55

the White Sea subtidal (Yakovis et al. 2013), we re- (=predator enclosures; similar to full cages but with
Author copy

peated the experiments in 2 consecutive years. A 3–4 individuals of Spirontocaris phippsii added), and
higher epibiont cover in consumer-removal experi- (4) ambient controls (see below). Plants were randomly
ments, if detected, would elucidate the role of top- distributed across the treatments. Plants and shrimp
down control in competition decrease and possible were placed inside the cages immediately before the
protection of the host from overgrowth. deployment. Shrimp carapace length was 5–8 mm,
and their total weight was similar in different shrimp
cages. The number of shrimp individuals per cage was
2. MATERIALS AND METHODS chosen based on the previous field manipulations
with S. phippsii in the White Sea (Yakovis & Arte-
To test the effect of mobile predators on sessile epi- mieva 2019, 2021). The sex of the shrimp used in the
fauna of Phycodrys rubens, we conducted 2 series of experiments was not identified. In July, the cages
field caging experiments in the shallow subtidal of the were anchored to concrete blocks and placed on the
White Sea (Velikaya Salma Strait between the Kare- bottom in a haphazard pattern. In September, we re-
lian shore and Velikiy Island in the western part of moved the cages and examined the epibiosis of each
Kandalaksha Bay) in 2014 and 2015. Both series lasted plant. In 2014, there were 8 full, 7 partial, and 7 shrimp
for 2 mo, from late July to late September. The expo- cages, while in 2015, there were 10 full, 7 partial, and
sure duration was chosen based on the seasonal tim- 8 shrimp cages. In 2014, cages were deployed at Site
ing of epifaunal recruitment (Chava et al. 2019) and V, and in 2015, they were equally divided between
algal growth pattern (Schoschina 1996). We exposed Site V and Site K. While collecting the cages, we also
cages at 2 sites: Site K (near Kamenukha island, 11 m sampled plants around the cages to use them as
deep, 66° 33.028’ N, 33° 9.295’ E) and Site V (near Veli- ambient controls. In 2014, we acquired 7 random
kiy Island, 12 m deep, 66° 33.437’ N, 33° 6.877’ E) with plants at Site V. In 2015, we used the largest plants
different hydrological regimes. Site V was located in from haphazardly placed 0.25 m2 frames (Сhava et al.
the narrowest part of the Velikaya Salma Strait, while 2019), with the total area of young blades exceeding
Site K was located further seaward, in the wider part 5800 mm2 (to match the sizes of the plants used in the
of the strait. The flood current at Site V was hence manipulations). As a result, in 2015, there were 4
much faster (up to 1.33 m s–1) than at Site K (up to ambient controls at Site V and 10 ambient controls at
0.75 m s–1). Detailed descriptions of the sites are Site K. See Table 1 for a summary of the treatments.
given by Chava et al. (2019). Seawater temperatures P. rubens thalli develop in such a way that young
are similar along the studied part of the Velikaya blades of the current year, which constitute the major
Salma Strait because of tidal mixing (Mileikovsky part of the thallus, can be easily distinguished from
1970). Temperature data were available for several the older part remaining from the previous 1–3 yr
dates in June, July, and sometimes August in 2009– (Schoschina 1996) (hereafter ‘young’ and ‘old’). In
2016 from an automatic weather station (Aanderaa July, young blades were almost free of any sessile epi-
Instruments, AW 2700) attached to a pier at the White fauna. Therefore, we did not document the initial
Sea Biological Station (66° 33.177’ N, 33° 6.181’ E) at state of the experimental plants and examined young
1 m depth. These measurements (Table A1 in the blades only at the end of the experiments. For each
Appendix) show that the summer of 2014 was much plant, we documented the surface area of young
warmer (by 2.6°C) than the summer of 2015. blades (accurate to 1 mm2), identified all of the sessile
One or 2 d prior to the set-up date, we collected organisms attached to young blades (except sponges)
P. rubens growing on small cobbles (<15 cm) and to species level, and counted them. We determined
S. phippsii individuals at Site V in 2014 and at both the area of each sponge (accurate to 0.1 mm2), calcu-
sites in 2015 and kept them in aerated aquaria until lated the number of units (zooids/polyps) in each
the set-up date. We selected large plants since most hydrozoan or bryozoan colony, and individually mea-
sessile epifauna is concentrated on them (see Chava sured the opercular diameter of serpulid polychaetes
et al. 2019). We fastened each cobble with a plant on it (accurate to 0.05 mm) and the shell length of bivalves.
inside a ‘cage’ (a cylindrical plastic basket 20 cm wide These values were used to estimate the areas covered
and 30 cm tall covered with 2 mm nylon mesh), 1 plant and the percent cover according to the size–area allo-
per cage. We used 4 types of treatments: (1) full cages metric relationships previously established from sub-
(=predator exclosures), (2) partial cages, to control samples (Chava et al. 2019). Undetermined bryozoan
for cage effect (similar to full cages but with two 70 × ancestrulae with individual areas less than 0.15 mm2
70 mm openings in the mesh), (3) shrimp cages were excluded from further analyses.
 56 Mar Ecol Prog Ser 744: 53–67, 2024

Natural density of S. phippsii in algal canopy is diffi- square-root transformed percent cover of high-order
Author copy

cult to assess accurately. Shrimp are traditionally sam- taxa using multivariate PERMANOVA based on Bray-
pled with push nets and beam trawls (Hiddink et al. Curtis dissimilarities. We also performed separate
2002, Schaffmeister et al. 2006). While these methods univariate PERMANOVAs on Euclidean distances to
have an efficiency of up to 50–60 % in sandy or silty test these effects for major fouling taxa in the commu-
habitats and seagrass beds, we could not use them be- nity, that is, sponges, hydroids, bivalves, bryozoans,
cause of the following limitations. Firstly, both of these and polychaetes. Where the treatment effect was sig-
methods entail the disturbance of the habitat (milder nificant and variances were homogeneous (see be-
in case of a push net and severe in case of a beam low), we examined pairwise differences between
trawl), which was unacceptable since P. rubens canopy treatment levels with the following set of contrasts:
is fragile and can be easily damaged or torn off the (1) ambient controls vs. partial cages to check for
gravel. Secondly, shrimp can easily escape both a the artifacts of caging procedure (hereafter ‘A vs. P’),
trawl and a push net (Schaffmeister et al. 2006), which (2) (ambient controls and partial cages) vs. full
increases the underestimation of their density. Con- cages (=predator exclosures) to test the effects of ex-
sidering these limitations, we decided to catch all vis- cluding any kind of predators larger than 2–3 mm
ible shrimp during dives using a hand net or a plastic (hereafter ‘(A & P) vs. F’),
jar. The efforts were standardized by making individual (3) (ambient controls and partial cages) vs. shrimp
dives as similar as possible. The same diver spent 45– enclosures to test the effect of S. phippsii predation
50 min moving in zigzags at 9–11 m depth during day- compared to natural predation level (hereafter ‘(A &
light hours 2–4 d prior to the deployment of the cages. P) vs. S’),
The dives were carried out when the strong tidal cur- (4) shrimp cages vs. full cages to test the effect of S.
rents were minimal. According to these rough esti- phippsii predation compared to predator absence
mates, the density of S. phippsii around Site V in July (hereafter ‘S vs. F’).
was 2.5 times higher in 2014 than in 2015. In July 2014, To check for homogeneity of variances between the
the mean number of shrimps per dive was 9 ± 1 indi- levels of the factors included in the analyses, we per-
viduals (n = 6 dives), while in July 2015 it was 3.5 ± 2 formed the PERMDISP test (Anderson 2006). In case
individuals (n = 8 dives). The difference was significant of significant heterogeneity of variances, we consid-
(p = 0.023, Mann-Whitney U-test). ered the results not reflecting the difference between
To elucidate possible differences in natural total means, since PERMANOVA does not distinguish the
abundances of epibionts in 2014 and in 2015, we per- difference between variation in location and scale.
formed univariate permutational analysis of variance In order to visualize the differences between treat-
(PERMANOVA, Anderson 2001) on Euclidean dis- ments and contributions of particular taxa, we used
tances calculated from total percent cover of epibionts non-metric multidimensional scaling (nMDS) based
on young blades of P. rubens at Site V in partial cages on the Bray-Curtis index similarity matrix. We per-
and ambient controls. Here and below, we used univari- formed the procedure separately for 2014 and 2015
ate PERMANOVA (see e.g. Bishop et al. 2012) to sub- and for the 2 years combined.
stitute univariate ANOVA where obtaining p-values Calculations were conducted in PRIMER 6.0 soft-
by permutation allows avoiding the assumption of ware with the PERMANOVA add-on and in R v.4.3.1
normality (Anderson 2017). Compared to other non- (R Core Team 2023). Means are reported ±SE.
parametric methods, it allows a direct additive parti-
tioning of variation for complex models with multiple
categorical predictors and covariates (Anderson 2001). 3. RESULTS
Surface area of young P. rubens blades significantly
affects total cover of epibionts (Chava et al. 2019), so 3.1. Background
we used the former as a covariate in the analysis.
To assess the effects of treatments on total cover of Mean abundances of all taxa by year, site, and treat-
epibionts, we also used univariate PERMANOVA on ment levels are summarized in Table 1. Total cover of
Euclidean distances. Separate analyses were con- sessile organisms in ambient controls and partial
ducted for 2014 (with one site) and 2015 (with 2 sites). cages was similar. It was also similar between years at
Treatment (fixed) and young plant part area (covari- Site V, which was sampled both in 2014 and in 2015
ate) were used as predictors for 2014; in addition, site (Tables 1 & 2). Mean ± SE total cover (ambient con-
(fixed) was used for 2015. To test the same effects on trols and partial cages pooled) at Site V was 7.41 ±
the multivariate community structure, we analyzed 1.79 % in 2014 and 6.20 ± 0.59 % in 2015; it was slightly
 Chava et al.: Top-down control in sub-arctic algal epibiosis 57
Author copy

