complex effects of ecosystem engineer loss on benthic ... · open 3 access freely available online

OPEN 3 ACCESS Freely available online <P PLOS I o h .

Complex Effects of Ecosystem Engineer Loss on Benthic Ecosystem Response to Detrital MacroalgaeFrancesca Rossi1*, Britta Gribsholt2, Frederic Gazeau3'4, Valentina Di Santo5, Jack J. M iddelburg2'61 Laboratoire Ecologie des Systèmes marins côtiers, Université Montpellier 2, Montpellier, France, 2 Department of Ecosystems, Royal Netherlands Institute for Sea Research, Yerseke, the Netherlands, 3 Centre National de la Recherche Scientifique-lnstitut National des Sciences de l'Univers, Laboratoire d ' Océanographie de Villefranche, Villefranche-sur-mer, France, 4 Université Pierre e t Marie Curie, Observatoire Océanologique de Villefranche, Villefranche-sur-mer, France, 5 Department of Biology, Boston University, Boston, Massachusetts, United States of America, 6 Department of Earth Sciences - Geochemistry, Faculty of Geosciences, Utrecht University, Utrecht, The Netherlands

AbstractEcosystem eng ineers c h a n g e abiotic conditions, c o m m u n ity assem bly an d ecosystem functioning. Consequently , their loss may modify th resholds o f ecosystem response to d is tu rbance and und e rm in e ecosystem stability. This s tu d y investigates ho w loss o f th e b io turba ting lugworm Arenicola marina modifies t h e response to macroalgal detrital en r ich m en t of se d im e n t b iogeochem ica l propert ies , m ic ro p h y to b en th o s and m acrofauna assem blages . A field manipulat ive ex per im en t was d o n e on an intertidal sandflat (Oosterschelde estuary, The Netherlands). Lugworms w ere deliberately excluded from 1 X m se d im e n t plots and different a m o u n ts o f detrital Ulva (0, 200 or 600 g W et Weight) w ere a d d e d twice. Sed im ent b iogeochem is try c h an g e s w ere evaluated th ro u g h ben th ic respiration, se d im e n t organ ic carbon c o n te n t a n d po rew ate r inorganic c arbon as well as detrital m acroalgae remaining in th e s e d im e n t o n e m o n th after enr ichm ent. Microalgal biomass and m acrofauna com pos it ion w ere m easured at th e sa m e time. Macroalgal c arbon mineralization a n d transfer to th e b en th ic co n su m ers w ere also inves tiga ted during d eco m p o s i t io n a t low en r ich m en t level (200 g WW). The interaction b e tw ee n lugw orm exclusion an d detr ital en r ich m en t did no t modify s e d im e n t organ ic c arbon o r ben th ic respiration. Weak b u t significant c h an g es w ere instead found for p o re w a te r inorganic carbon and microalgal biomass. Lugworm exclusion caused an increase o f p o rew ate r c arbon and a dec rease o f microalgal biomass, while detrital en r ich m en t drove th ese values back to values typical of lugw o rm -d o m in a ted sed im ents . Lugworm exclusion also d e creased th e a m o u n t of m acroalgae remaining into th e sed im e n t and accelera ted detrital c arbon mineralization an d C 0 2 release to th e w a te r colum n. Eventually, th e interaction b e tw e e n lugworm exclusion and detrital en r ich m en t affected m acrofauna a b u n d a n c e and diversity, which collapsed a t high level o f en r ich m en t only w h e n th e lugw orm s w ere present. This s tu d y reveals th a t in na ture th e role of this ecosystem e n g in e e r m ay be variable a n d so m e tim es have no o r even negative effects on stability, conversely to w h a t it should be ex p ec te d based on curren t research knowledge.

Citation: Rossi F, Gribsholt B, Gazeau F, Di Santo V, Middelburg JJ (2013) Complex Effects of Ecosystem Engineer Loss on Benthic Ecosystem Response to DetritalMacroalgae. PLoS ONE 8(6): e66650. doi:10.1371/journal.pone.0066650

Editor: Andrew Davies, Bangor University, United Kingdom

Received September 25, 2012; Accepted May 10, 2013; Published June 21, 2013

Copyright: © 2013 Rossi e t al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was funded by a PIONIER grant of the Netherlands Organization for Scientific Research to JJM. The funders had no role in study design, dataanalysis, decision to publish, or preparation of the manuscript.

Com peting Interests: The authors have declared that no competing interest exist.

* E-mail: francesca.rossi@univ-montp2

Introduction

It is widely recognized that species loss m ay affect ecosystem stability and, in turn , have im portan t social and ecological consequences [1—3]. Overall, w hen m ore species are present in a com m unity, ecosystem functions are m ore stable in tim e and increase their resistance o r resilience against disturbance, due to an increased variety o f life strategies, functional traits and responses to environm ental d isturbance [4—6], D ifferent species contributing to similar ecosystem functions can occur under different environm ental conditions an d ensure that functions are perform ed even w hen some o ther species are displaced [4,5,7].

Em pirical evidence has shown that sometimes species loss can have complex, often idiosyncratic effects on the stability o f ecosystem functions an d on bo th their resistance and resilience against d isturbance [8,9], This is especially true for the m arine coastal ecosystem, w here species identity, functional traits and environm ental context m ay rule ecosystem functions and their

response to disturbances [10-12]. O n e or few species can be key contributors to ecosystem functions to the extent that their loss can outweigh the effect o f species richness on ecosystem functions, resilience o r resistance [10,13,14], Such overwhelm ing effect can, however, vary with the environm ental context and change with increasing ecological complexity [15]. It seems thus evident that understand ing the role o f species loss in m arine systems m ight require a large body of knowledge o f the functional role o f individual species and o f the consequences o f losing particular species for ecosystem functions and for their response to disturbance.

Ecosystem engineer species strongly m odify resource availability and environm ental conditions, w hich in tu rn can alter com m unities and ecosystem functions [16,17]. T hese species are thus ideal candidates to test hypotheses on how local extinction o f key contributors to ecosystem functions can m odulate the response of ecosystems to disturbance. In m arine sediments and terrestrial soils, ecosystem engineers are often species that contribute to

PLOS ONE I www.plosone.org 1 June 2013 | Volume 8 | Issue 6 | e66650

http://www.plosone.org

Complex Effects of Ecosystem Engineer Loss

particle m ixing and solute penetration into the sedim ent through the process o f b io turbation , w hich includes sedim ent reworking and burrow ventilation [18], T hese b io tu rbating engineer species can be great contributors to diagenetic reactions, to the recycling o f sedim entary organic m atter, and to benthic com m unity structure and biodiversity [18-21]. T hey also have the potential to strongly m odify the response to d isturbance o f benthic com m unities and functions [22]. For instance b io turba tion by Marenzellaria viridis has been found to decrease hypoxic crises in the Baltic Sea [23]. It has also been found that in soil, reworking activity o f dom inan t contributors to b io turbation , such as earthw orm s, m odulates the response o f p lan t com m unities to invasions [24]. Surprisingly, however, ecosystem engineer loss has been poorly considered in studies investigating ecosystem response to disturbance.

In m arine coastal systems, a large portion o f seaweeds and seagrasses is no t consum ed by herbivores bu t returns to the environm ent as decaying organic m atter [25]. In shallow-water coastal habitats, detrital seaweeds an d seagrasses can supply im portan t am ounts o f organic m atter to benthic ecosystems. T hey represent an im portan t source o f d isturbance because during decom position they can alter sedim ent biogeochem istry and, in turn , benthic com m unities o f consum ers. Despite detrital seaweed and seagrass supply is often a natural event, proliferation and bloom s o f seaweeds as a consequence of eutrophication can dram atically increase detrital seaweed biomass and exacerbate the effects on benthic ecosystems [26-29]. Biomass decom position can induce hypoxia followed by the release o f reduced toxic com pounds such as sulphides, w hich are toxic for several benthic species and can therefore strongly reduce species abundance and diversity [26,27,29-33]. D etrital decom position can also provide food, directly for detritivores o r indirectly, by stim ulating bacterial m etabolism , thereby altering the benthic food web [28,34]. These effects strongly depend on the am ount o f detrital inputs and on its decom position patterns, w hich m ay strongly vary as a consequence o f benthic invertebrate com position and identity [35].

