Measuring Recovery of Benthic Species, Assessing Potential Spill-Over Effects and Socio-Economic Changes

Lyme Bay - A case study: measuring recovery of benthic species, assessing potential spill-over effects and socio-economic changes

Objective 4: Assessment the long-term effects of fishery area closures on long-lived and sessile species

17 December 2010

Defra Contract No: MB0101 Marine and Fisheries Science Unit, Nobel House, London

Author details:

O. Langmead MarLIN Biodiversity and Conservation Science Programme, Manager [email protected]

E.L. Jackson MarLIN Biodiversity and Conservation Science Programme, Manager [email protected]

D.T.I.Bayley MarLIN Biodiversity and Conservation Science Programme, Information Officer [email protected]

C.E. Marshall [email protected]

S.C. Gall University of Plymouth Marine Biology & Ecology Research Centre, Project Support Officer [email protected]

The Marine Life Information Network® for Britain and Ireland (MarLIN) The Marine Biological Association of the United Kingdom The Laboratory Citadel Hill Plymouth, PL1 2PB www.marlin.ac.uk

This report should be cited as:

Langmead, O., Jackson, E.L., Bayley, D.T.I., Marshall, C.E., Gall, S.C., 2010. Assessment of the long-term effects of fishery area closures on long-lived and sessile species. Report to Defra from the Marine Life Information Network (MarLIN). Plymouth: Marine Biological Association of the UK. Defra contract No.MB0101

Disclaimer: The content of this report does not necessarily reflect the views of Defra, nor is Defra liable for the accuracy of information provided, or responsible for any use of the reports content.

Executive Summary

The work presented in this report is part of a larger project funded by Defra which focuses on assessing the various changes that may ensue as a result of protection of a 60 nm2 area of Lyme Bay to mobile fishing gear, both in ecological and economic terms. The ecological monitoring includes monitoring of representative indicator species of the reef to examine recoverability and to “assess the long-term effects of fishery area closures on long lived and sessile benthic species”. The current report outlines the potential recoverability of fragile long-lived sessile species within the area through biological traits analysis and modelling, and also assesses the current level of current knowledge on long-term community recovery in temperate waters following marine area closures.

Complex hard substratum habitats recover slower from the impacts of fishing than other sites (Kaiser et al., 2006). This is related to the life history of the organisms that live there. Most of the sessile benthic taxa from the reefs of Lyme Bay are either fragile or of intermediate fragility. Species that can regenerate from fragments are likely to recover faster than those with a solitary habit (sponges, bryozoans and corals), but only if they have not suffered local extirpation and are reliant on recolonisation via larval recruitment. Taxa that cannot regenerate are likely to take longer since they are reliant on external supply of larvae and subsequent recruitment processes, which can be very variable, and in addition many are slow growing. Complete recovery has rarely been established, in part due to a lack of reference areas, which has confounded many experimental approaches (Engel & Kvitek, 1998).

Species interactions may mediate recovery and produce unexpected and dramatic responses and are hard to predict. Dramatic shifts have been documented in the literature from urchin barrens to macroalgal dominated communities, reversing trophic cascades e.g. (Salomon et al., 2008). These can occur over extended timescales. Nevertheless, analogous changes of a similarly dramatic nature are unlikely at Lyme Bay because there is no conspicuously dominant grazer equivalent to the urchins, nor is there any obviously grazer-dominated habitat. Secondly, this type of interaction is more likely to be seen in no-take reserves, where predators can recover, whereas the closure in Lyme Bay is closed only to demersal towed gears.

From the modelling work it is clear that all three approaches predicted that a large proportion of the closed area is suitable for supporting the pink sea fan. In reality, many of these areas are associated with absence records (false positives). The Maxent model outperformed the other models in all but one of the model performance indicators and on face value one might use this model in preference to the GLM output to support spatial management measures.

Overall this work identified a paucity of quantitative, comparable studies in the literature on which to predict recovery of sessile benthic communities. This highlights the importance of the monitoring work in Lyme Bay, not just to quantify patterns and rates of recovery of a priority UK habitat, and associated community that includes many species of conservation importance, but this has profound importance for adding to the global body of knowledge of reef systems and their recoverability from physical disturbance.

Contents Page

Executive Summary ...... 3 1 Introduction ...... 7 1.1 Project Background ...... 7 1.2 Assessing long-term effects of fishery closure ...... 8 1.3 Aims and specific objectives ...... 9 2 Methods ...... 10 2.1 Biological traits analysis of sensitivity and recoverability traits of long-lived sessile species ...... 10 2.1.1 Construction of a species list for Lyme Bay...... 10 2.1.2 Biological traits analysis ...... 11 2.2 Review of long-term effects of fishery area closures on long-lived sessile benthic organisms ...... 14 2.2.1 Construction of the database ...... 14 2.2.2 Literature search criteria ...... 14 2.2.3 Data collation and analyses ...... 15 2.3 Predictive modelling of Eunicella verrucosa distribution in Lyme Bay ...... 17 2.3.1 Distribution data ...... 17 2.3.2 Environmental data ...... 17 2.3.3 Modelling approach ...... 18 2.3.4 Model performance ...... 19 2.3.5 Model outputs ...... 21 2.3.6 Fishing activity data ...... 21 3 Results and discussion ...... 22 3.1 Analysis of sensitivity and recoverability traits of long-lived sessile species ...... 22 3.1.1 Fragile taxa ...... 26 3.1.2 Intermediate fragility taxa ...... 28 3.1.3 Robust taxa ...... 30 3.2 Long-tem effects of fishery area closures on long-lived sessile organisms ...... 31 3.3 Eunicella verrucosa distribution modelling in Lyme Bay ...... 40 3.3.1 Environmental parameters ...... 40 3.3.2 Model performance ...... 40 3.3.3 Historic fishing activity and long term monitoring ...... 45 3.3.4 Environmental parameters ...... 48 3.3.5 Model performance ...... 48 4 Conclusions ...... 50 5 References ...... 52 6 Appendices ...... 55 6.1 Appendix 1: Lyme Bay benthic species list ...... 55

List of Figures

Figure 1 Construction of the list of benthic species for Lyme Bay

Figure 2 Flow diagram of the steps used to filter Lyme Bay long-lived sessile taxa using life history trait analysis

Figure 3 Eunicella verrucosa presence (solid circles) and absence (hollow circles) records obtained through remote survey techniques

Figure 4 Age of marine reserves by geographic location; black bars indicate UK MPAs, green Mediterranean and blue temperate global

Figure 5 Proportion of protected areas by broad type

Figure 6 Size of protected area by MPA type

Figure 7 Predicted distribution of suitable pink sea fan based on the final Generalized Linear Model with substrate as the only environmental parameter

Figure 9 Predicted distribution of suitable pink sea fan based on the average predicted values from the GLM and Maxent model

Figure 10 Close up of the closed area with the Maxent predicted map

List of Tables

Table 1 Definitions of biological traits used in sensitivity and recoverability analyses

Table 2 Codes and categories for the seven traits used in the analysis

Table 3 Search terms used in combination to structure the literature search

Table 4 Variables used in the development of predictive distribution models for Eunicella verrucosa in Lyme Bay and their source

Table 5 Indicators selected for use in assessing the performance of the models. Note that the CCR, sensitivity and specificity were only used in Maxent when using independent test data because these indicators also require absence records

Table 6 Life-history traits for all long-lived sessile benthic species within Lyme bay. Traits include „Fragility‟, „Regeneration‟, „Maturity‟, „Fecundity‟, „Larval dispersal‟, „Lifespan‟ and „Growth‟, and are coded as specified in section

Table 7 Taxa lists of fragile, intermediate fragility and robust, long-lived sessile fauna within the Lyme Bay reserve

Table 8 The number of fragile taxa, which can disperse to discrete distances during their larval period

Table 9 Sensitivity and recoverability traits of group 1 taxa (long-lived, sessile, fragile taxa with no ability to regenerate)

Table 10 Sensitivity and recoverability traits of group 2 taxa (long-lived, sessile, fragile taxa with the ability to regenerate)

Table 11 Potential larval dispersal distances of taxa of intermediate fragility, these are divided regeneration ability

1 Introduction

1.1 Project Background

1.1 Early in 2008 Defra released a Partial Regulatory Impact Assessment and Consultation on “measures to protect biodiversity in Lyme Bay from the impact of fishing with dredges and other towed gear”. Following this consultation Defra designated a 60 nm2 closure to mobile fishing gears within Lyme Bay. The primary purpose for establishing the closed area is for the protection of marine biodiversity, namely, to ensure the structure of the reef system is maintained and to aid the recovery of the benthic habitats.

