rsosroyalsocietypublishingorg

ResearchCite this article Jones BL Unsworth RKF2016 The perilous state of seagrass in theBritish Isles R Soc open sci 3 150596httpdxdoiorg101098rsos150596

Received 3 November 2015Accepted 8 December 2015

Subject CategoryBiology (whole organism)

Subject Areasecologyenvironmental science

Keywordseutrophication indicators seagrassessubmerged aquatic vegetation water qualityZostera

Author for correspondenceRichard K F Unsworthe-mail rkfunsworthswanseaacuk

Electronic supplementary material is availableat httpdxdoiorg101098rsos150596 or viahttprsosroyalsocietypublishingorg

The perilous state ofseagrass in the British IslesBenjamin L Jones12 and Richard K F Unsworth23

1Sustainable Places Research Institute and 2Project Seagrass Sustainable PlacesResearch Institute Cardiff University Cardiff CF10 3BA UK3Seagrass Ecosystem Research Group College of Science Swansea UniversityWallace Building Swansea SA2 8PP UK

Seagrass ecosystems face widespread threat from reducedwater quality coastal development and poor land use Inrecent decades their distribution has declined rapidly and inthe British Isles this loss is thought to have been extensiveGiven increasing knowledge of how these ecosystems supportfisheries production the understanding of their potential rapidloss and the difficulty in restoring them it is vital wedevelop an understanding of the risks they are under so thatmanagement actions can be developed accordingly Developingan understanding of their environmental status and conditionis therefore critical to their long-term management This studyprovided to our knowledge the first examination of theenvironmental health of seagrass meadows around the BritishIsles This study used a bioindicator approach and involvedcollecting data on seagrass density and morphology alongsideanalysis of leaf biochemistry Our study provides to the bestof our knowledge the first strong quantitative evidence thatseagrass meadows of the British Isles are mostly in poorcondition in comparison with global averages with tissuenitrogen levels 75 higher than global values Such poor statusplaces their long-term resilience in doubt Elemental nutrientconcentrations and morphological change suggest conditionsof excess nitrogen and probable low light placing many ofthe meadows sampled in a perilous state although otherssituated away from human populations were perceived to behealthy Although some sites were of a high environmentalhealth all sites were considered at risk from anthropogenicimpacts particularly poor water quality and boating-baseddisturbances The findings of this study provide a warningof the need to take action with respect to water quality anddisturbance to prevent the further loss and degradation ofthese systems across the British Isles

1 IntroductionThe eelgrass Zostera marina forms a highly productive habitatwithin temperate coastal ecosystems of the Northern Hemisphere[1] In recent decades meadows of Z marina have declined rapidly

2016 The Authors Published by the Royal Society under the terms of the Creative CommonsAttribution License httpcreativecommonsorglicensesby40 which permits unrestricteduse provided the original author and source are credited

2

rsosroyalsocietypublishingorgRSocopensci3150596

[23] with anthropogenic activity thought to be accountable for the majority of the reported loses [34]Globally the loses of seagrass are occurring at an estimated rate of 7 yrminus1 [4] and with an estimated44 of the worldrsquos population now living within 150 km of the ocean these declines are only set toincrease [5] Learning how to protect these habitats is vital if we are to continue to reap the benefits ofthe ecosystem services they provide to humanity [6] This requires creating a far greater understandingof the safe environmental boundaries of their existence in order to develop thresholds of safety for theirmanagement

Increased nutrient input is one of the main forms of anthropogenic loading into the marineenvironment [7ndash9] and evidence suggests that this presents the biggest threat to seagrass meadows[410] This over-enrichment causes shifts in primary production away from seagrass towards faster-growing nutrient-limited macroalgae epiphytic microalgae and phytoplankton [11] Numerous studieshave reported how excessive growth of epiphytic algae on seagrass is a contributing factor to seagrassdegradation through smothering and decreasing their ability to absorb light [1213] Increased input ofnutrients enhances this excessive growth of epiphytes [14] reducing the resilience of seagrass and placingtheir viability at risk [10]

Owing to their susceptibility to changes in environmental conditions such as light and nutrientsseagrasses can be used as indicators of ecosystem quality and health [15] Within the European Union(EU) Water Framework Directive (WFD) seagrasses are increasingly used in assessing the ecologicalstatus of coastal areas where ecological status is classified by comparing characteristics mainly shootdensity percentage cover or depth limits to a reference level which reflects that of perceived lowanthropogenic influence [1617]

In the North Atlantic seagrass meadows have recently been highlighted for their value as a nurseryhabitat for juvenile fishes of commercial value such as plaice and Atlantic cod [18ndash20] and theirbiodiversity value is recognized by their inclusion in Marine Habitat Action Plans and as a focal part ofthe UK Biodiversity Action Plan [21]

In the British Isles (including the UK) seagrass meadows have historically been damaged anddegraded [2223] but detailed long-term data are spatially limited We know that multiple anthropogenicdrivers of change are present within these systems [22] and recent reviews [24] have categorized Zosteraas being nationally scarce Collation of sporadic datasets suggests populations are in severe decline (lossestimated at between 25 and 49 in the past 25 years) [25]

Available data on UK seagrass are also of limited scope and have limited capacity to reveal theenvironmental health of the meadow (eg mostly data on shoot density or total area) Althoughintertidal seagrasses in the UK have recently been included in government programmes these EU WFD-focused assessments do not assess environmental conditions neither via measurements of seagrassnutrient content or via physical-chemistry of the water column and therefore provide us very littleinformation pertinent to their long-term management Importantly there is little way of using thedata to act as an early warning of future loss Only the data collected within the Isles of Scillymonitoring programme presents information pertinent to understanding the status of the system as itis a long-term study that provides detailed data of shoot morphology density disease and epiphyticdata [26]

Despite growing recognition for their ecosystem service value and an appreciation for their loss thereexists little detailed information about the environmental status of seagrass meadows throughout theBritish Isles This is particularly the case with respect to nutrient enrichment which is highlighted asone of the key threats to these systems [22] and light availability which is a key driver of productivity[27] Given that there is strong evidence that the water quality is poor in many areas of the BritishIsles known to contain seagrass [28] new data are therefore required that can attempt to determinethe environmental health status of the seagrasses in the British Isles This is particularly important aswe currently have no means of knowing whether seagrass meadows may already be very close to theirenvironmental thresholds

In many parts of the world scientists and environmental managers are conducting seagrassmonitoring studies that include the use of bioindicators to determine the environmental health of theseagrass system [29ndash31] This approach uses the fact that seagrasses respond negatively to nutrientenrichment physiologically morphologically and biochemically [27] In fact by examining seagrassbioindicators there exists better scope to understand environmental condition as seagrass biochemicalmeasurements are time-integrated rather than based on commonly used point-sampling methods forwater quality assessments By integrating such measurements with detailed quantification of theirmorphometrics and epiphytic algae greater information about the environmental status of thesesensitive yet important seagrass meadows can be collected

3

rsosroyalsocietypublishingorgRSocopensci3150596

The aim of this study was to provide the first rapid assessment of the ecological and environmental

status of seagrass meadows (Z marina) over a wide spatial scale throughout the British Isles using aseries of established ecological and biochemical indicators This study provides us with to the best ofour knowledge the first use of biochemical indicators in seagrass meadows of the British Isles

2 Material and methods21 Study sitesSeagrass meadows at 11 locations around the British Isles were assessed during MayndashAugust 2013(table 1 and figure 1) Habitat data were collected in the field along with samples for further analysisin the laboratory At each site qualitative descriptive information was collected about the perceivedpresence of anthropogenic impacts in order to place the data in context All seagrass sites were withinthe range of 0ndash3 m depth

22 Plant collection and morphological measurementsAt all sampling sites three haphazardly placed 025 m2 seagrass plots containing Z marina were sampledShoot density counts and percentage cover estimates were taken on site and recorded All seagrassmaterial within the quadrats was then collected for subsequent laboratory analysis Large macroalgaeand Zostera noltii were not collected where it was present Morphological measurements (leaf width andlength number of leaves per shoot) were taken in the laboratory

Collected biomass was removed from sample bags and rinsed thoroughly in fresh water to removesalt water sediment and detritus Sheath length and width were taken from the longest intact leaf of eachshoot Length was measured using a rule to the nearest 10 mm from the meristem to the top of the leafand width was measured with callipers to the nearest 005 mm Number of leaves per shoot was alsorecorded for all shoots collected All epiphytes where present were carefully scraped from both sides ofthe leaf using a microscope slide

Cleaned leaf sections were dried at 60C for 24 h and ground until homogeneous before dry leaf masswas recorded using an Ohaus balance (max 100 g d = 01 mg Switzerland) Ground leaf tissue and shootdensity were used to calculate individual shoot biomass for comparison purposes Similarly the scrapedepiphytes were also dried at 60C until they were of constant weight before their dry mass was recordedEpiphyte mass was combined with shoot density to calculate epiphyte biomass per shoot

23 Leaf carbon nitrogen and phosphorus contentGround seagrass leaf tissue was analysed for carbon (C) nitrogen (N) and phosphorus (P) [29] Sampleswere analysed for per cent nitrogen and per cent carbon by weight using a continuous flow isotoperatio mass spectrometer (ANCA SL 20-20 Europa Scientific Crewe UK) and analysis of total P wasundertaken by inductively coupled plasma atomic emission spectrometry (Liberty AX Sequential ICP-AES Varian Inc CA USA) Dry matter determinations were carried out on each sample so values couldbe reported as per cent dry weight after which molar C N C P and N P values were calculated [2931]Our use of C N in this study is owing to its use as a robust early warning indicator of light reduction[27] Similarly C P has been identified as an indicator of environmental P limitation [30ndash33] and N P asan indicator of the balance in abundance of environmental N and P (the seagrass Redfield ratio definesN P values of 25ndash30 as balanced abundance of both N and P compared with light availability) [3234ndash36]

