Submitted 13 April 2017Accepted 14 July 2017Published 31 July 2017

Corresponding authorAtsuko Fukunagaatsukofukunaganoaagov

Academic editorRichard Taylor

Additional Information andDeclarations can be found onpage 19

DOI 107717peerj3651

Copyright2017 Fukunaga and Kosaki

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Use of multivariate control charts toassess the status of reef fish assemblagesin the Northwestern Hawaiian IslandsAtsuko Fukunaga12 and Randall K Kosaki2

1 Joint Institute for Marine and Atmospheric Research University of Hawailsquoi at Manoa Honolulu HIUnited States of America

2Papahanaumokuakea Marine National Monument National Ocean Service National Oceanic andAtmospheric Administration Honolulu HI United States of America

ABSTRACTA distance-based multivariate control chart is a useful tool for ecological monitoringto detect changes in biological community resulting from natural or anthropogenicdisturbances at permanent monitoring sites It is based on a matrix of any distances ordissimilarities among observations obtained from species composition and abundancedata and bootstrapping techniques are used to set upper confidence bounds that triggeran alarm for further investigationsWe extended the use ofmultivariate control charts tostratified random sampling and analyzed reef fishmonitoring data collected annually onshallow (le30 m) reefs across the Northwestern Hawaiian Islands (NWHI) part of thePapahanaumokuakeaMarineNationalMonument Fish assemblages in theNWHIweremostly stable with exceptions in the south region (NihoaMokumanamana and FrenchFrigate Shoals) in 2012 and 2015 where changes in the assemblage structure exceededthe upper confidence bounds of multivariate control charts However these were dueto changes in relative abundances of native species and potentially related to the smallnumbers of survey sites and relatively low coral covers at the sites particularly in 2015The present study showed that multivariate control charts can be used to evaluate thestatus of biological communities in a very large protected area Future monitoringof fish assemblages in the Papahanaumokuakea Marine National Monument shouldbe accompanied by specific habitat or environmental variables that are related topotential threats to its shallow-water ecosystems This should allow for more detailedinvestigations into potential causes and mechanisms of changes in fish assemblageswhen a multivariate control chart triggers an alarm

Subjects Biodiversity Ecology Marine BiologyKeywords Papahanaumokuakea Marine National Monument Multivariate control chartAssemblage structure Fish Ecological monitoring Stratified random sampling

INTRODUCTIONLong-term ecological monitoring programs generally focus on investigating the currentstatus of an ecosystem identifying trends or detecting changes in an ecosystem caused bynatural or anthropogenic disturbances (Lindenmayer amp Likens 2010) Disturbances can beclassified as short-term lsquolsquopulsersquorsquo and longer-term lsquolsquopressrsquorsquo disturbances although a seriesof pulse disturbances repeated frequently at short intervals with no time for the system

How to cite this article Fukunaga and Kosaki (2017) Use of multivariate control charts to assess the status of reef fish assemblages in theNorthwestern Hawaiian Islands PeerJ 5e3651 DOI 107717peerj3651

to recover is considered a press disturbance (Connell 1997) In a coral reef environmentpulse disturbances include tropical storms (Connell Hughes amp Wallace 1997) bleachingevents (Loya et al 2001) Acanthaster (crown-of-thorns starfish) outbreaks (Kayal et al2012) and sedimentation from dredging (Brown et al 1990) while press disturbancesinclude urbanization and long-term sewage discharges (Hunter et al 1995) and declineof herbivores (Bellwood et al 2004) It is often unknown where or when these natural oranthropogenic disturbances may occur thus requiring scientists and managers to monitorthe ecosystem for an impact resulting from these disturbances

A distance-based multivariate control chart is a useful tool for ecological monitoringto identify impacts quickly at individual sites when they occur (Anderson amp Thompson2004) This method is based on a matrix of any distances or dissimilarities (eg BrayndashCurtis dissimilarity chi-squared distance etc) among observations obtained from speciescomposition and abundance data For each site it calculates the distance between theobservation at time t and either the centroid of the previous (t minus 1) observations inmultivariate space or the centroid of observations at time points designated as a baselineThe distance from the centroid of all previous observations (dt ) is useful when identifyingpulse disturbances while the distance from the baseline centroid (dbt ) is sensitive to changesin species data caused by gradual press disturbances (Anderson amp Thompson 2004) Forboth dt and dbt confidence limits obtained by bootstrapping techniques are used to setupper confidence bounds within which the system is considered to be stable or lsquolsquoincontrolrsquorsquo Multivariate control charts thus trigger an alarm for further investigations whenvalues of dt andor dbt exceed these upper confidence bounds

The Northwestern Hawaiian Islands (NWHI) are located northwest of the inhabitedmain Hawaiian Islands and consist of ten major islands and atolls that span approximately2000 km from Nihoa (2304primeN 16155primeW) at the southeastern end to Kure atoll (2825primeN17820primeW) at the northwestern end (Fig 1) They are part of the PapahanaumokuakeaMarineNationalMonument amarine protected area that was originally established in 2006and expanded in 2016 to encompass 1508870 km2 it is one of the largest conservationareas in the world Despite their remote location and fully protected status shallow-waterreef habitats of the NWHI are still subject to natural andor anthropogenic disturbancessuch as marine debris ship grounding invasion of alien species (eg invertebrates algaeand fish) and increases in sea surface temperature due to climate change and resulting coralbleaching (Office of National Marine Sanctuaries 2009) Monitoring of shallow-water (le30m) coral reef fish has been conducted annually in the NWHI during the Reef Assessmentand Monitoring Program (RAMP) cruise each summer

Here we describe reef fish assemblages of the NWHI using the shallow-watermonitoringdata and demonstrate the use of distance-basedmultivariate control charts to detect changesin the structure of fish assemblage over time The method of multivariate control chart wasoriginally described using fish data collected over permanent transects at various sites ata certain time interval (Anderson amp Thompson 2004) allowing one to investigate naturaltemporal variability in a biological community at each site and to identify observationsoutside of this natural variability However due to the large size of the PapahanaumokuakeaMarineNationalMonument and thewide range of reef habitats it encompasses establishing

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 222

Figure 1 Map of the Hawaiian Archipelago including the Northwestern Hawaiian Islands (NWHI)The boundary of the Papahanaumokuakea Marine National Monument is shown by the dotted line

a sufficient number of permanent transects to assess the status of shallow-water (le30 m)habitats of the entire monument is difficult Islands and atolls being surveyed differ everyyear depending on that yearrsquos cruise schedule so survey sites are randomly selected fromsurvey domains (ie islandsatolls) stratified by reef-zone and depth range This stratifiedrandom sampling has an advantage over permanent transects in that it allows for moreprecise estimates of fish population density and abundance for the entire survey domain(Smith et al 2011) and is commonly applied to fisheries science (Cadima et al 2005) Ourmethod described here uses a distance-based linear model to account for variables that arepart of the stratified sampling design and obtains residuals which are in turn used formultivariate control charts This method expands multivariate control charts to ecologicalmonitoring programs where establishing permanent transects is not possible or practical