1.2 The work presented in this report is part of a larger project funded by Defra which focuses on assessing the various changes that may ensue as a result of protection, both in ecological and economic terms. In addition to monitoring the marine reef community as a whole, the monitoring of representative indicator species are required to examine recoverability and to “assess the long-term effects of fishery area closures on long lived and sessile benthic species”. Specifically the purpose of the project is to:

1. Identify and select a number of indicator species that represent the full range of life strategies used by benthic species in the study area [but selection of species should consider their wider application for monitoring of Marine Protected areas (MPA)]; 2. To develop a cost-effective sampling design for the monitoring of benthic recovery within the closure of an area of Lyme Bay; 3. To quantify the recovery of the identified species within the closure with the removal of towed gear1 compared to appropriate control areas; 4. To assess the long-term effects of fishery area closures on long lived and sessile benthic species; 5. To collect and store samples of selected benthic species for future DNA analysis; 6. To quantify and assess any effects on scallops (e.g. increased larval export and spill over) resulting from the area closure; and 7. To assess any socio-economic impacts2 (e.g. diversification, gear changes, changes to areas fished, effort changes) which result from the closure restrictions.

1.3 This information will be used to assess the effectiveness of marine protected areas in achieving conservation objectives; the socio-economic implications of MPAs; provide further detail on where fisheries management and conservation objectives could be integrated.

1 Demersal gear and dredgers 2 Other business operators such as the recreational industry etc should be considered in the scope of this assessment and not merely the fishing industry. 7

1.2 Assessing long-term effects of fishery closure

1.4 Rocky reefs supply food resources, nurseries and shelters to a variety of organisms and sustain high levels of biodiversity due to their heterogeneity and three dimensional complexity Turner et al. (1999). Complex hard substratum habitats recover slower from the impacts of fishing than other sites (Kaiser et al., 2006). This is related to the life history of the organisms that live there, particularly the rates of growth, lifespan and age of maturity. Other attributes that may play a role relate to reproduction and recruitment such as frequency and mode of reproduction, larval duration in the water column.

1.5 The process of recovery is more complex than regrowth of organisms from physical damage and replacement of locally extirpated species to previously known abundances. It can also involve complex interactions between species, sometimes different from those dominating the original assemblage, and may lead to unexpected outcomes (e.g. decreases in biomass or density of benthic organisms) but usually involves an increase in diversity (Halpern, 2003). In reef systems where there are organisms with complex life cycles (e.g. colonial) and that have sessile emergent growth forms that make them susceptible to mortality from physical disturbance, recovery is often difficult to predict (Henry et al., 2006).

1.6 Since the time-span of this project (3 years) is much shorter than the recovery time of many long-lived sessile species, it is not possible to directly assess the long-term effects of fishery area closures in Lyme Bay. So the approach proposed here is to make an informed assessment of the likely outcomes of fishery area closures using the best available knowledge from the scientific and grey literature.

1.7 This comprised two components: 1) an investigation into the sensitivity and recoverability of the long-lived sessile species recorded in the study area of Lyme Bay using biological traits analysis, and 2), a review of scientific literature into the long-term effects of fishery area closures on benthic fauna at a national, regional (NE Atlantic) and global scale to put findings into a wider context.

1.8 This work was supported by a third element, namely the predictive species distribution modelling of Eunicella verrucosa, to gain a spatial dynamic to the assessment of long term change by identifying priority areas for long-term monitoring. This work provides a spatial reference against which expected recovery can be measured, complementing the temporal scope of the first two tasks.

1.3 Aims and specific objectives

1.9 The overarching aim of this work is to assess the long-term effects of fishery area closures on long-lived and sessile species and build up a picture of the pattern of recovery that could be expected following the closure.

1.10 This was achieved by setting a number of specific objectives relating to the three tasks outlined in the previous section:

 To identify fragile long-lived sessile species found within Lyme Bay between 2007 and 2010, and assess their potential recoverability using biological traits analysis.

 To review the existing body of literature on recovery of temperate reef communities following MPA designation (and other closures) and draw out any common pattern regarding recovery of sessile species and relate this to attributes of protection where possible.

 To produce predictive maps of Eunicella verrucosa distribution within Lyme Bay using a number of modelling approaches to provide a spatial reference against which expected recovery can be measured and to identify priority areas for long-term monitoring.

2 Methods

2.1 Biological traits analysis of sensitivity and recoverability traits of long-lived sessile species

2.1.1 Construction of a species list for Lyme Bay

2.1 A list of benthic species recorded in Lyme Bay was constructed by combining existing national species records that were drawn together in Objective 1 of the MB0101 project (see Jackson et al. (2008) for details on sources and methods) together with survey data collected during the MB0101 project (from both University of Plymouth Marine Institute and Marine Bio-Images).

2.2 The additional records were filtered in the same way as for Objective 1; vertebrates were removed as they were considered outside the scope of this work except for fish, which were included where they met the following criteria:

 Environment: Fish with a close association with the benthos were included (Benthic, Demersal, Benthopelagic)  Habitat: Reef associated fish.  Fish were removed if they were nocturnal, predominately intertidal or small (< 5 cm length )

2.3 The species list comprised of 783 distinct taxa after matching with the WOrld Register of Marine Species (WoRMS, available from http://www.marinespecies.org), to ensure correct spellings were used throughout and no taxonomic synonyms were included (Appendix 1).

MB0101 Objective 1 benthic species list Species encountered during UoP monitoring work Species records collated from past studies that pre-date the start of the MB0101 project (drop-down camera and diver surveys)

Species encountered during fixed quadrat studies (24 surveys)

WoRMS match and merge

List of benthic species for Lyme Bay

Figure 1 Construction of the list of benthic species for Lyme Bay 10

2.1.2 Biological traits analysis

2.4 The list of benthic species recorded in Lyme Bay was used as a basis for the biological traits analysis. Firstly all non-sessile species were removed, reducing the original list of 783 benthic species to 425 sessile taxa. The definition of sessile used in this work was any organism which is „permanently attached to a substratum‟. This definition included:

 permanently attached organisms;  tube-dwelling organisms in permanent burrows; and  organisms attached to permanent substrate by byssus threads.

2.5 Information to determine whether a benthic organism was sessile was gathered from:

 the Biological Traits Information Catalogue (BIOTIC, www.marlin.ac.uk/biotic);  the MarLIN website and if neither of these proved fruitful; then  the considerable literature resources held in National Marine Biological Library (NMBL) hosted at the MBA were consulted.

2.6 BIOTIC3 is a database which contains information on over 40 biological trait categories for 685 selected benthic taxa, together with a bibliography of source literature. The emphasis is on benthic invertebrates and plants. In addition, MarLIN holds further traits information at generic level. Any gaps in species information in BIOTIC were researched primarily using the National Marine Biological Library. This is a unique world-class marine life and environmental sciences resource for the UK and Europe and collectively holds over 100,000 books, reprints, serials and reports, some dating back to before the 19th century. This wealth of information, along with web-based resources such as „Google Scholar‟ and „Web of Knowledge‟ were utilised to enable traits information for a large number of species to be included in analyses.

2.7 The list of sessile species for Lyme Bay was further filtered to retain only the long-lived species (longevity of >5 years) for further analysis. This comprised 62 taxa; 192 short-lived sessile taxa were removed along with an additional 172 taxa for which their longevity could not be assigned due to lack of information. In some instances, where information was lacking, entries were grouped according to similar taxa, up to the family level.