24 Data analysisAll morphological values where described are reported as mean plusmn sd One-way ANOVA wasconducted using SIGMAPLOT v 123 Prior to analysis data were tested for normality to meet theassumptions of parametric tests in the event of failure to adhere to normality and homogeneity ofvariance one-way ANOVA on ranks was conducted as an alternative

Each variable measured different aspects of seagrass status and to identify which of these presentedmost variability in the data a principal component analysis (PCA) was used [37ndash39] Principalcomponents (PCs) with eigenvalues greater than 10 were considered and variable coefficients of lessthan or equal to 03 and greater than 03 in each component were selected as dominant

4

rsosroyalsocietypublishingorgRSocopensci3150596

Table1SeagrasssamplingsitesaroundtheBritish

Islesdateofcollectionandinitialassessm

entonperceivedpresenceofanthropogenicimpacts

permanent

site

coordinates

date

industry

tourism

agriculture

catchment

population

anchoring

mooring

GelliswickBayWales(GB)

51 42prime 26primeprimeN53prime 42primeprimeW

26May2013

Southend-on-SeaEngland(SS)

51 32prime 21primeprimeN039

prime 04primeprimeE

17July2013

PrioryBayIsleofWight(PB)

50 42prime 23primeprimeN15prime 45primeprimeW

11June2013

RamseyBayIsleofMan(RB)

54 18prime38

primeprime N421

prime 18primeprime W

26August2013

StudlandBayEngland(SB)

50 38prime34

primeprime N156

prime 28primeprime W

17July2013

Kircubin

BayStrangfordLough

54 29prime02

primeprime N532

prime 20primeprime W

23August2013

NorthernIreland(KB)

Mannin

BayIreland(MB)

53 26prime45

primeprime N10

4prime 45primeprimeW

21August2013

LangnessIsleofMan(LA)

54 4

prime 27primeprime N

436

prime 48primeprimeW

18June2013

PorthdinllaenWales(PD)

51 44prime30

primeprime N516

prime 79primeprime W

27June2013

IslesofScilly

(IS)

LittleArthurEasternIsles

49 56prime59

primeprime N615

prime 59primeprime W

28July2013

HigherTown

BayStMartinrsquos

49 57prime 17primeprimeN616

prime 41primeprimeW

29July2013

GrimsbyHarbourTresco

49 57prime 41primeprimeN619

prime 53primeprime W

30July2013

BroadLedgesTresco

49 56prime29

primeprime N619

prime 41primeprimeW

31July2013

Skom

erMCZWales(SK)

51 44prime30

primeprime N516

prime 79primeprime W

2May2013

5

rsosroyalsocietypublishingorgRSocopensci3150596

MB

PD

SK

N

0 200 km

KBRB

LA

GB

SB PB

IS

SS

location of sampled seagrass beds

Figure 1 Seagrass meadows were sampled at 11 locations spread throughout the British Isles

3 Results31 Seagrass morphometricsLeaf length and width differed significantly between sites (figure 2 and table 2) with leaf length(p lt 0001 H10 = 322) ranging from 142 to 788 mm (mean 347 plusmn 201) and leaf width (p lt 0001H10 = 325) ranging from 3 to 11 mm (mean 5 plusmn 2) Longest leaves were observed in the Isles of Scillyand Mannin Bay and smallest in Porthdinllaen and Southend-on-Sea Similarly widest leaves wereobserved in the Isles of Scilly and Mannin Bay with the narrowest being observed in PorthdinllaenIn addition shoot characteristics were highly variable between sites Shoot biomass also significantly(p lt 0005 H10 = 279) differed with respect to site ranging from 01 to 26 g (mean 03 plusmn 08) with highestvalues in the Isles of Scilly and lowest values in Porthdinllaen Skomer MCZ and Kircubin Bay Shootdensity was again significantly (p lt 0005 H10 = 291) different between sites and ranged from 40 to 1017per 025 m2 (mean 483 plusmn 324) with highest values in Priory Bay and Porthdinllaen and lowest valuesin the Isles of Scilly and Ramsey Bay Seagrass cover ranged from 167 to 913 (mean 481 plusmn 214) withhighest values in the Isles of Scilly and lowest values in Gelliswick Bay and Kircubin Bay The numberof leaves per shoot ranged from 35 to 60 (mean 46 plusmn 09) with highest values in Priory Bay and lowestvalues in Mannin Bay and Porthdinllaen

32 Nutrient concentrations and ratiosSignificant differences in leaf tissue N (p lt 0001 F1032 = 377) and P (p lt 0001 F1032 = 298) wereobserved between locations Per cent N ranged from 205 to 554 (mean 358 plusmn 095) with highest valuesin Skomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Mannin Bay Thesefindings are surprising given that shoot density was highest at Porthdinllaen and Priory Bay two sitesof high N Pearsonrsquos correlation found no significant relationship between N and shoot density

6

rsosroyalsocietypublishingorgRSocopensci3150596

12

10

8

6

4

2

0

1000

800

600

400

200

0

leaf

wid

th (

mm

)

leaf

leng

th (

mm

)

Gel

lisw

ick

Prio

ry B

ay

Skom

er

Port

hdin

llaen

Lag

ness

Ram

sey

Stud

land

Sout

hend

Kir

cubi

n

Scill

y

Man

nin

leaf width (mm)leaf length (mm)

Figure 2 Leaf length and leaf width of seagrass at 11 locations spread throughout the British Isles

(electronic supplementary material appendix 1) The N values exhibited at sites in this study weremostly very high relative to other global values (table 3) Gelliswick Bay Priory Bay Skomer MCZPorthdinllaen Langness Ramsey Bay Southend-on-Sea and Kircubin Bay all had values above 3 andonly Studland Bay the Isles of Scilly and Mannin Bay exhibited values below 3 No sites exhibited Nvalues lower than the global average of 204

Per cent P ranged from 011 to 038 (mean 021 plusmn 007) where highest values were recorded atSkomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Studland Bay PrioryBay Porthdinllaen Langness Ramsey Bay Studland Bay Kircubin Bay the Isles of Scilly and ManninBay all exhibited P values lower than the global average of 027 and only Gelliswick Bay Skomer MCZand Southend-on-Sea exhibited higher than average values

Levels of the nutrient ratio N P significantly differed with respect to site (p lt 0001 F1032 = 575) Allsites were found to be highly imbalanced in terms of available nutrients with N P values more than30 (figure 3) Specifically Priory Bay Langness Studland Bay Kircubin Bay and the Isles of Scilly allhad highly elevated N P ratios (more than 40) suggesting high nutrient imbalance and a limitation inP These sites all with P values lower than the global average also exhibited high C P ratios (morethan 500) indicating that the plants are growing in environments with a relatively low P pool (figure 4)C P ratios also differed significantly with respect to site (p lt 0001 F1032 = 150) In comparison SkomerMCZ Southend-on-Sea and Gelliswick Bay all with high P levels elevated N P and C P ratios ofless than 500 indicate that the plants are growing in a nutrient-rich environment with a relativelylarge P pool

The ratio of C N can provide a potential indication of light limitation in seagrass [47] wherelevels below 20 suggest reduced light availability and levels below 15 being of potential limitation(figure 5) [2733] Levels of C N differed significantly with respect to site (p lt 0001 F1032 = 228) Onthis basis seagrass at Skomer MCZ Southend-on-Sea and Gelliswick Bay all suffered from potentiallight limitation Only the seagrass at the Isles of Scilly Mannin Bay and to a lesser extent Studland Bayhad levels suggesting sufficient light availability

33 Principal component analysisIn order to determine which characteristics contributed most to the variability within and betweensites a PCA was used In the PCA of the measured seagrass characteristics there were three PCAsthat had eigenvalues greater than 1 and together accounted for 784 of the variance (table 4) Thefirst principal component (PC1) and the component with the most importance had an eigenvalue of537 and contributed to 537 of the variance This component was dominated by molar C N molarC P percentage cover shoot biomass leaf width and leaf length This axis is suggested to be looselyrepresentative of growth and productivity and thus overall health in relation to available nutrients andlight This is a useful way of ranking sites based on their leaf characteristics where an increase in PC1represents an increase in overall leaf size attributed to an increase in light availability (high C N) and adecrease in P load (high C P)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

3

rsosroyalsocietypublishingorgRSocopensci3150596

The aim of this study was to provide the first rapid assessment of the ecological and environmental

status of seagrass meadows (Z marina) over a wide spatial scale throughout the British Isles using aseries of established ecological and biochemical indicators This study provides us with to the best ofour knowledge the first use of biochemical indicators in seagrass meadows of the British Isles

2 Material and methods21 Study sitesSeagrass meadows at 11 locations around the British Isles were assessed during MayndashAugust 2013(table 1 and figure 1) Habitat data were collected in the field along with samples for further analysisin the laboratory At each site qualitative descriptive information was collected about the perceivedpresence of anthropogenic impacts in order to place the data in context All seagrass sites were withinthe range of 0ndash3 m depth

22 Plant collection and morphological measurementsAt all sampling sites three haphazardly placed 025 m2 seagrass plots containing Z marina were sampledShoot density counts and percentage cover estimates were taken on site and recorded All seagrassmaterial within the quadrats was then collected for subsequent laboratory analysis Large macroalgaeand Zostera noltii were not collected where it was present Morphological measurements (leaf width andlength number of leaves per shoot) were taken in the laboratory