MATERIALS AND METHODSSurvey designMonitoring of reef fish was annually performed using a stationary point count (SPC)method (Williams et al 2015) a modification of the stationary visual census technique(Bohnsack amp Bannerot 1986) from 2007 to 2016 except for 2008 (surveyors used belttransects) and 2013 Surveyed islandsatolls of the NWHI (Fig 1) were Nihoa (2304primeN16155primeW) Mokumanamana (2334primeN 16442primeW) French Frigate Shoals (FFS 2352primeN16617primeW) Gardner Pinnacles (Gardner 2501primeN 16759primeW) Maro Reef (Maro 2525primeN17035primeW) Laysan Island (Laysan 2542primeN 17144primeW) Lisianski Island (Lisianski2604primeN 17358primeW) Pearl and Hermes Atoll (PHA 2756primeN 17544primeW) Midway Atoll(Midway 2812primeN 17721primeW) and Kure Atoll (Kure 2825primeN 17820primeW) Survey sites ateach island or atoll were chosen each year by generating random points (coordinates) on

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 322

Table 1 Stratification scheme for SPC fish surveys at each islandatoll

Depth (m) Locations

0ndash6 6ndash18 18ndash30

Forereef FRF-S FRF-M FRF-D All islands and atollsLagoon LAG-S LAG-M LAG-D FFS Maro PHA Midway and KureBackreef BKR-S ndash ndash FFS Maro PHA Midway and Kure

GIS bathymetric and bottom composition maps in which target hard-bottom habitats formonitoring at depths le30 m were stratified by reef zone (backreef forereef and lagoon)and depth range (shallow 0ndash6 m mid 6ndash18 m and deep 18ndash30 m) Only one site eachwas surveyed from the backreef-mid and backreef-deep strata in the entire surveys sothese sites were omitted from analysis resulting in seven strata in this study (Table 1)backreef-shallow (BKR-S) forereef-shallow (FRF-S) forereef-mid (FRF-M) forereef-deep(FRF-D) lagoon-shallow (LAG-S) lagoon-mid (LAG-M) and lagoon-deep (LAG-D)Nihoa Mokumanamana Gardner Laysan and Lisianski do not have lagoon and backreefhabitats so shallow reefs of these locations were categorized based on depth into FRF-SFRF-M or FRF-D

At each site one or two pairs of divers simultaneously performed a reef fish surveyBriefly divers laid a 30-m transect line along a depth contour and counted fish insideadjacent SPC cylinders 15 m in diameter For the first five minutes (species enumerationperiod) divers recorded all species observed in their cylinders and during a followingtallying period they systematically worked through their lists to record the number ofindividuals for each species If a species was observed during the enumeration periodbut absent when recording the number of individuals during the tallying period diversrecorded the number present in their cylinders when it was first observed during theenumeration period Divers remained stationary at the centers of their cylinders to avoiddisturbance as much as possible but at the end of the tallying period they swam throughtheir plots to count small semi-cryptic species Prior to participating in SPC surveyseach year all surveyors went through rigorous training for fish identification sizing andcount Observer biases were also assessed on a daily basis throughout the cruises forquality assurance and control of survey data In addition starting 2010 divers performedrapid visual assessments of benthic cover by recording percent cover of major functionalcategories (eg hard corals macroalgae and crustose coralline algae) and characterizationof survey sites into reef types aggregate reef aggregate patch reef aggregate patch reefsmixed habitat pavement pavement with patch reefs pavement with sand channels rockboulder reef rubble spur and groove and scattered coralrock

Field work in the NWHI was conducted under Papahanaumokuakea Marine NationalMonument research permits PMNM-2007-048 PMNM-2009-058 PMNM-2010-052PMNM-2011-022 PMNM-2012-034 PMNM-2014-018 PMNM-2015-012 and PMNM-2016-023

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 422

Data analysesMultivariate analyses of reef fish assemblages were done using the software packagePRIMER 7 (Clarke amp Gorley 2015) with the PERMANOVA+ add-on (Anderson Gorleyamp Clarke 2008) Fish counts were first averaged to obtain fish abundance per survey sitePelagic and semi-pelagic fishes including sharks rays tunas jacks sardines anchoviesand herrings were excluded in order to focus on non-transient resident reef fish at eachsite as the survey method was designed specifically for reef fish monitoring Less-commonspecies that did not account for at least 10 of the total fish abundance at any of the surveysites were also excluded in order to reduce noise in data analyses The structure of fishassemblages as a whole was computed using the BrayndashCurtis measure after square-roottransformation of the abundance data

In order to characterize fish assemblages in the NWHI we used similarity percentageanalysis (SIMPER Clarke amp Warwick 2001) to determine fish species that typified eachstratum across all years The BrayndashCurtis similarities calculated among observationswithin each stratum were broken down into contribution from each species Similaritycontributions of each specieswere then averaged A fish specieswith a high average similaritycontribution and a low standard deviation indicates that the species is found at a high andconsistent abundance and therefore it typifies that particular stratum Trophic habits of thetop eight fishes that typified each stratumwere then determined based on various databasesand references including FishBase (ver 062016 httpwwwfishbaseorg) the NOAAPacific Islands Fisheries Science Centerrsquos database (wwwpifscnoaagovcredfishphp)Randall (2007) Hiatt amp Strasburg (1960) and Hobson (1974)

Formal tests to examine relationships between the structure of fish assemblages onthe basis of the BrayndashCurtis dissimilarities and explanatory variables were done using adistance-based linear model (DISTLM Legendre amp Anderson 1999 McArdle amp Anderson2001) with 4999 permutations of residuals under a reduced model Due to differences inlatitude and thus water temperature in winter reef fish assemblages in theHawaiian Islandschange along the island chain (Grigg et al 2008) It is therefore important to account forpotential effects of these geographical location as well as strata when examining changes infish assemblages over time Specifically we used three sets of explanatory variables locationstratum and year in the order of these variables being added to the model and performedDISTLM conditional tests to examine the amount of additional variability explained byan explanatory variable after fitting one or more other explanatory variables consideringoverlap in the variability explained by multiple explanatory variables The rationale forthis analysis was to examine the effects of habitat (ie stratum) on the structure of fishassemblages after large-scale spatial variability (ie geographical locations within theNWHI) was taken into account and then the presence of temporal variability (year-to-yeardifferences) after large-scale spatial variability and the effects of strata were considered Thelocation variable was obtained by performing principal component analysis of latitudinaland longitudinal GPS coordinates of each survey site The first principal component (PC1)axis explained 995 of the variance in the original coordinates so centered PC1 scoreswere used as a single location variable The stratum and year variables were both categoricalvariables with seven and eight levels respectively

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 522

Multivariate control chartsFor multivariate control charts we computed the assemblage structure on the basis ofthe BrayndashCurtis dissimilarity using the same square-root transformed abundance dataafter removing pelagic and semi-pelagic fishes and less-common species Removingless-common species could eliminate the possibility of capturing a species that was new tothe NWHI in multivariate control charts unless it occurred in a relatively high number atany survey sites but new species would have been readily noted by divers during surveysor while entering data into a database Note that the process of removing less-commonspecies is not required for construction of multivariate control charts so these speciesmay be retained if changes in their abundances are of interest Similarly as previouslystated multivariate control charts can be constructed using a matrix of any distances ordissimilarities (ie not limited to the BrayndashCurtis dissimilarity) so an appropriate measureof ecological structure should be chosen based on the nature of each study