2.8 The list of 62 sessile long-lived species was then investigated for a further six life history traits: fragility, regeneration, maturity, fecundity, larval dispersal and growth. Definitions of each of these traits are given in Table 1.

3 All additional traits researched during this contract have been added to BIOTIC to increase the species coverage and traits holdings and are available online. 11

Table 1 Definitions of biological traits used in sensitivity and recoverability analyses (sources: BIOTIC, www.marlin.ac.uk/biotic) Trait Definition Fragility The propensity to suffer damage from a physical impact. Regeneration The capacity for partial or whole regrowth or regeneration. Maturity The time taken to reach reproductive maturity from birth. Fecundity The average number of offspring per reproductive episode. Larval dispersal potential The potential horizontal distance larvae may travel before settling. Lifespan The potential maximum time from birth to death. Growth rate The average increase in width/length per unit time over the whole lifespan.

2.9 The sensitivity and recoverability traits were grouped into categories and values assigned for each level of each trait (Table 2) allowing a matrix to be produced.

Table 2 Codes and categories for the seven traits used in the analysis. Coded scores for analysis 1 2 3 4 5 Recoverability traits Fragility Fragile Intermediate Robust Regeneration No Yes Maturity (years) < 1 1 - 2 3 – 5 6 – 10 Fecundity < 2 k 2 k – 200 k > 200 k DispPotLarvae < 0.1 km 0.1 – 1 km 1 – 10 km > 10 km Lifestyle traits Lifespan (years) < 1 1 - 2 3 – 5 6 – 10 ≥ 11 Growth rate ≤ 1 cm/yr 1 – 3 cm/yr 3 – 5 cm/yr > 5 cm/yr

2.10 Fragile species were identified (those considered to have a high propensity to suffer damage from a physical impact). The rationale behind this split was that recovery would only be seen in taxa that had previously been damaged by the mechanical action of dredging. This shorter list of fragile taxa group was again divided according to the potential ability for larval dispersal, measured as the horizontal distance that larvae may travel before settling. This was measured in discrete categories: <0.1 km, 0.1-1 km, and >10 km (the category of 1-10 km was unpopulated). From these three groups the resulting taxa were rated on a scale of 1 – 3 according to their growth and maturity traits within a matrix. The steps followed during this process are outlined in Figure 2 below.

Figure 2. Flow diagram of the steps used to filter Lyme Bay long-lived sessile taxa using life- history trait analysis.

Long-lived sessile

Fragile Intermediate Robust

YES Regenerative ability NO

Larval dispersal distance

1 2 3 4

Growth/maturity Growth/maturity Growth/maturity Growth/maturity

2.2 Review of long-term effects of fishery area closures on long- lived sessile benthic organisms

2.2.1 Construction of the database

2.11 An existing database was modified to meet the requirements of this project http://www.marlin.ac.uk/fisheriesmanagement/index.html (Sewell et al., 2007). The database was customised to hold information on:

 reference source;  details of the closure (location, size, duration, fishery restrictions);  specific information on the habitat and environment;  information on the type of fishing activity; and  recovery of species following protection measures focussing on long-lived sessile species.

2.12 The latter section of the database was designed to hold details on changes that occurred after an area was closed (i.e. in cover, density, health) of species. Because these changes at the level of species can be manifest at a community level (in terms of changes in the functional relationship between species) and also as ecosystem level responses, additional provision was made to the database to allow recording of these higher organisational level responses.

2.2.2 Literature search criteria

2.13 A number of criteria were developed to focus the literature search and ensure that only relevant and informative literature would be included in the database. This included restricting the studies to those:

 from temperate reefs (i.e. excluding all research and monitoring related to sedimentary substratum and tropical systems such as coral reefs);  that considered effects of closure on benthic invertebrates, macroalgae and seagrass. Studies that were restricted to mobile species were excluded unless changes to mobile species had documented effects on benthic communities;  that were conducted between 1970 and 2009;  that included recovery (i.e. studies that were restricted to impacts from trawling or other activities resulting in physical disturbance only did not qualify);  that comprised primary research; review papers were excluded but their references consulted to identify the original research cited;  with a temporal dimension (i.e. excluding those that solely compared between closed and open areas at single point in time);  closures without installations, wrecks or other anthropogenic structures; and  studies that documented permanent closures and not temporary, seasonal or recreational fishing restrictions.

2.14 A list of search terms were defined to frame the literature search and these were used in every possible combination (term 1 + term 2+ term 3+ term 4), as shown in Table 3, in Web of Knowledge Web of ScienceTM.

Table 3 Search terms used in combination to structure the literature search

Term 1 Term 2 Term 3 Term 4 fisher* closure impact benthic restriction effect benthos recovery recover* seabed conservation disturbance seafloor protect demersal licence seabottom mpa reserve ntz exclusion management * indicates „wildcard‟

2.15 The list of search terms was used to interrogate online databases such as Web of Knowledge in Web of Science, Google „Scholar‟ and the reference database held by UK MPA Centre (hosted at the MBA) http://www.ukmpas.org/mpareferences/search.php. Following the collection of relevant papers, their reference list was also searched in an attempt to find a truly comprehensive list, this was particularly important for review papers, which were not included directly.

2.16 Since the purpose was to identify individual incidents of recovery of benthic organisms following closure, repeat publications on the same site were excluded and only the most recent was included to prevent double counting.

2.2.3 Data collation and analyses

2.17 Attributes of each of the reserves was collected from each literature references. Where there was more than one publication on a marine site, the most recent was used (unless earlier work showed more detail on benthic effects).

2.18 Age of reserve at the time of the work was not always given in the publication, but invariably the year of establishment of the reserve was. In such cases an assumption was made that the work was done 2 years before the publication date, and the age of the reserve calculated from the year of establishment.

2.19 The size of the site was converted to km2 to ensure consistency, and this was estimated for reports where the size was not explicitly given.

2.20 The main response variable used was the percent difference in abundance of individual species over the course of the study (sometimes from the time when protection was implemented). This was not always possible since different workers measured different parameters, so biomass, density and production were also recorded where there were significant differences indicative of recovery. In addition any higher level effects (community to ecosystem) were also recorded where given.

2.3 Predictive modelling of Eunicella verrucosa distribution in Lyme Bay

2.3.1 Distribution data

2.21 More than 200 distribution records were available for the pink sea fan in Lyme Bay but in order to reduce the potential variability between data from different sources, only data from the Devon Wildlife Trust (DWT) pink sea fan survey and University of Plymouth surveys for DWT (Black, 2007, Stevens et al., 2007) were used in the model building process.

2.22 These two data sets provided presence and absence records for the pink sea fan which increased the number of modeling options available because we were not limited to presence only modeling methods. In total, 87 records were used in the analysis (for the GLM/GAM models) of which 35 were presence records. Only presence records are used in Maxent analysis. The pink sea fan records that were used in the model building process are illustrated in Figure 3.

Figure 3 Eunicella verrucosa presence (solid circles) and absence (hollow circles) records obtained through remote survey techniques.

2.3.2 Environmental data

2.23 Environmental variables used to build the models were chosen based on factors known to influence gorgonian distribution. The environmental data used in the study is presented in Table 4 along with its source.

2.24 The DWT substrate categories were regrouped into three larger classes (rock, mixed and sediment) so that each class had a greater number of distribution records in it. 2.25 The remotely sensed data (namely sea surface temperature, suspended particulate matter and chlorophyll a concentration) was obtained for a year period from July 2007 to June 2008. A minimum, mean and maximum value was calculated for each cell over the year period and each of these parameters was modelled individually.

Table 4 Variables used in the development of predictive distribution models for Eunicella verrucosa in Lyme Bay and their source.