B ioturbating fauna can m odulate the recycling of detrital m acroalgae enhancing organic m atter m ineralization and m icrobial activity [36-38]. Som e burrow ing b io turbating species, such as certain crabs o r polychaetes, can also incorporate detritus into their burrow wall lining and represent im portan t sinks for detrital m acroalgae [39]. In addition, often b io tu rbating infauna are also detritivores that can consum e directly detritus and contribute to its recycling in the food web. In a laboratory experim ent, for instance, it was found that nitrogen and carbon fluxes at the sedim ent-water interface changed following b o th the addition o f the detrital green m acroalgae Chaetomorpha linum and the presence of the b io tu rbating polychaete Nereis diversicolor. This polychaete ingested m acroalgal fragm ents, enhanced m icrobial m etabolism and con tribu ted to diffuse solutes in the interstitial sediment and at the sediment- w ater interface [36],

O n e com m on well-recognized b io tu rbating species typical o f tem perate waters is the burrow ing lugworm Arenicola marina. This w orm lives head dow n in J-shaped burrow s w here it ingests sedim ent and defecates on the surface. Its activity has been recognized to destabilize sediment and counteract the effect o f m icrophytobenthos, increase sedim ent m icrohabitats and alter sedim ent biogeochem istry, by enhancing solute penetration into sedim ent dep th an d re-distributing organic m aterial along the vertical profile [40], C onsequently, this species m ay alter o ther fauna and vegetation [36,37,41], This study reports on how the loss o f this b io tu rbating lugw orm m ay alter the response to detrital m acroalgae of intertidal benthic m acrofauna and sediment biogeochem istry. It was hypothesized that the deliberate exclusion

o f lugworm s w ould increase the m agnitude o f changes caused by the d isturbance o f m acroalgal detritus for (i) carbon availability and benthic respiration; (ii) m icroalgal biom ass and (iii) m acrofauna assemblages. Recycling o f the organic carbon derived from detrital decom position was also investigated to explore some of the m echanism s that m ight regulate the effect o f el. marina on benthic responses to m acroalgal detritus.

Materials and Methods

S tud y Site a n d E xperim en ta l D esignTw o m anipulative experim ents were done on an intertidal flat

o f the Oosterschelde estuary (The N etherlands) dom inated by the b io turbating lugworm Arenicola marina, during sum m er 2006. T he study intertidal flat, nam ed Slikken van V iane (51" 37 ' 0 0 " N, 4" 01 ' 0 0 " E), was a sheltered a rea situated on the upper intertidal close to a salt m arsh. Sedim ent grain-size was 65% fine sand and 35% very fine sand (Rossi F., unpublished data). T he O osterschelde estuary is a national park designated as zone V I by the In ternational U nion for C onservation of N ature (IUCN). This IU C N classification has am ong its objectives the facilitation o f scientific research and environm ental m onitoring, m ainly related to the conservation and sustainable use o f natural resources (h ttp ://w w w .iu cn .o rg ). No specific perm its w ere therefore required for perform ing these scientific experim ents, w hich did not involve endangered or pro tected species. T h e first experim ent tested the effect o f el. marina exclusion on the response of m acrofauna assemblages and sediment biogeochem istry to different am ounts o f detrital m acroalgae. T he second experim ent m easured the changes in detrital recycling following el. marina exclusion. For the first experim ent (Fig. 1), all lugworm s were excluded from 15 random ly chosen sedim ent plots o f l x l m, some m eters apart, by burying m osquito nets to the dep th o f approxim ately 7 cm (hereafter exclusion A — treatm ent), following the m ethod proposed by [41] in a m ore extensive area. A procedural control for the effect o f the net an d o f the burial was done for o ther 15 sedim ent plots o f 1 x i m by burying m osquito nets w ith 1 0x10 cm holes (hereafter p rocedural control A+ treatm ent).

O n e m onth after lugw orm exclusion, the sedim ent plots were organically-enriched by addition o f detrital Ulva spp. Fresh Ulva thalli w ere previously collected in the Oosterschelde estuary and w ashed to rem ove anim als and sediment. Thalli were then w eighted an d frozen in separate plastic bags. Freezing is

■ Lugworm excluded plots (A -) □ Procedural control plots (A +)

600 0 200 600 0 200

200 0 600 600 0 200

0 200 0 200 600 600

200 600 0 0 600 200

600 0 200 600 0 200

Figure 1. Schematic experimental design for experim ent 1. Each rectangle corresponds to 1 m2 plot. Numbers refers to the detrital enrichm ent. (0: no addition; 200:200 g WW w ere added twice a t a 1 w eek interval; 600:600 g WW w ere added twice with a 1 w eek interval). Distance am ong plots are no t in scale. doi:10.1371/journal.pone.0066650.g001


http://www.iucn.org



com m only used in experim ents th a t aims at testing effects of detrital m acroalgae and decom position patterns. This treatm ent is necessary in order to create detritus and ensure th a t Ulva is dead when buried since this seaweed can survive for very long periods in the dark [30]. T h e thalli were left defrosting a t air tem perature and then added to the top 1 cm o f the sedim ent by gently hand- churning the detritus with the sediment. E nrichm ent was done twice, on Ju ly 3 and Ju ly 10, 2006. Each tim e either 200 or 600 g of w et w eighted Ulva (hereafter treatm ents 200 and 600, respectively) were added to 10 exclusion (A—) and 10 procedural control (A—) plots. These am ounts corresponded to 40 and 120 g dry weight. O th er 10 plots (5 A — and 5 A+) were no t enriched (hereafter trea tm ent 0), bu t the sedim ent was hand-churned to control for the m anipulation. T h ere was no procedural control for the physical presence of detritus as previous experim ents had dem onstrated no effects [28-30]. T h e am ount o f Ulva spp. was based on the available literature to cause m edium and high levels o f disturbance (e.g. [27,30,32,33]).

T h e second experim ent was ru n at the sam e tim e o f experim ent 1. Tw o additional p rocedural control plots for burial and 2 exclusion plots were enriched with 200 g loC-labelled Ulva twice, resem bling the trea tm en t “ 200” . These four plots were used to trace the fate of the carbon released from detrital Ulva. T he labeling of Ulva was done in the laboratory. Leaves were washed and carefully cleaned to elim inate epiphytes and animals. T hen , thalli were transferred to aquaria containing filtered seawater supplem ented w ith N a H loC 0 3 (99% loC) to the concentration of 0.2 g L \ T h e aquaria were oxygenated and covered with transparen t PV C to limit 1 'C bacterial respiration.

Field S am plingFor the first experim ent, sam pling was done on Ju ly 30 and

August 1, 2006, one m onth after detrital enrichm ent. This tim ing was chosen because previous studies suggested th a t effects o f algal m ats on m acrofauna occur within 2 to 10 weeks from the deposition [31-33,42] and that decom position o f Ulva occurs w ithin a m onth (e.g. [43,44]).

Sedim ent samples were taken for organic carbon (hereafter OC), porew ater total inorganic carbon (DIC — H 2C O 3+ H C O 3 + C O 32 ), m icroalgal biomass (chlorophyll a) and m acro fauna species com position, abundance and diversity. T h e sedim ent for chlorophyll a and O C was collected using a cut syringe of

3 111111 inner diam eter (i.d.). For chlorophyll a, the sedim ent was collected to 1 cm depth, kept in the dark and preserved at — 80°C. T h e sedim ent for O C was collected to the depth of 6 cm, sectioned into 0- 1, 1-2 and 2 -6 cm depths and preserved freeze- dried. Ulva detritus was rem oved from this sedim ent un d er the stereo-m icroscope. T h e sedim ent for porew ater D IC extraction was collected with cores o f 5 cm i.d. and sectioned into 0 -1 , 1-2 and 2 -6 cm depth intervals at the lowest and highest levels of organic enrichm ent (0 and 600) in both the control and the exclusion treatm ents. Tw o cores were collected from each plot and then pooled to reach the am ount o f w ater needed for the analyses. Porew ater was extracted by centrifugation (1344 g, 15 min) and the supernatan t was filtered on M illipore 45 |Im . T he extracted w ater was stored in glass vials, sealed, im m ediately poisoned with saturated H gC F to stop bacterial activity and analyzed w ithin 3 d. O nly 3 out o f 5 random ly-chosen replicate plots were sam pled for O C and T C O 2.

T h e sedim ent for m acrofauna was collected using 11 cm i.d. PV C core. T his sedim ent was sieved (0.5 n in i mesh) and m acrofauna fixed in 4% buffered formol. Before sampling, all plots were pho tographed to estim ate the abundance o f A. marina using visible m ounds. In the laboratory, all plots were independently exam ined by four researchers to reduce estim ate bias. Five random ly-chosen na tura l plots were also photographed and then sam pled using 11 cm i.d. cores to test for the effect o f the p rocedural control A+ on the na tura l abundance of A. marina and o f the o ther benthic m acrofauna. N o differences were detected betw een na tura l and procedural control treatm ents (see Results section) and this latter was used for evaluating the effect of detrital enrichm ent and lugw orm exclusion.