We used regional centroids in multivariate space to calculate dt and dbt by dividing theNWHI into three regions based on their latitudes south (23N Nihoa Mokumanamanaand FFS) mid (25ndash26N Gardner Maro Lisianski and Laysan) and north (27ndash28NPHA Midway and Kure) Rationale for this use of regional (rather than islandatoll orthe entire NWHI) centroids was that at the islandatoll level there were many missingdata as none of the islandsatolls were surveyed every year and at an archipelagic scaleit could be difficult to detect any impacts that occurred in more or less localized areaThus this grouping was for the purpose of detecting impacts at a spatial scale smaller thanthe entire NWHI while having consistent data points for each time period and effects oflatitudinal and longitudinal differences on the structure of fish assemblages among surveysites were explicitly taken into account prior to construction of control charts (see below)Examples of potential localized effects specific to the NWHI include passing storms severelocalized bleaching (see Couch et al 2016) and introduced fishes particularly the snappersLutjanus kasmira and Lutjanus fulvus and the grouper Cephalopholis argus which were allintentionally introduced to the main Hawaiian Island of Olsquoahu in the 1950rsquos and becameestablished (see Randall 1987)

Multivariate control charts were constructed using the statistical software R version331 (R Core Team 2016) In order to account for the effects of geographical location andstrata on the structure of fish assemblages we used DISTLM with the two explanatoryvariables location (ie centered PC1 scores described above) and stratum and obtainedresiduals and these residuals are in turn used for multivariate control charts SpecificallyGowerrsquos centered matrix (G) was first calculated from the BrayndashCurtis dissimilarity matrixamong survey sites The hat matrix (H) was also calculated as H=X(XprimeX)minus1Xprime in whichX was a model matrix of the explanatory variables (ie location and stratum) constructedusing the function modelmatrix() in the stats package Then the matrix of fitted values isHGH while that of residuals is (IminusH)G(IminusH) in which I is the identity matrix (McArdleamp Anderson 2001) We performed eigenvalue decomposition on (IminusH)G(IminusH) insteadof on G as in Anderson amp Thompson (2004) and scaled eigenvectors by square-root ofrespective eigenvalues in order to obtain principal coordinates of the residuals for eachsite Locations of regional centroids in multivariate space were obtained by averaging the

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 622

Table 2 Numbers of sites surveyed per stratum per year Strata are backreef-shallow (BKR-S) forereef-shallow (FRF-S) forereef-mid (FRF-M) forereef-deep (FRF-D) lagoon-shallow (LAG-S) lagoon-mid(LAG-M) and lagoon-deep (LAG-D)

2007 2009 2010 2011 2012 2014 2015 2016

BKR-S 9 10 10 5 4 7 4 0FRF-S 9 21 14 26 5 15 18 22FRF-M 29 73 43 41 34 42 41 89FRF-D 16 43 19 42 20 25 33 60LAG-S 22 27 13 20 10 0 0 6LAG-M 27 4 15 4 15 0 0 5LAG-D 5 1 4 3 3 0 0 0

principal coordinates per year per region For each region Euclidean distances amongcentroids of different survey years were calculated from the averaged principal coordinateswhich in turn were used to calculate dt and dbt (see Eqs (4)ndash(7) in Anderson amp Thompson2004 for details) For calculation of dbt the first two years of data (2007 and 2009) weredesignated as a baseline in order to capture any gradual shifts in the assemblage structuresince the establishment of Papahanaumokuakea Marine National Monument in 2006Bootstrapping procedure followed the method described in Anderson amp Thompson (2004)but treating the regional centroids as a population of sites The procedure was repeated1000 times in order to obtain 50 and 95 confidence bounds for both dt and dbt If anyregions had dt andor dbt above the 95 confidence bounds at any time points SIMPERanalyses were done in order to identify which species typified each stratum in thoseparticular years and compare them to those that typified each stratum in all other years

RESULTSIn total 268 species of reef fish were identified at 1013 sites from the seven strata of theNWHI (Table 2)Many of these species were represented only by a few individuals at a smallnumber of sites and removing less-common species resulted in 84 species for data analysesAverage BrayndashCurtis similarities of the structure of fish assemblages within each stratumgroupmostly ranged from 40 to 50 (possible range 0ndash100) although the stratum FRF-D hada lower average similarity of 358 (Table 3) The top eight species of fish that typified eachstratumaccounted for approximately 50ormore of thewithin-group similarities andweremostly herbivores and invertivores (Table 3) Overall the invertivore Thalassoma duperreywas the most consistently abundant species in the NWHI having the highest averagesimilarity contribution among all reef fish species in all but LAG-D stratum T duperreyalso had the highest mean abundance in all of the strata within shallow or mid depths

The top eight typifying species were the same between the two shallow strata LAG-Sand BKR-S and similar among all three shallow strata with invertivores T duperreyThalassoma ballieui and Stethojulis balteata and herbivores Acanthurus triostegus andStegastes marginatus being consistently present in high abundanceWithin forereef habitatthere was a gradual species turnover with increasing depths with invertivores Parupeneusmultifasciatus and Bodianus albotaeniatus and an herbivore Ctenochaetus strigosus being

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 722

Table 3 Top eight fish species that typified each stratum based on SIMPER analysis on the BrayndashCurtis similarity calculated from square-root transformed abundance The average similarity of eachstratum is in parentheses

Species AvAbund AvSim SimSD Contrib

BKR-S (500)Thalassoma duperrey (I) 360 1054 354 2108Stethojulis balteata (I) 252 620 195 1240Acanthurus triostegus (H) 244 547 160 1094Stegastes marginatus (H) 261 530 136 1059Coris venusta (I) 182 419 111 838Thalassoma ballieui (I) 107 200 096 400Macropharyngodon geoffroy (I) 109 198 102 396Parupeneus multifasciatus (I) 081 163 119 326FRF-S (414)Thalassoma duperrey (I) 471 936 285 2258Stegastes marginatus (H) 253 384 138 927Acanthurus triostegus (H) 257 370 137 894Thalassoma ballieui (I) 119 209 152 503Acanthurus nigroris (H) 152 201 099 486Chromis vanderbilti (P) 262 189 053 455Chlorurus perspicillatus (H) 121 180 108 434Stethojulis balteata (I) 139 178 076 430FRF-M (444)Thalassoma duperrey (I) 371 707 307 1594Stegastes marginatus (H) 213 294 106 663Acanthurus nigroris (H) 156 215 119 485Parupeneus multifasciatus (I) 127 213 170 481Chromis vanderbilti (P) 236 213 063 481Ctenochaetus strigosus (H) 158 189 089 426Bodianus albotaeniatus (I) 100 186 206 420Acanthurus triostegus (H) 150 171 097 386FRF-D (358)Thalassoma duperrey (I) 207 363 135 1012Centropyge potteri (H) 202 336 133 938Chromis hanui (P) 272 312 088 870Parupeneus multifasciatus (I) 123 253 162 706Ctenochaetus strigosus (H) 167 231 084 644Bodianus albotaeniatus (I) 096 183 130 512Chromis ovalis (P) 183 147 053 410Acanthurus nigroris (H) 112 145 081 405LAG-S (429)Thalassoma duperrey (I) 355 879 274 2052Stethojulis balteata (I) 195 438 159 1021Stegastes marginatus (H) 218 377 127 879