Environmental Variable Data source Substrate DWT Depth* Atlas of UK Marine Renewable Energy Spring peak current velocity (m/s)* Resources Sea Surface Temperature (°C) (AVHRR sensor)** Suspended Particulate Matter (g m3) NEODAAS (MERIS sensor) ** Chlorophyll A concentration (mg m3) (MODIS sensor) ** * © Crown Copyright. All rights reserved 2008 **Source: NERC Earth Observation Data Acquisition and Analysis Service (NEODAAS)

2.26 In addition to the variables listed in the table, the interaction between maximum suspended sediment and current (referred to as the scour proxy) was investigated in the GLM/GAM modelling process. The scour proxy was considered relevant because it is possible that the combination of high levels of suspended inorganic material and strong currents could have a negative impact on the gorgonians both in the short term, in the form of reduced feeding potential and damage to the soft tissues, and in the long term by preventing the settlement of planulae larvae.

2.3.3 Modelling approach

2.27 Despite the fact that both presence and absence records were available, a presence only modelling approach (Maxent) was also adopted in order to compare the outputs and predictive accuracy from the different models. Furthermore, presence records are often much more widely available than absence records.

2.28 Further details and practical explanations of the modelling methods used in this study can be found in Phillips et al. (2006) (Maxent) and Wood (2006) (GLM/GAMs).

Generalized Linear Models and Generalized Additive Models 2.29 All GLM and GAM analyses were performed in R (R Development Core Team, 2009) using the mgcv library (Wood, 2008). Binomial models with a logit link were used to fit bivariate models with each individual environmental variable. Once significant variables had been identified, a forward stepwise selection process was used to produce the final model. Variables that were highly correlated (Pearson‟s correlation coefficient >0.7) were not included together in the final model building process.

Maxent 2.30 The Maxent analysis was carried out using Maxent version 3.3.1. Maxent does not check for correlations between variables. Correlations were investigated in R and, after examining the results of preliminary Maxent analyses, the mean values for the three remotely sensed variables were excluded from the analysis, leaving only the minimum and maximum values.

2.31 To make a more fair comparison with the GAM analysis, the substrate categories were also merged in the Maxent analysis.

2.3.4 Model performance

2.32 The potential application of a predictive model‟s outputs to management lies in their ability to accurately predict the distribution of the species in question and managers will seek models that can demonstrate excellent performance in this respect. Four metrics were employed to assess the performance of the models and these are listed in Table 5.

2.33 For most of these performance indicators, it is necessary to define a threshold. A threshold is the point on the logistic probability scale (that ranges from 0 to 1) at which probabilities above it are defined as presence and probabilities below it are defined as absence. There are various objectives methods of defining a threshold (in addition to subjective methods). In this study, a prevalence method was chosen where the threshold is defined as (Cramer, 2003):

2.34 Of course, this method only applies to the GAM/GLM modelling. For this reason, a threshold independent indicator, the area under the curve (AUC

score), was also chosen, the results for which are built into the Maxent modelling package anyway and are easy to calculate for other modelling approaches. 2.35 A different approach was required in order to define the threshold for assessing the performance of Maxent based on independent test data. Another approach defined by Cramer (2003) uses predicted probabilities to define thresholds. Here, the average predicted probability of the Maxent model building occurrences was used as a threshold.

Defined threshold Indicator Application required?

A measure of the overall predictive Correct Classification Rate (CCR) Yes accuracy of the model

A measure of the model‟s ability to Sensitivity Yes correctly classify presence

A measure of the models‟ ability to Specificity Yes correctly classify absence

Another measure of the overall Area under the curve (AUC) No predictive accuracy of the model

2.36 The performance of the Lyme Bay Eunicella models was assessed in three ways:  The predicted values for the final model were compared with the observed values (GAM/GLM);  Internal re-sampling, where two thirds of the data were randomly selected and fitted to the final model leaving the final third on which to test the predictions, was repeated 100 times (Maxent) or 1000 times (GLM/GAM) in order to gain confidence intervals for the performance indicators, and  The predictions of the final model were applied to independent presence and absence data from „Seasearch‟ dive surveys and the 2008 video transects from the closed area monitoring program. Absence data from the DWT surveys, which were not used in the Maxent model building, were used as additional test data for Maxent.

2.3.5 Model outputs

2.37 A predicted probability layer was produced for each of the two modelling approaches. In addition, the two predictive layers were combined in order to try and identify areas which both methods had selected as suitable for the presence of pink sea fans. The Maxent output is a raster grid (with a cell size of 500m for this analysis). In order to combine the outputs from the two modelling approaches, both layers were converted into a 500m polygon from which the average predicted value for each cell could be calculated. For the combined approach, the threshold was defined as the:

2.3.6 Fishing activity data

2.38 Scallop dredging has been reported as a fishing method that can negatively and significantly influence benthic communities (e.g. Blyth et al. (2004) Hiddink et al. (2007b)). To add another dimension to the overall spatial picture of predicted pink sea fan distribution in light of benthic recovery, spatial data for scallop dredging activity prior to the closure was superimposed on the model predictions. In order to provide a more realistic footprint of scallop fishing, 275m buffers were applied to the point data for scallop dredging activity (see Marshall et al. (in prep.) for a detailed explanation). Point data for scallop fishing were obtained from the Devon Sea Fisheries Committee.

3 Results and discussion

3.1 Analysis of sensitivity and recoverability traits of long-lived sessile species

3.1 The sensitivity and recoverability life-history traits for the 62 taxa identified as long-lived and sessile (from the overall list of 425 benthic taxa occurring in Lyme Bay) are shown in Table 6. Whilst most trait characteristics were populated, a number of traits, particularly fecundity, growth and age at maturity values were frequently absent from research records.

Table 6 Life-history traits for all long-lived sessile benthic species within Lyme bay. Traits include ‘Fragility’, ‘Regeneration’, ‘Maturity’, ‘Fecundity’, ‘Larval dispersal’, ‘Lifespan’ and ‘Growth’, and are coded as specified in section 2. Larval dispersal Species name Fragility Regeneration Maturity Fecundity Lifespan Growth potential Actinia equina 2 1 2 - 1 5 2 Actinia fragacea 2 1 2 - 1 5 2 Actinothoe sphyrodeta 2 1 1 1 3 5 1 Adamsia carciniopados 2 1 2 - 1 4 - Alcyonidium spp. 2 1 2 2 1 4 1 Alcyonidium diaphanum 2 1 2 2 1 4 1 Alcyonidium gelatinosum 2 1 2 2 1 4 1 Alcyonium spp. 1 2 3 2 4 5 1 Alcyonium digitatum 1 2 3 2 4 5 1 Alcyonium glomeratum 1 2 3 2 4 5 1 Amphianthus dohrnii 1 1 2 - 1 5 - Anemonia viridis 2 1 2 - 1 5 2 Aulactinia verrucosa 2 1 2 - 1 5 2 Axinella damicornis 1 2 - - - 5 1 Axinella dissimilis 1 2 - - - 5 1 Calliactis spp. 2 1 2 - 1 4 - Calliactis parasitica 2 1 2 - 1 4 - Caryophyllia spp. 1 1 2 - 4 5 1 Caryophyllia (Caryophyllia) inornata 1 1 2 - 4 5 1 Caryophyllia (Caryophyllia) smithii 1 1 2 - 4 5 1 Cereus pedunculatus 2 1 1 1 3 5 1 Cerianthus spp. 2 1 - - 4 5 2 Cerianthus lloydii 2 1 - - 4 5 2 Chondracanthus acicularis 3 1 2 3 - 4 4 Chondrus crispus 3 1 2 3 - 4 4 Cirratulus cirratus 1 1 2 1 2 4 2 Cliona celata 2 2 3 2 2 5 4 Corallina officinalis 2 1 3 - 4 5 2 Delesseria sanguinea 2 1 - - - 4 - Eunicella spp. 2 1 3 2 2 5 1

