Supplementary Material – Community ecological monitoring in MPA networks
Paper title: Community ecological modelling as an alternative to physiographic classifications for marine conservation planning.
Authors: Emily Rubidge1,2*, Katie SP Gale1,2, and Janelle Curtis1
1Institute of Oceans Sciences, Fisheries and Oceans Canada, P.O. Box 6000, Sidney, B.C. V8L 4B2
2Pacific Biological Station, Fisheries and Oceans Canada, 3190 Hammond Bay Rd., Nanaimo, BC, V9T 6N7, Canada
*Corresponding author: Emily Rubidge, or
Supplementary Material
Contents
Supplementary Material
1. Biological data collection
2. Justification of species and sites to include
3. List of species included in analysis
4. Correlation analysis
5. PERMANOVA Results Tables
6. Dendrogram cut-off
7. Boxplots of environmental variation of ecological units
8. Indicator species analysis comparison
References
1. Biological data collection
Crab by Trap (Dunham et al 2011, DFO 2015)
For stock assessment purposes, Fisheries and Oceans Canada (DFO) carries out trap surveys in areas with commercial crab fisheries, using commercial-style circular metal Dungeness traps (90 cm diameter, 26 cm high) with closed escape ports. Traps are baited with herring and soaked overnight for approximately 24 hours. Target and non-target species are recorded. Data from several other research programs including an established moult timingmonitoring program in Queen Charlotte Sound, Hecate Strait, and HaidaGwaiiare also included in this dataset.
Tanner Crab(Gillespie et al 2004)
An experimental commercial tanner crab (Chionocetestanneri) fishery ran along the continental slope of British Columbia between 1999-2003, and was accompanied by research trawl and trap surveys. Trawls were carried out using a Campelen 1800 shrimp trawl with rockhopper footgear, and supplementary trap surveys were carried out between 1999-2001 with top-loading conical traps with 40 mm mesh. Data was collected on target and non-target species.
Groundfish
Themajority of the groundfish data used in the cluster analysis came from the Pacific synoptic surveys, which cover the continental shelf of BC (Olsen et al. 2009b). The surveys do not include inlets, enclosed waters, sensitive habitats (e.g., Hecate Strait sponge reefs, Learmonth Bank red tree corals, Rockfish Conservation Areas), or areas otherwise untrawlable(Workman et al. 2008a, Olsen et al. 2009b). The surveys follow a stratified random design, with each of the regions being split into depth and area strata. A proportional random sample of 4 km2 (2km x 2km) grid cells within each strata are selected to be surveyed (Olsen et al. 2009b). Depths between 50 and 500 m are divided into 4 depth strata (Olsen et al. 2009a, Stanley et al. 2007) in all regions except West Coast Queen Charlotte Island, where the depth range is 180–1300 m (Workman et al. 2008b). Tow length is 20 minutes, except for deep Hecate Strait tows (>500 m) which are 40 minutes (Olsen et al. 2009b, Workman et al. 2008a) and for the deepest tows off of west coast of HaidaGwaii (>800 m), which were 40 minutes until 2008, after which time they were 20 minutes. The standard gear is an Atlantic Western IIA box trawl with 5 inch mesh. Catch is sorted by lowest possible taxonomic group and weighed or counted; subsamples are sometimes taken, and most fishes are identified to species, while invertebrates are often recorded to higher taxa. Biological data (size, sex, age) are also collected for species of interest (Workman et al. 2008b, Olsen et al. 2009b). The remaining data in the groundfish dataset came from a variety of research and stock assessment sources, including surveys from the International Pacific Halibut Commission, Pacific Halibut Management Association, and DFO surveys for sablefish, thornyhead, hake, Pacific cod, and rockfish. Most of these non-synoptic surveys were trap and longline surveys, with some trawls. All surveys contained records of target and non-target species.