Table 1. Mean ± SE percent cover of epibionts on young Phycodrys rubens blades by year, site, and treatment level in the field experi-
ments. Site K: near Kamenukha Island; Site V: near Velikiy Island (see Section 2 for details); N: number of replicates. Epiphytic algae
and cirripeds were extremely scarce, and their mean cover (never exceeding 0.0002 and 0.009 %, respectively) is omitted. Note
that only a single polychaete species (Circeis armoricana) was recorded

Treatment Site Year N Porifera Coelenterata Polychaeta Bivalvia Bryozoa Tunicata Total

Ambient control V 2014 7 0.21 ± 0.21 0.22 ± 0.16 1.21 ± 0.42 0.02 ± 0.02 6.29 ± 1.83 <0.01 7.95 ± 2.58
Ambient control K 2015 10 0.41 ± 0.20 0.14 ± 0.02 1.48 ± 0.15 0.14 ± 0.05 3.10 ± 0.44 0.02 ± 0.01 5.30 ± 0.59
Ambient control V 2015 4 0.30 ± 0.10 0.62 ± 0.18 1.23 ± 0.27 0.50 ± 0.16 4.61 ± 0.68 0 7.26 ± 0.51
Full cage V 2014 8 0.42 ± 0.42 0.06 ± 0.04 1.60 ± 0.27 0 12.55 ± 0.99 0.05 ± 0.03 14.68 ± 1.21
Full cage K 2015 5 0.12 ± 0.09 0.07 ± 0.03 0.92 ± 0.18 0.02 ± 0.02 5.19 ± 0.86 0.05 ± 0.04 6.38 ± 0.97
Full cage V 2015 5 0.08 ± 0.06 0.39 ± 0.14 0.68 ± 0.06 0.07 ± 0.05 4.67 ± 0.62 0.02 ± 0.02 5.91 ± 0.59
Partial cage V 2014 7 0 0.04 ± 0.03 0.92 ± 0.26 0.01 ± 0.004 5.91 ± 2.39 <0.01 6.88 ± 2.67
Partial cage K 2015 4 0.04 ± 0.01 0.01 ± 0.01 0.62 ± 0.10 0.01 ± 0.01 3.19 ± 0.78 <0.01 3.88 ± 0.87
Partial cage V 2015 3 0.07 ± 0.07 0.24 ± 0.07 0.85 ± 0.17 0.07 ± 0.04 3.54 ± 0.20 0.01 ± 0.01 4.79 ± 0.43
Shrimp cage V 2014 7 0.04 ± 0.03 0.25 ± 0.16 0.72 ± 0.16 0.03 ± 0.02 6.21 ± 2.21 0.02 ± 0.02 7.28 ± 2.42
Shrimp cage K 2015 4 0.16 ± 0.09 0.06 ± 0.03 0.12 ± 0.05 0.01 ± 0.01 0.31 ± 0.16 <0.01 0.66 ± 0.26
Shrimp cage V 2015 4 0.01 ± 0.01 0.47 ± 0.13 0.09 ± 0.02 0.10 ± 0.06 0.77 ± 0.21 0 1.45 ± 0.36

lower at Site K in 2015 (4.89 ± 0.50%). Bryozoans dom- Whitney U-test), while serpulid polychaetes were
inated ambient controls and partial cages, contributing more abundant at Site K (25 ± 2 %) than at Site V (17 ±
on average from 58 to 84 % of the total cover depend- 2 %, p = 0.025). Relative abundances of bryozoans,
ing on site and year, with the most abundant species sponges, bivalves, and other taxa did not differ signif-
being Juxtacribrilina annulata, Celleporella hyalina, icantly at these 2 sites.
and Electra pilosa. Serpulid polychaetes (particularly
Circeis armoricana) were the second most abundant
group, contributing from 14 to 29 % of the total cover. 3.2. Experiment
Sponges, hydroids, and bivalves each contributed up
to 8 %, while all other taxa taken together contributed Total cover of sessile organisms displayed contrast-
less than 1 %. At Site V (ambient controls and partial ing responses to manipulations in 2014 and 2015. In
cages pooled), there were notably more bryozoans in 2014, when ambient abundance of shrimp was higher,
2014 (83 ± 2%) than in 2015 (68 ± 4%, p = 0.002, Mann- the exclusion of predators had a strong impact: the
Whitney U-test), while the relative abundance of ser- mean total cover in full cages was twice as high (15 ±
pulid polychaetes was similar in the 2 years (15 ± 2 % 1 %) as in all other treatments, where it was similar
and 17 ± 2 %, respectively, p = 0.322). (Fig. 1, Tables 1 & 3). The variances were homogeneous
Sites V and K (compared in 2015) differed slightly in (PERMDISP, F = 0.64, p = 0.835). Treatment effect was
species composition (ambient controls and partial significant. Both contrasts involving full cages were
cages pooled): there were more hydrozoans at Site V significant, while the contrasts comparing ambient
(7 ± 1 %) than at Site K (2 ± 1 %, p = 0.004, Mann- controls with partial cages and both of them with
shrimp cages were insignificant (Table 3).
Table 2. Effect of year on the total epibiont cover on young Phycodrys rubens In 2015, when ambient abundance of
blades at Site V. Univariate PERMANOVA on Euclidean distances between shrimp was lower, the exclusion of pred-
total cover of sessile organisms on young blades of P. rubens in partial cages ators made no difference: shrimp cages
and ambient controls in 2014 and 2015. A: ambient controls; P: partial cages;
demonstrated a significantly lower total
Area: total area of young blades of a plant
cover than all other treatments, which
were similar (Fig. 1, Tables 1 & 4). The
Source of df SS Pseudo-F p Unique
variation permutations
variances were homogeneous (PERM-
DISP, F = 1.87, p = 0.188). Treatment
Area (fixed, covariate) 1 0.0028 0.8026 0.295 9778 effect was significant regardless of site.
Treatment (fixed, A vs. P) 1 <0.0001 0.0015 0.974 9865 Both contrasts involving shrimp cages
Year (fixed, 2014 vs. 2015) 1 0.0011 0.3114 0.638 9868
were significant, while the contrasts
Treatment × Year 1 0.0004 0.1064 0.745 9841
Residuals 16 0.0555 comparing ambient controls with par-
Total 20 0.0604 tial cages and both with full cages
were insignificant (Table 4).
 58 Mar Ecol Prog Ser 744: 53–67, 2024