Larval dispersal Species name Fragility Regeneration Maturity Fecundity Lifespan Growth potential Eunicella verrucosa 2 1 3 2 2 5 1 Flustra foliacea 1 2 2 2 2 4 2 Hippoporina pertusa 1 1 2 2 2 4 2 Hoplangia durotrix 1 1 2 - 4 5 1 Laminaria digitata 3 1 2 3 2 4 4 Laminaria hyperborea 3 1 2 3 2 4 4 Lithophyllum incrustans 2 1 3 - 4 5 1 Lithothamnium (Lithophyllum) 2 1 3 - 4 5 1 Maerl 1 1 - - 1 5 1 Metridium senile 2 1 - - 4 5 4 Modiolus modiolus 2 1 4 3 4 5 1 Mytilus edulis 2 1 2 3 4 5 1 Parerythropodium coralloides 1 2 3 2 4 5 1 Pentapora spp. 1 1 2 2 2 4 2 Pentapora fascialis 1 1 2 2 2 4 2 Phymatolithon calcareum 1 1 3 - 4 5 1 Sagartia spp. 2 1 1 1 3 5 1 Sagartia elegans 2 1 1 1 3 5 1 Sagartia troglodytes 2 1 1 1 3 5 1 Sagartiogeton laceratus 2 1 1 1 3 5 1 Sagartiogeton undatus 2 1 1 1 3 5 1 Schizomavella spp. 1 1 2 2 2 4 2 Schizomavella auriculata 1 1 2 2 2 4 2 Schizomavella linearis 1 1 2 2 2 4 2 Securiflustra spp. 1 2 2 2 2 4 2 Securiflustra securifrons 1 2 2 2 2 4 2 Tethya spp. 2 2 2 2 3 5 2 Tethya aurantium 2 2 2 2 3 5 2 Urticina eques 2 1 2 - 1 5 2 Urticina felina 2 1 2 - 1 5 2 Veneridae 2 1 2 3 4 4 2 Venerupis senegalensis 2 1 2 3 4 4 2

3.2 Of the 62 long-lived sessile taxa recorded in Lyme Bay, 24 (37.5%) of these were fragile, 36 (56.3%) had an intermediate level of fragility, and only 4 (6.3%) were classed as robust. Table 7 shows the taxa found within each fragility category.

Table 7 Taxa lists of fragile, intermediate fragility and robust, long-lived sessile fauna within the Lyme Bay reserve. Fragile Intermediate Robust Alcyonium spp. Actinia equina Chondracanthus acicularis Alcyonium digitatum Actinia fragacea Chondrus crispus Alcyonium glomeratum Actinothoe sphyrodeta Laminaria digitata Amphianthus dohrnii Adamsia carciniopados Laminaria hyperborea Axinella damicornis Alcyonidium spp. Axinella dissimilis Alcyonidium diaphanum Caryophyllia spp. Alcyonidium gelatinosum Caryophyllia inornata Anemonia viridis Caryophyllia smithii Aulactinia verrucosa Cirratulus cirratus Calliactis spp. Flustra foliacea Calliactis parasitica Hippoporina pertusa Cereus pedunculatus Hoplangia durotrix Cerianthus spp. Lithothamnion glaciale Cerianthus lloydii Parerythropodium coralloides Cliona celata Pentapora spp. Corallina officinalis Pentapora fascialis Delesseria sanguinea Phymatolithon calcareum Eunicella spp. Schizomavella spp. Eunicella verrucosa Schizomavella auriculata Lithophyllum incrustans Schizomavella linearis Lithothamnion coralloides Securiflustra spp. Lithothamnium (Lithophyllum) Securiflustra securifrons Metridium senile Modiolus modiolus Mytilus edulis Sagartia spp. Sagartia elegans Sagartia troglodytes Sagartiogeton laceratus Sagartiogeton undatus Tethya spp. Tethya aurantium Urticina eques Urticina felina Veneridae Venerupis senegalensis

3.1.1 Fragile taxa

3.3 Within the context of reef recovery, the fragile taxa – those that are most sensitive to physical disturbance – are most likely to show recovery following the cessation of this disturbance.

3.4 Within the fragile taxa, more than half (58.3%) did not possess the ability to regenerate (group 1), while ten (41.7%) did have this ability (group 2). The potential for larval dispersal within both group 1 and group 2 ranged between <0.1 km and >10 km.

3.5 Within group 1 (fragile species without the ability to regenerate), two taxa (14.3%) have an extremely low larval dispersal potential of less than 0.1 km (Amphianthus dohrnii, the sea fan anemone and Lithothamnion glaciale, a type of maerl). Both of these species are priority species for conservation with the former being a UK Biodiversity Action Plan species and the latter a biogenic reef forming species. A further seven taxa (50%) have the potential disperse to between 0.1-1 km, and only five taxa (35.7%) have the potential to disperse distances greater than 10 km from their initial release location.

3.6 Within group 2 (fragile species that can regenerate), three bryozoan taxa have a low dispersal potential (between 0.1-1 km): Flustra foliacea, Securiflustra securifrons and Securiflustra spp. Out of the ten taxa in group 2, only four have the potential to disperse to areas greater than 10 km away, while the larval dispersal ability of 2 branching sponges is currently unknown (Table 8).

3.7 Species with the ability to regenerate from fragments are more likely to recover than those which cannot regenerate. Local sources of recolonisation by lateral growth and vegetative propagation from fragments that survived the physical disturbance have been shown to play a more important role than external larval supply in shaping the structure of communities following disturbances (Connell & Keough, 1985). Species without the ability to regenerate and have low larval dispersal potential may take a long time to even reappear in the protected area if they have undergone local extirpation. Factors such as local hydrography and distance from sources of recruitment may then come into play with very difficult to predict consequences.

Table 8 The number of fragile taxa, which can disperse to discrete distances during their larval period. The table is split into taxa which can or cannot regenerate.

Dispersal distance Total

< 0.1 0.1 - 1 taxa km km 1 - 10 km > 10 km Group 1 14.3% 50% - 35.7% 14 (no) Regeneration Group 2 10% 30% - 40% 10 (yes)* * the total for this group is 80% because the two sponge species, Axinella damicornis and Axinella dissimilis do not have larval dispersal information available.

3.8 Within the group of fragile taxa that are not able regenerate (group 1), growth is generally slow (<1 and up to 3 cm yr-1), and the age at maturity predominantly between one and two years, with the exception of Phymatolithon calcareum which takes between three and five years (Table 9). These species are not likely to show recovery from populations that escaped physical damage for some years following closure, and will be reliant on incoming larvae as space cannot be rapidly recolonised by regeneration. And while the recovery of maerl has been reported as 3-5 years, the physical alterations to biogenic reef habitats following the death of the reef-forming species can mean that recovery is not possible, or takes extremely long timescales (Hall-Spencer & Moore, 2000a).

Table 9 Sensitivity and recoverability traits of group 1 taxa (long-lived, sessile, fragile taxa with no ability to regenerate) Disp. Species name Larval Growth Maturity Fecundity Lifespan Amphianthus dohrnii 1 - 2 - 5 Lithothamnion glaciale 1 1 - - 5 Cirratulus cirratus 2 2 2 1 4 Hippoporina pertusa 2 2 2 2 4 Pentapora fascialis 2 2 2 2 4 Pentapora spp. 2 2 2 2 4 Schizomavella auriculata 2 2 2 2 4 Schizomavella linearis 2 2 2 2 4 Schizomavella spp. 2 2 2 2 4 Caryophyllia inornata 4 1 2 - 5 Caryophyllia smithii 4 1 2 - 5 Caryophyllia spp. 4 1 2 - 5 Hoplangia durotrix 4 1 2 - 5 Phymatolithon calcareum 4 1 3 - 5

3.9 Within the group of taxa that are able to regenerate, growth rate is generally slow to medium (<1cm and up to 5 cm yr-1), and the age at maturity prolonged, at between one and five years (Table 10). Thus the advantage

conferred by having an ability to regenerate may be offset by slow rates of growth and prolonged age at maturity, meaning a low recoverability.