Sedim ent cores were also used to m easure fluxes o f inorganic carbon (TCCF) and oxygen (CF) at the sedim ent-w ater interface. In tact sedim ent cores were collected by pushing Plexiglas cores (11 cm i.d.) into the sedim ent to a depth o f 20 cm at each o f three replicate plots for the lowest and the highest levels o f organic enrichm ent (0 and 600) in bo th the control and the exclusion treatm ents. T h e net was carefully cut a round the cores to allow the core penetration into the sediment. Cores were sealed w ith rubber stoppers and dug out o f the sediment. All cores were b rought to a clim ate controlled room within 2 h, and kept in the dark at in situ tem perature (10°C) before further handling. T h e same day, the cores were incubated in the dark for 12 h. Incubation in the dark

Arenicola-Qxeluded sediment(A-)

Procedural control sediment(A+)

Figure 2. Comparison of sediment surface appearance after Arenicola marina exclusion at low tide. A - represent th e sedim ent after 1 m onth from the burial of net to exclude burrowing species. A+ is th e procedural control trea tm en t (see Materials and M ethods for details). doi:10.1371/journal.pone.0066650.g002




□ 0 ■ 200 ■ 600

A + A - A + A - A +

lcm 2 cm 6 cm

(b)

y 1 Id Id iA + A +

lcm 2 cm

A - A +

6 cm

Figure 3. Mean (+1SE) for (a) bulk sediment organic carbon (OC) and (b) porewater dissolved inorganic carbon (DIC) at different sediment depths { I cm surface sediment to 1 cm depth; 2 cm. from the depth of 1 to 2 cm; 6 cm: from the depth of 2 to 6 cm). In th e graphs, A—: A. marina excluded; A+: procedural control for burial; th e legend indicate the levels o f detrital addition (0: no addition; 200:200 g WW w ere added twice a t a 1 w eek interval; 600:600 g WW w ere added tw ice with a 1 w eek interval). Data were sam pled one m onth after detrital addition. Asterisk indicates significant differences.doi:10.1371/journal.pone.0066650.g003

did no t allow to gather additional inform ation concerning the (auto)trophic state o f the sediment, w hich w ould need a light incubations o f o ther sediment cores. O u r system was however lim ited to carry 12 cores a t a tim e and no further analyses could be done. P rior to incubations, cores were inundated w ith well- oxygenated O osterschelde w ater collected the same day ([N H 4+] = 20 pm ol L", [ N 0 3~] = 110 pm ol L “ 1, salinity = 22) and left to acclim atize in the dark for 6 -8 h. D uring incubation, the overlying w ater-colum n (25-30 cm) was continuously stirred with a m agnetic stirrer, m aintain ing continuous w ater circulation at a ra te well below the resuspension limit. Dissolved 0 2 never decreased below 65% o f saturation. W ater samples for T C 0 2 and 0 2 were collected in glass vials and sealed avoiding a ir bubbles. All samples were im m ediately poisoned w ith saturated H gC l2 to stop bacterial activity and then analyzed w ithin 3 d.

For experim ent 2, sedim ent samples were taken from the four plots w ith labelled Ulva. Sam pling was done on four occasions, 6 , 8 , 15 an d 17 days after detrital enrichm ent. E ach time, one plexiglass intact sedim ent core (11 cm i.d.) was collected to 20 cm depth, following the m odality used to sample the cores for estim ating benthic respiration. T o reduce plot dam age during core collection, the central pa rt o f the plot (20 cm from the edges) was divided into 4 quadrants o f 3 0 x 3 0 cm. E ach sam pling occasion, a sedim ent core was taken from a different quadran t. An area o f 15x15 cm was thus dam ages each tim e, corresponding to 2.25% o f the entire plot area. T h e sam pled sedim ent was replaced with plexiglass tubes to avoid any collapse an d to leave the rem aining

Table 1. Two-way analyses o f variance for s e d im e n t organic carbon (OC) an d p o rew ate r DIC a t different s e d im e n t d e p th s o n e m o n th after detrital enr ichm ent.

0 -1 cm 1 -2 cm

2 -6 cm

Sedim ent OC d f MS F MS F MS F

Exclusion = E 1 0.67 10-3

0.40 1.60 10~3 3.80 0.0910-3

0.52

Detritalenrichment = D

2 0.61 10-3

0.96 0.09 10~3 0.40 0.6210-3

2.22

ExD 2 1.7 10" 3 2.64 0.42 10~3 1.90 0.1710-3

0.62

Residual 12 0.63 10-3

0.22 10~3 0.2810-3

Porewater DIC

Exclusion = E 1 1.89 17.50 5.06 1.03 10.15 4.80


1 0.06 2.00 1.28 5.12 0.15 0.28

ExD 1 0.11 3.43 4.93 19.71 0.30 0.57*

Residual 8 0.03 1.00 2.11

Significant values (P<0.05) are n bold. Data were not-transformed (PC-cochran>0.05). Asterisk indicates pooling of interaction when P>0.25.

(a) 150.00

•3 100.00

P

o<1

50.00

0.00

-50.00

- 100.00

Œ IA- A+

(b) 2.5

1.5

0.5

A- A+

Figure 4. Mean (+1 SE) for (a) benthic oxygen (0 2) consumption and total carbon dioxide (TC 02) released during 12 hours dark incubation; (b) Respiratory quotient (RQ). In the graphs, A : A. marina excluded; A+: procedural control for burial; the legend indicate the levels of detrital addition (0: no addition; 200:200 g WW were added twice a t a 1 w eek interval; 600:600 g WW w ere added twice with a 1 w eek interval). Data w ere sam pled one m onth after detrital addition. doi:10.1371/journal.pone.0066650.g004




Table 2. Two-way analyses o f variance for se d im en t-w a te r ex ch an g es of 0 2 and T C 02 and for th e ben th ic respiratory coefficient RQ.

o 2 t c o 2 RQ

df MS F MS F MS F

Exclusion = E 1 8.8 10-6 2.38 252.08 0.78 0.29 2.39


1 8.0 10-9 0.00 102.08 0.31 0.02 0.18

ExD 1 4.38 1 ~7 0.11* 70.08 0.22* 0.00 0.01*

Residual 8 4.10 1-6 325.08 0.14

Significant values (P<0.05) are in bold. Data were not-transformed (Pc-cochran >0.05), except for O2, which was arc-tangent transformed to hom ogenise the residual variances. Asterisk indicates pooling of interaction when P>0.25. doi:10.1371 /journal.pone.0066650.t002

p art o f the plots undisturbed. T h e cores were handled like the cores used for evaluating oxygen and carbon exchange at the sedim ent-w ater interface for the experim ent 1. How ever, before incubating for wet respiration and determ ining release o f T C 0 2 in the w ater colum n, the labelled cores w ere incubated for dry fluxes during 6 h. A t the beginning an d a t the end o f the incubation, air samples were collected using syringes an d pum ped into tubes contain ing soda lim e to trap C 0 2 and m easure the

C 0 2 derived from Ulva decom position. A L IC O R autom ated C 0 2 analyzer was used to quantify C 0 2 release every hour. Sedim ent-w ater exchange ra te and m acrofauna da ta were generated as for experim ent 1 (see “Analytical m easures” section).

Following flux incubations for bo th experim ents, the cores were sieved at 0.5 nin i to collect the m acrofauna. No lugworm was included in the cores, m eaning that any observed patte rn represented the result o f sedim ent and m acrofauna alteration due to Arenicola, bu t not the direct activity o f the lugw orm during incubation.

Analytical M easu resIn the laboratory, m acrofauna were identified to the species

level, counted and dried a t 60"C for 48 h to obtain dry weight biomass to be used for stable isotope analyses relative to the experim ent 2. T h e detrital Ulva rem aining in the samples was also collected, dried and weighed. Chlorophyll a was extracted in darkness for 24 h a t 0 -4 "C using a 90% acetone solution. T he sedim ent was hom ogenized an d sub-samples o f about 1 g dry weight were taken. Pigm ents w ere m easured spectrophotom etri- cally before and after 0.1 N HC1 acidification. Pre-com busted (450"C for 3 h) sedim ent was used as a blank. M easurem ents for bulk sedim ent O C were m ade using a Perkin-Elm er C H N elem ent analyzer on freeze-dried m aterial, after carbonate rem oval with HC1.