(continued on next page)

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 822

Table 3 (continued)

Species AvAbund AvSim SimSD Contrib

Coris venusta (I) 160 348 085 813Acanthurus triostegus (H) 192 294 111 687Thalassoma ballieui (I) 098 159 097 372Parupeneus multifasciatus (I) 090 154 097 360Macropharyngodon geoffroy (I) 087 151 080 351LAG-M (405)Thalassoma duperrey (I) 250 589 290 1455Chlorurus sordidus (H) 200 345 116 852Scarus dubius (H) 208 343 092 848Ctenochaetus strigosus (H) 192 258 082 637Parupeneus multifasciatus (I) 101 221 190 545Bodianus albotaeniatus (I) 087 181 150 447Chaetodon miliaris (P) 087 141 085 348Stegastes marginatus (H) 120 140 061 346LAG-D (426)Ctenochaetus strigosus (H) 332 634 128 1488Chlorurus sordidus (H) 196 349 105 818Scarus dubius (H) 193 314 092 736Chromis hanui (P) 192 300 116 703Centropyge potteri (H) 154 291 148 684Parupeneus multifasciatus (I) 112 274 202 643Thalassoma duperrey (I) 159 274 124 643Bodianus albotaeniatus (I) 099 239 162 562

NotesTrophic categories are in parentheses H herbivore I invertivore P planktivore The heading of each column showslsquolsquoAvAbundrsquorsquo average square-root transformed abundance lsquolsquoAvSimrsquorsquo average similarity contribution lsquolsquoSimSDrsquorsquo ratioof the average similarity contribution to the standard deviation of similarity contribution lsquolsquoContribrsquorsquo percentage of thecontribution by the species to the within-group similarity

consistently abundant in the stratum FRF-M and planktivoresChromis hanui andChromisovalis and an herbivore Centropyge potteri in the stratum FRF-D Species turnover wasrelatively similar in lagoon with C strigosus B albotaeniatus C hanui and C potteribecoming more abundant with increasing depths but two species of parrotfish Chlorurussordidus and Scarus dubius were also consistently abundant at depths ge6 m (LAG-M andLAG-D)

DISTLM conditional tests showed that the location variable explained a significantproportion (47) of the variation in the structure of fish assemblages as a whole(P = 00002 Table 4) The stratum variable also explained significant and the largestproportion of the variation in the assemblage structure with an additional 138(P = 00002 Table 4) Together these two variables explained 185 of the variationin the assemblage structure The year variable added only 33 to the explained variationafter the location and stratum variables were fitted However this was also statisticallysignificant (P = 00002 Table 4) confirming the presence of temporal variability in thestructure of fish assemblages after the effects of geographical locations and strata weretaken into account

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 922

Table 4 Results of DISTLM conditional tests for the structure of fish assemblages based on the BrayndashCurtis measure Explanatory variables used were location stratum and year

Explained SS Pseudo-F P Proportion R2

Location 100230 5030 00002 0047 0047+ Stratum 291880 2838 00002 0138 0185+ Year 70038 604 00002 0033 0218

Table 5 Numbers of sites surveyed per islandatoll per year

2007 2009 2010 2011 2012 2014 2015 2016

Nihoa 0 0 0 8 0 0 0 0Mokumanamana 0 13 0 8 0 0 0 0SouthFFS 57 0 27 8 15 27 8 47Gardner 0 0 0 12 0 0 0 0Maro 0 39 0 25 0 0 17 0Laysan 0 14 0 23 0 0 8 0

Mid

Lisianski 8 19 25 9 25 28 18 40PHA 52 0 41 18 31 0 23 56Midway 0 51 0 30 0 34 14 0NorthKure 0 43 25 0 20 0 8 39

The grouping of islandsatolls into the three regions south mid and north resulted inconsistent data points for each region for each time period although the total numbers ofsurvey sites were still relatively low in some years particularly in the south region (Table 5)Multivariate control charts showed both dt and dbt above the 95 confidence boundsin the south region in 2012 and 2015 (Fig 2) A high value of dt was also observed inthe south region in 2009 but this is likely an artifact as the dt in 2009 was simply thedistance between the centroids of observations in 2007 and those in 2009 (see Anderson ampThompson 2004 for details) In the south region in 2012 identities of the top eight speciesthat typified sites from two lagoon strata LAG-M (seven sites) and LAG-D (three sites)were relatively similar to those that typified these strata in other years (Tables 6 and 7) Onthe other hand two forereef strata FRF-M (three sites) and FRF-D (two sites) had moredifferent assemblages of typifying species compared to those in other years (Tables 6 and 7)Specifically two invertivores Pseudojuloides cerasinus andOxycheilinus bimaculatus whichwere not among the top eight typifying species for the entire NWHI (Table 3) or in allother years for the south region (Table 6) were consistently present in high abundance inthe stratum FRF-M in 2012 Similarly two butterflyfish Chaetodon fremblii and Chaetodontrifascialis and an herbivore Zebrasoma flavescenswere consistently abundant in the stratumFRF-D in 2012 (Table 7) but these species were never among the top eight typifying speciesfor the entire NWHI (Table 3) or in all other years for the south region (Table 6) Therewere no surveys from the shallow strata (ie BKR-S FRF-S and LAG-S) in the south regionin 2012

In the south region in 2015 fishes that typified sites from the forereef strata FRF-S(two sites) FRF-M (three sites) and FRF-D (three sites) were very different from species

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1022

(a)

0

10

20

30

2009 2010 2011 2012 2013 2014 2015 2016

Year

d t

North

Mid

South

(b)

0

10

20

30

2009 2010 2011 2012 2013 2014 2015 2016

Year

dbt North

Mid

South

Figure 2 Distance-based multivariate control charts dt (a) and dbt (b) for the south mid and north re-

gions of the NWHI constructed based on DISTLM residuals (IminusH)G(IminusH) The dashed line indicatesthe 95th percentile and the dotted line indicates the 50th percentile both estimated from 1000 bootstrapsamples For calculation of db

t the first two years of data (2007 and 2009) were designated as a baseline

that typified these strata in other years in terms of both species identities and abundances(Tables 6 and 8) While the identities of the top eight typifying species were more similarbetween 2015 and other years at depths lt6 m (FRF-S) than at depths ge6 m abundance ofa planktivore Chromis vanderbilti was on average higher for 2015 than other years withmean square-root transformed abundances of 623 and 499 respectively (Tables 6 and 8)Increased abundances of small-bodied planktivores were also observed at 6ndash18 m depths(FRF-M) with C vanderbilti C hanui and Chromis acares being the most numericallyabundant species (Table 8) At 18ndash30 m depths (FRF-D) C hanui was consistentlyabundant in 2015 but themean square-root transformed abundance was 094 and relativelylow (Table 8) this fish had the mean square-root transformed abundance of 232 and wasthe most numerically abundant species in this stratum in other years (Table 6) In additionsimilar to the stratum FRF-M in 2012 two invertivores P cerasinus and O bimaculatuswere consistently present in high abundance in the stratum FRF-D in 2015 There wereno surveys from either the backreef or lagoon stratum in the south region in 2015