Table 10 Sensitivity and recoverability traits of group 2 taxa (long-lived, sessile, fragile taxa with the ability to regenerate) Disp. Species name Larval Growth Maturity Fecundity Lifespan Flustra foliacea 2 2 2 2 4 Securiflustra securifrons 2 2 2 2 4 Securiflustra spp. 2 2 2 2 4 Alcyonium digitatum 4 1 3 2 5 Alcyonium glomeratum 4 1 3 2 5 Alcyonium spp. 4 1 3 2 5 Parerythropodium coralloides 4 1 3 2 5 Axinella damicornis - 1 - - 5 Axinella dissimilis - 1 - - 5

3.1.2 Intermediate fragility taxa

3.10 Within the intermediate fragility taxa, the majority 33 (91.7%) did not have the ability to regenerate (group 1), while only three (8.3%) had this ability (group 2). Larval dispersal potential within both group 1 and group 2, ranged between <0.1 km and >10 km (separated into four discrete groups of, <0.1 km, 0.1 – 1 km, 1 - 10 km and >10 km).

3.11 Within group 1 (taxa of intermediate fragility with no ability to regenerate), 13 taxa (39.4 %) had a highly restricted dispersal potential (< 0.1 km): a group comprised of almost entirely of cnidarians, both sea anemones and soft corals together with the maerl Lithothamnion corallioides (which is both a UK BAP species and an EC Habitats Directive listed species). A further two taxa (6.1%) had limited dispersal potential (between 0.1-1 km: Eunicella verrucosa, the pink sea fan and Eunicella spp. (which would also be E. verrucosa since there are no other congenerics in the Bay, but the double counting is an artefact of different levels of taxonomic resolution from surveys. A total of seven taxa (21.2%) have the potential to disperse to between 1 and 10 km, and ten taxa (30.3%) could potentially disperse distances of greater than 10 km from the larval source.

3.12 Within group 2 (taxa of intermediate fragility with the ability to regenerate, one taxon (33.3%) has a low dispersal potential (0.1 – 1 km): the sponge Cliona celata. The other two taxa (66.7%) have the potential to disperse between 1 and 10 km: the golf ball sponge Tethya aurantium and Tethya spp. (Table 11).

Table 11 Potential larval dispersal distances of taxa of intermediate fragility, these are divided regeneration ability.

Dispersal distance Total taxa < 0.1 km 0.1 - 1 km 1 - 10 km > 10 km Group 1 (no) 39.40% 6.10% 21.20% 30.30% 33 Regeneration Group 2 (yes) - 33.30% 66.70% - 3

3.13 Within the group of taxa that are not able regenerate, growth rates are generally slow (<1 to 3 cm yr-1), with the exception of the plumose anemone, Metridium senile, which grows at a rate of >5 cm yr-1. The ages at maturity for this group were highly variable, ranging from <1 - 5 years. An exception was the horse mussel, Modiolus modiolus, which has an extremely prolonged maturity period of between six and 10 years. Shown in Table 12

Table 12 Sensitivity and recoverability traits of group 1 taxa (long-lived, sessile taxa of intermediate fragility with no ability to regenerate) Disp. Species name Larval Growth Maturity Fecundity Lifespan Actinia equina 1 2 2 - 5 Actinia fragacea 1 2 2 - 5 Adamsia carciniopados 1 - 2 - 4 Alcyonidium diaphanum 1 1 2 2 4 Alcyonidium gelatinosum 1 1 2 2 4 Alcyonidium spp. 1 1 2 2 4 Anemonia viridis 1 2 2 - 5 Aulactinia verrucosa 1 2 2 - 5 Calliactis parasitica 1 - 2 - 4 Calliactis spp. 1 - 2 - 4 Lithothamnion coralloides 1 1 3 - 5 Urticina eques 1 2 2 - 5 Urticina felina 1 2 2 - 5 Eunicella spp. 2 1 3 2 5 Eunicella verrucosa 2 1 3 2 5 Actinothoe sphyrodeta 3 1 1 1 5 Cereus pedunculatus 3 1 1 1 5 Sagartia elegans 3 1 1 1 5 Sagartia spp. 3 1 1 1 5 Sagartia troglodytes 3 1 1 1 5 Sagartiogeton laceratus 3 1 1 1 5 Sagartiogeton undatus 3 1 1 1 5 Cerianthus lloydii 4 2 - - 5 Cerianthus spp. 4 2 - - 5 Corallina officinalis 4 2 3 - 5 Lithophyllum incrustans 4 1 3 - 5 Lithothamnium (Lithophyllum) 4 1 3 - 5 29

Disp. Species name Larval Growth Maturity Fecundity Lifespan Metridium senile 4 4 - - 5 Modiolus modiolus 4 1 4 3 5 Mytilus edulis 4 1 2 3 5 Veneridae 4 2 2 3 4 Venerupis senegalensis 4 2 2 3 4 Delesseria sanguinea - - - - 4

3.14 Within the group of taxa that could regenerate, growth was at a medium to rapid rate of one up to more than five cm yr-1, and the age at maturity was prolonged at between two and five years (Table 13).

Table 13 The larval dispersal, growth, maturity, fecundity and lifespan values of all long-lived, sessile, intermediate fragility taxa with the ability to regenerate. Disp. Species name Larval Growth Maturity Fecundity Lifespan Cliona celata 2 4 3 2 5 Tethya aurantium 3 2 2 2 5 Tethya spp. 3 2 2 2 5

3.1.3 Robust taxa

3.15 None of the four robust taxa possessed the ability to regenerate: Chondracanthus acicularis, Chondrus crispus, Laminaria digitata, and Laminaria hyperborean, and where known, had a restricted dispersal potential (0.1-1 km). All taxa have rapid growth rates (>5 cm yr-1), and a slightly prolonged maturity period of one to two years.

3.16 These species are unlikely to show recovery since they are not sensitive to physical disturbance, and unless there are species interactions that play out along rather difficult to predict trajectories, it would be expected that these species will remain at their current abundances.

3.2 Long-tem effects of fishery area closures on long-lived sessile organisms

3.17 The literature searches produced roughly 340 papers in total. The majority of papers fell outside the strict criteria outlined in the methods: from tropical regions, reported general impacts of fisheries or focussed on communities from soft sediment. Only 29 sites fully met the criteria and were included in the database.

3.18 Because of the low numbers of sites that fulfil the criteria, an elaborate meta- analysis was not possible, but instead a more qualitative analysis simply showing the findings was all that could be conducted. This clearly shows the lack of information at a global scale on the recovery of benthic reef communities following closures and highlights the importance and uniqueness of the Lyme Bay study.

3.19 Most MPAs reported were less than 10 years old, with only 31% reported as older than ten (Figure 4). The majority (66%) came from temperate global waters (non-European), while 17% were studies carried out in UK protected areas and the same number again in the Mediterranean.

Figure 4. Age of marine reserves by geographic location; black bars indicate UK MPAs, green Mediterranean and blue temperate global (n=29).

3.20 The marine protected areas varied in type from strict no-take zones to MPAs designed to aid fisheries management, experimental closures (generally to study the effect of removal of fishing pressure) and a general MPA category which were essentially those which did not fit into any of the other categories, but notably were established for the function of biodiversity conservation not fisheries management (Figure 5).

Figure 5 Proportion of protected areas by broad type

3.21 For the 27 sites for which information on the area protected was available, it is clear that there are a great deal more small sites than large ones (Figure 6). Experimental closures, obviously, fall within the smallest category, and no take zones and other MPAs fall within the three smallest categories. The largest size categories comprise closures for fisheries management and these include sites from the North Western Atlantic (Gulf of Maine, NE Georges Bank, E Scotian Shelf) and California (Monterey Bay) and the UK (South West Inshore Potting area).

Figure 6 Size of protected area by MPA type

3.22 In 24 out of the 29 marine protected sites, some evidence of recovery of sessile species was reported. Out of the 6 reports where changes in the benthic communities were not observed, two are from UK waters: the first documents an experimental closure in the Firth of Clyde, maerl did not recover during the 4 year duration of the study (Hall-Spencer & Moore, 2000b), while the second is from Lundy no-take zone where benthic communities were unchanged after 5 years (Hoskin et al., 2009), although they were not considered to have been impacted by the fishery (potting) that was excluded from the area. Other cited reasons for lack of recovery include local hydrographic regimes (Salomon et al., 2008) and also trophic changes leading to increases in predatory fish leading to decreased abundance of eipifaunal species (Moreno et al., 1986) (Table 14).