Flux rates o f T C 0 2 and 0 2 betw een the sedim ent and the overlying w ater w ere estim ated as difference betw een the initial and the final T C 0 2 or 0 2 concentrations during dark incubation, assum ing constant solute exchange with time. T h e oxygen dissolved into the w ater samples was analyzed using the standard m ethod of W inkler titration, while w ater T C 0 2 and porew ater D IC w ere determ ined by the flow in jection/diffusion cell technique, w hich is based on the diffusion o f C 0 2 th rough a hydrophobic m em brane into a flow o f deionized water, thus generating a gradient o f conductivity p roportional to the concentration of C 0 2.

30 [ 0 D200 ■ 600

m 25 M)

A - A+

Figure 5. Mean (+1SE) for chlorophyll a concentration at the surface sediment (1 cm depth). In the graphs, A - : A. marina excluded; A+: procedural control for burial; th e legend indicate the levels of detrital addition (0: no addition; 200:200 g WW w ere added twice a t a 1 w eek interval; 600:600 g WW were added twice with a 1 w eek interval). Data were sam pled one m onth after detrital addition. Asterisk indicates significant differences. doi:10.1371/journal.pone.0066650.g005

W ater T C 0 2 (wet fluxes), gaseous C 0 2 (dry flux), bulk sediment and m acrofauna collected in the labelled plots were analyzed for carbon isotope com position (8 C). T h e gaseous C 0 2 (dry flux) cap tured using soda lime was analyzed after acidification. T h e 13C isotope signature for T C 0 2 dry and wet fluxes was analyzed using a GC colum n coupled to a Finningan delta S mass spectrom eter IR M S. T h e carbon isotopic com position for sediment, residual detritus an d anim als was determ ined using a Fisons elem ental analyzer coupled to a F inningan delta S mass spectrom eter. 13C to *~C ratio was referred to V ienna PDB and expressed in units o f %o. R eproducibility o f the m easurem ents is bette r than 0.2%o [45]. T h e excess o f the heavy isotope o f carbon (above background) was expressed as total uptake (I) in m illigrams of 13C. I was calculated as the p roduct o f excess 13C (E) and carbon or biomass (C):

I = E X C

E = Fcontrol F¡¡„„pie

w here F indicates the fraction o f isotope added an d it is calculatedas:

F = R /R + 1

w here R can be calculated from the m easured S13C using Rreference of 0.0112372:

4 f \Rsam ple / ̂ reference I X 10'

Statistical A nalysesFor experim ent 1, da ta were analyzed with two-way orthogonal

m ixed m odel o f analysis o f variance (ANOVA), w ith exclusion (2 levels: p rocedural control A+ and exclusion o f .4. marina, A —) and detrital enrichm ent (3 o r 2 levels: 0, 200 an d 600) as fixed and random factors, respectively. D etrital enrichm ent was considered a random factor because the am ount o f detritus to add was chosen w ithin a range of values to simulate conditions o f low and high




Table 3. Two-way analyses o f variance for chlorophyll a, m acrofauna a b u n d a n c e and n u m b er (N) of species.

Chlorophyll a Abundance N o f species

df MS F MS F MS F

Exclusion = E 1 72.23 1.27 0.48 0.69 0.03 0.00

Detrital enrichment = D 2 53.74 4.02 0.29 0.56 30.00 7.11

ExD 2 56.80 4.25 2.34 4.81 17.73 4.21

Residual 24 13.37 0.49 4.22

Significant values (P<0.05) are in bold. doi:10.1371 /journal.pone.0066650.t003

Data were sqrt-transformed for macrofauna abundance.

enrichm ent, bu t no specific hypothesis concerned the chosen am ounts. A significant interaction betw een exclusion an d detrital enrichm ent was expected if the exclusion o f A. marina changed the response to detrital enrichm ent. These analyses were done for the total abundance o f m acrofauna individuals, num ber o f m acrofaun a species, chlorophyll a, sedim ent O C , porew ater D IC , w ater 0 2 and T C 0 2 exchange rate. C-cochran test was used to test for residual variance hom ogeneity before runn ing the analyses. W hen this test was significant (iP<0.05), da ta were transform ed. Pooling was done w hen P > 0.25. A posteriori m ultiple com parison SN K test was run w hen interaction term was significant.

T h e variables collected for evaluating the recycling of detrital 13C (experim ent 2) were analyzed w ith repeated m easure A N O V A , w ith tim e (6 , 8, 15 and 17 sam pling days from the end o f detrital addition) as the repeated m easure random factor and A. marina exclusion as fixed factor. Plots (2 levels) was nested in exclusion. T he interaction of tim e with the exclusion treatm ent was m easured over the interaction tim exp lo t. W hen significant interaction was found, the m ain effect exclusion was not interpreted . A ssum ption about the correlation am ong the different observations o f the repeated m easure factor (time) was tested with M auchley’s sphericity test.

D ecom position rate o f m acroalgal detritus was estim ated by fitting the first-order model:

B, = B0ek'

W here Bt is the g o f detrital biomass a t tim e t (days), B0 is the initial biomass, a t tim e 0 , a t the m om ent o f the second detrital addition and k is the specific decom position constant. T h e two detrital additions were done at a distance o f one week, w hich m ay greatly affect the estim ates o f B0. W e therefore considered the estimates o f residual Ulva as the sum o f two decom position curves, w ith the same constant k bu t different decom posing time:

Bt = B (0_ 1)ekl't+1) + B (0)ekt

T h e curves o f 13C benthic incorporation (including loss due to respiration) w ere fitted to the logistic curve:

I t = Y / ( \ + q e - " )

q = ( Y - I 0) / I 0

W here Y is the 13C loss from m acroalgal decom position and r is the capacity o f benthos to incorporate C.

Results

After the exclusion of A. marina, no burrow s w ere visible in the excluded plots (Fig. 2) and the num ber o f burrow s in the p rocedural control (treatm ent A+) was com parable to th a t o f na tu ra l sedim ent (m ean± S E , n = 5:8.7 2 ± 1.77 an d 10 .33±1.17 burrow s m 3 for the natural and the p rocedural control, respectively). Overall, there w ere no differences in m acrofauna species com position betw een the natural sedim ent and the p rocedural control plots. M acrofauna abundance (mean ± SE, n = 5) was 2 9 4 ± 4 2 and 2 2 0 ± 5 3 individual m 3, w hereas the num ber o f species was 10 ± 0 .2 and 10±0 .5 for A+ and natural sediment, respectively. Therefore, A+ could be used as a control for the effect o f lugworm exclusion and detrital m acroalgae.

Overall, in the natural sediments an d in the A+ plots the gastropod Hydrobia ulvae was num erically dom inant, representing 90% of m acrofauna abundance. T h e rem aining 10% was m ade up by oligochaetes (3%), deposit feeder polychaetes (3%) such as spionids, nereids, capitellids an d cirratulids and bivalves (3%), m ainly Macoma balthica an d Cerastoderma edule.

C arb o n Availabili ty a n d B enthic RespirationA t the tim e o f sampling, sediment organic carbon (OC) did not

vary following detrital enrichm ent, the exclusion o f A. marina or their interactions (Table 1, Fig. 3a; A N O V A , T > 0 .05). M acroalgal debris was still present in the sedim ent w here detrital m acroalgae were added a t the highest enrichm ent level (treatm ent 600). T he biomass o f this m acroalgal debris was smaller in the lugworm - excluded sedim ent th an in p rocedural controls (m ea n ± SE: lugworm exclusion, treatm ent A —: 2 .3 ± 1 .4 gD W m 3; A+: 1 2 .0± 6 .6 gD W m - 3 ; AN O V A : F lt „ = 6.19, P = 0.038).

T h ere was a significant effect o f the in teraction betw een lugworm exclusion an d detrital enrichm ent on porew ater D IC to the dep th o f 1-2 cm (Table 1; Fig. 3b). In the non-enriched sediment, porew ater D IC concentra tion was h igher in the lugworm exclusion th an in the control plots (SNK test at P= 0.05: non-enriched plots, treatm ent 0: A —>A +). M oreover, in the lugworm -exclusion plots, porew ater D IC was lower in the enriched than in the non-enriched plots (SNK test at i 5= 0 .0 5 : Arenicola exclusion, A —: 0> 600). V alues in detrital-enriched, lugw orm -excluded sedim ent w ere similar to those m easured for the A+ treatm ent (Fig. 3b). A similar pa tte rn was observed for superficial (0-1 cm) and deeper sedim ent (2-6 cm in Fig. 3b), bu t no significant changes w ere m easured.