18
other treatments (Figs. 3 & 4). The variances, how-
Author copy

2014 2015 B
16 ever, were heterogeneous (PERMDISP, F = 6.65, p =
0.004), rendering the results of PERMANOVA unin-
14 formative (Table 6).
Both in 2014 and 2015, bryozoans were a major
Mean total cover (%)

12
A contributor to the differences in epibiont cover
A A between treatments, while the proportion of hydro-
10
zoans and bivalves mostly varied between sites and
8
B
years (Figs. 2–4). Consistently, univariate analyses
B revealed the significant effect of manipulations on
6 B the mean percent cover only in bryozoans (Tables 1,
7, & 8). In polychaetes (2014), bivalves (2014),
4
hydroids (both years), and sponges (both years),
2 A the treatment effect was insignificant, while poly-
chaetes and bivalves in 2015 showed heterogeneous
0 variances (PERMDISP, F = 3.88, p = 0.032 and F =
Ambient control Partial cages Full cages Shrimp cages 6.87, p = 0.030, respectively). In 2014, predator
Fig. 1. Mean total cover of epibionts (±SE) on young Phycod- exclusion made mean bryozoan cover in full cages
rys rubens blades by treatment in 2014 and 2015. Fill pattern twice as high as in all other treatments (Table 1,
and capital letters indicate homogeneous groups within the Fig. 5). The variances were homogeneous (PERM-
years according to PERMANOVA contrasts (Tables 3 & 4).
DISP, F = 0.88, p = 0.769). Treatment effect was
Partial fill denotes the groups with the lower mean total cover.
Sites pooled in 2015 significant, and so were both contrasts involving
full cages, while the contrasts comparing ambient
controls with partial cages and both of the latter
Multivariate community structure showed a gen- with shrimp cages were insignificant (Table 7). In
erally similar response. In 2014, predator exclusion 2015, full cages also displayed a higher mean bryo-
altered the taxonomic composition so that full cages zoan cover than partial cages and ambient controls,
grouped separately from the other treatments on while, in contrast, shrimp enclosures had much
nMDS plots (Fig. 2). The variances were homoge- fewer bryozoans than any other treatment (Table 1,
neous (PERMDISP, F = 3.15, p = 0.117). Treatment Fig. 5). The variances were homogeneous (PERM-
effect and both contrasts involving full cages were DISP, F = 2.26, p = 0.130). Treatment effect was
significant, while the contrasts comparing ambient significant, and the only insignificant contrast was
controls with partial cages and both of them with the one comparing partial cages and ambient con-
shrimp cages were insignificant (Table 5). In 2015, trols (Table 8).
shrimp cages clearly grouped separately from all

Table 3. Effect of predator abundance manipulations on the total epibiont cover 4. DISCUSSION
on young Phycodrys rubens blades in 2014. Univariate PERMANOVA on Eucli-
dean distances between total cover of sessile organisms on young blades of P.
rubens in the 2014 field experiment. Significant effects and contrasts are high-
Although we found dramatic inter-
lighted in bold. A: ambient controls; P: partial cages; F: full cages; S: shrimp en- annual variation in the consumer con-
closures; Area: total area of young blades of a plant; **p < 0.01; ***p < 0.001. trol of sessile organisms covering young
PERMDISP: F = 0.64, p = 0.835 blades of Phycodrys rubens, their mean
total cover never exceeded 15 % even
Source of df SS Pseudo-F p Unique after consumer removal. This clearly
variation permutations shows that, in contrast to a compa-
Area (fixed, covariate) 1 0.0174 5.7137 0.0281** 9800
rable epibiosis of Chondrus crispus in
Treatment (fixed) 3 0.0371 4.0519 0.0158** 9953 temperate waters (Stachowicz & Whit-
Contrast A vs. P 1 0.0001 0.0291 0.8599 9823 latch 2005), it is not top-down regula-
Contrast (A & P) vs. F 1 0.0299 9.3622 0.0025** 9852 tion that prevents young P. rubens
Contrast (A & P) vs. S 1 0.0001 0.0291 0.8987 9848
blades from heavy overgrowth. In the
Contrast S vs. F 1 0.0223 23.719 0.0001*** 9829
Residuals 24 0.0732 warm year (2014), mobile consumers
Total 28 0.1221 substantially shaped the community,
and their removal doubled the total
 Chava et al.: Top-down control in sub-arctic algal epibiosis 59

in 2015. At the same time, bryozoans

Author copy

Table 4. Effect of predator abundance manipulations on the total epibiont cover

on young Phycodrys rubens blades in 2015. Univariate PERMANOVA on Eucli-were much less abundant in 2015, when
dean distances between total cover of sessile epifauna on young blades of P.
there were more sponges, hydroids,
rubens in the 2015 field experiment. Significant effects and contrasts are high-
and bivalves, than in 2014. Greater con-
lighted in bold. A: ambient controls; P: partial cages; F: full cages; S: shrimp
enclosures; Area: total area of young blades of a plant; ***p < 0.001. sumer control in the warm year and low
PERMDISP: F = 1.87, p = 0.188 consumer control, combined with less
potential prey, in the cold year re-
Source of df SS Pseudo-F p Unique sulted in a remarkably similar total per-
variation permutations cent cover in a natural community.
Selective predation is a powerful
Area (fixed, covariate) 1 0.0007 3.40 0.0702 9832
driver of community composition, but
Treatment (fixed) 3 0.0096 14.74 0.0001*** 9958
Contrast A vs. P 1 0.0002 0.92 0.3439 9816 much of its power comes from the indi-
Contrast (A & P) vs. F 1 0.0002 0.84 0.3588 9811 rect positive effects on inferior compet-
Contrast (A & P) vs. S 1 0.0075 42.48 0.0001*** 9826 itors in the presence of strong competi-
Contrast S vs. F 1 0.0054 24.81 0.0004*** 9873 tion (Paine 1974). In temperate waters,
Site 1 0.0006 2.68 0.1088 9825
Treatment × Site 3 0.0007 1.01 0.4000 9948 generalist mesopredatory shrimp can
Residuals 30 0.0065 selectively target bryozoans, prevent-
Total 38 0.0234 ing them from monopolizing artificial
hard substrates (Dumont et al. 2011).
Mesograzer snails remove dominating
abundances of sessile organisms. According to our bryozoans from the red seaweed (Stachowicz & Whit-
rough estimates, in 2014 there were also more meso- latch 2005), while grazing chitons similarly restrict
predatory shrimp in the surrounding habitat. In the bryozoan presence in marina dock fouling (Nydam &
extremely cold year (2015), when there were ap- Stachowicz 2007). In our study, bryozoans dominated
parently fewer shrimp around, consumer removals on young Phycodrys blades and were a primary target
had an almost negligible effect on the total cover at for consumers, whereas subdominant tubeworms and
both sites. Consumers unequally impacted different other less abundant groups were less targeted. Ho-
sessile epibenthic taxa, targeting mostly bryozoans, wever, indirect benefits for inferior competitors in this
which were heavily preyed upon both in 2014 and particular microhabitat are improbable, because the

Mollusca = 0.61 Hydrozoa = 0.59

0.6

Polychaeta = 0.72
0.4
Total = 0.92
MDS2

0.2

Bryozoa = 0.94
0.0

−0.2

−0.5 0.0 0.5 1.0 1.5 2.0

MDS1

Ambient control Full cage Partial cage Spirontocaris cage

log(Area of young plant parts) 9.0 10.0 11.0

Fig. 2. Multivariate structure of the epibiosis on young Phycodrys rubens blades by treatment in 2014. Non-metric multidi-
mensional scaling (nMDS) on square-root transformed standardized total cover of high-order taxa in 2014. Bray-Curtis dis-
similarity. Diamonds are centroids, ellipses denote SE. Vectors show significant correlations of standardized percent cover
by taxa with nMDS axes
 60 Mar Ecol Prog Ser 744: 53–67, 2024

large shrimp individuals (Chava et al.