Collected biomass was removed from sample bags and rinsed thoroughly in fresh water to removesalt water sediment and detritus Sheath length and width were taken from the longest intact leaf of eachshoot Length was measured using a rule to the nearest 10 mm from the meristem to the top of the leafand width was measured with callipers to the nearest 005 mm Number of leaves per shoot was alsorecorded for all shoots collected All epiphytes where present were carefully scraped from both sides ofthe leaf using a microscope slide

Cleaned leaf sections were dried at 60C for 24 h and ground until homogeneous before dry leaf masswas recorded using an Ohaus balance (max 100 g d = 01 mg Switzerland) Ground leaf tissue and shootdensity were used to calculate individual shoot biomass for comparison purposes Similarly the scrapedepiphytes were also dried at 60C until they were of constant weight before their dry mass was recordedEpiphyte mass was combined with shoot density to calculate epiphyte biomass per shoot

23 Leaf carbon nitrogen and phosphorus contentGround seagrass leaf tissue was analysed for carbon (C) nitrogen (N) and phosphorus (P) [29] Sampleswere analysed for per cent nitrogen and per cent carbon by weight using a continuous flow isotoperatio mass spectrometer (ANCA SL 20-20 Europa Scientific Crewe UK) and analysis of total P wasundertaken by inductively coupled plasma atomic emission spectrometry (Liberty AX Sequential ICP-AES Varian Inc CA USA) Dry matter determinations were carried out on each sample so values couldbe reported as per cent dry weight after which molar C N C P and N P values were calculated [2931]Our use of C N in this study is owing to its use as a robust early warning indicator of light reduction[27] Similarly C P has been identified as an indicator of environmental P limitation [30ndash33] and N P asan indicator of the balance in abundance of environmental N and P (the seagrass Redfield ratio definesN P values of 25ndash30 as balanced abundance of both N and P compared with light availability) [3234ndash36]

24 Data analysisAll morphological values where described are reported as mean plusmn sd One-way ANOVA wasconducted using SIGMAPLOT v 123 Prior to analysis data were tested for normality to meet theassumptions of parametric tests in the event of failure to adhere to normality and homogeneity ofvariance one-way ANOVA on ranks was conducted as an alternative

Each variable measured different aspects of seagrass status and to identify which of these presentedmost variability in the data a principal component analysis (PCA) was used [37ndash39] Principalcomponents (PCs) with eigenvalues greater than 10 were considered and variable coefficients of lessthan or equal to 03 and greater than 03 in each component were selected as dominant

4

rsosroyalsocietypublishingorgRSocopensci3150596

Table1SeagrasssamplingsitesaroundtheBritish

Islesdateofcollectionandinitialassessm

entonperceivedpresenceofanthropogenicimpacts

permanent

site

coordinates

date

industry

tourism

agriculture

catchment

population

anchoring

mooring

GelliswickBayWales(GB)

51 42prime 26primeprimeN53prime 42primeprimeW

26May2013

Southend-on-SeaEngland(SS)

51 32prime 21primeprimeN039

prime 04primeprimeE

17July2013

PrioryBayIsleofWight(PB)

50 42prime 23primeprimeN15prime 45primeprimeW

11June2013

RamseyBayIsleofMan(RB)

54 18prime38

primeprime N421

prime 18primeprime W

26August2013

StudlandBayEngland(SB)

50 38prime34

primeprime N156

prime 28primeprime W

17July2013

Kircubin

BayStrangfordLough

54 29prime02

primeprime N532

prime 20primeprime W

23August2013

NorthernIreland(KB)

Mannin

BayIreland(MB)

53 26prime45

primeprime N10

4prime 45primeprimeW

21August2013

LangnessIsleofMan(LA)

54 4

prime 27primeprime N

436

prime 48primeprimeW

18June2013

PorthdinllaenWales(PD)

51 44prime30

primeprime N516

prime 79primeprime W

27June2013

IslesofScilly

(IS)

LittleArthurEasternIsles

49 56prime59

primeprime N615

prime 59primeprime W

28July2013

HigherTown

BayStMartinrsquos

49 57prime 17primeprimeN616

prime 41primeprimeW

29July2013

GrimsbyHarbourTresco

49 57prime 41primeprimeN619

prime 53primeprime W

30July2013

BroadLedgesTresco

49 56prime29

primeprime N619

prime 41primeprimeW

31July2013

Skom

erMCZWales(SK)

51 44prime30

primeprime N516

prime 79primeprime W

2May2013

5

rsosroyalsocietypublishingorgRSocopensci3150596

MB

PD

SK

N

0 200 km

KBRB

LA

GB

SB PB

IS

SS

location of sampled seagrass beds

Figure 1 Seagrass meadows were sampled at 11 locations spread throughout the British Isles

3 Results31 Seagrass morphometricsLeaf length and width differed significantly between sites (figure 2 and table 2) with leaf length(p lt 0001 H10 = 322) ranging from 142 to 788 mm (mean 347 plusmn 201) and leaf width (p lt 0001H10 = 325) ranging from 3 to 11 mm (mean 5 plusmn 2) Longest leaves were observed in the Isles of Scillyand Mannin Bay and smallest in Porthdinllaen and Southend-on-Sea Similarly widest leaves wereobserved in the Isles of Scilly and Mannin Bay with the narrowest being observed in PorthdinllaenIn addition shoot characteristics were highly variable between sites Shoot biomass also significantly(p lt 0005 H10 = 279) differed with respect to site ranging from 01 to 26 g (mean 03 plusmn 08) with highestvalues in the Isles of Scilly and lowest values in Porthdinllaen Skomer MCZ and Kircubin Bay Shootdensity was again significantly (p lt 0005 H10 = 291) different between sites and ranged from 40 to 1017per 025 m2 (mean 483 plusmn 324) with highest values in Priory Bay and Porthdinllaen and lowest valuesin the Isles of Scilly and Ramsey Bay Seagrass cover ranged from 167 to 913 (mean 481 plusmn 214) withhighest values in the Isles of Scilly and lowest values in Gelliswick Bay and Kircubin Bay The numberof leaves per shoot ranged from 35 to 60 (mean 46 plusmn 09) with highest values in Priory Bay and lowestvalues in Mannin Bay and Porthdinllaen

32 Nutrient concentrations and ratiosSignificant differences in leaf tissue N (p lt 0001 F1032 = 377) and P (p lt 0001 F1032 = 298) wereobserved between locations Per cent N ranged from 205 to 554 (mean 358 plusmn 095) with highest valuesin Skomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Mannin Bay Thesefindings are surprising given that shoot density was highest at Porthdinllaen and Priory Bay two sitesof high N Pearsonrsquos correlation found no significant relationship between N and shoot density

6

rsosroyalsocietypublishingorgRSocopensci3150596

12

10

8

6

4

2

0

1000

800

600

400

200

0

leaf

wid

th (

mm

)

leaf

leng

th (

mm

)

Gel

lisw

ick

Prio

ry B

ay

Skom

er

Port

hdin

llaen

Lag

ness

Ram

sey

Stud

land

Sout

hend

Kir

cubi

n

Scill

y

Man

nin

leaf width (mm)leaf length (mm)

Figure 2 Leaf length and leaf width of seagrass at 11 locations spread throughout the British Isles

(electronic supplementary material appendix 1) The N values exhibited at sites in this study weremostly very high relative to other global values (table 3) Gelliswick Bay Priory Bay Skomer MCZPorthdinllaen Langness Ramsey Bay Southend-on-Sea and Kircubin Bay all had values above 3 andonly Studland Bay the Isles of Scilly and Mannin Bay exhibited values below 3 No sites exhibited Nvalues lower than the global average of 204

Per cent P ranged from 011 to 038 (mean 021 plusmn 007) where highest values were recorded atSkomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Studland Bay PrioryBay Porthdinllaen Langness Ramsey Bay Studland Bay Kircubin Bay the Isles of Scilly and ManninBay all exhibited P values lower than the global average of 027 and only Gelliswick Bay Skomer MCZand Southend-on-Sea exhibited higher than average values

Levels of the nutrient ratio N P significantly differed with respect to site (p lt 0001 F1032 = 575) Allsites were found to be highly imbalanced in terms of available nutrients with N P values more than30 (figure 3) Specifically Priory Bay Langness Studland Bay Kircubin Bay and the Isles of Scilly allhad highly elevated N P ratios (more than 40) suggesting high nutrient imbalance and a limitation inP These sites all with P values lower than the global average also exhibited high C P ratios (morethan 500) indicating that the plants are growing in environments with a relatively low P pool (figure 4)C P ratios also differed significantly with respect to site (p lt 0001 F1032 = 150) In comparison SkomerMCZ Southend-on-Sea and Gelliswick Bay all with high P levels elevated N P and C P ratios ofless than 500 indicate that the plants are growing in a nutrient-rich environment with a relativelylarge P pool

The ratio of C N can provide a potential indication of light limitation in seagrass [47] wherelevels below 20 suggest reduced light availability and levels below 15 being of potential limitation(figure 5) [2733] Levels of C N differed significantly with respect to site (p lt 0001 F1032 = 228) Onthis basis seagrass at Skomer MCZ Southend-on-Sea and Gelliswick Bay all suffered from potentiallight limitation Only the seagrass at the Isles of Scilly Mannin Bay and to a lesser extent Studland Bayhad levels suggesting sufficient light availability

33 Principal component analysisIn order to determine which characteristics contributed most to the variability within and betweensites a PCA was used In the PCA of the measured seagrass characteristics there were three PCAsthat had eigenvalues greater than 1 and together accounted for 784 of the variance (table 4) Thefirst principal component (PC1) and the component with the most importance had an eigenvalue of537 and contributed to 537 of the variance This component was dominated by molar C N molarC P percentage cover shoot biomass leaf width and leaf length This axis is suggested to be looselyrepresentative of growth and productivity and thus overall health in relation to available nutrients andlight This is a useful way of ranking sites based on their leaf characteristics where an increase in PC1represents an increase in overall leaf size attributed to an increase in light availability (high C N) and adecrease in P load (high C P)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