DISCUSSIONThe structure of fish assemblages on shallow (le30 m) reefs of the NWHI was affectedby both geographical location and habitat types Despite the long distance of sim2000 kmfrom the southeastern to the northwestern end of the NWHI and the latitudinal gradientthat affects seawater temperature and in turn the structure of fish assemblages (Mundy2005) the variation in the assemblage structure explained by these two variables wasmuch larger for the habitat (ie stratum) variable (Table 4) Changes in the structureof fish assemblages along depth were also evident for both forereef and lagoon habitatswhen species that typified the strata across the NWHI were identified in SIMPER analysis

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1122

Table 6 Top eight fish species that typified each stratum in the south region based on SIMPER analysison the BrayndashCurtis similarity calculated from square-root transformed abundance excluding observa-tions from 2012 and 2015 The average similarity of each stratum is in parentheses

Species AvAbund AvSim SimSD Contrib

FRF-S (3626)Thalassoma duperrey (I) 360 696 223 1919Chromis vanderbilti (P) 499 544 108 1501Acanthurus nigrofuscus (H) 299 395 119 1089Plectroglyphidodon imparipennis (I) 151 218 116 601Acanthurus triostegus (H) 173 186 090 513Stegastes marginatus (H) 117 160 115 442Coris venusta (I) 114 137 033 378Acanthurus olivaceus (H) 125 100 050 276FRF-M (4324)Thalassoma duperrey (I) 282 629 278 1455Chromis vanderbilti (P) 344 445 087 1028Acanthurus nigrofuscus (H) 189 285 117 658Parupeneus multifasciatus (I) 124 255 215 590Bodianus bilunulatus (I) 092 187 192 433Acanthurus olivaceus (H) 129 185 094 429Paracirrhites arcatus (I) 118 162 080 375Macropharyngodon geoffroy (I) 104 160 101 369FRF-D (3460)Thalassoma duperrey (I) 162 293 142 847Centropyge potteri (H) 171 278 113 803Chromis hanui (P) 232 269 086 777Parupeneus multifasciatus (I) 118 264 163 762Acanthurus nigrofuscus (H) 161 232 109 671Ctenochaetus strigosus (H) 182 213 072 616Sufflamen bursa (I) 089 177 112 511Bodianus bilunulatus (I) 086 172 127 497LAG-M (4747)Ctenochaetus strigosus (H) 294 577 158 1217Thalassoma duperrey (I) 202 492 425 1036Chlorurus sordidus (H) 232 464 192 978Scarus dubius (H) 226 413 118 869Parupeneus multifasciatus (I) 122 306 415 644Bodianus bilunulatus (I) 102 242 285 510Acanthurus nigrofuscus (H) 135 203 104 429Centropyge potteri (H) 102 159 098 335LAG-D (3904)Ctenochaetus strigosus (H) 314 579 106 1482Chlorurus sordidus (H) 181 333 104 852Parupeneus multifasciatus (I) 095 268 173 687

(continued on next page)

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1222

Table 6 (continued)

Species AvAbund AvSim SimSD Contrib

Scarus dubius (H) 173 261 072 669Centropyge potteri (H) 121 250 126 640Chromis hanui (P) 135 242 096 619Bodianus bilunulatus (I) 088 230 136 589Thalassoma duperrey (I) 117 221 102 566

NotesTrophic categories are in parentheses after species names H herbivore I invertivore P planktivore The heading of each col-umn shows lsquolsquoAvAbundrsquorsquo average square-root transformed abundance lsquolsquoAvSimrsquorsquo average similarity contribution lsquolsquoSimSDrsquorsquoratio of the average similarity contribution to the standard deviation of similarity contribution lsquolsquoContribrsquorsquo percentage of thecontribution by the species to the within-group similarity

(Table 3) The effect of habitat types on coral reef fish assemblages potentially resultingfrom differences in water movement energy have been previously described (McGehee1994) Differences in fish assemblages between barrier and patch reef habitats have alsobeen reported from FFS and Midway in the NWHI (DeMartini Parrish amp Parrish 1996)The present study confirms the importance of including habitat types and depth in thesurvey design of fish monitoring in the NWHI and accounting for their effects whenanalyzing the data

On shallow-water reefs of the NWHI wrasses particularly Thalassoma duperrey werethe most consistently abundant taxon they accounted for all invertivores that typified eachstratum with an exception of the goatfish Parupeneus multifasciatus (Table 3) In lagoonhabitat six out of the top eight typifying species were wrasses at 0ndash6 m depths but theybecame less abundant at deeper depths being replaced by P multifasciatus and herbivoroussurgeonfish and parrotfish at 6ndash18 m and then by Centropyge potteri and Chromis hanuiat 18ndash30 m depths In forereef habitat herbivorous surgeonfish and damselfish wereconsistently abundant in addition to invertivores at 0ndash18 m depths but similar to lagoonhabitatC potteri andC hanui also became abundant at 18ndash30m depths Overall numericalabundance of herbivores and invertivores on shallow reefs is consistent with the resultsof a previous study in the NWHI that investigated changes in the trophic structure offish assemblages from euphotic (1ndash30 m) to mesophotic (gt30 m) depths (Fukunaga et al2016) In that study C potteri and C hanui were two of the top three species that typified27ndash40 m depths Therefore these two species seem to characterize lower euphotic andupper mesophotic depths in the NWHI

Herbivory plays an important role in shaping benthic communities on shallow-watercoral reefs by regulating algal biomass (McCook 1999) In the main Hawaiian Islandof Olsquoahu the territorial damselfish Stegastes marginatus (previously referred to asS fasciolatus see Randall 2007) has been documented to defend a small patch of thebottom (sim1 m2) and to graze algal mats inside while parrotfish and surgeonfish grazingall erect algae outside the territory (Hixon amp Brostoff 1996) In the present study thedamselfish S marginatus was consistently abundant in all the strata within 0ndash18 m depth(Table 3) Herbivores that typified these strata also included surgeonfish (Acanthurustriostegus Acanthurus nigroris Ctenochaetus strigosus) parrotfish (Chlorurus perspicillatusChlorurus sordidus Scarus dubius) or both (Table 3) Thus the same grazing pattern could

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1322

Table 7 Top eight fish species that typified each stratum in the south region in 2012 based on SIMPERanalysis on the BrayndashCurtis similarity calculated from square-root transformed abundance The aver-age similarity of each stratum is in parentheses There were no surveys in the strata FRF-S and BKR-S inthe south region in 2012