Both A. marina exclusion and detrital enrichm ent o r their interaction did no t significantly change sedim ent-w ater exchanges o f T C 0 2 and 0 2 nor the RQ^ respiration coefficient ( T C 0 2: 0 2 ratio) (Table 2; Fig. 4).




(a)400

300

200

100

200

100

(c)

(d)

□ 0

■ 200 ■ 600

Hi12

10ooC/3

•2h"o(UQhGO

<Na 20

£ 150b2 10

tw0 5

0 U lA- A+

Figure 6. Mean (+1SE) for (a) total num ber of macrofauna individuals, (b) abundance of Hydrobia ulvae, (c) num ber of macrofauna species and (d) numbers of Arenicola marina burrows in the plots where the mosquito net was buried (A—) and in the procedural control plots (A+) at the end of the experim ent, e.g; 1 month from the second addition of detrital Ulva (0: no addition; 200:200 g WW were added twice at a 1 w eek interval; 600:600 g WW were added twice with a 1 week interval). D ata w e re sam p led o n e m o n th a fte r de trita l ad d itio n . A sterisk in d ica tes s ign ifican t d ifferences. do i:10 .1371/jo u rn a l.pone .0 0 6 6 6 5 0 .g 0 0 6

Microalgal Biom ass a n d M acro faun aT here was a significant interaction betw een A. marina exclusion

and detrital enrichm ent for m icroalgal biomass (Table 3; Fig. 5). Lugworm exclusion had a negative effect on m icroalgal biomass in the non-enriched plots (SNK test: 0: A + > A —). In the lugworm- excluded sedim ent, detrital enrichm ent enhanced m icroalgal biomass (SNK test: A —: 0 < 2 0 0 = 600). Values were similar to those m easured for the A+ control plots (treatm ents 200 and 600 in Fig. 5).

60

^ 50

bù 40

I 30 £

H? 20

S io

-0.06

-0.08

- 0.1

- 0.2

10 12 14 16 18

Time (days)

Figure 7. Decomposing curves for Ulva detritus in lugworm excluded (treatm ent A - ; square symbols in the graphs) or procedural control sediments (treatm ent A+; triangle symbols in the graph). Sym bols Ind ica te m e a n (± SE) for m acroa lgal b iom ass an d a re re la tive to e x p er im e n ts 2. do i:10 .13 7 1 /jo u rn a l.pone .0 0 6 6 6 5 0 .g 0 0 7

A significant interaction betw een A. marina exclusion and detrital enrichm ent was also found for m acrofauna abundance and num ber o f species (Table 3). A bundance was lowered by the highest level o f detrital enrichm ent in the plots w here A. marina was present (Fig. 6a; SN K test: A+: 0 = 200>600). Overall, bo th the dom inant gastropod Hydrobia ulvae and all o ther rem aining m acrofauna decreased in abundance, especially nereids, spionids and bivalves (ANOVA: F2, 94 = 4.26, P — 0.026 for H. ulvae, Fig. 6b; F2, 24 = 4.65, P — 0.019 for rem aining taxa, da ta were square-root transform ed; SN K tests for both: A+: 0 = 200>600). N um ber o f species showed a sim ilar pattern , decreasing in response to the highest level o f enrichm ent in the A+ plots (Fi. 6G; SN K test: A+: 0 = 200>600). D etrital enrichm ent seemed to not affect the abundance of the lugworms as the num ber o f Arenicola burrow s rem ained relatively stable at different levels o f detrital enrichm ent (2 way AN O V A : F2t 12 = 0.88, P= 0.34, Fig. 6d).

T he Fate o f Detrital C arbo nT h ere were differences in the rate o f Ulva decom position

following A. marina exclusion. T h e decom position constant k was — 0.2 for the detritus added to the A. marina excluded sedim ent (A— plots in Fig. 7). Interestingly, the A+ plots did not fit well the exponential curve using the same constant k through the decom position. R ather, at the beginning o f decom position, they fitted k values betw een —0.06 and —0.08, w hereas at the end they fitted a K constant o f —0.1 (Fig. 7). T h ere was a significant interaction betw een exclusion treatm ent and tim e after detrital enrichm ent for the flux rate of T C 0 2 from the sedim ent to the w ater colum n (Table 4). M ore T 13C 0 2 was released by the sedim ent in the exclusion treatm ent w ithin 6 days from burial (Fig. 8a). R espiration rates were (mean ± SE) 0 .8 4 ± 0 .2 5 and 0 .1 2 ± 0 .0 3 m g 13C m 2 day 1 for dry and wet fluxes, respectively.

Part of the 13C rem ained in the benthos and it was incorporated in the bulk sedim ent and in the m acrofauna. T h e am ount incorporated was variable am ong plots, but it did not significantly vary betw een exclusion treatm ents or tim e (Table 4; Fig. 8 b—c). H.ulvae incorporated the largest am ount o f G am ong all m acrofauna taxa (m ean ± SE: 3 .2 ± 0 .8 and 1 .2± 0 .4 m g 13C m 2, for H. ulvae and rem aining fauna, respectively).




Table 4. R epea ted m easu re analyses of variance an d sphericity tes ts for th e 13C incorporation in m acrofauna an d bulk sed im ent, and for t h e 1SC 0 2 released from th e b e n th o s (benthic respiration) during detrital d eco m p o s i t io n a t an in te rm ed ia te level of detrital e n r ich m en t (200 gWW p lo t_1a d d e d twice).

M acrofauna Bulk sedim ent Benthic respiration

d f MS F MS F MS F

Exclusion = E 1 5.97 1CT7 0.12 4.27 1CT4 0.22 1.52 1CT5 9.73

Time = T 3 3.13 10 -7 0.60 8.78 1CT4 1.02 1.84 1CT5 14.45

Residual (Plots) 2 4.96 1CT6 1.96 10 -3 1.56 1CT6

TxE 3 6.21 IO-7 0.12 8.79 1CT4 1.02 8.98 1CT6 7.07

Residual (TxPlots) 6 5.21 10 -6 8.59 1CT4 1.27 1CT6

Sphericity test (y2) n = 6 7.80 4.83 8.06

Significant values (P<0.05) are in bold. Sphericity test was not significant. doi:10.1371 /journal.pone.0066650.t004

Discussion

B ioturbation can enhance solute penetration in deeper sediments, their exchange betw een the sedim ent and the overlying w ater an d re-distribute organic m aterial along the vertical profile [36,37,41], As such, the loss o f species that are im portant contributors to b io turbation m ay greatly change the response to detrital decom position o f sedim ent biogeochem istry. In laboratory experim ents, for instance, A. marina and the polychaete Nereis diversicolor were found to increase tu rnover o f carbon an d nutrients during detrital m acroalgal decom position [19,36], T h e exclusion o f A. marina from intertidal sediments was thus expected to modify the biogeochem ical response to m acroalgal detrital enrichm ent. Conversely, a t the end o f decom position (1 m onth after the second enrichm ent) benthic respiration ( 0 2 and T C 0 2 sedim ent-water exchanges) and sedim ent organic carbon content did no t change as an effect o f lugworm exclusion.

Ulva am oun t an d tim e o f sam pling (one m onth after enrichm ent) allowed decom position an d release o f a large quantity o f organic m atter, w hich could be m ineralized or accum ulated in the sediment, thereby modifying sedim ent organic carbon, inorganic carbon dissolved in the interstitial w ater and fluxes a t the sedim ent-w ater interface. Ulva halftime decom position rate is w ithin 15 days [30], By considering the range o f k constant calculated for this study (between —0.06 an d —0.2), a t least 85% Ulva w ould be decom posed at the tim e o f sam pling for bo th enrichm ent levels. If one considers that d ried sediment weighs about 1600 g I- an d Ulva organic carbon content is a round 40% o f dried tissues [34], this decom posed detritus corresponds to organic carbon concentrations ranging betw een 0.15 and 1.2% sedim ent dry weight for the top first cm, roughly 2 -1 0 times the concentration o f organic carbon found in natu ral sediments for low an d high enrichm ent, respectively (see the non-enriched sedim ent in Fig. 2a). T h e lack of sedim ent carbon accum ulation and changes in T C 0 2 fluxes a t the sedim ent-w ater interface indicated that the benthic system was able to rem ove the carbon derived from in situ decom position an d that b io turbation did not play a central role in detrital carbon recycling.