4

rsosroyalsocietypublishingorgRSocopensci3150596

Table1SeagrasssamplingsitesaroundtheBritish

Islesdateofcollectionandinitialassessm

entonperceivedpresenceofanthropogenicimpacts

permanent

site

coordinates

date

industry

tourism

agriculture

catchment

population

anchoring

mooring

GelliswickBayWales(GB)

51 42prime 26primeprimeN53prime 42primeprimeW

26May2013

Southend-on-SeaEngland(SS)

51 32prime 21primeprimeN039

prime 04primeprimeE

17July2013

PrioryBayIsleofWight(PB)

50 42prime 23primeprimeN15prime 45primeprimeW

11June2013

RamseyBayIsleofMan(RB)

54 18prime38

primeprime N421

prime 18primeprime W

26August2013

StudlandBayEngland(SB)

50 38prime34

primeprime N156

prime 28primeprime W

17July2013

Kircubin

BayStrangfordLough

54 29prime02

primeprime N532

prime 20primeprime W

23August2013

NorthernIreland(KB)

Mannin

BayIreland(MB)

53 26prime45

primeprime N10

4prime 45primeprimeW

21August2013

LangnessIsleofMan(LA)

54 4

prime 27primeprime N

436

prime 48primeprimeW

18June2013

PorthdinllaenWales(PD)

51 44prime30

primeprime N516

prime 79primeprime W

27June2013

IslesofScilly

(IS)

LittleArthurEasternIsles

49 56prime59

primeprime N615

prime 59primeprime W

28July2013

HigherTown

BayStMartinrsquos

49 57prime 17primeprimeN616

prime 41primeprimeW

29July2013

GrimsbyHarbourTresco

49 57prime 41primeprimeN619

prime 53primeprime W

30July2013

BroadLedgesTresco

49 56prime29

primeprime N619

prime 41primeprimeW

31July2013

Skom

erMCZWales(SK)

51 44prime30

primeprime N516

prime 79primeprime W

2May2013

5

rsosroyalsocietypublishingorgRSocopensci3150596

MB

PD

SK

N

0 200 km

KBRB

LA

GB

SB PB

IS

SS

location of sampled seagrass beds

Figure 1 Seagrass meadows were sampled at 11 locations spread throughout the British Isles

3 Results31 Seagrass morphometricsLeaf length and width differed significantly between sites (figure 2 and table 2) with leaf length(p lt 0001 H10 = 322) ranging from 142 to 788 mm (mean 347 plusmn 201) and leaf width (p lt 0001H10 = 325) ranging from 3 to 11 mm (mean 5 plusmn 2) Longest leaves were observed in the Isles of Scillyand Mannin Bay and smallest in Porthdinllaen and Southend-on-Sea Similarly widest leaves wereobserved in the Isles of Scilly and Mannin Bay with the narrowest being observed in PorthdinllaenIn addition shoot characteristics were highly variable between sites Shoot biomass also significantly(p lt 0005 H10 = 279) differed with respect to site ranging from 01 to 26 g (mean 03 plusmn 08) with highestvalues in the Isles of Scilly and lowest values in Porthdinllaen Skomer MCZ and Kircubin Bay Shootdensity was again significantly (p lt 0005 H10 = 291) different between sites and ranged from 40 to 1017per 025 m2 (mean 483 plusmn 324) with highest values in Priory Bay and Porthdinllaen and lowest valuesin the Isles of Scilly and Ramsey Bay Seagrass cover ranged from 167 to 913 (mean 481 plusmn 214) withhighest values in the Isles of Scilly and lowest values in Gelliswick Bay and Kircubin Bay The numberof leaves per shoot ranged from 35 to 60 (mean 46 plusmn 09) with highest values in Priory Bay and lowestvalues in Mannin Bay and Porthdinllaen

32 Nutrient concentrations and ratiosSignificant differences in leaf tissue N (p lt 0001 F1032 = 377) and P (p lt 0001 F1032 = 298) wereobserved between locations Per cent N ranged from 205 to 554 (mean 358 plusmn 095) with highest valuesin Skomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Mannin Bay Thesefindings are surprising given that shoot density was highest at Porthdinllaen and Priory Bay two sitesof high N Pearsonrsquos correlation found no significant relationship between N and shoot density

6

rsosroyalsocietypublishingorgRSocopensci3150596

12

10

8

6

4

2

0

1000

800

600

400

200

0

leaf

wid

th (

mm

)

leaf

leng

th (

mm

)

Gel

lisw

ick

Prio

ry B

ay

Skom

er

Port

hdin

llaen

Lag

ness

Ram

sey

Stud

land

Sout

hend

Kir

cubi

n

Scill

y

Man

nin

leaf width (mm)leaf length (mm)

Figure 2 Leaf length and leaf width of seagrass at 11 locations spread throughout the British Isles

(electronic supplementary material appendix 1) The N values exhibited at sites in this study weremostly very high relative to other global values (table 3) Gelliswick Bay Priory Bay Skomer MCZPorthdinllaen Langness Ramsey Bay Southend-on-Sea and Kircubin Bay all had values above 3 andonly Studland Bay the Isles of Scilly and Mannin Bay exhibited values below 3 No sites exhibited Nvalues lower than the global average of 204

Per cent P ranged from 011 to 038 (mean 021 plusmn 007) where highest values were recorded atSkomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Studland Bay PrioryBay Porthdinllaen Langness Ramsey Bay Studland Bay Kircubin Bay the Isles of Scilly and ManninBay all exhibited P values lower than the global average of 027 and only Gelliswick Bay Skomer MCZand Southend-on-Sea exhibited higher than average values

Levels of the nutrient ratio N P significantly differed with respect to site (p lt 0001 F1032 = 575) Allsites were found to be highly imbalanced in terms of available nutrients with N P values more than30 (figure 3) Specifically Priory Bay Langness Studland Bay Kircubin Bay and the Isles of Scilly allhad highly elevated N P ratios (more than 40) suggesting high nutrient imbalance and a limitation inP These sites all with P values lower than the global average also exhibited high C P ratios (morethan 500) indicating that the plants are growing in environments with a relatively low P pool (figure 4)C P ratios also differed significantly with respect to site (p lt 0001 F1032 = 150) In comparison SkomerMCZ Southend-on-Sea and Gelliswick Bay all with high P levels elevated N P and C P ratios ofless than 500 indicate that the plants are growing in a nutrient-rich environment with a relativelylarge P pool

The ratio of C N can provide a potential indication of light limitation in seagrass [47] wherelevels below 20 suggest reduced light availability and levels below 15 being of potential limitation(figure 5) [2733] Levels of C N differed significantly with respect to site (p lt 0001 F1032 = 228) Onthis basis seagrass at Skomer MCZ Southend-on-Sea and Gelliswick Bay all suffered from potentiallight limitation Only the seagrass at the Isles of Scilly Mannin Bay and to a lesser extent Studland Bayhad levels suggesting sufficient light availability

33 Principal component analysisIn order to determine which characteristics contributed most to the variability within and betweensites a PCA was used In the PCA of the measured seagrass characteristics there were three PCAsthat had eigenvalues greater than 1 and together accounted for 784 of the variance (table 4) Thefirst principal component (PC1) and the component with the most importance had an eigenvalue of537 and contributed to 537 of the variance This component was dominated by molar C N molarC P percentage cover shoot biomass leaf width and leaf length This axis is suggested to be looselyrepresentative of growth and productivity and thus overall health in relation to available nutrients andlight This is a useful way of ranking sites based on their leaf characteristics where an increase in PC1represents an increase in overall leaf size attributed to an increase in light availability (high C N) and adecrease in P load (high C P)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

5

rsosroyalsocietypublishingorgRSocopensci3150596

MB

PD

SK

N

0 200 km

KBRB

LA

GB

SB PB

IS

SS

location of sampled seagrass beds

Figure 1 Seagrass meadows were sampled at 11 locations spread throughout the British Isles

3 Results31 Seagrass morphometricsLeaf length and width differed significantly between sites (figure 2 and table 2) with leaf length(p lt 0001 H10 = 322) ranging from 142 to 788 mm (mean 347 plusmn 201) and leaf width (p lt 0001H10 = 325) ranging from 3 to 11 mm (mean 5 plusmn 2) Longest leaves were observed in the Isles of Scillyand Mannin Bay and smallest in Porthdinllaen and Southend-on-Sea Similarly widest leaves wereobserved in the Isles of Scilly and Mannin Bay with the narrowest being observed in PorthdinllaenIn addition shoot characteristics were highly variable between sites Shoot biomass also significantly(p lt 0005 H10 = 279) differed with respect to site ranging from 01 to 26 g (mean 03 plusmn 08) with highestvalues in the Isles of Scilly and lowest values in Porthdinllaen Skomer MCZ and Kircubin Bay Shootdensity was again significantly (p lt 0005 H10 = 291) different between sites and ranged from 40 to 1017per 025 m2 (mean 483 plusmn 324) with highest values in Priory Bay and Porthdinllaen and lowest valuesin the Isles of Scilly and Ramsey Bay Seagrass cover ranged from 167 to 913 (mean 481 plusmn 214) withhighest values in the Isles of Scilly and lowest values in Gelliswick Bay and Kircubin Bay The numberof leaves per shoot ranged from 35 to 60 (mean 46 plusmn 09) with highest values in Priory Bay and lowestvalues in Mannin Bay and Porthdinllaen

32 Nutrient concentrations and ratiosSignificant differences in leaf tissue N (p lt 0001 F1032 = 377) and P (p lt 0001 F1032 = 298) wereobserved between locations Per cent N ranged from 205 to 554 (mean 358 plusmn 095) with highest valuesin Skomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Mannin Bay Thesefindings are surprising given that shoot density was highest at Porthdinllaen and Priory Bay two sitesof high N Pearsonrsquos correlation found no significant relationship between N and shoot density