2. Justification of species and sites to include
We removed species with low frequency in the dataset because they can add noise to multivariate analyses and provide little information in addition to that obtained from more common species (Gauch 1982; McCune and Grace 2002). The legitimacy of removing such species depends upon the objective of the study (Poos and Jackson 2012). Removing species with low frequency in the dataset is in line with our objective to examine broad-scale patterns in biological communities that are somewhat stable over time and that are representative of the diversity of the area. Literature-reported “rules of thumb” for exclusion are species found in less than 5% of sites (McGarigal et al. 2000), 10% of sites (Marchant 1990; McCune and Grace 2002) or even higher (Marchant et al. 1997). To maximize the inclusion of species while also reducing noise and potential biases in the analysis, we chose a conservative exclusion threshold and removed species reported in less than 1% of sites.
3. List of species included in analysis
Table 2.Species included in cluster analysis.Threshold for inclusion was species found in more than 1% of sites (37 sites).Habitats were obtained from FishBase, SeaLifeBase, and other sources and were accurate from those sources as of December 2014.
Phylum / Species / Habitat / Number of sites presentDecapods (crabs, shrimp, 26 species) / Argisalaskensis / Benthic / 51
Argislevior / Benthic / 40
Chionoecetesangulatus / Benthic / 50
Chionoecetestanneri / Benthic / 482
Chorilialongipes / Benthic / 87
Crangonalaskensis / Benthic / 114
Eualusbiunguis / Benthic / 79
Eualusmacrophthalmus / Benthic / 77
Lithodescouesi / Benthic / 272
Lopholithodesforaminatus / Benthic / 42
Metacarcinus magister / Benthic / 192
Munidaquadrispina / Benthic / 126
Neocrangoncommunis / Benthic / 100
Notostomusjaponicus / Benthic / 92
Oregoniagracilis / Benthic / 108
Pandalopsisdispar / Benthic / 610
Pandalus borealis / Benthic / 228
Pandalusdanae / Benthic / 48
Pandalusjordani / Benthic / 768
Pandalusplatyceros / Benthic / 631
Pandalustridens / Benthic / 137
Paracrangonechinata / Benthic / 97
Paralomismultispina / Benthic / 93
Pasiphaeapacifica / Benthic / 209
Pasiphaeatarda / Benthic / 70
Romaleonbranneri / Benthic / 55
Fish (96 species) / Albatrossiapectoralis / Demersal / 474
Ammodyteshexapterus / Benthopelagic / 246
Anarrhichthysocellatus / Demersal / 80
Anoplopoma fimbria / Demersal / 2167
Apristurusbrunneus / Demersal / 109
Atheresthesstomias / Demersal / 2503
Bathyagonusnigripinnis / Demersal / 119
Bathyagonuspentacanthus / Demersal / 258
Bathyrajaabyssicola / Demersal / 45
Bathyrajaaleutica / Demersal / 125
Bathyrajainterrupta / Demersal / 443
Bathyrajaparmifera / Demersal / 46
Bathyrajatrachura / Demersal / 128
Bothrocarabrunneum / Demersal / 143
Careproctusmelanurus / Demersal / 104
Chitonotuspugetensis / Demersal / 141
Citharichthyssordidus / Demersal / 551
Coryphaenoidesacrolepis / Demersal / 462
Coryphaenoidescinereus / Demersal / 157
Coryphaenoidesfilifer / Demersal / 77
Cymatogasteraggregata / Demersal / 115
Embassichthysbathybius / Demersal / 156
Enophrys bison / Demersal / 52
Entosphenustridentatus / Demersal / 62
Eopsettajordani / Demersal / 1164
Eptatretusdeani / Demersal / 72
Gadusmacrocephalus / Demersal / 1720
Glyptocephaluszachirus / Demersal / 1860
Hemilepidotushemilepidotus / Demersal / 38
Hemitripterusbolini / Demersal / 61