Species AvAbund AvSim SimSD Contrib

FRF-M (3470)Thalassoma duperrey (I) 192 434 548 1251Parupeneus multifasciatus (I) 172 424 200 1223Chromis vanderbilti (P) 235 408 398 1174Chromis hanui (P) 231 260 491 749Coris venusta (I) 144 239 058 689Pseudojuloides cerasinus (I) 120 223 058 642Oxycheilinus bimaculatus (I) 126 214 058 617Bodianus bilunulatus (I) 076 204 398 587FRF-D (4097)Chromis hanui (P) 263 519 ndash 1267Bodianus bilunulatus (I) 141 424 ndash 1035Chaetodon fremblii (I) 150 424 ndash 1035Chlorurus sordidus (H) 271 424 ndash 1035Zebrasoma flavescens (H) 193 424 ndash 1035Ctenochaetus strigosus (H) 264 367 ndash 896Thalassoma duperrey (I) 148 367 ndash 896Chaetodon trifascialis (C) 111 300 ndash 732LAG-M (5320)Ctenochaetus strigosus (H) 345 540 126 1015Chlorurus sordidus (H) 247 535 251 1005Thalassoma duperrey (I) 217 458 401 860Scarus dubius (H) 171 304 139 572Acanthurus nigrofuscus (H) 149 278 199 523Bodianus bilunulatus (I) 114 277 580 520Chaetodon trifascialis (C) 130 271 232 510Chaetodon fremblii (I) 110 269 584 506LAG-D (6536)Ctenochaetus strigosus (H) 424 908 730 1390Chlorurus sordidus (H) 320 760 1321 1163Thalassoma duperrey (I) 216 538 1515 823Chromis hanui (P) 284 503 217 770Scarus dubius (H) 272 485 457 742Centropyge potteri (H) 205 452 910 691Zebrasoma flavescens (H) 226 444 807 679Parupeneus multifasciatus (I) 160 340 634 521

NotesTrophic categories are in parentheses after species names H herbivore I invertivore P planktivore C Corallivore The head-ing of each column shows lsquolsquoAvAbundrsquorsquo average square-root transformed abundance lsquolsquoAvSimrsquorsquo average similarity contribu-tion lsquolsquoSimSDrsquorsquo ratio of the average similarity contribution to the standard deviation of similarity contribution lsquolsquoContribrsquorsquopercentage of the contribution by the species to the within-group similarity

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1422

Table 8 Top eight fish species that typified each stratum in the south region in 2015 based on SIMPERanalysis on the BrayndashCurtis similarity calculated from square-root transformed abundance The aver-age similarity of each stratum is in parentheses There were no surveys in the stratum BKR-S and lagoonhabitat in the south region in 2015

Species AvAbund AvSim SimSD Contrib

FRF-S (5289)Chromis vanderbilti (P) 632 998 ndash 1887Thalassoma duperrey (I) 468 727 ndash 1375Paracirrhites arcatus (I) 268 529 ndash 1001Acanthurus nigrofuscus (H) 216 467 ndash 882Acanthurus olivaceus (H) 187 432 ndash 817Plectroglyphidodon johnstonianus (C) 187 432 ndash 817Acanthurus triostegus (H) 111 249 ndash 472Macropharyngodon geoffroy (I) 156 249 ndash 472FRF-M (3642)Chromis vanderbilti (P) 563 813 301 2233Chromis hanui (P) 524 408 165 1120Chromis acares (P) 567 391 058 1073Macropharyngodon geoffroy (I) 301 370 377 1017Parupeneus multifasciatus (I) 198 333 597 913Sufflamen bursa (I) 182 324 380 890Thalassoma duperrey (I) 237 161 058 441Acanthurus triostegus (H) 162 091 058 249FRF-D (3275)Oxycheilinus bimaculatus (I) 365 960 229 2931Parupeneus multifasciatus (I) 198 433 179 1322Sufflamen fraenatum (I) 100 383 691 1168Pseudojuloides cerasinus (I) 151 363 299 1110Centropyge fisheri (H) 216 249 058 761Acanthurus olivaceus (H) 202 213 058 649Chaetodon kleinii (P) 177 213 058 649Chromis hanui (P) 094 150 058 459

NotesTrophic categories are in parentheses after species names H herbivore I invertivore P planktivore C Corallivore The head-ing of each column shows lsquolsquoAvAbundrsquorsquo average square-root transformed abundance lsquolsquoAvSimrsquorsquo average similarity contribu-tion lsquolsquoSimSDrsquorsquo ratio of the average similarity contribution to the standard deviation of similarity contribution lsquolsquoContribrsquorsquopercentage of the contribution by the species to the within-group similarity

potentially exist in the NWHI at lt18 m depths where S marginatus surgeonfish andparrotfish were consistently abundant At greater depths however S marginatus was notnumerically abundant in either the forereef habitat or lagoon herbivores were numericallydominated by the angelfish C potteri and parrotfish in the FRF-D and LAG-D stratarespectively

While the presence of temporal variability in the structure of reef fish assemblages canbe statistically tested through multivariate analyses such as analysis of similarity (Clarke ampWarwick 2001) permutational multivariate analysis of variance (Anderson 2001 McArdleamp Anderson 2001) and DISTLM multivariate control charts pinpoint when a change in

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1522

the assemblage structure is more than what would be expected from natural temporalvariability (Anderson amp Thompson 2004) In the present study dt and dbt in the southregion in 2012 and 2015 were both above the control chartsrsquo 95 confidence bounds(Fig 2) Examining species that typified each stratum in these years and comparing themwith those in other years using SIMPER analyses revealed how fish assemblages differed inthese two years the increases in dt and dbt in both 2012 and 2015 were due to changes inrelative abundances of fish species The species identified to typify each stratum in thesetwo years included Oxycheilinus bimaculatus Chaetodon fremblii and Zebrasoma flavescensthat have previously been reported to be relatively common in shallow reef habitats inthe NWHI (DeMartini Parrish amp Parrish 1996 Parrish amp Boland 2004) Changes in thestructure of fish assemblages in 2015 were also partly due to increases in abundances ofsmall-bodied planktivores (Chromis vanderbilti Chromis acares and C hanui) in forereefhabitat at 0ndash18 m depths and a decrease in abundance of C hanui at 18ndash30 m depthsMore importantly the typifying species of these two years did not include any introducedspecies thus there was no evidence of introduced species displacing native fishes

There are some important limitations on the interpretation of multivariate controlcharts in the present study where surveys were performed at different sites every yearFish monitoring surveys can be accompanied by habitat surveys to record variables suchas benthic composition (Floeter et al 2007) and habitat complexity (Gratwicke amp Speight2005) that are known to influence fish assemblages For surveys using permanent transectsdifferences in fish assemblages or habitat likely reflect actual changes in these variables overtime at each site On the other hand observations based on stratified random sampling aredependent on site selection at each time point and changes in fish assemblages or habitatvariables are not representative of any specific sites multivariate control charts identifychanges in the multivariate centroids (or averages) of response variables for each groupof sites (ie region in our case) Interpretation of temporal changes becomes particularlydifficult when the number of sites surveyed from each stratum is small as is the case inthe present study in the south region in 2012 and 2015 While differences in habitat or anyother environmental variables may be directly or indirectly linked to unusual observationsof fish assemblages identified by control charts such differences could be either actualchanges in the habitat over time or simply results of site selection where a small numberof lsquolsquounusualrsquorsquo sites happened to be chosen In the present study although habitat variableswere not collected consistently over time percent covers of hard corals (recorded since2010) were relatively low at all of the survey sites in the strata FRF-M and FRF-D in thesouth region in 2015 (Fig 3) In particular all three sites surveyed in the stratum FRF-Dhad le5 coral covers and their reef types were recorded as rubble Similarly two of thethree sites surveyed in the stratum FRF-M in the south region in 2012 also had a reeftype of pavement with le5 coral covers These were the strata where two invertivoresPseudojuloides cerasinus and O bimaculatus were consistently abundant Therefore thealarm triggered by the control charts could be at least in part due to the small numbers ofsurvey sites and low coral covers at these sites particularly in 2015