6

rsosroyalsocietypublishingorgRSocopensci3150596

12

10

8

6

4

2

0

1000

800

600

400

200

0

leaf

wid

th (

mm

)

leaf

leng

th (

mm

)

Gel

lisw

ick

Prio

ry B

ay

Skom

er

Port

hdin

llaen

Lag

ness

Ram

sey

Stud

land

Sout

hend

Kir

cubi

n

Scill

y

Man

nin

leaf width (mm)leaf length (mm)

Figure 2 Leaf length and leaf width of seagrass at 11 locations spread throughout the British Isles

(electronic supplementary material appendix 1) The N values exhibited at sites in this study weremostly very high relative to other global values (table 3) Gelliswick Bay Priory Bay Skomer MCZPorthdinllaen Langness Ramsey Bay Southend-on-Sea and Kircubin Bay all had values above 3 andonly Studland Bay the Isles of Scilly and Mannin Bay exhibited values below 3 No sites exhibited Nvalues lower than the global average of 204

Per cent P ranged from 011 to 038 (mean 021 plusmn 007) where highest values were recorded atSkomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Studland Bay PrioryBay Porthdinllaen Langness Ramsey Bay Studland Bay Kircubin Bay the Isles of Scilly and ManninBay all exhibited P values lower than the global average of 027 and only Gelliswick Bay Skomer MCZand Southend-on-Sea exhibited higher than average values

Levels of the nutrient ratio N P significantly differed with respect to site (p lt 0001 F1032 = 575) Allsites were found to be highly imbalanced in terms of available nutrients with N P values more than30 (figure 3) Specifically Priory Bay Langness Studland Bay Kircubin Bay and the Isles of Scilly allhad highly elevated N P ratios (more than 40) suggesting high nutrient imbalance and a limitation inP These sites all with P values lower than the global average also exhibited high C P ratios (morethan 500) indicating that the plants are growing in environments with a relatively low P pool (figure 4)C P ratios also differed significantly with respect to site (p lt 0001 F1032 = 150) In comparison SkomerMCZ Southend-on-Sea and Gelliswick Bay all with high P levels elevated N P and C P ratios ofless than 500 indicate that the plants are growing in a nutrient-rich environment with a relativelylarge P pool

The ratio of C N can provide a potential indication of light limitation in seagrass [47] wherelevels below 20 suggest reduced light availability and levels below 15 being of potential limitation(figure 5) [2733] Levels of C N differed significantly with respect to site (p lt 0001 F1032 = 228) Onthis basis seagrass at Skomer MCZ Southend-on-Sea and Gelliswick Bay all suffered from potentiallight limitation Only the seagrass at the Isles of Scilly Mannin Bay and to a lesser extent Studland Bayhad levels suggesting sufficient light availability

33 Principal component analysisIn order to determine which characteristics contributed most to the variability within and betweensites a PCA was used In the PCA of the measured seagrass characteristics there were three PCAsthat had eigenvalues greater than 1 and together accounted for 784 of the variance (table 4) Thefirst principal component (PC1) and the component with the most importance had an eigenvalue of537 and contributed to 537 of the variance This component was dominated by molar C N molarC P percentage cover shoot biomass leaf width and leaf length This axis is suggested to be looselyrepresentative of growth and productivity and thus overall health in relation to available nutrients andlight This is a useful way of ranking sites based on their leaf characteristics where an increase in PC1represents an increase in overall leaf size attributed to an increase in light availability (high C N) and adecrease in P load (high C P)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

6

rsosroyalsocietypublishingorgRSocopensci3150596

12

10

8

6

4

2

0

1000

800

600

400

200

0

leaf

wid

th (

mm

)

leaf

leng

th (

mm

)

Gel

lisw

ick

Prio

ry B

ay

Skom

er

Port

hdin

llaen

Lag

ness

Ram

sey

Stud

land

Sout

hend

Kir

cubi

n

Scill

y

Man

nin

leaf width (mm)leaf length (mm)

Figure 2 Leaf length and leaf width of seagrass at 11 locations spread throughout the British Isles

(electronic supplementary material appendix 1) The N values exhibited at sites in this study weremostly very high relative to other global values (table 3) Gelliswick Bay Priory Bay Skomer MCZPorthdinllaen Langness Ramsey Bay Southend-on-Sea and Kircubin Bay all had values above 3 andonly Studland Bay the Isles of Scilly and Mannin Bay exhibited values below 3 No sites exhibited Nvalues lower than the global average of 204

Per cent P ranged from 011 to 038 (mean 021 plusmn 007) where highest values were recorded atSkomer MCZ and Southend-on-Sea and lowest values in the Isles of Scilly and Studland Bay PrioryBay Porthdinllaen Langness Ramsey Bay Studland Bay Kircubin Bay the Isles of Scilly and ManninBay all exhibited P values lower than the global average of 027 and only Gelliswick Bay Skomer MCZand Southend-on-Sea exhibited higher than average values

Levels of the nutrient ratio N P significantly differed with respect to site (p lt 0001 F1032 = 575) Allsites were found to be highly imbalanced in terms of available nutrients with N P values more than30 (figure 3) Specifically Priory Bay Langness Studland Bay Kircubin Bay and the Isles of Scilly allhad highly elevated N P ratios (more than 40) suggesting high nutrient imbalance and a limitation inP These sites all with P values lower than the global average also exhibited high C P ratios (morethan 500) indicating that the plants are growing in environments with a relatively low P pool (figure 4)C P ratios also differed significantly with respect to site (p lt 0001 F1032 = 150) In comparison SkomerMCZ Southend-on-Sea and Gelliswick Bay all with high P levels elevated N P and C P ratios ofless than 500 indicate that the plants are growing in a nutrient-rich environment with a relativelylarge P pool

The ratio of C N can provide a potential indication of light limitation in seagrass [47] wherelevels below 20 suggest reduced light availability and levels below 15 being of potential limitation(figure 5) [2733] Levels of C N differed significantly with respect to site (p lt 0001 F1032 = 228) Onthis basis seagrass at Skomer MCZ Southend-on-Sea and Gelliswick Bay all suffered from potentiallight limitation Only the seagrass at the Isles of Scilly Mannin Bay and to a lesser extent Studland Bayhad levels suggesting sufficient light availability

33 Principal component analysisIn order to determine which characteristics contributed most to the variability within and betweensites a PCA was used In the PCA of the measured seagrass characteristics there were three PCAsthat had eigenvalues greater than 1 and together accounted for 784 of the variance (table 4) Thefirst principal component (PC1) and the component with the most importance had an eigenvalue of537 and contributed to 537 of the variance This component was dominated by molar C N molarC P percentage cover shoot biomass leaf width and leaf length This axis is suggested to be looselyrepresentative of growth and productivity and thus overall health in relation to available nutrients andlight This is a useful way of ranking sites based on their leaf characteristics where an increase in PC1represents an increase in overall leaf size attributed to an increase in light availability (high C N) and adecrease in P load (high C P)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

7

rsosroyalsocietypublishingorgRSocopensci3150596

Table2M

easuredparametersofseagrasssampledat11locationsspreadthroughouttheBritishIsles

shoot

shoot

seagrass

leaf

leaf

epiphytes

leaves

site

N

P

density(per025

m2 )

biomass(g)

cover()

length(mm)

width(mm)

(gpershoot)

pershoot

GelliswickBay(GB)

437plusmn

036

027plusmn

006

440

plusmn104

009plusmn

004

167

plusmn29

218plusmn

440

plusmn00

006plusmn

002

40plusmn

00

Southend-on-Sea(SS)

500plusmn

018

033plusmn

001

317

plusmn35

008plusmn

001

500

plusmn00

156plusmn

738

plusmn00

000plusmn

000

37plusmn

01

PrioryBay(PB)

380plusmn

004

021plusmn

000

1017

plusmn253

009plusmn

001

483

plusmn152

296plusmn

467

plusmn00

000plusmn

000

60plusmn

01

RamseyBay(RB)

318plusmn

021

019plusmn

003

187

plusmn12

009plusmn

002

550

plusmn50

367plusmn

2443

plusmn05

000plusmn

000

49plusmn

04

StudlandBay(SB)

290plusmn

019

014plusmn

001

360

plusmn98

009plusmn

003

333

plusmn29

417plusmn

2059

plusmn03

000plusmn

000

51plusmn

03

Kircubin

Bay(KB)

335plusmn

021

018plusmn

003

850

plusmn380

006plusmn

001

200

plusmn87

153plusmn

941

plusmn01

000plusmn

000

43plusmn

02

Mannin

Bay(MB)

227plusmn

022

014plusmn

003

420

plusmn147

019plusmn

012

650

plusmn265

594plusmn

7767

plusmn06

001plusmn

001

35plusmn

03

Langness(LA)

343plusmn

051

017plusmn

002

250

plusmn30

014plusmn

001

617

plusmn29

405plusmn

1455

plusmn03

001plusmn

000

56plusmn

01

Porthdinllaen(PD)

351plusmn

013

025plusmn

002

973

plusmn393

003plusmn

001

350

plusmn87

142plusmn

2032

plusmn01

004plusmn

004

36plusmn

05

IslesofScilly

(IS)

263plusmn

029

013plusmn

001

40plusmn

14261plusmn

140

913

plusmn25

788plusmn

49107

plusmn05

000plusmn

001

54plusmn

03

Skom

erMCZ(SK)