Hexagrammosdecagrammus / Demersal / 170
Hippoglossoideselassodon / Demersal / 772
Hippoglossusstenolepis / Demersal / 2076
Hydrolaguscolliei / Demersal / 2347
Icelinusfilamentosus / Demersal / 313
Icelinus tenuis / Demersal / 64
Isopsettaisolepis / Demersal / 174
Lepidopsettabilineata / Demersal / 678
Lumpenussagitta / Benthopelagic / 149
Lycenchelyscrotalinus / Demersal / 59
Lycodesbrevipes / Demersal / 61
Lycodescortezianus / Demersal / 190
Lycodesdiapterus / Demersal / 276
Lycodespacificus / Demersal / 578
Lycodespalearis / Demersal / 68
Lyopsettaexilis / Demersal / 1088
Malacocottuskincaidi / Demersal / 244
Malacocottuszonurus / Demersal / 251
Microgadusproximus / Demersal / 249
Microstomuspacificus / Demersal / 1831
Oncorhynchusketa / Benthopelagic / 75
Oncorhynchustshawytscha / Benthopelagic / 83
Ophiodonelongatus / Demersal / 1644
Parophrysvetulus / Demersal / 1117
Platichthysstellatus / Demersal / 40
Pleuronichthysdecurrens / Demersal / 289
Podothecusaccipenserinus / Demersal / 230
Poroclinusrothrocki / Demersal / 53
Psettichthysmelanostictus / Demersal / 196
Psychrolutesparadoxus / Demersal / 61
Radulinusasprellus / Demersal / 120
Raja binoculata / Demersal / 613
Raja rhina / Demersal / 1645
Ronquilusjordani / Demersal / 145
Sebastesaleutianus / Demersal / 869
Sebastesalutus / Demersal / 1248
Sebastes aurora / Demersal / 84
Sebastesbabcocki / Demersal / 1303
Sebastes borealis / Demersal / 346
Sebastesbrevispinis / Demersal / 1437
Sebastescaurinus / Demersal / 139
Sebastescrameri / Demersal / 298
Sebastesdiploproa / Demersal / 391
Sebasteselongatus / Demersal / 843
Sebastesflavidus / Demersal / 798
Sebasteshelvomaculatus / Demersal / 719
Sebastesjordani / Demersal / 41
Sebastesmaliger / Demersal / 672
Sebastesminiatus / Reef-associated / 45
Sebastes nebulosus / Reef-associated / 96
Sebastesnigrocinctus / Reef-associated / 103
Sebastespaucispinis / Reef-associated / 384
Sebastespinniger / Demersal / 902
Sebastesproriger / Demersal / 664
Sebastesreedi / Demersal / 503
Sebastesruberrimus / Reef-associated / 935
Sebastesvariegatus / Demersal / 162
Sebasteswilsoni / Demersal / 229
Sebasteszacentrus / Demersal / 794
Sebastolobusalascanus / Demersal / 1265
Sebastolobusaltivelis / Demersal / 335
Somniosuspacificus / Benthopelagic / 42
Theragrachalcogramma / Benthopelagic / 1206
Triglopsmacellus / Demersal / 49
Xeneretmuslatifrons / Demersal / 51
Xeneretmusleiops / Demersal / 103
Anthozoans (sea pens, corals, 3 species) / Halipteriswillemoesi / Benthic / 287
Paragorgiaarborea / Benthic / 47
Ptilosarcusgurneyi / Benthic / 88
Asteroids (sea stars, 22 species) / Ceramasterpatagonicus / Benthic / 53
Cheiraster (Luidiaster) dawsoni / Benthic / 140
Crossaster borealis / Benthic / 112
Crossasterpapposus / Benthic / 246
Ctenodiscuscrispatus / Benthic / 235
Dermasteriasimbricata / Benthic / 50
Heterozoniasalternatus / Benthic / 66
Hippasteriacalifornica / Benthic / 86
Hippasteriaphrygiana / Benthic / 183
Lophasterfurcilliger / Benthic / 84
Luidiafoliolata / Benthic / 245
Mediasteraequalis / Benthic / 116
Nearchaster (Nearchaster) aciculosus / Benthic / 56
Orthasteriaskoehleri / Benthic / 102
Pisasterbrevispinus / Benthic / 169
Poraniopsisinflata / Benthic / 91
Pterastertesselatus / Benthic / 176
Pycnopodiahelianthoides / Benthic / 434
Sagenasterevermanni / Benthic / 76
Solasterdawsoni / Benthic / 44
Stylasteriasforreri / Benthic / 400
Thrissacanthiaspenicillatus / Benthic / 38