Author copy

Table 5. Effect of predator abundance manipulations on the multivariate com-

munity structure of the epibiosis on young Phycodrys rubens blades in 2014. 2024). This is consistent with our results
PERMANOVA on square-root transformed Bray-Curtis dissimilarities be- showing the relative vulnerability of
tween standardized cover of high-order taxa in the 2014 field experiment. Sig-
bryozoans in the field. Caged shrimp,
nificant effects and contrasts are highlighted in bold. A: ambient controls; P:
in fact, had a similar impact on the
partial cages; F: full cages; S: shrimp enclosures; Area: total area of young
blades of a plant; ***p < 0.001. PERMDISP: F = 3.15, p = 0.117 community in 2014 and 2015, reducing
the average bryozoan percent cover by
Source of df SS Pseudo-F p Unique 4–6 % compared to predator exclo-
variation permutations sures, but in 2015 this meant a near
elimination of bryozoans (Fig. 5). The
Area (fixed, covariate) 1 1607 3.22 0.0202 9945
4–6 % reduction closely matched the
Treatment (fixed) 3 4906 3.28 0.0004*** 9926
Contrast A vs. P 1 777 1.31 0.2639 9949 natural predation rate in 2014. In 2015,
Contrast (A & P) vs. F 1 2777 6.06 0.0008*** 9951 the reduction of bryozoan cover in am-
Contrast (A & P) vs. S 1 431 0.67 0.6223 9953 bient controls and partial cages com-
Contrast S vs. F 1 2651 6.64 0.0001*** 9954 pared to exclosure cages was about 2 %,
Residuals 24 11958
consistent with the estimated inter-
Total 28 18007
annual difference in ambient shrimp
abundance.
low total cover clearly indicated the lack of competi- Our design has a limited capability to assess spatial
tion. While Spirontocaris phippsii is unlikely to be the variation of consumer control, since the latter was
only mesopredator in the system, our results obtained much stronger in 2014, when we exposed all cages at
in 2014 indicate that it is a key one. Although this a single site. However, bryozoans were impacted by
shrimp is omnivorous (Birkely & Gulliksen 2003, Ya- consumers in both years, and neither location nor its
kovis & Artemieva 2019, 2021), a recent microcosm interaction with treatment affected bryozoan cover in
study (Chava et al. 2024) revealed that, owing to its 2015. Thus, compared to its interannual variation and
feeding habits, it has a much stronger impact on bryo- taxonomic selectivity, the difference in the consumer
zoans than on co-dominant serpulid tubeworms. control strength between the 2 sites was negligible.
Unitary tubeworms outperform modular bryozoans Chondrus epibiosis shows extensive spatial variation
against shrimp because of the superior ability of tube- of the consumer control strength associated with
worms to outgrow predation pressure, especially from patchy distribution of mesograzer snails (Stachowicz

Mollusca = 0.58
Hydrozoa = 0.47

0.2
Total = 0.85
Polychaeta = 0.70
MDS2

0.0

Bryozоа = 0.88
−0.2

−0.4

−0.5 0.0 0.5 1.0 1.5 2.0

MDS1

Ambient control Full cage Partial cage Spirontocaris cage

Site V Site K log(Area of young plant parts) 8.8 9.2 10.0

Fig. 3. Multivariate structure of the epibiosis on young Phycodrys rubens blades by treatment in 2015. Non-metric MDS (nMDS)
on square-root transformed standardized total cover of high-order taxa in 2015. Bray-Curtis dissimilarity. Diamonds are cen-
troids, ellipses denote SE. Vectors show significant correlations of standardized percent cover by taxa with nMDS axes
 Chava et al.: Top-down control in sub-arctic algal epibiosis 61
Author copy

Hydrozoa = 0.55 S S

Mollusca = 0.46 Shrimp cage

1.0 Polychaeta = 0.70 2014 vs 2015
Total = 0.73

Bryozoa = 0.86
F F
Full cage

0.5
MDS2

A A
Ambient control
A
S
0.0 F F

P S A
P P P
Partial cage

−0.5

−1.0 −0.5 0.0 0.5

MDS1
Fig. 4. Multivariate structure of the epibiosis on young Phycodrys rubens blades by treatment in 2014 and 2015. Non-metric
MDS (nMDS) on square-root transformed standardized total cover of high-order taxa in 2014. Bray-Curtis dissimilarity.
Squares with letters inside are centroids, ellipses denote SE. Vectors show significant correlations of standardized percent
cover by taxa with nMDS axes

predators (Shinomiya et al. 2017), they

Table 6. Effect of predator abundance manipulations on the multivariate com-
munity structure of the epibiosis on young Phycodrys rubens blades in 2015.
generally tend to show spatial abun-
PERMANOVA on square-root transformed Bray-Curtis dissimilarities be-
dance variation at a larger scale (Bod-
tween standardized cover of high-order taxa in the 2015 field experiment. Sig-
deke 1976, Koeller 2000). Since we
nificant effects and contrasts are highlighted in bold. A: ambient controls; P:
studied only 2 sites, there is a good
partial cages; F: full cages; S: shrimp enclosures; Area: total area of young
blades of a plant; **p < 0.01, ***p < 0.001. PERMDISP: F = 6.65, p = 0.004**
chance that, despite the difference in
the tidal flow, they were too similar
Source of df SS Pseudo-F p Unique in terms of consumer abundance to
variation permutations detect spatial variation in top-down
Area (fixed, covariate) 1 370 0.89 0.5102 9940 control of Phycodrys epibiosis.
Treatment (fixed) 3 11225 8.99 0.0001*** 9907 To our knowledge, substantial inter-
Contrast A vs. P 1 1018 3.58 0.0004*** 9953 annual variation in consumer control
Contrast (A & P) vs. F 1 1527 4.47 0.0002*** 9935 strength has never been previously de-
Contrast (A & P) vs. S 1 7552 16.45 0.0001*** 9953
Contrast S vs. F 1 5424 9.81 0.0001*** 9945 tected in a seaweed epibiosis, although
Site 1 5403 12.98 0.0001*** 9952 it is common in various other freshwater
Treatment × Site 3 1657 1.33 0.1611 9914 and marine systems in tropical and tem-
Residuals 30 12489
perate zones (Heck & Wilson 1987,
Total 38 33440 Hairston 1988, Posey et al. 1995, Navar-
rete 1996). The known mechanisms of
& Whitlatch 2005). However, shrimp are much more this variation range from relatively infrequent climatic
mobile than snails. While the spatial distribution of events such as El Niño–Southern Oscillation to grad-
shrimp can also be patchy, at least when shaped by ual long-term trends like oligotrophication (Meserve
the patches of FS providing prey and refuge from et al. 2003, Kerimoglu et al. 2013). Ambient abiotic
 62 Mar Ecol Prog Ser 744: 53–67, 2024