530plusmn

032

036plusmn

001

457

plusmn100

005plusmn

002

533

plusmn76

279plusmn

539

plusmn00

000plusmn

000

40plusmn

02

average

358plusmn

095

021plusmn

007

483

plusmn324

03plusmn

08481

plusmn214

346plusmn

201

55plusmn

23001plusmn

002

46plusmn

09

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

8

rsosroyalsocietypublishingorgRSocopensci3150596

60

50

40

30

20

10

0

balanced nutrientsP-limited

highly P-limited

GB PB SK PD LA RB SB SS KB IS MB

NP

rat

io

Figure 3 Leaf N P ratio variability within seagrass at 11 locations spread throughout the British Isles (balanced nutrients = 0ndash20P-limited= 20ndash40 highly P-limitedge40) (values for differentiation between states derived from [303133])

Table 3 Zostera marina leaf carbon nitrogen and phosphorus content from literature sources

location C N P C N N P C P reference

North Carolina USA 3284 184 022 2081 1849 38494 [40]

North Carolina USA 3367 153 025 2566 1353 34731 [41]

Virginia USA 3750 220 022 1988 2211 43957 [42]

Denmark 3840 192 2332 [43]

California USA 3840 237 034 1890 1541 29125 [29]

South Korea 3528 273 1509 [44]

Denmark 3078 194 1944 [45]

Oregon USA 3400 270 040 1469 1493 21920 [46]

Oregon USA 3500 130 020 3140 1437 45129 [46]

Oregon USA 2900 140 010 2416 3096 74785 [46]

various 3600a 250b 039c 1679 1418 23804 [32]

global average 3462 204 027 2092 1800 38993

study averages 4773 358 021 1654 389 65453

aForty-six measurementsbForty-five measurementscThirty-six measurements

Molar N P shoot biomass shoot density epiphytes per shoot and leaves per shoot were the dominantvariables for the second principal component (PC2) which had a much smaller eigenvalue of 139 andcontributed to an additional 139 of the variance We identified PC2 as being representative of relativenutrient balance owing to the high weighting of molar N P which when low is indicative of balancednutrients The third and final principal component (PC3) which was responsible for a smaller 108 ofthe variance and had an eigenvalue of 108 was dominated by molar C N molar C P per cent coverepiphytes per shoot and leaves per shoot We can identify that lower light availability and increasednutrient load are indicative of lower seagrass coverage and higher epiphyte biomass but contribute lessto the variability over each location

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

9

rsosroyalsocietypublishingorgRSocopensci3150596

1200

1000

800

600

400

200

0

low Pelivated Phigh P

GB PB SK PD LA RB SB SS KB IS MB

CP

rat

io

Figure 4 Leaf C P ratio variability within seagrass at 11 locations spread throughout the British Isles (low P ge600 elevatedP= 400ndash600 high Ple400) (values for differentiation between states derived from [303133])

30

25

20

15

10

5

0

high lightreduced lightlow light

GB PB SK PD LA RB SB SS KB IS MB

CN

rat

io

Figure 5 Leaf C N ratio variability within seagrass at 11 locations spread throughout the British Isles (high lightge20 reduced light=14ndash20 low lightle14) (values for differentiation between states derived from [303133])

Figure 6 illustrates the variability of seagrasses at sites in relation to PC1 and PC2 The spread of dataalong the PC1 axis (representative of growth as a measure of seagrass health) suggests that seagrass inthe Isles of Scilly are in an ecologically healthy state compared with seagrasses located elsewhere in theBritish Isles specifically that found in Southend-on-Sea Gelliswick Bay Skomer MCZ and Porthdinllaen

34 Anecdotal observationsAt all sites available information and personal observations of the authors provided evidence of thepresence of risk factors to seagrass The majority of the sites assessed were found to have issues that havethe potential to result in elevated nutrients conditions and disturbance These risks may ultimately resultin reduced water quality and light availability Six of the sites were observed to have boat anchoring andpermanent moorings present or in close proximity to the seagrass

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

10

rsosroyalsocietypublishingorgRSocopensci3150596

2

0

ndash2

ndash4

Gelliswick BayPriory BaySkomer IslandPorthdinllaenLangnessRamsey BayStudland BaySouthend-on-SeaKircubin BayIsles of ScillyMannin Bay

ndash4 ndash2 0 2 4 6PC1

seagrass health

PC2

nutr

ient

bal

ance

Figure 6 Principal component analysis (PCA) of seagrass characteristics from 11 locations spread throughout the British Isles

Table 4 PCA of seagrass sampled at 11 locations spread throughout the British Isles (Coefficients shown in italics represent dominantvariables in each PC)

principal component

PC1 PC2 PC3

summary values

eigenvalues 537 139 108

per cent variation 537 139 108

cumulative per cent variation 537 676 784

seagrass characteristics

C N 0344 minus0024 0369

C P 0373 minus0223 0375

N P 0275 minus0495 0253

per cent cover 0313 0271 minus0453

shoot biomass 0323 0314 minus0049

shoot density minus0211 minus0422 minus0017

leaf length (mm) 0406 0209 0032

leaf width (mm) 0403 0086 minus0034

epiphytes (g per shoot) minus0175 0363 0592

leaves per shoot 025 minus0414 minus0313

4 DiscussionSeagrass meadows are productive coastal habitats of high ecosystem service value that are threatenedglobally [2448] One of the main threats is that of declining water quality [411] This study providesto our knowledge the first quantitative evidence of how such declining water quality is negativelyimpacting the health status of seagrass meadows throughout the British Isles Although the data onlyprovide a rapid and limited snap shot of the environmental status of these meadows it is sufficientto clearly indicate that many seagrass meadows in the British Isles are under anthropogenic stress andprobably in a poor state of health many of which are in sites of apparent conservation protection (eg

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

11

rsosroyalsocietypublishingorgRSocopensci3150596

within special areas of conservation) The study also highlights the presence of some healthy seagrassmeadows but these are far from large human populations (Isles of Scilly and Mannin Bay Ireland)Long-term ecological data from the Isles of Scilly seagrass confirm the relatively healthy status of thesesystems [26]

The tissue nutrient data for seagrasses of the British Isles collected in this study indicate that thesehabitats are highly over-enriched with N levels over 75 higher than the global average for Z marinaOur British Isles sites had N levels consistently higher than the global average with the exception ofIsles of Scilly and Mannin Bay in the west of Ireland

In comparison with N levels seagrasses were found to be on average low in P with levels waybelow the global average suggesting P-limitation and a largely unbalanced nutrient regime P limitationmay be exacerbated by the abundance of calcium carbonate sediments at some of these sites [49]particularly those in the Isles of Scilly Studland Bay and Mannin Bay Only seagrasses at Skomer MCZand Southend-on-Sea had P levels above the global average

Specific sites of concern with high nitrogen values were those in Gelliswick Bay (Milford Haven)Priory Bay (Isle of Wight) Southend-on-Sea and Skomer MCZ Although the first three of these arewithin or close to water bodies of known eutrophication problems (Milford Haven Waterway ThamesEstuary and SolentIsle of Wight) the site at Skomer MCZ was initially thought to be removed fromsources of high nutrient input The capacity for leaves to absorb nutrients declines with leaf age wherehighest N values have been observed in April (younger leaves) after which they rapidly decline totheir lowest values in July [50] Skomer MCZ and Gelliswick Bay were sampled at the start and endof May respectively thus high leaf N levels may partly be owing to early sampling however subsequentsamples taken at different times of the year have revealed similar levels of N (R Unsworth 2014unpublished data) suggesting this is not the case Additionally we hypothesize that the high nutrientconcentrations recorded at Skomer MCZ may be the result of the high populations of seals [51] birds[52] and potentially a small level of untreated effluent from nearby buildings present at the site Theseissues may be exacerbated by potential low mixing of the water within the bay As is increasinglybeing realized globally seagrasses that contain a healthy and balanced associated fauna may actuallybe buffered from the impacts of elevated nutrients through the process of grazing [1053] The enhancedprotection afforded to this site by MCZ designation may contribute towards buffering of these nutrientconcerns

The use of the C N ratio as an indicator of light availability [2754] also suggests that seagrassmeadows of the British Isles are under anthropogenic stress with values 25 lower than the globalaverage and at levels that other studies have suggested indicate light limitation [273054] Particularconcern exists for seagrass at Skomer MCZ Priory Bay Gelliswick Bay Porthdinllaen and Southend-on-Sea which all have C N ratios equal to or lower than 15 which is a level potentially below which isindicative of light limitation [273031] Not surprisingly seagrass at the Isles of Scilly and Mannin Bayall had C N levels higher than 20 suggesting a high light environment suitable for productive seagrassmeadows Studland Bay was also found to have a suitable light environment

Low-impact (high C N high C P) sites such as the Isles of Scilly and Mannin Bay were characteristicof high biomass larger leaves high numbers of leaves per shoot high percentage cover generally lowshoot density and lower epiphyte biomass whereas high-impact (low C N low C P) sites such asGelliswick Bay Southend-on-Sea and Skomer MCZ were characteristic of low biomass smaller leaveslow percentage cover generally high shoot density and high epiphyte biomass showing similarities toprevious findings elsewhere with the exception of shoot density [55ndash57]

Our use of multivariate analysis enabled a generalized assessment of the lsquohealthrsquo of these 11 seagrasssites in the British Isles All three sites in Wales (Gelliswick Bay Skomer MCZ and Porthdinllaen)as well as that at Southend-on-Sea performed the worst Although the present assessment providesan important warning about the status of seagrass at those sites the limited data extent provided bythis rapid assessment requires that these findings are interpreted with caution The data in this studywere collected over a summer period between May and August such unbalanced sampling may haveinfluenced the data to a small degree owing to seasonality Future comparative studies between sitesusing the indicators described in this study need to be undertaken with seasonal comparison andorusing consistent sampling times and use consistent methodologies