Table 14. Reports where no recovery was observed in sessile benthic communities Size of Years Location Closure type closure Response Comments Reference closed 2 (km ) Experimental Clyde, 70% reduction in maerl; no recovery (Hall-Spencer & trawl in a 0.175 Scotland 4 during 4 years of monitoring. Moore, 2000b) closed area Long bay - No difference in kelp and urchin This was attributed to the (Salomon et al., Okura, New MPA 9.8 abundance inside and outside the hydrographic characteristics of the 13 2008) Zealand closed. site. This was attributed to lack of No change in sessile epifauna in first 5 (Hoskin et al., Lundy, UK No-take zone 4 significant impact from potting fishery 5 years of monitoring. 2009) prior to closure. Mehuin, Reduced numbers of barnacles and This was due to increase in herbivores (Moreno et al., MPA 0.006 Chile 4 macroalgae. and carnivorous gastropods. 1986) Ninepin No change in the richness of Effectiveness of reserves correlated (Edgar & Barrett, point, No-take zone 0.59 6 1999) Tasmania invertebrates or plants. with reserve size. No change in the richness of Tinderbox, Effectiveness of reserves correlated (Edgar & Barrett, No-take zone 6 0.53 invertebrates or plants, but increasing Tasmania with reserve size. 1999) densities in large fish.

3.23 Out of the incidences of recovery it was not possible to explore any commonality of recovery patterns on a quantitative basis because of the low numbers of samples, different response variables measured and the enormous variability in the size, age and levels of protection in the protected areas. The latter ranged from complete protection (no take zones) to low levels of demersal trawling (approximately a quarter of the comparative areas outside the protected area, (Engel & Kvitek, 1998)) to areas that were protected from demersal mobile gear but off-bottom fishing took place along with static gears e.g. (Hermsen et al., 2003), (Bradshaw et al., 2000).

3.24 The sites that showed recovery within 1 to 5 years varied widely in their responses: increases in bull kelp density of 200-400% were observed in Las Cruces, Chile (Castilla & Duran, 1985), whereas on NE Georges Bank 3-fold increases were reported in tube-building worm species (Hermsen et al., 2003) indicating rapid recovery of species with high larval transport (Table 15). This is also reflected in highly dispersive mobile species like echinoderms that have been reported to recover to pre-disturbance levels after just one recruitment cycle (Robinson et al., 2001). Recovery was extremely fast for sponges and corals at one site following experimental trawling, no damage could be identified and the previous abundances were reported after just one year of the disturbance (Van Dolah et al., 1987).

Table 15 Reports where recovery was observed in sessile benthic communities between 1-5 years following the cessation of the physical disturbance Size of Years Location Closure type closure Response Reference closed 2 (km ) Experimental dredge Only short-term effects were monitored; differences in heavily fished and through area (Bradshaw et al., Isle of Man 5 2 unfished epifaunal communities were unexpected e.g. slow growing closed to 2000) Alcyonium digitatum present in abundance on high effort grounds. demersal gear Las Cruces, 2-4 fold increase in intertidal bull-kelp Durvillea antarctica density in the (Castilla & Duran, No-take zone 5 0.0482 Chile closed area as it is prohibited from harvest. 1985) Closed to Increased abundance of sessile benthic species: 75% Eunicella verrucosa, (Hiddink et al., Lyme Bay* demersal 1 44.59 60% Alcyonium digitatum, 100% Pentapora fascialis. 2007a) towed gear Lightly Monterey trawled area Bay Natural 7-fold difference in fragile anemones (Urticina spp.) and 9-fold difference in (Engel & Kvitek, (1/4 of 2 60,801 Marine sea pens (Ptilosarcus spp.0 compared with heavily trawled area. 1998) outside Sanctuary intensity) Closed to N Georges Increase in abundance of benthic fauna, especially fragile (e.g. Urticina demersal 4 (Collie et al., 2005) Bank spp.) in protected areas. towed gear Closed to NE Georges (Hermsen et al., demersal 5 10000 3-fold increases in tube-building worm (Thelepus cincinnatus) at deep sites. Bank 2003) towed gear Recovery St examined Catherine's following 0.005 (Van Dolah et al., Island, 31 Full recovery in terms of abundance of sponges and corals within 1 year. experimental (approx) 1987) Georgia, trawl in USA closed area

* Lyme Bay refers to the voluntary closure that pre-dates the statutory closure of 60 nm2 implemented in 2009

3.25 Recovery over larger timescales in some cases involved complete shifts in the structure of ecosystems from urchin barrens to macroagal dominated communities with the reversal of trophic cascades (i.e. recovery of predatory fish populations that control key herbivore groups such as echinoderms). This has been reported from Leigh, Poor Knights, Mokohinau, and Tawharanui marine reserves in New Zealand (Salomon et al., 2008), from Torre Guaceto Marine Reserve, Italy (Guidetti, 2007) and also from Anacapa Island Reserve, California (Behrens & Lafferty, 2004). These effects occurred over long time scales (8-33 years) and were from reserves that had complete protection. These dramatic effects mediated by species interactions are hard to predict and only possible when other fisheries activities are prohibited (including recreational) so that urchin predators (fish and crustacean) populations can recover.

Table 16 Reports where recovery was observed in sessile benthic communities over long time periods (>5 years) following the cessation of the physical disturbance

Size of Years Location Closure type closure Response Reference closed 2 (km ) Structurally complex taxa of colonial epifaunal (sponges, bushy CAII, Georges Closed to bryozoans) more abundant at undisturbed sites. It took at least 2 demersal (Asch & Collie, 2008) Bank, 6 years for these species to become established following closure. Non towed gear USA/Canada colonial species proved less responsive. Investigation on annual repeated trawling impacts on colonial Closed to E Scotian shelf epifauna but starting conditions vs trawled state show decrease in demersal (Henry et al., 2006) 10 13719.62 numbers of taxa and biomass, but effects not significant or lasting (4TVW) towed gear beyond one year. Gulf of Maine, Closed to Benthic communities were dominated by more disturbance intolerant, demersal 6 3772.894 sessile families compared with unprotected area which was (Knight, 2005) USA towed gear dominated by disturbance tolerant, opportunistic families. Maria island, Increase in invertebrate and algal abundances, increase in fish and (Edgar & Barrett, No-take zone 6 7 Tasmania invertebrate species richness and reduction in abalone and urchins. 1999) Poor knights & Significant increase in kelp abundance and reduction in sea urchin Mokohinau, MPA 8 18.9 density. Ecosystem effects in terms of increases in primary (Salomon et al., 2008) NZ productivity and reversal of trophic cascades. Experimental Punta date mussel Recovery faster in MPA with accelerated recolonisation. Some species did not recolonise though (e.g. Codium bursa, Tricleocarpa (Bevilacqua et al., campanella, fishing in 6 12 fragilis, chrondrsia reniformis, Hemimycale columella, Iricina 2006) Italy established variabilis, Petrosia ficiformis, Serpulorbis sp. Ciona intestinalis). No-take zone Western Gulf Closed to Sites dominated by gravel (which included boulders in some areas), of Maine demersal showed no effects on infauna but strong effects on epifaunal density (Grizzle et al., 2009) towed gear 7 515 closure area, and taxonomic richness between closed and open areas. USA and gillnets Anacapa Shifts to kelp dominated communities inside protected areas as (Behrens & Lafferty, urchins were reduced in density (also by disease), reversal of trophic Island 2004) Reserve, CA No-take zone 16 cascades where urchin predators are restored in protected areas.