with respect to the summer water tem-

Author copy

Table 7. Effect of predator abundance manipulations on the percent cover of

bryozoans on young Phycodrys rubens blades in 2014. Univariate PERM- perature range (Fig. 6). Below we dis-
ANOVA on Euclidean distances between total cover of bryozoans on young
cuss the processes shaping epibiosis
blades of P. rubens in the 2014 field experiment. Significant effects and con-
trasts are highlighted in bold. A: ambient controls; P: partial cages; F: full and the possible effects of climatic dif-
cages; S: shrimp enclosures; Area: total area of young blades of a plant; *p < ferences between years.
0.05; **p < 0.01; ***p < 0.001. PERMDISP: F = 0.88, p = 0.769 Community composition and abun-
dances of sessile epibenthic organisms
Source of df SS Pseudo-F p(perm) Unique colonizing newly developed seaweed
variation permutations blades generally result from larval
supply (recruitment) and subsequent
Area (fixed, covariate) 1 0.0133 6.28 0.022* 9832
Treatment (fixed) 3 0.0291 4.60 0.011 9957
growth of epibionts, their competition,
Contrast A vs. P 1 0.0005 0.15 0.718 9857 substrate provision by the host, con-
Contrast (A & P) vs. F 1 0.0231 10.99 0.001** 9863 sumer control, and physical disturb-
Contrast (A & P) vs. S 1 0.0001 0.05 0.827 9862 ance. Physical disturbance frequency
Contrast S vs. F 1 0.0163 17.31 <0.001*** 9850
is low in subtidal habitats (Palmer et al.
Residuals 24 0.0507
1996); there were no traces of wave
Total 28 0.0879
action at our 11–12 m deep sites.
Shielding associated organisms from
Table 8. Effect of predator abundance manipulations on the percent cover of abiotic stress, which is commonly pro-
bryozoans on young Phycodrys rubens blades in 2015. Univariate PERM- vided by FS in the intertidal (Bruno et
ANOVA on Euclidean distances between total cover of bryozoans on young al. 2003, Olafsson 2016), does not ap-
blades of P. rubens in the 2015 field experiment. Significant effects and contrasts pear to be important in a subtidal hab-
are highlighted in bold. A: ambient controls; P: partial cages; F: full cages; S:
shrimp enclosures; Area: total area of young blades of a plant; ***p < 0.001.
itat. In fact, the assemblages associated
PERMDISP: F = 2.26, p = 0.130 with Phycodrys blades may be even
more accessible to mesopredators than
Source of df SS Pseudo-F p Unique those on underlying gravel and rocks
variation permutations shielded by Phycodrys canopy.
Competition for space is also uncom-
Area (fixed, covariate) 1 0.0003 1.49 0.2296 9825 mon on young Phycodrys blades, as
Treatment (fixed) 3 0.0070 13.64 0.0001*** 9957
indicated by 8 % mean total cover on
Contrast A vs. P 1 <0.0001 0.02 0.9006 9829
Contrast (A & P) vs. F 1 0.0011 5.39 0.0273 9835 random natural plants in 2014 (Table 1)
Contrast (A & P) vs. S 1 0.0041 32.55 0.0001*** 9831 and 7–10 % in 2015 (Table 2 in Chava et
Contrast S vs. F 1 0.0039 21.57 0.0005*** 9852 al. 2019). This is in contrast with com-
Site 1 0.0002 1.07 0.3111 9803 parable systems from tropical and tem-
Treatment × Site 3 0.0005 0.87 0.4584 9945
Residuals 30 0.0052 perate waters (Stachowicz & Whitlatch
Total 38 0.0150 2005, da Gama et al. 2008). At the same
time, red seaweed productivity is
highly sensitive to water temperature
conditions can switch a system from top-down to bot- (Paalme et al. 2011). Phycodrys grows twice as fast at
tom-up regulation and back (Hoekman 2010). Tempo- 10°C as at 4°C (Gordillo et al. 2022). The area of
ral limitations of our experiments make it impossible young Phycodrys blades per square meter of the bot-
to distinguish a causation from fluctuation, but given tom at Sites V and K was, accordingly, 2–4 times
the contrasting water temperatures in 2014 and 2015, lower after the extremely cold summer of 2015 than
the abiotic regulation of consumer control seems after the contrastingly warm following summer
highly probable. Importantly, according to the long- (Fig. 6), while the density of the plants remained sim-
term records of water temperatures in the White Sea ilar (Chava et al. 2019). Given that the summer of 2014
(at Station D1 in the Chupa Inlet of the Kandalaksha was also extremely warm, Phycodrys growth at our
Bay, 66° 19.836’ N, 33° 40.098’ E; data from: Usov et al., research sites supplied substrate at a rate well
https://w w w.st.nmfs.noaa.gov/copepod/time- exceeding the cumulative recruitment and growth
series/ru-10101/), the summer of 2015 was the coldest capacity of the epibionts both in high-productive and
since 2010, while the summer of 2014 was the sixth low-productive years. However, this excess is appar-
warmest since 1991. This means that our experiments ently limited to large and average-sized Phycodrys
were by chance made under contrasting conditions plants. Total cover on individual hosts is negatively
 Chava et al.: Top-down control in sub-arctic algal epibiosis 63

16 correlated with their size, reaching 25–40 % on young

Author copy

2014 2015

B blades of small plants (Chava et al. 2019; Fig. 3).

14 Heavy overgrowth and its regulation by consumers
Mean percent cover of bryozoans (%)

may hence cause size-selective pressure on the pop-

12 ulation structure of the host. Competition for space
and consequent indirect effects of consumers target-
10
A
ing dominant competitors seem important on old Phy-
A A
codrys blades and underlying rocks where the cover
8 of sessile organisms is several times higher (Chava et
al. 2019, V. Verzhbitsky pers. comm.).
B Our results show that lower natural bryozoan abun-
6
A dances in 2015 were not caused by large mesopred-
A
4
ators. There are still some other potential consumers
that can pass through a 2 mm mesh, e.g. turbellarians,
2
nudibranchs, and polychaetes (Lidgard 2008). Since
C we did not observe many of these, we assume that in
years with weak consumer control, the community as-
0
Ambient control Partial cages Full cages Shrimp cages sembly is mostly driven by supply-side processes (i.e.
recruitment, see Underwood & Keough 2001 for re-
Fig. 5. Mean percent cover of bryozoans (%, ±SE) on young view) and further shaped by growth rate variations in
Phycodrys rubens blades by treatment in 2014 and 2015. Fill response to ambient conditions. Larvae and juveniles
pattern and capital letters indicate homogeneous groups of sessile organisms are usually sensitive to fluctu-
within the years according to PERMANOVA contrasts
(Tables 7 & 8). Solid fill denotes the groups with the highest
ations in temperature (Ushakova 2003, Saunders &
mean bryozoan cover, crosshatched partial fill denotes Metaxas 2007, Bitschofsky et al. 2011). Lower temp-
the group with the lowest mean bryozoan cover. Sites eratures can also cause limited nutrient supply to ma-
pooled in 2015 ture epibionts restricting both their growth and repro-
ductive potential (Seed & Hughes 1992,
Gibson et al. 2001). Substantial tax-
15 Substantial onomic variation of seasonal recruit-
consumer
control ment timing in Arctic waters (Meyer et
al. 2017) suggests that climatic vari-
14
ables can unequally affect different
Mean summer temperature (°C)

taxa, causing interannual shifts in com-

13 munity composition. The established
residents surviving from the previous
years on the neighboring stable sub-
12
strates, like old Phycodrys parts and
rocks, must also affect larval supply
11 and settlement patterns (Meyer et al.
2017). Consequently, the abiotic drivers
No
consumer from previous years imprint in the
10
control community history of the adjacent less
ephemeral microhabitats and would
9 also shape the epibiosis of young Phy-
codrys blades.
The observed difference in the com-
8
munity composition in the 2 study
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023