It is im portan t to re-iterate th a t fluxes were m easured during 12 h dark incubation o f sedim ent cores w ithout lugworms, which indicated the overall indirect effect o f lugworm s on detrital benthic m ineralization, bu t did not include any direct effects o f lugworm s due to burrow ventilation during the period of incubation. T he im portance o f ventilation was instead assessed in the field by

m easuring porew ater D IC , w hich increased after lugworm rem oval, indicating a decreased solute exchange due to A. marina loss. Interestingly, these changes were variable along the vertical sedim ent profile and were offset by the addition o f detrital m acroalgae, w hich supported the idea that fast m ineralization o f detrital carbon in these sediments is only in pa rt due to b io turbation . T h e com plexity o f the relationships betw een species, functions an d response to d isturbance can increase w ith increasing ecological realism [15,46], O u r experim ents were done in situ and they were designed to test effects o f A. marina under natural environm ental conditions. D etrital m acroalgae were deliberately added to the sediment, w ithout using artificial nets th a t could limit detrital loss. This m ethod exposed decom posing detritus to tides and waves, especially w hen detritus rem ained at the sediment surface. T h e role o f b io turbation on sedim ent biogeochem istry has been m ainly assessed th rough laboratory studies. T h e few field studies on the role o f Arenicola on porew ater, fluxes and organic m atter decom position have revealed th a t its role m ay vary and decrease considerably in na tu re [19,37,47]. By com paring field to laborato ry experim ents, it was found, for instance, that solute exchange was m ore variable in the field because it was no t only ruled by b io turbation bu t also by advection related to waves and tides and by differences in the reactivity o f sedim ent organic m atter related to the com plexity o f biological com m unities o f p rim ary producers and consum ers [47]. T his is particularly true w hen experim ents are done in perm eable sediment, w here A. marina is often a dom inant species [48], In these sediments, hydrodynam ics can enhance advective flow from the porew ater to the overlying w ater colum n, rule the exchange of solutes and organic m atter a t the sedim ent-w ater interface an d overwhelm the biological effect o f b io turbation [49]. A lthough our experim ents were done in a relatively sheltered sand flat, as indicated by the dom inance o f fine-grained sand, num erous ripple m arks were visible at low tide (Fig. 2) an d they probably indicated strong hydrodynam ics generated by waves and tides. In these sediments, hydrodynam ics could accelerate the exchange o f solutes produced during decom position, the rem oval o f m acroalgal detritus from the sedim ent and override the effect o f b io tu rbation on the response o f sedim ent biogeochem istry to detrital m acroalgae.

T h e decom position ra te o f Ulva in these sediments was fast, as indicated by the estim ated k constants (k = —0.2 an d —0.1 for A — and A+, respectively; Fig. 7). These values were at the lim it o r even exceeded the range o f values typical o f Ulva detritus, w hich varies betw een —0.04 an d —0.1 [34], M ore im portantly, the estim ated k

PLOS ONE I www.plosone.org June 2013 | Volume 8 | Issue 6 | e66650



( a )

" b 5.0(N

a 4.0ÖD 3.0

(N 2.0OO 1.0m

H 0.01

(c)

Oco

bû

Ö¡3

f o3H—Iouo03

LO 12 14 16 18

0 2 4 6 8 10 12 14 16 18

10

8

6

4

2

00 2 4 6 8 10 12 14 16 18

Time (days)

Figure 8. Mean ± SE for (a) T13C 0 2 released from the sediment to the water column (sum of dry and w et fluxes); (b) am ount of 13C organic carbon in the bulk sediment (averaged over depth profile) and (c) am ount o f 13C carbon incorporated by macrofauna. D ata a re fitted to th e LOESS sm o o th in g w ith an interval o f 4 a n d 1 po lynom ial d e g re e . V alues a re av e rag e d o v er th e tw o p lo ts w h e re A. marina w as exc lu d ed ( tre a tm e n t A —; sq u are sym bo ls in th e g raphs) an d th e tw o p ro ced u ra l con tro l p lo ts ( tre a tm e n t A+; trian g le sym bo ls in th e g raph ). D ata refers to th e ex p e r im e n t 2. A sterisk in d ica tes s ign ifican t d ifferences. doi:10 .13 7 1 /jo u rn a l.p o n e .0066650 .g008

constant for the lugworm -exclusion sedim ent was twice th a t o f the control plots. In addition, the biomass o f m acroalgal debris was, on average, 6 times larger in controls than in lugworm -excluded sediments (2 and 12 g D W per p lot 1 for A+ and A —, respectively). T his m ay bring to the conclusion th a t despite hydrodynam ics could play an im portan t role in determ ining

benthic response to detrital m acroalgae, A. marina could represent an im portan t sink for detritus and it could m odify in situ decom position rate , as found for o ther burrow ing bioturbators [39,50]. These species can m echanically bury m acroalgae in their funnels, thereby decreasing detritus availability to surface sedim ent and slowing down decom position rate, while m aintain ing a pool of organic m atter in the sediment. B uried detritus and slowed decom position rate could, in turn , m odify m ineralization processes. In this study, a peak in T ’CCE release to the w ater colum n was identified at an early stage o f decom position in A. marina- excluded sedim ent plots, suggesting th at the loss o f A. marina accelerated the m ineralization o f detrital Ulva a t an early stage of decom position. This explanation o f increased m ineralization following A. marina loss was supported by the finding th at porew ater D IC decreased in response to detrital addition in A. marina excluded sediments. In addition, sim ultaneously to this decrease, m acroalgal biomass increased. Probably, w ithout the m echanical action o f burying m acroalgal fragm ents by Arenicola, detrital m acroalgae rem aining on sedim ent surface and decom posing faster released nutrients th at enhanced bacterial m etabolism and their carbon consum ption. Evidence for increased carbon consum ption during biomass decom position has been found for p lan t litter in freshwater ecosystems [51,52]. In addition, benthic m icroalgae are often regulated by bottom -up processes [53] and they could greatly benefit from nu trien t release, thereby contrib u ting to recycling the m ineralized detrital carbon. Interestingly, detrital enrichm ent and lugworm s h ad similar effects on m icroalgal biomass. Probably, both lugworm bioirrigation and detrital decom position can provide nutrients necessary to regulate m icroalgal growth [19,28,37,47].

A. marina is believed to influence the assembly rules and the distribution o f m acrofauna and, as such, their patterns o f response to disturbances by altering the sedim entary hab ita t (e. g. [37,41]). O u r results showed a negative effect o f A. marina on m acrofauna response to detrital addition since both m acrofauna abundance and diversity decreased following the highest level o f enrichm ent in presence o f lugworms. T he increased detrital burial b y T . marina could play a central role in determ ining this response. Patches of detrital m acroalgae in sediments can govern the spatial and tem poral variability o f m acrofauna patterns o f distribution. T he presence o f detritus m ay decrease num bers o f individuals and diversity because o f hypoxia during decom position. O nce detritus has decom posed, fast recolonizing taxa can prom ptly occupy the previously d isturbed patches and benefit from increased food supply [29]. D etrital burial due to lugworm activity probably prolonged hypoxic conditions and slowed down m acrofauna recovery. T he decrease in abundance com m on to all taxa and the loss o f diversity could corroborate this explanation as anoxic conditions related to detrital decom position deplete m ost taxa abundance and diversity [29]. M oreover, burial could decrease detrital availability to surface consum ers and therefore prevent their colonization, once hypoxia ceases. M acroalgal supply to sedim ent surface can represent an im portan t source o f food for surface grazers and detritivores, such as the gastropod Hydrobia ulvae [34,54], Part o f detrital carbon was transferred to m acrofauna, especially to H. ulvae (experim ent 2, Fig. 7c), w hich was the num erically dom inant species in these area. Its im portan t role for detrital carbon transport to the food web was found previously in the surrounding area [34], T his species is highly mobile and it crawls on sedim ent surface for feeding. Its local abundance m ight thus vary rapidly following food availability on surface sediment.

In sum m ary, our field experim ent showed th a t A. marina can be an im portan t sink for detrital Ulva. How ever, conversely to w hat it should be expected based on curren t research knowledge, this




species has m oderate, sometimes negative effects on the response to detrital enrichm ent o f sedim ent biogeochem istry and m acrofauna. A. marina loss causes the sedim entary ecosystem to becom e a faster recycler o f detrital m acroalgae, enhancing detrital loss, carbon consum ption and m ineralization. This m ay bring to the conclusion th a t the loss o f this species m ight be un im portan t or som ew hat beneficial for certain intertidal ecosystems. However, fast recycling and detrital dispersal could underm ine carbon storage capacity o f benthos and, in turn , the stability o f whole estuarine ecosystem facing organic enrichm ent and eutroph ication.