Echinoids (sea urchins, 2 species) / Strongylocentrotusfragilis / Benthic / 1064
Strongylocentrotuspallidus / Benthic / 123
Holothuroids (sea cucumbers, 7 species) / Apostichopuscalifornicus / Benthic / 138
Apostichopusleukothele / Benthic / 244
Capheiramollis / Benthic / 41
Pannychiamoseleyi / Benthic / 84
Pseudostichopusmollis / Benthic / 176
Psolussquamatus / Benthic / 38
Synallacteschallengeri / Benthic / 43
Ophiuroids (brittle stars, 5 species) / Asteronyxloveni / Benthic / 45
Gorgonocephaluseucnemis / Benthic / 198
Ophiomusiumlymani / Benthic / 39
Ophiurasarsii / Benthic / 149
Stegophiura ponderosa / Benthic / 70
Bivalves (2 species) / Chlamyshastata / Benthic / 75
Chlamysrubida / Benthic / 145
Cephalopods (8 species) / Chiroteuthis calyx / Benthic / 58
Enteroctopusdofleini / Benthic / 83
Galiteuthisphyllura / Benthic / 71
Graneledoneboreopacifica / Benthic / 39
Octopoteuthisdeletron / Benthic / 96
Onykiarobusta / Benthic / 57
Opisthoteuthiscaliforniana / Benthic / 97
Rossiapacifica / Benthic / 475
Gastropods (3 species) / Arminacalifornica / Benthic / 46
Fusitritonoregonensis / Benthic / 501
Tochuinatetraquetra / Benthic / 59
Total (174 species) / 3615
4. Correlation analysis
Table 3. List of environmental data layers compiled and/or generated for use in environmental analysis
Feature / Derivatives / Source / Native Resolution / Number of layersChlorophyll A / Annual and seasonal (3month) minimum, maximum, mean, and range / NASA (AquaMODIS) / 4 km / 20
Sea Surface Temperature (SST) / Seasonal (3month) and 13year (“overall”) composites / NASA (AquaMODIS) / 4 km / 5
Depth / Depth, aspect, slope, rugosity (surface to planar area) / BCMCA/CHS / 100 m / 4
Bottom temperature / Spring, summer, winter, fall, max, min, range / Mike Foreman, DFO (Foreman et al. 2008) / 500 m raster; initial node spacing varies from 100 m nearshore to >50 km offshore / 7
Bottom salinity / Spring, summer, winter, fall, max, min, range / Mike Foreman, DFO (Foreman et al. 2008) / 500 m raster; initial node spacing varies from 100 m nearshore to >50 km offshore) / 7
Flow (non-tidal current velocity) / Summer (overall, north-south, east-west) and winter (overall, north-south, east-west) / Mike Foreman, DFO (Foreman et al. 2008) / 500 m raster; initial node spacing varies from 100 m nearshore to >50 km offshore / 6
Tidal speed and direction / Bottom RMS speed over 29 days of the M2 component / Mike Foreman, DFO (Foreman et al. 2008) / 500 m raster; initial node spacing varies from 100 m nearshore to >50 km offshore / 2
Grain Size / - / Gregr et al. (in press) / 100 m / 1
Adversity / - / Gregr et al. (in press) / 100 m / 1
Disturbance / - / Gregr et al. (in press) / 100 m / 1
Nitrate / - / Bio-ORACLE / 9.2 km / 1
Silicate / - / Bio-ORACLE / 9.2 km / 1
Phosphate / - / Bio-ORACLE / 9.2 km / 1
pH / - / Bio-ORACLE / 9.2 km / 1
Dissolved oxygen / - / Bio-ORACLE / 9.2 km / 1
Total / - / - / - / 59
Of the 59 environmental variables prepared for analysis, many were highly correlated (-0.82 to 0.99, Figure 1). We removed highly correlated variables (R2 >0.7) prior to running the random forest analysis. Variables were retained in the analysis based on two criteria 1) evidence supporting their role in community turnover (as reviewed in Harris 2012b) and 2) whether the data were available for our entire study extent... We retained 12 environmental variables (Figure 2) with correlation values ranging from -0.68 to 0.62.