years encompasses a lower percent

cover of predation-sensitive bryozoans
Fig. 6. Mean summer water temperatures in the White Sea in 1991–2023. Mea- associated with the decreased meso-
sured at Station D1 in the Chupa Inlet of Kandalaksha Bay (66° 19.836’ N,
predator abundance and consumer con-
33° 40.098’ E), averaged across 0.5–10 m depths (data from Usov et al., https://
www.st.nmfs.noaa.gov/copepod/time-series/ru-10101/). The 2 years of the trol strength in the colder year. This
field experiments are highlighted in bold may be either a coincidence or the re-
 64 Mar Ecol Prog Ser 744: 53–67, 2024

sult of a causal relationship. Shrimp population dy- vealed in our study indicates that the data obtained in
Author copy

namics in temperate and polar zones is commonly one year of observations or experiments in polar
driven by water temperatures, which affect hatching waters should be treated with caution.
and juveniles (van der Veer & Bergman 1987, Beukema
1992, Oh et al. 1999, Aschan & Ingvaldsen 2009,
Koeller et al. 2009, Richards et al. 2012, Beukema & 5. CONCLUSIONS
Dekker 2014, 2020), and is bottom-up controlled by
food availability (Salama & Hartnoll 1992, Hufnagl et We can draw the following conclusions from this
al. 2010). Fish predation (Tiews 1978, Koeller 2000, study. (1) Consumer control of community composi-
Wieland et al. 2007, Jónsdóttir 2017), which is in turn tion in a sub-arctic seaweed epibiosis can show strong
regulated by marine mammals (Temming & Hufnagl interannual variation. Exclusion of consumers dou-
2015), can also affect shrimp. However, based on the bled the total cover in 2014 but had no effect on it in
limited available data, Spirontocaris is not known to be 2015, when the summer was about 3°C colder than in
a part of fish diet in the White Sea (Gerasimova & the previous year. (2) Consumers, particularly the
Podrazhanskaya 1991, Ershov 2010). Colder water in mesopredatory shrimp, primarily affected bryozoans
summer 2015 is thus consistent with lower shrimp both in 2014 and 2015. There were fewer bryozoans on
abundance. Yet, considering that Spirontocaris has a unmanipulated plants in the colder year, apparently
lifespan up to 5 yr (Vogt 2019 based on Węsławski because of the lower recruitment rate. (3) Compared
1987), its population structure can accumulate the ef- to temperate and tropical waters, consumer control in
fects of climatic variables from different years. While a polar sea is not critical for the protection of an algal
the variation in relative abundance of bryozoans may host from being heavily overgrown. However, top-
cause bottom-up control of shrimp abundance, it is down regulation of epibiosis appears to be triggered
also likely that both are driven by the same climatic by higher water temperatures. In the face of climate
variables (see Saunders & Metaxas 2007, Beukema & change, top-down regulation may become a more
Dekker 2014). At least, the reproductive cycles of important factor, especially in smaller host plants
other organisms from higher trophic layers, including with higher epibiont cover.
predatory shrimp, have apparently evolved to use en-
vironmental triggers correlated with the abundance of
Acknowledgements. Our heartfelt thanks go to the volun-
potential prey, minimizing direct bottom-up trophic teers who helped us in the field, especially Ksenia Shunkina,
control (Koeller et al. 2009). However, a much greater Vladimir Krapivin, Nikolay Neretin, Artem Isachenko, Vla-
observation effort would be necessary to detect the in- dimir Chava, and Dmitry Ozerov. We are grateful to Alex-
terannual correlation between climatic variables and ander Tsetlin and the staff of the White Sea Biological Sta-
tion, who created incredible working conditions in the fairly
abundances of shrimp and bryozoans.
harsh environment. This paper owes much to comments and
The strength of biotic interactions, including con- suggestions by Natalia Lentsman and 3 anonymous re-
sumer control, generally decreases with latitude viewers. Thanks are due to our families and friends for sup-
(Schemske et al. 2009, Baskett & Schemske 2018), port in these troubled times. The research was funded by
although the supporting data are geographically lim- RSF (project No. 23-24-00191).
ited. Among the communities shaped by FS, tropical
and temperate seagrass beds show the latitudinal gra- LITERATURE CITED
dient of predation strength (Freestone et al. 2020).
However, direct consumer control and trophic cas- Amsler CD, McClintock JB, Baker BJ (2014) Chemical medi-
ation of mutualistic interactions between macroalgae and
cades also affect at least some communities associ-
mesograzers structure unique coastal communities along
ated with FS in polar waters (Yakovis & Artemieva the western Antarctic Peninsula. J Phycol 50:1–10
2019, 2021). We showed that consumer control is not a Anderson MJ (2001) A new method for non-parametric
critical factor preventing a sub-polar seaweed host multivariate analysis of variance. Austral Ecol 26:32–46
from extensive overgrowth. Our findings suggest that Anderson MJ (2006) Distance-based tests for homogeneity
of multivariate dispersions. Biometrics 62:245–253
consumer control of epibiosis is triggered by higher Anderson MJ (2017) Permutational multivariate analysis of
water temperatures. This means that a future increase variance (PERMANOVA). In: Balakrishnan N, Colton T,
of top-down regulation due to the ongoing rapid cli- Everitt B, Piegorsch W, Ruggeri F, Teugels JL (eds) Wiley
matic changes in Arctic and sub-Arctic areas StatsRef: statistics reference online. https://doi.org/10.
1002/9781118445112.stat07841
(McCrystall et al. 2021) may be predicted. In ad- Aschan M, Ingvaldsen R (2009) Recruitment of shrimp (Pan-
dition, a large interannual variation range of con- dalus borealis) in the Barents Sea related to spawning
sumer control (from substantial to negligible) re- stock and environment. Deep Sea Res II 56:2012–2022
 Chava et al.: Top-down control in sub-arctic algal epibiosis 65

Baskett CA, Schemske DW (2018) Latitudinal patterns of Dumont CP, Harris LG, Gaymer CF (2011) Anthropogenic
Author copy