T hese findings clearly dem onstrate th a t under na tura l conditions, the local extinction of an ecosystem engineer m ay have com plex effects on the ecological stability o f m arine sediment biogeochem istry an d benthic com m unities o f consum ers. T hey also highlight the im portance o f perform ing experim ents under

References1. G rim e J P (1997) Ecology - B iodiversity a n d ecosystem fu nction : T h e d eb a te

d eepens. Science 277: 1 2 6 0 -1 2 6 1 .2. M c C a n n K S (2000) T h e diversity-stab ility d eb a te . N a tu re 405: 2 2 8 -2 3 3 .3. W o rm B, B arb ie r EB , B e a u m o n t N , D uffy J E , Folke C , e t al. (2006) Im p ac ts o f

b io d iversity loss o n o cean ecosystem services. Science 314: 787—790.4. T ilm a n D , D o w n in g AT. (1994) B iodiversity a n d stab ility in grasslands. N a tu re

367: 3 6 3 -3 6 5 .5. Y ach i S, L o re au M (1999) B iodiversity a n d ecosystem p ro d u ctiv ity in a f luc tating

e n v iro n m en t: T h e in su ran ce hypothesis . P ro ceed in g s o f th e N a tio n a l A cad em y o f Sciences o f th e U n ited S tates o f A m erica 96: 1463—1468.

6. D o w n in g A , L e ib o ld M (2010) Species richness facilitates ecosystem resilience in aq u a tic fo o d w ebs. F resh w a te r Biology 55: 2123—2137.

7. Ives A , C a rp e n te r S (2007) S tability a n d d iversity o f ecosystem s. Science: 58—62.8. P fiste rer A, S ch m id B (2002) D iv e rs ity -d ep en d en t p ro d u c tio n c a n decrease the

stab ility o f ecosystem fun ctio n in g . N a tu re 416: 84—86.9. Allison G (2004) T h e in fluence o f species diversity a n d stress in tensity o n

c o m m u n ity resistance a n d resilience. E co log ical M o n o g ra p h s 74: 117—134.10. E m m e rso n M C , S olan M , Ernes C , P a terso n D M , R affaelli D (2001) C o n s is ten t

p a tte rn s a n d th e id iosyncra tic effects o f b io d iversity in m arin e ecosystem s. N a tu re 411: 7 3 -7 7 .

11. G oodsell I J , U n d e rw o o d A J (2008) C o m p lex ity a n d id iosyncrasy in the responses o f a lgae to d is tu rb an c e in m o n o - a n d m ulti-species assem blages. O eco lo g ia 157: 5 0 9 -5 1 9 .

12. R ossi F , V o s M , M id d e lb u rg J (2009) Species id en tity , d iversity a n d m icro b ia l c a rb o n flow in reassem b lin g m ac ro b e n th ic com m unities. O ikos: 50 3 —512.

13. S o lan M , C a rd in a le BJ, D o w n in g A L , E n g e lh a rd t K A M , R u esin k J L , e t al. (2004) E x tin c tio n a n d ecosystem fu n c tio n in th e m arin e b en th o s . Science 306: 1 1 7 7 -1 1 8 0 .

14. B o lam SG , F e rn an d es T F , H u x h a m M (2002) D iversity , b io m ass , a n d ecosystem processes in the m arin e b en th o s . E co log ical M o n o g ra p h s 72: 5 9 9 —615.

15. R o m a n u k T , V o g t R , K o lasa J (2009) E co log ical rea lism a n d m ech an ism s by w h ich diversity begets stability . O ikos 118: 8 1 9 -8 2 8 .

16. J o n e s C G , L aw to n J H , S h ach ak M (1994) O rg an ism s as ecosystem engneers. O ikos 69: 3 7 3 -3 8 6 .

17. B erke S (2010) F u n c tio n a l G ro u p s o f E cosystem E ngineers: A P ro p o sed C lassification w ith C o m m e n ts o n C u r re n t Issues. In teg ra tiv e a n d C o m p a ra tiv e Biology 50: 14 7 -1 5 7 .

18. K ris te n sen E , P en h a-L o p es G , D elefosse M , V a ld e m a rse n T , Q u in ta n a C O , e t al. (2012) W h a t is b io tu rb a tio n ? T h e n e e d for a p rec ise d efin ition for fau n a in aq u a tic sciences. M a rin e E co logy P rogress Series 446: 285—302.

19. K ris te n sen E (2001) Im p a c t o f p o ly ch ae tes {Nereis a n d A renicola) o n sed im en t b io g eo ch em istry in coasta l a reas: P ast, p rese n t, a n d fu tu re developm ents. A bstrac ts o f P ap e rs o f th e A m e ric a n C h e m ic a l Society 221: U 5 3 8 —U 538.

20. B raeck m an U , P ro v o o st P , M o en s T , S o e tae rt K , M id d e lb u rg J , e t al. (2011) B iological vs. P hysica l M ix in g Effects o n B en th ic F o o d W eb D ynam ics. Plos O n e 6: : -e l8 0 7 8 . d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 1 8 0 7 8 .

21. L o h re r A M , T h ru s h SF, G ibbs M M (2004) B io tu rb a to rs e n h a n ce ecosystem fu n ctio n th ro u g h co m p lex b io g eo ch em ica l in te rac tio n s . N a tu re 431: 1092—1095.

22. E k lo fJ , v a n d e r H e id e T , D o n a d i S, v a n d e r Z ee E , O ’H a ra R , e t al. (2011) H a b i ta t-M e d ia te d F a c ilita tio n a n d C o u n te ra c t in g E co sy stem E n g in e e r in g In te rac tiv e ly In flu en ce E cosystem R esponses to D is tu rb an ce . Plos O n e 6: e 2 3 2 2 9 .-d o i: 1 0 .1 3 7 1 /jo u rn a l.p o n e .0023229 .

23. N o rk k o J , R e e d D C , T im m e rm a n n K , N orkko A , G usta fsson B G , e t al. {2012) A w elco m e c a n o f w orm s? H y p o x ia m itig a tio n b y a n invasive species. G lo b al C h a n g e Biology 18: 4 2 2 -4 3 4 . d o k lO .l 1 1 1 / j . 1 3 6 5 -2 4 8 6 .2 0 1 1.02513.x.

24. E isen h a u e r N , M ilcu A , S abais A , S cheu S {2008) A n im al E cosystem E ng in eers M o d u la te th e D iversity-Invasib ility R ela tio n sh ip . Plos O n e 3.

25. C e b r ia n J {2004) R o le o f f irs t-o rd er consum ers in ecosystem c a rb o n flow. Ecology L ette rs 7: 2 3 2 -2 4 0 .

natura l conditions if we are to understand an d pred ic t the response o f ecosystems to pertu rba tion and species loss.

AcknowledgmentsThe authors wish to thank the technicians for their invaluable help in the chemical analyses and E. W eerman, JJ Jansen and F M ontserrat and numerous other students for their help in the field. S. Thrush is kindly thanked for his critical revision o f an early version of this MS. This study is part of the BIOFUSE project, within the M ARBEF network of excellence.

Author ContributionsConceived and designed the experiments: FR BG. Perform ed the experiments: FR BG FG VDS. Analyzed the data: FR BG FG VDS. Contributed reagents/m aterials/analysis tools: TTM. W rote the paper: FR BG FG V D SJJM .

26. R affaelli D G , R a v e n J A , P oo le LJ {1998) E cological im p a c t o f g ree n m acro a lg a l b loom s. O c e a n o g ra p h y a n d M a rin e Biology, a n A n n u a l R ev iew 36: 97—125.

27. F o rd R B , T h ru s h SF, P ro b e r t P K {1999) M a c ro b e n th ic c o lo n isa d o n o f d istu rb an ces o n a n in te rtid a l sandflat: th e in fluence o f season a n d b u r ie d algae. M a rin e Ecology P rogress Series 191: 163—174.

28. R ossi F , U n d e rw o o d A J {2002) Sm all-scale d is tu rb an c e a n d in creased n u trien ts as influences o n in te rtid a l m ac ro b e n th ic assem blages: ex p e rim e n ta l b u r ia l o f w rack in d iffe ren t in te rtid a l e n v iro n m en ts . M a rin e Ecology P rogress Series 241: 2 9 -3 9 .