Present research also highlights the value of using seagrass tissue nutrients and morphometricmeasurements as indicators of environmental change in these important ecosystems reflecting ambientnutrient levels better than point sampling alone Daily seasonal and annual variability in water qualityis caused both by pulsed events (eg outfall discharges storms) and mixing through the action of tidesand currents [58] This in addition to the relatively rapid uptake of nutrients by algae microbes and

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

12

rsosroyalsocietypublishingorgRSocopensci3150596

plants [58ndash60] means that early signs of enrichment are difficult to detect and that point measurementsof nutrients (conventional water samples) do not always describe the levels of nutrients available tothe biota and its impacts Seagrasses absorb nutrients from the water column and sediments henceseagrass leaf tissue nutrients can give a time-integrated relative measure of the available environmentalnutrients [116162]

Although the use of seagrass as an indicator has been readily applied as such a tool elsewhere inEurope [17] current seagrass monitoring strategies in the UK which are routinely carried out lackcontext and do not give indication of environment status as a whole [1617] The development of an indexof seagrass health based on the main PC from PCA revealed that there was extensive variability betweensites in the UK driven by parameters not considered in existing monitoring programmes Our analysisidentified that molar C N and C P along with shoot biomass leaf length and leaf width contributedto over 50 of the variability over the spread of data Molar C N which indicates light availability(and N availability) and C P which is indicative of P availability are useful indicators of environmentquality [32] Thus by combining these with the morphological characteristics of Z marina we canbegin to form a picture of how changes in environment quality owing to nutrient enrichment affectZ marina health

The present dataset suggests that C N C P leaf length and width as well as percentage cover arebest used to determine ecological status of seagrass meadows around the British Isles Shoot density hasbeen readily applied as a measurement for meadow health [63] and is reported to respond negativelyto nutrient enrichment [64ndash66] This study however in addition to on-site observations has shown thatthis partly does not hold true for the British Isles The seagrass at Porthdinllaen which shows signs ofoverenrichment has relatively high shoot density and forms a large meadow However leaf length andwidth were the lowest recorded and epiphyte coverage was high When compared with the seagrass atthe Isles of Scilly which has the lowest shoot density but highest leaf length and width and low epiphytecoverage it is clear that multiple seagrass metrics are required to assess the environmental and ecologicalhealth of the meadow

In addition to the quantitative assessment of their environmental health our study also revealedextensive qualitative anecdotal evidence of the impacts placed on seagrass meadows around the BritishIsles Seagrass meadows at all sites except for Priory Bay and Skomer MCZ were subject to eitherpermanent moorings extensive anchoring or both The extensive nature of the anthropogenic activitysuggested other impacts (trampling and bait digging) were also present

This study provided to our knowledge the first rapid environmental assessment of seagrassmeadows throughout the British Isles In conclusion we find strong evidence of the perilous state ofthe majority of seagrass meadows sampled in the British Isles These meadows were mostly subjected toexcess nutrients and turbid conditions and even those not subject to poor water quality were subjectedto other anthropogenic risks (eg moorings and anchoring) With increasing sea surface temperaturesowing to climatic change it is imperative that seagrass has sufficient light availability to maintain apositive carbon balance owing to elevated respiration [67] What we are observing here suggests manyseagrass meadows are potentially already close to their light thresholds and therefore the resilience ofthese systems to such climatic change is being placed in doubt [10]

This study also shows that the seagrass Z marina can readily be used to provide a valid representationof ecosystem status more so than that of just water sampling alone It is suggested that C N C P leaflength and leaf width as well as per cent cover and shoot biomass are used for future monitoring ofseagrass in the British Isles to give a time-integrated representation of meadow status

Given the increasing body of evidence of the high fisheries nursery value of seagrass meadowsthroughout Northern Europe and the consequences of this for food security this study provides evidencefor the increasing need to manage the environmental condition of seagrass systems in the British IslesReducing the impact of poor water quality and disturbance will enhance the resilience of these systemsinto the future

Data accessibility All data collected in the present study are available within the electronic supplementary materialappendix 2Competing interests We declare we have no competing interestsFunding The authors acknowledge the financial support of the Welsh Government Ecosystem Resilience FundSEACAMS and the Milford Haven Port AuthorityAcknowledgements The authors thank the following for their assistance with sample collection Phil Newman MarkBurton Kate Lock Amy Marsden Robert Hughes Neil Garrick-Maidment Jade Berman James Bull Tony Glen andNicole Esteban We also thank the anonymous reviewers for their invaluable advice

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

13

rsosroyalsocietypublishingorgRSocopensci3150596

References1 den Hartog C 1970 The seagrasses of the world

Amsterdam The Netherlands North HollandPublishing

2 Orth RJ et al 2006 A global crisis for seagrassecosystems Bioscience 56 987ndash996 (doi1016410006-3568(2006)56[987AGCFSE]20CO2)

3 Short F Wyllie-Echeverria S 1996 Natural andhuman-induced disturbance of seagrassesEnviron Conserv 23 17ndash27 (doi101017S0376892900038212)

4 Waycott M et al 2009 Accelerating loss ofseagrasses across the globe threatens coastalecosystems Proc Natl Acad Sci USA 10612 377ndash12 381 (doi101073pnas0905620106)

5 United Nations 2014 United Nations Atlas of theOceans See httpwwwoceansatlasorgindexjsp

6 Cullen-Unsworth LC Nordlund L Paddock J BakerS McKenzie LJ Unsworth RKF 2014 Seagrassmeadows globally as a coupled social-ecologicalsystem implications for human wellbeingMarPollut Bull 83 387ndash397 (doi101016jmarpolbul201306001)

7 Short FT Burdick DM 1996 Quantifying eelgrasshabitat loss in relation to housing development andnitrogen loading in Waquoit Bay MassachusettsEstuaries 19 730ndash739 (doi1023071352532)

8 Tomasko DA Dawes CJ Hall MO 1996 The effects ofanthropogenic nutrient enrichment on turtle grass(Thalassia testudinum) in Sarasota Bay FloridaEstuaries 19 448ndash456 (doi1023071352462)

9 Vitousek PM Mooney HA Lubchenco J Melillo JM1997 Human domination of Earthrsquos ecosystemsScience 277 494ndash499 (doi101126science2775325494)

10 Unsworth RKF Collier CJ Waycott M McKenzie LJCullen-Unsworth LC 2015 A framework for theresilience of seagrass ecosystemsMar Pollut Bull100 34ndash46 (doi101016jmarpolbul201508016)

11 Burkholder JM Tomasko DA Touchette BW 2007Seagrasses and eutrophication J Exp Mar Biol Ecol350 46ndash72 (doi101016jjembe200706024)

12 Lavery PS Vanderklift MA 2002 A comparison ofspatial and temporal patterns in epiphyticmacroalgal assemblages of the seagrassesAmphibolis griffithii and Posidonia coriaceaMarEcol Prog Ser 236 99ndash112 (doi103354meps236099)

13 Neverauskas VP 1987 Monitoring seagrass bedsaround a sewage-sludge outfall in South AustraliaMar Pollut Bull 18 158ndash164 (doi1010160025-326X(87)90239-6)

14 McClelland JW Valiela I 1998 Linking nitrogen inestuarine producers to land-derived sourcesLimnol Oceanogr 43 577ndash585 (doi104319lo19984340577)

15 Dennison WC Longstaff BJ OrsquoDonohue MJ 1997Seagrasses as bio-indicators In Karumba dredging1996 - environmental monitoring report EcoPortsMonograph Series No 6 (eds S HillmanS Raaymakers) p 255 Brisbane Australia PortsCorporation of Queensland

16 Foden J Brazier DP 2007 Angiosperms (seagrass)within the EU water framework directive a UKperspectiveMar Pollut Bull 55 181ndash195(doi101016jmarpolbul200608021)

17 Marbagrave N et al 2013 Diversity of European seagrassindicators patterns within and across regionsHydrobiologia 704 265ndash278 (doi101007s10750-012-1403-7)

18 Bertelli CM Unsworth RKF 2014 Protecting thehand that feeds us seagrass (Zostera marina) servesas commercial juvenile fish habitatMar Pollut Bull83 425ndash429 (doi101016jmarpolbul201308011)

19 Lilley RJ Unsworth RKF 2014 Atlantic cod (Gadusmorhua) benefits from the availability of seagrass(Zostera marina) nursery habitat Glob Ecol Conserv2 367ndash377 (doi101016jgecco201410002)

20 Peters JR McCloskey RM Hinder SL Unsworth RKF2014 Motile fauna of sub-tidal Zostera marinameadows in England andWalesMar Biodivers 45647ndash654 (doi101007s12526-014-0264-x)

21 Jackson EL Attrill MJ Jones MB 2006 Habitatcharacteristics and spatial arrangement affectingthe diversity of fish and decapod assemblages ofseagrass (Zostera marina) beds around the coast ofJersey (English Channel) Estuarine Coast Shelf Sci68 421ndash432 (doi101016jecss200601024)

22 Jackson EL Griffiths CA Durkin O 2013 A guide toassessing and managing anthropogenic impact onmarine angiosperm habitat - part 1 literaturereview Natural England Commissioned Reports

23 Kay QON 1998 A review of the existing state ofknowledge of the ecology and distribution ofseagrass beds around the coast of WalesCountryside Council for Wales Science Report 296Bangor Wales

24 Davison DM Hughes DJ 1998 Zostera biotopes(volume I) An overview of dynamics and sensitivitycharacteristics for conservation management ofmarine SACs Scottish Association for Marine Science95 UK Marine SACs Project

25 Hiscock K Sewell J Oakley J 2005 The marine healthcheck 2005 a report to gauge the health of the UKrsquossea life Godalming UK WWF-UK