herbivore pressure in a temperate herb support the biotic structures as a spatial refuge from predation for the inva-
interactions hypothesis. Ecol Lett 21:578–587 sive bryozoan Bugula neritina. Mar Ecol Prog Ser 427:
Beukema JJ (1992) Expected changes in the Wadden Sea 95–103
benthos in a warmer world: lessons from periods with Ellison AM (2019) Foundation species, non-trophic interac-
mild winters. Neth J Sea Res 30:73–79 tions, and the value of being common. iScience 13:
Beukema JJ, Dekker R (2014) Variability in predator abun- 254–268
dance links winter temperatures and bivalve recruitment: Ershov P (2010) Changes in the diet of the coastal cod Gadus
correlative evidence from long-term data in a tidal flat. morhua in the Kandalaksha Gulf of the White Sea under
Mar Ecol Prog Ser 513:1–15 conditions of increased abundance of three-spined stick-
Beukema JJ, Dekker R (2020) Half a century of monitoring leback Gasterosteus aculeatus. J Ichthyol 50:84–88
macrobenthic animals on tidal flats in the Dutch Wadden Estes JA, Palmisano JF (1974) Sea otters: their role in struc-
Sea. Mar Ecol Prog Ser 656:1–18 turing nearshore communities. Science 185:1058–1060
Birkely SR, Gulliksen B (2003) Feeding ecology in five Freestone AL, Carroll EW, Papacostas KJ, Ruiz GM, Torchin
shrimp species (Decapoda, Caridea) from an Arctic fjord ME, Sewall BJ (2020) Predation shapes invertebrate
(Isfjorden, Svalbard), with emphasis on Sclerocrangon diversity in tropical but not temperate seagrass commu-
boreas (Phipps, 1774). Crustaceana 76:699–715 nities. J Anim Ecol 89:323–333
Bishop MJ, Byers JE, Marcek BJ, Gribben PE (2012) Density- Gerasimova O, Podrazhanskaya S (1991) Living conditions
dependent facilitation cascades determine epifaunal and peculiarities of trophic relations of commercial fishes
community structure in temperate Australian man- in the White Sea. In: Biological resources: state, prospects
groves. Ecology 93:1388–1401 and problems of their rational exploitation. Biotopic basis
Bitschofsky F, Forster S, Scholz J (2011) Regional and tem- of the distribution of commercial and food marine ani-
poral changes in epizoobiontic bryozoan-communities of mals: collected papers. VNIRO, Moscow, p 116–125
Flustra foliacea (Linnaeus, 1758) and implications for Gibson RN, Barnes M, Atkinson RJA (2001) Life-history pat-
North Sea ecology. Estuar Coast Shelf Sci 91:423–433 terns in serpulimorph polychaetes: ecological and evolu-
Boddeke R (1976) The seasonal migration of the brown tionary perspectives. Oceanogr Mar Biol Annu Rev 39:
shrimp Crangon crangon. Neth J Sea Res 10:103–130 1–101
Bruno JF, Stachowicz JJ, Bertness MD (2003) Inclusion of Gordillo FJL, Carmona R, Jiménez C (2022) A warmer Arctic
facilitation into ecological theory. Trends Ecol Evol 18: compromises winter survival of habitat-forming sea-
119–125 weeds. Front Mar Sci 8:750209
Cabrera R, Areces AJ, Díaz-Larrea J, Sahu SK, Cruz-Aviña Grishankov AV, Ninbourg EA, Artemieva AV, Khaitov VM,
JR, García LGN (2022) Population dynamics of coloniz- Yakovis EL (1997) Benthos of Solovetskiy bay (Onega Bay,
ing fauna and its effect on growth rates of the farmed red the White Sea). Vestn St Petersbg Univ Ser 3 Biol 1:3–11
alga Alsidium triquetrum (SG Gmelin) Trevisan. Nat Sci Hairston NG Jr (1988) Interannual variation in seasonal pre-
14:42–55 dation: its origin and ecological importance. Limnol
Chadwick NE, Morrow KM (2011) Competition among ses- Oceanogr 33:1245–1253
sile organisms on coral reefs. In: Dubinsky Z, Stambler N Heck KL Jr, Wilson KA (1987) Predation rates on decapod
(eds) Coral reefs: an ecosystem in transition. Springer, crustaceans in latitudinally separated seagrass commu-
Dordrecht, p 347–371 nities: a study of spatial and temporal variation using
Chava A, Artemieva A, Yakovis E (2019) Plant part age and tethering techniques. J Exp Mar Biol Ecol 107:87–100
size affect sessile macrobenthic assemblages associated Hiddink JG, Marijnissen SAE, Troost K, Wolff WJ (2002)
with a foliose red algae Phycodrys rubens in the White Predation on 0-group and older year classes of the bivalve
Sea. Diversity 11:80 Macoma balthica: interaction of size selection and inter-
Chava A, Artemieva A, Yakovis E (2024) Effect of a generalist tidal distribution of epibenthic predators. J Exp Mar Biol
mesopredator on modular and unitary sessile prey associ- Ecol 269:223–248
ated with a foundation species. Ecol Evol 14:e11413 Hoekman D (2010) Turning up the heat: temperature
Connolly SR, Roughgarden J (1999) Theory of marine commu- influences the relative importance of top-down and bot-
nities: competition, predation, and recruitment-dependent tom-up effects. Ecology 91:2819–2825
interaction strength. Ecol Monogr 69:277–296 Honkanen T, Jormalainen V (2005) Genotypic variation in
D’Antonio C (1985) Epiphytes on the rocky intertidal red tolerance and resistance to fouling in the brown alga
alga Rhodomela larix (Turner) C. Agardh: negative Fucus vesiculosus. Oecologia 144:196–205
effects on the host and food for herbivores? J Exp Mar Hufnagl M, Temming A, Dänhardt A, Perger R (2010) Is
Biol Ecol 86:197–218 Crangon crangon (L. 1758, Decapoda, Caridea) food lim-
da Gama BAP, Santos RP, Pereira RC (2008) The effect of epi- ited in the Wadden Sea? J Sea Res 64:386–400
bionts on the susceptibility of the red seaweed Cryptone- Jónsdóttir IG (2017) Predation on northern shrimp (Pandalus
mia seminervis to herbivory and fouling. Biofouling 24: borealis) by three gadoid species. Mar Biol Res 13:
209–218 447–455
Díaz-Tapia P, Maggs CA, Macaya EC, Verbruggen H (2018) Kerimoglu O, Straile D, Peeters F (2013) Seasonal, inter-
Widely distributed red algae often represent hidden intro- annual and long-term variation in top–down versus bot-
ductions, complexes of cryptic species or species with tom–up regulation of primary production. Oikos 122:
strong phylogeographic structure. J Phycol 54:829–839 223–234
Duffy JE (1990) Amphipods on seaweeds: partners or pests? Koeller PA (2000) Relative importance of abiotic and biotic
Oecologia 83:267–276 factors to the management of the northern shrimp (Pan-
Duffy JE (2006) Biodiversity and the functioning of seagrass dalus borealis) fishery on the Scotian Shelf. J Northwest
ecosystems. Mar Ecol Prog Ser 311:233–250 Atl Fish Sci 27:21–34
 66 Mar Ecol Prog Ser 744: 53–67, 2024

Koeller P, Fuentes-Yaco C, Platt T, Sathyendranath S and Posey M, Powell C, Cahoon L, Lindquist D (1995) Top down
Author copy