Figure 2.Correlation analysis of environmental variables
Figure 3.Correlation plot of environmental variables retained for random forest analysis
5. PERMANOVA Results Tables
Table 4a.PERMANOVA analysis of variance table for Ecosections analysis, including 5 Ecosections (Continental Slope, Dixon Entrance,Hecate Strait, Queen Charlotte Sound, and Vancouver Island Shelf).
PERMANOVA Analysis of Variance / Df / Sums Of Sqs / Mean Sq / F.Model / R2 / Pr(>F)Ecosections / 4 / 97.702 / 24.4255 / 166.83 / 0.32279 / 0.001
Residuals / 1400 / 204.976 / 0.1464 / 0.67721
Total / 1404 / 302.677 / 1.0000
Table 4b.PERMDISP analysis of variance table for Ecosections analysis, including 5 Ecosections (Continental Slope, Dixon Entrance,Hecate Strait, Queen Charlotte Sound, and Vancouver Island Shelf).
PERMDISPAnalysis of Variance / Df / SumSq / Mean Sq / F-Value / Pr(>F)
Ecosections / 4 / 0.4601 / 0.115033 / 5.8241 / < 0.001
Residuals / 1400 / 27.6518 / 0.019751
Table 4c PERMANOVA analysis of variance table after removing Continental Slope Ecosection (4 Ecosections:Dixon Entrance,Hecate Strait, Queen Charlotte Sound, and Vancouver Island Shelf).
PERMANOVA Analysis of Variance / Df / Sums Of Sqs / Mean Sq / F.Model / R2 / Pr(>F)Ecosections / 3 / 34.342 / 11.4474 / 80.306 / 0.17703 / 0.001
Residuals / 1120 / 159.652 / 0.1425 / 0.82297
Total / 1123 / 193.994 / 1.0000
Table 5a.PERMANOVA analysis of variance table for community based ecological units.
PERMANOVA Analysis of Variance / Df / Sums Of Sqs / MeanSqs / F.Model / R2 / Pr(>F)Ecological units / 4 / 97.796 / 24.4491 / 218.65 / 0.59716 / 0.001
Residuals / 590 / 65.972 / 0.1118 / 0.40284
Total / 594 / 163.769 / 1.0000
Table 5bPERMDISP analysis of variance table for community based ecological units.
PERMDISPAnalysis of Variance / Df / SumSq / MeanSq / F Value / Pr(>F)
Ecological units / 4 / 0.6172 / 0.154308 / 9.8531 / < 0.001
Residuals / 590 / 9.2400 / 0.015661
6. Dendrogram cut-off
Figure 4. Number of sites in each of the top 8 clusters for dendrograms cut at increasing βsimvalues. The box bounded by dotted lines show cluster distribution at βsimcutoffs of 0.55. Dendrograms cut with a βsimof 1 contain only one cluster, which contains all sites. Lower βsimvalues result in more clusters, but fewer sites retained in the most populated clusters
Figure 5.Variance (standard deviation) of the number of sites in the top 8 clusters compared to the number of sites retained in those clusters for βsim cut-off values of 0.55. Higher βsim cut-off values increase the number of sites included in the top clusters but reduces the evenness of those clusters.