Despite the relatively small proportion of variation in the structure of fish assemblagesexplained by the location and stratum variables (lt20 Table 4) the use of DISTLM

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1622

(a)

(b)

(c)

(d)

(e)

(f)

Figure 3 Percent cover of hard corals at each survey site in the south region Nihoa Mokumanamanaand FFSHorizontal jitters were added for overlapping values

residuals rather than the original data matrix made some differences to the control chartswhen compared to those constructed using the original data matrix (Fig 4) For examplevalues of dt and dbt for the south region in 2012 were relatively comparable to and higherthan respectively those in 2015 when the original data matrix was used (Fig 4) On theother hand the use of DISTLM residuals considerably reduced both dt and dbt in 2012resulting in these values being lower than those in 2015 (Fig 2) This is consistent withthe results of SIMPER analysis when comparing the top eight species that typified eachstratum in the south region in 2012 (Table 7) and 2015 (Table 8) those in 2012 are moresimilar to the eight typifying species from all other years (Table 6) Our use of DISTLMto obtain residuals for multivariate control charts is parallel to the concept of detrendingin which a temporal or spatial trend in ecological data is extracted prior to data analyses(Legendre amp Legendre 2012) Here we limited explanatory variables for DISTLM to thosethat were part of our survey design (ie strata) and a potential source of spatial correlation(ie geographical location) While it is possible to include other explanatory variablessuch as reef types and coral covers for detrending successively extracting multiple trendsrequires some caution as each extraction distorts the residuals (Legendre amp Legendre 2012)

For monitoring of a protected area BACI (Before-After-Control-Impact) designs areoften used to compare biological communities inside and outside the protected area overtime but lack of comparable sites can preclude such a monitoring design while scientists

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1722

(a)

0

10

20

30

40

2009 2010 2011 2012 2013 2014 2015 2016

Year

d t

North

Mid

South

(b)

0

10

20

30

40

2009 2010 2011 2012 2013 2014 2015 2016

Year

dbt North

Mid

South

Figure 4 Distance-based multivariate control charts dt (a) and dbt (b) for the south mid and north re-

gions of the NWHI constructed based on the original data matrix G The dashed line indicates the 95thpercentile and the dotted line indicates the 50th percentile both estimated from 1000 bootstrap samplesFor calculation of db

t the first two years of data (2007 and 2009) were designated as a baseline Values ofdt and db

t after DISTLM (as shown in Fig 2) are overlaid in gray for comparison Note that the 95th per-centile lines are specific to the dt and db

t constructed based on the original data matrix G so they do notapply to the dt and db

t after DISTLM shown in gray

and managers try to evaluate whether the protected area is achieving its objectives (Stringellet al 2013) Multivariate control charts allow for evaluation of protected areas even in theabsence of appropriate control sites (Anderson amp Thompson 2004 Stringell et al 2013)The present study showed that multivariate control charts can be extended to a stratifiedrandom sampling design and used to evaluate the status of biological communities in avery large protected area Here we focused our analyses on non-transient resident reeffish species in the NWHI The use of the BrayndashCurtis measure for multivariate controlcharts followed by SIMPER analyses to identify typifying species for each stratum put moreemphasis on common species so less-common species were removed prior to the analysesEach user should adjust however pre-treatments of data (eg data transformationand removal of rare species) choice of dissimilaritydistance measures and methods ofinvestigation following multivariate control charts according to the nature of their studyand ecological questions they have

In conclusion reef fish assemblages in the NWHI were mostly stable with exceptionsin the south region (Nihoa Mokumanamana and FFS) in 2012 and 2015 The potentialsensitivity of multivariate control charts to a small sample size combined with low coralcover when applied to stratified random sampling calls for some caution during the processof survey site selection where selected sites should cover a wide range of coral reef habitatrepresentative of the NWHI In addition multivariate control charts do not provideunderstanding of the causes mechanisms or ecological processes of observed changes inbiological communities Therefore for ecological monitoring of the PapahanaumokuakeaMarine National Monument it is critical for researchers and managers to identify potential

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1822

threats to shallow reef habitats of theNWHI and ensure that future efforts of fishmonitoringare accompanied by measures of specific environmental variables that can be used to assesswhether changes in fish assemblages are associated with any of these threats

ACKNOWLEDGEMENTSWe thank the officers and crew of the NOAA ship Hilsquoialakai for logistic support and fieldassistance and RAMP fish divers for data collection S Godwin served as chief scientist onnumerous RAMP cruises and facilitated collection of these data NOAA Pacific IslandsFisheries Science Centerrsquos Coral Reef Ecosystem Program contributed to field work andsurvey operations as a partner agency K Fujii prepared themapof theHawaiianArchipelagofor this publication We also thank M Pawley M Green and an anonymous reviewer whoimproved themanuscript through their input The scientific results and conclusions as wellas any views or opinions expressed herein are those of the authors and do not necessarilyreflect the views of NOAA or the Department of Commerce

ADDITIONAL INFORMATION AND DECLARATIONS

FundingThis work was funded by NOAArsquos Office of National Marine Sanctuaries (through thePapahanaumokuakea Marine National Monument) and by a grant to NOAA FisheriesrsquoPacific Islands Fisheries Science Center from NOAArsquos Coral Reef Conservation ProgramNo additional external non-NOAA funding was received for this study The funders hadno role in study design data collection and analysis decision to publish or preparation ofthe manuscript

Grant DisclosuresThe following grant information was disclosed by the authorsNOAArsquos Office of National Marine SanctuariesNOAA Fisheriesrsquo Pacific Islands Fisheries Science Center

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Atsuko Fukunaga analyzed the data wrote the paper prepared figures andor tablesreviewed drafts of the paperbull Randall K Kosaki contributed reagentsmaterialsanalysis tools wrote the paperreviewed drafts of the paper

Field Study PermissionsThe following information was supplied relating to field study approvals (ie approvingbody and any reference numbers)

Field work in the Papahanaumokuakea Marine National Monument (NorthwesternHawaiian Islands) was conducted under PapahanaumokuakeaMarineNationalMonumentresearch permits

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 1922

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj3651supplemental-information

REFERENCESAndersonMJ 2001 A new method for non-parametric multivariate analysis of variance

Austral Ecology 2632ndash46 DOI 101111j1442-9993200101070ppxAndersonMJ Gorley RN Clarke KR 2008 PERMANOVA+ for PRIMER guide to

software and statistical methods Plymouth PRIMER-EAndersonMJ Thompson AA 2004Multivariate control charts for ecological and envi-

ronmental monitoring Ecological Applications 141921ndash1935 DOI 10189003-5379Bellwood DR Hughes TP Folke C NystroumlmM 2004 Confronting the coral reef crisis