26 Potouroglou M Kenyon EJ Gall A Cook KJ Bull JC2014 The roles of flowering overwinter survival andsea surface temperature in the long-termpopulation dynamics of Zostera marina around theIsles of Scilly UKMar Pollut Bull 83 500ndash507(doi101016jmarpolbul201403035)

27 McMahon K Collier C Lavery PS 2013 Identifyingrobust bioindicators of light stress in seagrasses ameta-analysis Ecol Indic 30 7ndash15 (doi101016jecolind201301030)

28 Nedwell DB Dong LF Sage A Underwood GJC 2002Variations of the nutrients loads to the mainlandUK estuaries correlation with catchment areasurbanization and coastal eutrophication EstuarineCoast Shelf Sci 54 951ndash970 (doi101006ecss20010867)

29 Fourqurean JW Moore TO Fry B Hollibaugh JT1997 Spatial and temporal variation in CNP ratiosgamma 15 N and gamma 13 C of eelgrass Zosteramarina as indicators of ecosystem processesTomales Bay California USAMar Ecol Prog Ser157 147ndash157 (doi103354meps157147)

30 McKenzie L Collier CJ Waycott M Unsworth RKFYoshida RL Smith N 2012 Monitoring inshoreseagrasses of the GBR and responses to waterquality Proc 12th Int Coral Reef Symp 15B 4

31 McKenzie LJ Unsworth RKF 2009 Great barrier reefwater quality protection plan (reef rescue) - marinemonitoring program intertidal seagrass FinalReport for the Sampling Period 1st September2008ndash31st May 2009 p 127 Cairns FisheriesQueensland

32 Duarte CM 1990 Seagrass nutrient contentMarEcol Prog Ser 67 201ndash207 (doi103354meps067201)

33 McKenzie LJ Collier C Waycott M 2014 Reef rescuemarine monitoring program - inshore seagrassannual report for the sampling period 1 July 2011ndash31May 2012 TropWATER James Cook UniversityCairns

34 Atkinson MJ 1983 CNP ratios of benthic marineplants Limnol Oceanogr 28 568ndash574(doi104319lo19832830568)

35 Ferdie M Fourqurean JW 2004 Responses ofseagrass communities to fertilization along agradient of relative availability of nitrogen andphosphorus in a carbonate environment LimnolOceanogr 49 2082ndash2094 (doi104319lo20044962082)

36 Fourqurean JW Manuel SA Coates KA KenworthyWJ Boyer JN 2015 Water quality isoscapes andstoichioscapes of seagrasses indicate general Plimitation and unique N cycling in shallow waterbenthos of Bermuda Biogeosci Discuss 129751ndash9791 (doi105194bgd-12-9751-2015)

37 Johnson L Gage S 1997 Landscape approaches tothe analysis of aquatic ecosystems Freshw Biol 37113ndash132 (doi101046j1365-2427199700156x)

38 Salita JT Ekau W Saint-Paul U 2003 Field evidenceon the influence of seagrass landscapes on fishabundance in Bolinao northern PhilippinesMarEcol Prog Ser 247 183ndash195 (doi103354meps247183)

39 Zar JH 1984 Biostatistical Analysis SydneyAustralia Prentice Hall

40 Burkholder JM Mason KM Glasgow HB Jr 1992Water-column nitrate enrichment promotes declineof eelgrass (Zostera marina) evidence fromseasonal mesocosm experimentsMar Ecol ProgSer 81 163ndash178

41 Touchette B Burkholder J Glasgow H 2003Variations in eelgrass (Zostera marina L)morphology and internal nutrient composition asinfluenced by inreased temperature and watercolumn nitrate Estuaries 26 142ndash155(doi101007BF02691701)

42 Moore KA Wetzel RL 2000 Seasonal variations ineelgrass (Zostera marina L) responses to nutrientenrichment and reduced light availability inexperimental ecosystems J Exp Mar Biol Ecol 2441ndash28 (doi101016S0022-0981(99)00135-5)

43 Pedersen MF Borum J 1992 Nitrogen dynamics ofeelgrass Zostera marina during a late summerperiod of high growth and low nutrient availabiltyMar Ecol Prog Ser 80 65ndash73(doi103354meps080065)

44 Kim YK Kim J-H Kim SH Kim JW Park SR Lee K-S2012 Growth dynamics of the seagrass Zosteramarina in Jindong Bay on the southern coast ofKorea Algae 27 215ndash224(doi104490algae2012273215)

14

rsosroyalsocietypublishingorgRSocopensci3150596

45 Papadimitriou S Kennedy H Kennedy DP Borum J

2005 Seasonal and spatial variation in the organiccarbon and nitrogen concentration and their stableisotopic composition in Zostera marina (Denmark)Limnol Oceanogr 50 1084ndash1095(doi104319lo20055041084)

46 Kaldy JE 2006 Production ecology of thenon-indigenous seagrass dwarf eelgrass (Zosterajaponica Ascher and Graeb) in a Pacific northwestestuary USA Hydrobiologica 553 201ndash217(doi101007s10750-005-5764-z)

47 Abal EG Loneragan N Bowen P Perry CJ Udy JWDennison WC 1994 Physiological andmorphological responses of the seagrass Zosteracapricorni Aschers to light intensity J Exp Mar BiolEcol 178 113ndash129 (doi1010160022-0981(94)90228-3)

48 Cullen-Unsworth LC Unsworth RKF 2013 Seagrassmeadows ecosystem services and sustainabilityEnvironment 55 14ndash27 (doi101080001391572013785864)

49 McGlathery K Marino R Howarth R 1994 Variablerates of phosphate uptake by shallowmarinecarbonate sediments mechanisms and ecologicalsignificance Biogeochemistry 25 127ndash146 (doi101007BF00000882)

50 Pedersen MF Borum J 1993 An annual nitrogenbudget for a seagrass Zostera marina populationMar Ecol Prog Ser 101 169ndash177 (doi103354meps101169)

51 Roman J McCarthy JJ 2010 The whale pumpmarine mammals enhance primary productivity ina coastal basin PLoS ONE 5 e13255 (doi101371journalpone0013255)

52 Ganning B Wulff F 1969 The effects of birddroppings on chemical and biological dynamics inbrackish water rockpools Oikos 20 274ndash286(doi1023073543194)

53 Bakker ES Pagegraves JF Arthur R Alcoverro T In pressAssessing the role of large herbivores in thestructuring and functioning of freshwater andmarine angiosperm ecosystems Ecography(doi101111ecog01651)

54 Collier CJ Waycott M McKenzie LJ 2012 Lightthresholds derived from seagrass loss in the coastalzone of the northern Great Barrier Reef AustraliaEcol Indic 23 211ndash219 (doi101016jecolind201204005)

55 Hauxwell J Cebrian J Furlong C Valiela I 2001Macroalgal canopies contribute to eelgrass (Zosteramarina) decline in temperate estuarine ecosystemsEcology 82 1007ndash1022 (doi1018900012-9658(2001)082[1007MCCTEZ]20CO2)

56 Neckles HA Wetzel RL Orth RJ 1993 Relative effectsof nutrient enrichment and grazing onepiphyte-macrophyte (Zostera marina) dynamicsOecologia 93 285ndash295 (doi101007BF00317683)

57 Udy JW Dennison WC 1997 Growth andphysiological responses of three seagrass species toelevated nutrients in Moreton Bay Australia J ExpMar Biol Ecol 217 253ndash277 (doi101016S0022-0981(97)00060-9)

58 Lee KS Short FT Burdick DM 2004 Development ofa nutrient pollution indicator using the seagrassZostera marina along nutrient gradients in threeNew England estuaries Aquat Bot 78 197ndash216(doi101016jaquabot200309010)

59 Stapel J Aarts TL vanDuynhoven BHM deGroot JDvandenHoogen PHW Hemminga MA 1996 Nutrientuptake by leaves and roots of the seagrass Thalassiahemprichii in the Spermonde ArchipelagoIndonesiaMar Ecol Prog Ser 134 195ndash206(doi103354meps134195)

60 Terrados J Williams S 1997 Leaf versus rootnitrogen uptake by the surfgrass PhyllospadixtorreyiMar Ecol Prog Ser 149 267ndash277(doi103354meps149267)

61 Bulthuis DA Woelkerling WML 1981 Effects of insitu nitrogen and phosphotus enrichment of thesediments on the seagrass Heterozostera tasmanica(Martens ex Aschers) den Hartog in WesternPort Victoria Australia J Exp Mar Biol Ecol 53193ndash207 (doi1010160022-0981(81)90019-8)

62 Hemminga MA Duarte CM 2000 Seagrassecology Cambridge UK Cambridge UniversityPress

63 Dale AL Chesworth JC 2013 Inventory of eelgrassbeds in Hampshire and the Isle of Wight SectionOne report Hampshire and Isle of Wight WildlifeTrust Hampshire

64 Cabaccedilo S et al 2013 Effects of nutrient enrichmenton seagrass population dynamics evidence andsynthesis from the biomassndashdensity relationshipsJ Ecol 101 1552ndash1562 (doi1011111365-274512134)

65 Robinson CLK Yakimishyn J Dearden P 2011Habitat heterogeneity in eelgrass fish assemblagediversity and turnover Aquat Conserv MarFreshw Ecosyst 21 625ndash635 (doi101002aqc1227)

66 Schmidt AL Wysmyk JKC Craig SE Lotze HK 2012Regional-scale effects of eutrophication onecosystem structure and services of seagrass bedsLimnol Oceanogr 57 1389ndash1402 (doi104319lo20125751389)

67 Koch M Bowes G Ross C Zhang X-H 2013 Climatechange and ocean acidification effects onseagrasses and marine macroalgae Glob ChangeBiol 19 103ndash132 (doi101111j1365-2486201202791x)