others (2009) Basin-scale coherence in phenology of vs. bottom up control of benthic community composition
shrimps and phytoplankton in the North Atlantic Ocean. on an intertidal tideflat. J Exp Mar Biol Ecol 185:19–31
Science 324:791–793 Quijón PA, Snelgrove PVR (2005) Differential regulatory
Konar B, Iken K (2005) Competitive dominance among ses- roles of crustacean predators in a sub-arctic, soft-sed-
sile marine organisms in a high Arctic boulder commu- iment system. Mar Ecol Prog Ser 285:137–149
nity. Polar Biol 29:61–64 R Core Team (2023) R: a language and environment for statis-
Lamy T, Koenigs C, Holbrook SJ, Miller RJ, Stier AC, Reed tical computing. R Foundation for Statistical Computing,
DC (2020) Foundation species promote community sta- Vienna
bility by increasing diversity in a giant kelp forest. Eco- Richards RA, Fogarty MJ, Mountain DG, Taylor MH (2012)
logy 101:e02987 Climate change and northern shrimp recruitment variabil-
Lidgard S (2008) Predation on marine bryozoan colonies: ity in the Gulf of Maine. Mar Ecol Prog Ser 464:167–178
taxa, traits and trophic groups. Mar Ecol Prog Ser 359: Salama AJ, Hartnoll RG (1992) Effects of food and feeding
117–131 regime on the growth and survival of the prawn Palaemon
McCrystall MR, Stroeve J, Serreze M, Forbes BC, Screen JA elegans Rathke, 1837 (Decapoda, Caridea). Crustaceana
(2021) New climate models reveal faster and larger 63:11–22
increases in Arctic precipitation than previously pro- Saunders M, Metaxas A (2007) Temperature explains settle-
jected. Nat Commun 12:6765 ment patterns of the introduced bryozoan Membranipora
Menge BA, Sutherland JP (1976) Species diversity gradients: membranacea in Nova Scotia, Canada. Mar Ecol Prog Ser
synthesis of the roles of predation, competition, and tem- 344:95–106
poral heterogeneity. Am Nat 110:351–369 Schaffmeister BE, Hiddink JG, Wolff WJ (2006) Habitat use
Meserve PL, Kelt DA, Milstead WB, Gutiérrez JR (2003) of shrimps in the intertidal and shallow subtidal seagrass
Thirteen years of shifting top-down and bottom-up con- beds of the tropical Banc d’Arguin, Mauritania. J Sea Res
trol. Bioscience 53:633–646 55:230–243
Meyer KS, Sweetman AK, Kuklinski P, Leopold P and others Schemske DW, Mittelbach GG, Cornell HV, Sobel JM, Roy
(2017) Recruitment of benthic invertebrates in high Arc- K (2009) Is there a latitudinal gradient in the importance
tic fjords: relation to temperature, depth, and season. of biotic interactions? Annu Rev Ecol Evol Syst 40:
Limnol Oceanogr 62:2732–2744 245–269
Mileikovsky SA (1970) Seasonal and daily dynamics in pela- Schoschina EV (1996) Seasonal and age dynamics of growth
gic larvae of marine shelf bottom invertebrates in near- and reproduction of Phycodrys rubens (Rhodophyta) in
shore waters of Kandalaksha Bay (White Sea). Mar Biol 5: the Barents and White Seas. Aquat Bot 55:13–30
180–194 Seed R, Harris S (1980) The epifauna of the fronds of Lam-
Miller RJ, Page HM, Reed DC (2015) Trophic versus struc- inaria digitata Lamour in Strangford Lough, Northern
tural effects of a marine foundation species, giant kelp Ireland. Proc Royal Ir Acad B 80:91–106
(Macrocystis pyrifera). Oecologia 179:1199–1209 Seed R, Hughes RN (1992) Reproductive strategies of epial-
Moore AFP, Duffy JE (2016) Foundation species identity and gal bryozoans. Invertebr Reprod Dev 22:291–300
trophic complexity affect experimental seagrass commu- Shinomiya Y, Chiba S, Kanamori M, Hashizume S, Yoshino
nities. Mar Ecol Prog Ser 556:105–121 K, Goshima S (2017) Importance of patch size variation
Nandakumar K (1996) Importance of timing of panel expo- for the population persistence of a decapod crustacean in
sure on the competitive outcome and succession of ses- seagrass beds. Mar Ecol Prog Ser 570:157–171
sile organisms. Mar Ecol Prog Ser 131:191–203 Stachowicz JJ (2001) Mutualism, facilitation, and the struc-
Navarrete SA (1996) Variable predation: effects of whelks on ture of ecological communities: Positive interactions play
a mid-intertidal successional community. Ecol Monogr a critical, but underappreciated, role in ecological com-
66:301–321 munities by reducing physical or biotic stresses in exist-
Nydam M, Stachowicz JJ (2007) Predator effects on fouling ing habitats and by creating new habitats on which many
community development. Mar Ecol Prog Ser 337:93–101 species depend. Bioscience 51:235–246
Oh CW, Hartnoll RG, Nash RDM (1999) Population Stachowicz JJ, Whitlatch RB (2005) Multiple mutualists
dynamics of the common shrimp, Crangon crangon (L.), provide complementary benefits to their seaweed host.
in port Erin Bay, Isle of Man, Irish Sea. ICES J Mar Sci 56: Ecology 86:2418–2427
718–733 Teagle H, Hawkins SJ, Moore PJ, Smale DA (2017) The role
Olafsson E (ed) (2016) Marine macrophytes as foundation of kelp species as biogenic habitat formers in coastal mar-
species. CRC Press, Boca Raton, FL ine ecosystems. J Exp Mar Biol Ecol 492:81–98
Paalme T, Kotta J, Kersen P, Martin G, Kukk H, Torn K Temming A, Hufnagl M (2015) Decreasing predation levels
(2011) Inter-annual variations in biomass of loose lying and increasing landings challenge the paradigm of non-
algae Furcellaria–Coccotylus community: the relative management of North Sea brown shrimp (Crangon cran-
importance of local versus regional environmental fac- gon). ICES J Mar Sci 72:804–823
tors in the West Estonian Archipelago. Aquat Bot 95: Tiews K (1978) The predator–prey relationship between fish
146–152 populations and the stock of brown shrimp (Crangon
Paine RT (1974) Intertidal community structure: experimen- crangon L.) in German coastal waters. Rapp P-V Réun
tal studies on the relationship between a dominant com- Cons Int Explor Mer 172:250–258
petitor and its principal predator. Oecologia 15:93–120 Underwood AJ, Keough MJ (2001) Supply-side ecology: the
Palmer MA, Allan JD, Butman CA (1996) Dispersal as a nature and consequences of variations in recruitment
regional process affecting the local dynamics of marine of intertidal organisms. In: Bertness MD, Gaines SD,
and stream benthic invertebrates. Trends Ecol Evol 11: Hay ME (eds) Marine community ecology. Sinauer Asso-
322–326 ciates, Sunderland, MA, p 183–200
 Chava et al.: Top-down control in sub-arctic algal epibiosis 67

Ushakova OO (2003) Combined effect of salinity and tem- Whalen MA, Duffy JE, Grace JB (2013) Temporal shifts in
Author copy

perature on Spirorbis spirorbis L. and Circeis spirillum L. top-down vs. bottom-up control of epiphytic algae in a
larvae from the White Sea. J Exp Mar Biol Ecol 296:23–33 seagrass ecosystem. Ecology 94:510–520
van der Veer HW, Bergman MJN (1987) Predation by crusta- Wieland K, Storr-Paulsen M, Sünksen K (2007) Response in
ceans on a newly settled 0-group plaice Pleuronectes pla- stock size and recruitment of northern shrimp (Pandalus
tessa population in the western Wadden Sea. Mar Ecol borealis) to changes in predator biomass and distribution
Prog Ser 35:203–215 in West Greenland waters. J Northwest Atl Fish Sci 39:
Vogt G (2019) A compilation of longevity data in decapod 21–33
crustaceans. Nauplius 27:e2019011 Yakovis E, Artemieva A (2019) Epibenthic predators con-
Wahl M, Hay ME (1995) Associational resistance and shared trol mobile macrofauna associated with a foundation
doom: effects of epibiosis on herbivory. Oecologia 102: species in a subarctic subtidal community. Ecol Evol 9:
329–340 10499–10512
Ware C, Dijkstra JA, Mello K, Stevens A, O’Brien B, Ikedo W Yakovis E, Artemieva A (2021) Effects of a trophic cascade
(2019) A novel three-dimensional analysis of functional on a multi-level facilitation cascade. J Anim Ecol 90:
architecture that describes the properties of macroalgae 2462–2470
as a refuge. Mar Ecol Prog Ser 608:93–103 Yakovis E, Artemieva A, Fokin M, Varfolomeeva M, Shuna-
Węsławski JM (1987) Distribution of Decapoda (Crustacea) tova N (2013) Synchronous annual recruitment variation
in South Spitsbergen coastal waters with remarks on their in barnacles and ascidians in the White Sea shallow sub-
ecology and breeding biology. Pol Polar Res 8:121–134 tidal 1999–2010. Hydrobiologia 706:69–79

Appendix
Table A1. Water temperatures (°C) near the experimental sites in summer
2009–2016. Measured at 66° 33.177’ N, 33° 6.181’ E in Velikaya Salma Strait,
Kandalaksha Bay, at 1 m depth. The 2 years of the field experiments are high-
lighted in bold. The warmest and coldest values for each date are highlighted
in yellow and blue, respectively

Year 5 June 25 June 5 July 25 July 5 August 25 August

2009 6.1 9.4 11.1 11.7 14.2 11.6

2010 6.1 10.3 11.5 12.3 14.7 9.4
2011 6.9 12.2 13.1 15.2 14.8
2012 5.4 11.1 10.3 10.2 12.3 11.1
2013 9.6 12.6 15.2 13.7
2014 7.9 10.1 10.0 16.3 16.5 13.1
2015 5.3 7.2 9.4 10.6 12.1 12.2
2016 8.5 11.7 13.7 14.6

Editorial responsibility: Just Cebrian, Submitted: November 15, 2023

San Francisco, California, USA Accepted: July 17, 2024
Reviewed by: 3 anonymous referees Proofs received from author(s): August 18, 2024