Nature 429827ndash833 DOI 101038nature02691Bohnsack JA Bannerot SP 1986 A stationary visual census technique for quantitatively

assessing community structure of coral reef fishes NOAA Technical Report NMFS41 US Department of Commerce National Oceanic and Atmospheric Administra-tion National Marine Fisheries Service Washington DC 15

Brown BE Le Tissier MDA Scoffin TP Tudhope AW 1990 Evaluation of the environ-mental impact of dredging on intertidal coral reefs at Ko Phuket Thailand usingecological and physiological parametersMarine Ecology Progress Series 65273ndash281DOI 103354meps065273

Cadima EL Caramelo AM Afonso-Dias M Conte de Barros P TandstadMO de Leiva-Moreno JI 2005 Sampling methods applied to fisheries science a manual FAOFisheries Technical Paper No 434 FAO Rome 88

Clarke KR Gorley RN 2015 PRIMER v7 user manualtutorial Plymouth PRIMER-EClarke KRWarwick RM 2001 Change in marine communities an approach to statistical

analysis and interpretation Plymouth PRIMER-EConnell JH 1997 Disturbance and recovery of coral assemblages Coral Reefs

16S101ndashS113 DOI 101007s003380050246Connell JH Hughes TPWallace CC 1997 A 30-year study of coral abundance recruit-

ment and disturbance at several scales in space and time Ecological Monographs67461ndash488 DOI 1018900012-9615(1997)067[0461AYSOCA]20CO2

Couch CS Burns JHR Steward K Gutlay TN Liu G Geiger EF Eakin CM KosakiRK 2016 Causes and consequences of the 2014 mass coral bleaching event inPapahanaumokuakea Marine National Monument Technical Report for NOAAPapahanaumokuakea Marine National Monument 28

DeMartini EE Parrish FA Parrish JD 1996 Interdecadal change in reef fish populationsat French Frigate Shoals and Midway Atoll Northwestern Hawaiian Islandsstatistical power in retrospect Bulletin of Marine Science 58804ndash825

Floeter SR KrohlingW Gasparini JL Ferreira CEL Zalmon IR 2007 Reef fishcommunity structure on coastal islands of the southeastern Brazil the influence

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 2022

of exposure and benthic cover Environmental Biology of Fishes 78147ndash160DOI 101007s10641-006-9084-6

Fukunaga A Kosaki RKWagner D Kane C 2016 Structure of mesophotic reef fishassemblages in the Northwestern Hawaiian Islands PLOS ONE 11e0157861DOI 101371journalpone0157861

Gratwicke B Speight MR 2005 The relationship between fish species richness abun-dance and habitat complexity in a range of shallow tropical marine habitats Journalof Fish Biology 66650ndash667 DOI 101111j0022-1112200500629x

Grigg RW Polovina J Friedlander AM Rohmann SO 2008 Biology of coral reefs inthe Northwestern Hawaiian Islands In Riegl BM Dodge RE eds Coral reefs of theUSA Dordrecht Springer 573ndash594

Hiatt RW Strasburg DW 1960 Ecological relationships of the fish fauna on coral reefsof the Marshall Islands Ecological Monographs 3065ndash127 DOI 1023071942181

HixonMA Brostoff WN 1996 Succession and herbivory effects of differentialfish grazing on Hawaiian coral-reef algae Ecological Monographs 6667ndash90DOI 1023072963481

Hobson ES 1974 Feeding relationships of teleostean fishes on coral reefs in KonaHawaii Fishery Bulletin 72915ndash1031

Hunter CL StephensonMD Tjeerdema RS Crosby DG Ichikawa GS GoetzlJD Paulson KS Crane DB Martin M Newman JW 1995 Contaminantsin oysters in Kaneohe Bay HawaiiMarine Pollution Bulletin 30646ndash654DOI 1010160025-326X(95)00039-P

Kayal M Vercelloni J De Loma TL Bosserelle P Chancerelle Y Geoffroy S Stieve-nart C Michonneau F Penin L Planes S AdjeroudM 2012 Predator crown-of-thorns starfish (Acanthaster planci) outbreak mass mortality of corals andcascading effects on reef fish and benthic communities PLOS ONE 7e47363DOI 101371journalpone0047363

Legendre P AndersonMJ 1999 Distance-based redundancy analysis testing multi-species responses in multifactorial ecological experiments Ecological Monographs691ndash24 DOI 1018900012-9615(1999)069[0001DBRATM]20CO2

Legendre P Legendre L 2012Numerical ecology Oxford ElsevierLindenmayer DB Likens GE 2010 The science and application of ecological monitor-

ing Biological Conservation 1431317ndash1328 DOI 101016jbiocon201002013Loya Y Sakai K Yamazato K Nakano Y Sambali H VanWoesik R 2001 Coral

bleaching the winners and the losers Ecology Letters 4122ndash131DOI 101046j1461-0248200100203x

McArdle BH AndersonMJ 2001 Fitting multivariate models to communitydata a comment on distance-based redundancy analysis Ecology 82290ndash297DOI 1018900012-9658(2001)082[0290FMMTCD]20CO2

McCook LJ 1999Macroalgae nutrients and phase shifts on coral reefs scientific issuesand management consequences for the Great Barrier Reef Coral Reefs 18357ndash367DOI 101007s003380050213

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 2122

McGeheeMA 1994 Correspondence between assemblages of coral reef fishes andgradients of water motion depth and substrate size off Puerto RicoMarine EcologyProgress Series 105243ndash255 DOI 103354meps105243

Mundy BC 2005 Checklist of the fishes of the Hawaiian Archipelago Honolulu BishopMuseum Press

Office of National Marine Sanctuaries 2009 Papahanaumokuakea marine nationalmonument condition report 2009 Silver Spring US Department of CommerceNational Oceanic and Atmospheric Administration Office of National MarineSanctuaries 54

Parrish FA Boland RC 2004Habitat and reef-fish assemblages of banks in the North-western Hawaiian IslandsMarine Biology 1441065ndash1073DOI 101007s00227-003-1288-0

R Core Team 2016 R a language and environment for statistical computing V 331Vienna R Foundation for Statistical Computing Available at httpwwwR-projectorg

Randall JE 1987 Introductions of marine fishes to the Hawaiian Islands Bulletin ofMarine Science 41490ndash502

Randall JE 2007 Reef and shore fishes of the Hawaiian Islands Honolulu Sea GrantCollege Program University of Hawaii

Smith SG Ault JS Bohnsack JA Harper DE Luo J McClella DB 2011Multi-species survey design for assessing reef-fish stocks spatially explicit manage-ment performance and ecosystem condition Fisheries Research 10925ndash41DOI 101016jfishres201101012

Stringell TB Bamber RN BurtonM LindenbaumC Skates LR SandersonWG 2013A tool for protected area management multivariate control charts lsquocopersquo with rarevariable communities Ecology and Evolution 31667ndash1676 DOI 101002ece3585

Williams ID Baum JK Heenan A Hanson KM NadonMO Brainard RE 2015Human oceanographic and habitat drivers of central and western Pacific coral reeffish assemblages PLOS ONE 10e0120516 DOI 101371journalpone0120516

Fukunaga and Kosaki (2017) PeerJ DOI 107717peerj3651 2222