7. Boxplots of environmental variation of ecological units
Figure 6.Boxplots showing environmental differences among ecological units delineated by random forest modelling, for the top three differentiating variables (depth, salinity range, and temperature range).
8. Indicator species analysis comparison
Table 6.Mean and standard deviation of IndValvalues between BCMEC Ecosections and biological clusters
Analysis / Section Name / N / Mean IndVal / SD of IndValEcosections / Continental Slope / 7 / 0.348 / 0.087
Dixon Entrance / 1 / 0.264 / N/A
Queen Charlotte Sound / 7 / 0.304 / 0.059
Overall / 15 / 0.322 / 0.074
Ecological units / Dogfish Bank / 13 / 0398 / 0.116
Other Banks / 4 / 0.319 / 0.074
Shelf / 2 / 0.265 / 0.020
Slope / 7 / 0.500 / 0.199
Troughs / 10 / 0.415 / 0.090
Overall / 36 / 0.406 / 0.133
References
DFO (2015) Pacific Region integrated fisheries management plan: crab by trap, January 1 to December 31, 2015
Dunham JS, Phillips A, Morrison J, Jorgensen G (2011) A manual for Dungeness Crab surveys in British Columbia. Can Tech Rep Fish AquatSci 2964: 68 pp.
Foreman MGG, Crawford WR, Cherniawsky JY, Galbraith J (2008) Dynamic ocean topography for the northeast Pacific and its continental margins. Geophys Res Let 35(22): 5 pp.
Gauch HG(1982) Multivariate analysis in community ecology. Cambridge University Press, Cambridge
Gillespie GE, Fong KH, Phillips AC, Meyer GR, Boutillier JA (2004) Development of a new fishery for Tanner Crabs (ChionoecetestanneriRathbun, 1893) off British Columbia: 2003 Status Report. Can SciAdvis Sec Res Doc 2004/132
Marchant R (1990) Robustness of classification and ordination techniques applied to macroinvertebrate communities from the La Trobe River, Victoria. Australian J Mar Freshwater Res 41: 493–504
Marchant R, Hirst A, Norris RH et al (1997) Classification and ordination of macroinvertebrate assemblages from running waters in Victoria, Australia. J N Am BenthoSoc 16: 664–681
McCune B, Grace J (2002) Analysis of ecological communities. MjM Software Design, Gleneden Beach, Oregon
McGarigal K, Cushman S, and Stafford S (2000) Multivariate Statistics for Wildlife and Ecology Research. Springer-Verlag, New York, NY, USA
Olsen N, Rutherford KL, Stanley RD, Wyeth M(2009a) West Coast Vancouver Island Groundfish Bottom Trawl Survey, May 26th to June 22nd, 2008. Can Man Rep Fish AquatSci 2092: 50 pp.
Olsen N, Rutherford KL, Stanley RD, Wyeth, M (2009b) Hecate Strait Groundfish Bottom Trawl Survey, May 26th to June 21st, 2009. Can Man Rep FishAquatSci 2901: 49 pp.
Poos MS, Jackson DA (2012) Addressing the removal of rare species in multivariate bioassessments: The impact of methodological choices.EcolInd 18: 82-90
Stanley RD, Olsen N, Workman GD, Cleary J, de la Mare W(2007). A review of the Queen Charlotte Sound Groundfish Bottom Trawl Survey (2003-2005). Can Tech Rep Fish AquatSci 2709: 59 pp.
Workman GD, Olsen N, Fargo J, Stanley RD(2008a) West Coast Vancouver Island Groundfish Bottom Trawl Survey, R/V WE RICKER, May 25th to June 10th, 2004. Can Man Rep Fish AquatSci 2826: 64 pp.
Workman GD, Stanley RD, Olsen N, Rutherford KL (2008b) West Coast Queen Charlotte Islands Groundfish Bottom Trawl Survey, September 11th to October 17th, 2007. Can Man Rep Fish AquatSci 2823: 45 pp.
1