29. K e la h e r B P , L ev in to n J S {2003) V a r ia tio n in d e tr ita l e n ric h m e n t causes spatio- tem p o ra l v a ria tio n in so ft-sed im en t assem blages. M a rin e Ecology P rogress Series 261: 8 5 -9 7 .

30. R ossi F {2006) S m all-scale b u r ia l o f m acro a lg a l de tritu s in m arin e sedim ents: Effects o f U lva spp. o n th e sp a tia l d istr ib u tio n o f m ac ro fa u n a assem blages.

J o u rn a l o f E x p e rim e n ta l M a rin e Biology a n d Ecology 332: 84—95.31. B o lam SG , F e rn an d e z T F , R e a d P , R affaelli D {2000) Effects o f m ac ro a lg a l m ats

o n in te rtid a l sandflats: a n ex p e rim e n ta l study. J o u rn a l o f E x p e rim e n ta l M arin e B iology a n d E cology 249: 123—137.

32. H u ll SC {1987) M a cro a lg a l m ats a n d species ab u n d a n ce : a field ex p e rim en t. E s tu arin e C o a s ta l a n d Shelf Science 25: 5 1 9 -5 3 2 .

33. T h ru s h SF {1986) T h e su b litto ra l m ac ro b e n th ic co m m u n tiy s tru c tu re o f a n irish sea-lough : effec t o f d e c o m p o s in g a c cu m u la to in s o f seaw eed . J o u r n a l o f E x p e r im e n ta l M a rin e Biology a n d Ecology 96: 199—212.

34. R ossi F {2007) R ecycle o f b u r ie d m ac ro a lg a l d e tr itu s in sedim ents: use o f d u a llab e llin g ex p e rim en ts in th e field. M a rin e B io logy 150: 1073—1081.

35. G o d b o ld J A , S o lan M , K illh a m K {2009) C o n su m e r a n d reso u rce d iversity effects o n m arin e m acro a lg a l deco m p o sitio n . O ikos 118: 77—86.

36. H a n se n K , K risten sen E {1998) T h e im p a c t o f th e p o ly ch ae te N ere is d iversico lor a n d e n ric h m e n t w ith m ac ro a lg a l {C h ae to m o rp h a linum ) d e tr itu s o n b en th ic m etab o lism a n d n u tr ie n t d y n am ics in o rg an ic -p o o r a n d o rg an ic -r ich sed im en t.

J o u rn a l o f E x p e rim e n ta l M a rin e Biology a n d Ecology 231: 201—223.37. O ’B rien A , V o lk e n b o rn N , v a n B eusekom J , M o rris L , K e o u g h M {2009)

In te rac tiv e effects o f p o rew a te r n u trie n t en ric h m e n t, b io tu rb a tio n a n d sed im en t c h a ra c te r is t ic s o n b e n th ic a s sem b lag es in s a n d y se d im e n ts . J o u r n a l o f E x p e r im e n ta l M a rin e Biology a n d Ecology: 51—59.

38. D ’A n d re a A, D e W itt T {2009) G eo ch em ica l ecosystem e n g in e e rin g b y th e m u d sh rim p U p o g e b ia pug e tten sis {Crustacea: T h a lass in id ae ) in Y a q u in a Bay, O re g o n : D e n s ity -d e p en d e n t effects o n o rg an ic m a tte r rem in era liz a tio n a n d n u trie n t cycling. L im nology a n d O c e a n o g ra p h y 54: 1911—1932.

39. V o n k J , K n e e r D , S tap el J , A sm us H {2008) S h rim p b u r ro w in tro p ica l seagrass m eadow s: A n im p o rta n t sink for litter. E s tu arin e C o asta l a n d S he lf Science 79: 7 9 -8 5 .

40. M ey sm an F JR , G a lak tio n o v E S , G rib sh o lt B, M id d e lb u rg J J {2006) B io irrigation in p e rm e a b le sedim ents: A dvective p o re -w a te r t ra n s p o rt in d u ce d b y b u rro w ven tila tio n . L im nology a n d O c e a n o g ra p h y 51: 142—156.

41. V o lk e n b o rn N , R eise K {2006) L u g w o rm exclusion ex p erim en t: R esponses by d e p o s it f ee d in g w o rm s to b io g en ic h a b ita t tra n s fo rm a tio n s . J o u r n a l O f E x p e r im e n ta l M a rin e Biology A n d Ecology 330: 169—179.

42. N o rk k o A, B o n sd o rff E {1996) P o p u la tio n responses o f coasta l z o o b en th o s to stress in d u ce d b y d riftin g a lgal m ats. M a rin e Eco logy P rogress Series 140: 141— 151.

43. B u ch sb au m R , V alie la I, Sw ain T , D zierzesk i M , A llen S {1991) A v ailab le A n d R e fra c to ry N itro g e n In D e tritu s O f C o a s ta l V a sc u la r P lan ts A n d M acro a lg ae . M a rin e Eco logy-P rogress Series 72: 131—143.

44. N e d e rg a a rd R I , R isg aard -P e te rso n N , F in ste r K {2002) T h e im p o rta n c e o f su lfate red u c tio n a ssoc ia ted w ith U lv a lac tu ca thalli d u r in g d ecom position : a m esocosm ex p e rim en t. J o u rn a l o f E x p e rim e n ta l M a rin e Biology a n d Ecology 275: 15-29 .




45. H e rm a n P M J. M id d e lb u rg J J . W iddow s J , L ucas C H . H e ip C H R (2000) S table iso topes as tro p h ic tracers: c o m b in in g field sam p lin g a n d m an ip u la tiv e labelling o f fo o d resources fo r m acro b en th o s . M a rin e E cology P rogress Series 204: 79—92.

46. M ck ie B. S ch in d ler M . G essner M . M alm q v is t B (2009) P lacin g b iod iversity a n d ecosystem fu n c tio n in g in con tex t: en v iro n m en ta l p e rtu rb a tio n s a n d the effects o f species richness in a s tre a m field ex p e rim en t. O eco log ia: 757—770.

4 7 . P a p a s p y ro u S, K r is te n s e n E , C h r is te n s e n B (2007) A re n ic o la m a r in a (Polychaeta) a n d o rg an ic m a tte r m in e ra lisa tio n in san d y m arin e sedim ents: In situ a n d m ic ro co sm co m p ariso n . E s tu arin e C o a s ta l a n d S h e lf Science 72: 213— 2 2 2 .

48. N e e d h a m H , P ild itc h C , L o h re r A , T h ru s h S (2011) C o n te x t-S p e c ific B io tu rb a tio n M ed ia tes C h a n g e s to E cosystem F u n ctio n in g . E cosystem s 14: 1 0 9 6 -1 1 0 9 .

49. H u e tte l M , R o y H . P re ch t E. E h re n h au ss S (2003) H y d ro d y n a m ic a l im p a c t o n b io g eo ch em ica l processes in a q u a tic sedim ents. H y d ro b io lo g ia 494: 231—236.

50. R e tra u b u n A , D aw so n M , E v an s S (1996) T h e ro le o f th e b u r ro w fu n n el in feed in g p rocesses in th e lu g w o rm A ren ico la m a rin a (L). Jo u rn a l o f E x p erim en ta l M a rin e Biology a n d Eco logy 202: 10 7 -1 1 8 .

51. F o n ta in e S, B a rd o u x G. A b b ad ie L. M a rio tti A (2004) C a rb o n in p u t to soil m ay d ecrease soil c a rb o n co n ten t. Ecology L ette rs 7: 314—320.

52. L e n n o n J T (2004) E x p e rim e n ta l ev idence th a t te rres tria l c a rb o n subsidies in crease C 0 2 flux fro m lake ecosystem s. O eco lo g ia 138: 5 8 4 —591.

53. H ille b ran d H . W o rm B. L o tze H (2000) M a rin e m ic ro b en th ic c o m m u n ity s tru c tu re reg u la te d b y n itro g en lo ad in g a n d g raz in g p ressu re . M a rin e Ecology P rogress Series 204: 2 7 -3 8 .

54. R ie ra P (2010) T ro p h ic p lastic ity o f th e g as tro p o d H y d ro b ia u lvae w ith in a n in te rtid a l b ay (Roscoff. France): A stab le iso tope evidence. J o u rn a l o f Sea R e se a rc h 63: 78—83.



complex effects of ecosystem engineer loss on benthic ... · open 3 access freely available online

Documents