Study focussed on red coral only. Demographic information on recovery taking around 30 years for this commercially important (Linares et al., 2010) Carry-le species. Larger, older and more dense colonies than in nearby Rouet, France MPA 22 0.85 unprotected sites. Study focussed on red coral only. Demographic information on Cerbere- recovery taking around 30 years for this commercially important (Linares et al., 2010) Banyuls, species. Larger, older and more dense colonies than in nearby France MPA 31 0.72 unprotected sites. Significant increase in kelp abundance and reduction in sea urchin Leigh marine density. Ecosystem effects in terms of increases in primary (Salomon et al., 2008) reserve, NZ MPA 33 5.26 productivity and reversal of trophic cascades. Study focussed on red coral only. Demographic information on recovery taking around 30 years for this commercially important (Linares et al., 2010) Scandola, species. Larger, older and more dense colonies than in nearby Corsica MPA 30 0.65 unprotected sites. Closed to SW UK trawling Area closed to demersal trawled gear (static gears only) had a significantly greater benthic community biomass, and it took >2 years (Blyth et al., 2004) (Brixham- (inshore for this to be distinct from the trawled areas. Plymouth) IPA potting area) 22 500 Significant increase in kelp abundance and reduction in sea urchin Tawharanui, density. Ecosystem effects in terms of increases in primary (Salomon et al., 2008) NZ MPA 24 5 productivity and reversal of trophic cascades. Torre Guaceto Significant increase in macroalgal abundance and reduction in sea (Guidetti, 2007) Marine urchin density Reserve, Italy No-take zone 11 22.2

3.3 Eunicella verrucosa distribution modelling in Lyme Bay

3.3.1 Environmental parameters

3.26 Despite differences between the model outputs, it is simple to identify which variables are contributing the most to the observed pattern in the data by way of examining the percentage of deviance explained (GLM/GAM) and examining jackknife plots (Maxent).

3.27 In the GAM/GLM modelling, substrate, current, minimum SST, mean Chl. A, the scour proxy and maximum Chl. A were found to be significant in order of descending amount of deviance explained. Substrate alone was included in the final model which was a GLM because there was no longer a need for smoothing terms. Rocky areas were associated with increased probability of pink sea fan presence and, in the GLM, were the only areas above the threshold. Current explained 29% deviance and predicted presence was predicted to be associated with currents up to approximately 0.5 m s-1.

3.28 In the Maxent model substrate and current are also the top two factors contributing most to the model. Minimum SPM, maximum Chl. A and depth were also listed as being in the top five factors (in order of descending contribution to the model) although this is only heuristically estimated and looking at the jacknife plots it appears that depth is more important than either minimum SPM or maximum Chl. A in terms of model gain. The plots produced during the Maxent model run indicate that increased current is associated with a reduced probability of presence and the same pattern follows for maximum Chl. A and minimum SPM.

3.29 Both models recognise Chlorophyll A as an important influence on the distribution of pink sea fans. The GAM with mean Chl. A showed lower concentrations (<1.6 mg m-3) to be associated with low probabilities. In Maxent, maximum Chl. A concentration was associated with increased probability between ca. 6-17 mg m-3 although exact values were difficult to determine from the graphical outputs.

3.30 Depth was found to be important in the Maxent model although it was found to be highly correlated to current in the GAM.

3.31 SPM featured as an important variable in both Maxent and a GLM (the latter as part of the scour proxy) suggesting that high levels of suspended inorganic particles are able to negatively influence its distribution. In Maxent, probability starts to decline sharply above ~0.213 g m-3.

3.3.2 Model performance

3.32 Figure 7, Figure 8 and Figure 9 illustrate areas predicted to be suitable for supporting pink sea fans from the GLM, Maxent and combined approaches respectively. Based on total area alone, the GLM predicted the highest

amount of suitable habitat (148.75 km2) although a small proportion of this can be attributed to the fact that because substrate alone was included in the final model, the original DWT substrate layer could be used. This layer covered a greater area than the area for which data for all variables was available. Predicted suitable area was reduced to 54.75 km2 and 68.84 km2 for the Maxent and combined approaches respectively.

Figure 7 Predicted distribution of suitable pink sea fan based on the final Generalized Linear Model with substrate as the only environmental parameter. The bright red areas are those areas above the threshold, that is, they are predicted to be suitable for pink sea fans. These are the rocky areas of Lyme Bay. The rock category encompassed ‘rock and mixed’ and ‘rock’ substrates. Although no distribution records were available for the ‘rock’ areas, they are assumed to be as suitable for pink sea fans as the ‘rock and mixed’ areas and so are also shaded bright red but hatched.

Figure 8 Predicted distribution of suitable pink sea fan based on the Maxent model including substrate, current, minimum SPM, maximum Chl. A, depth, minimum, Chl. A, maximum SPM, minimum SST and maximum SST as environmental parameters. The bright red areas are those areas above the threshold i.e. predicted to be suitable for pink sea fans. 43

Figure 9 Predicted distribution of suitable pink sea fan based on the average predicted values from the GLM and Maxent model. The bright red areas are those areas above the threshold i.e. predicted to be suitable for pink sea fans.

3.33 The performance of the three different modelling approaches is outlined in Table 17. Based on independent test data, no one modelling approach consistently outperformed the others although the GLM was the worst performer for three of the indicators. The combined approach is not considered further because it is not the highest performer for any of the performance indicators.

3.34 The GLM performed poorly with regard to specificity because of the high number of false positives. False positives, in modelling terms, occur when absence records in the validation data occur in areas which the model has predicted to be suitable for the species in question. Conversely, false negatives are positive sightings in areas the model deems unsuitable for the target species. The relatively high number of false negatives has contributed to the low sensitivity scores for the Maxent and combined approaches.

3.3.3 Historic fishing activity and long term monitoring

3.35 31.3% of the GLM false positives in the closed area coincided spatially with the scallop fishing buffer. An almost identical amount (31.6%) of false positives was located in the buffer zones for Maxent. The combined approach had the lowest number of false positives that were in the same area as past fishing activity (27.3%). Three of the 2008 monitoring station false positives were consistent between all three approaches (Figure 10).

Table 17 Performance of the three different modelling approaches based on internal re-sampling and independent test data. The combined prediction performance is based on the average prediction for each cell in the Lyme Bay grid. Performance indicator results to two decimal places.

Model GLM GLM GLM Maxent Maxent Maxent Combined (full (internal (test data) (full dataset) (internal (test data) prediction dataset) resampling) resampling) (test data) Factors Substrate, current, minimum SPM, maximum Chl. A, included in Substrate depth, minimum, Chl. A, maximum SPM, minimum SST, - final model maximum SST 0.86 ± CCR 0.86 0.59 - - 0.75 0.65 0.004 0.94 ± Sensitivity 0.94 0.87 - - 0.52 0.43 0.003 0.80 ± Specificity 0.81 0.40 - - 0.83 0.79 0.006 0.89 ± AUC 0.89 0.62 0.974 0.963 ± 0.009 0.76 0.74 0.003

Figure 10 Close up of the closed area with the Maxent predicted map. The large light blue circles show the potential footprint of 2005-2006 scallop dredging activity based on data from Devon Sea Fisheries Committee. The small dark blue points show the location of the towed video sampling stations for the current monitoring program. The three diamonds represent those monitoring stations that were false positives for both the Maxent and GLM modelling approaches that coincided spatially with past scallop dredging activity and are recommended for inclusion in extended monitoring program design.

3.3.4 Environmental parameters

3.36 The environmental parameters identified as having a significant influence on the distribution of pink sea fans within Lyme Bay are consistent with gorgonian ecological studies e.g. Barham & Davies (1968, Carpine & Grasshoff, 1975, Bryan & Metaxas, 2007). That substrate was found to be the most important predictor is to be expected and the availability of hard substratum is commonly reported as a limiting factor for coral species.

3.37 The importance of substrate in the GLM led to the exclusion of other significant parameters in the final model and, ultimately, affected model performance because the large areas of rocky substrate were not differentiated by the addition of supplementary environmental variables. However, the Maxent model included a significantly higher number of environmental parameters and was therefore able to better discriminate different areas within Lyme Bay. Areas of predicted suitability for the Maxent and combined model approaches were almost exclusively within the closed area. For the GLM, any rock area within the Bay was predicted to be suitable and amounted to more than double that of the suitable area predicted by Maxent or the combined approach. Furthermore, the total area predicted to be suitable for pink sea fans outside the closed area was only about 25% smaller than the total area predicted to be suitable inside it for the GLM (Figure 3).

3.3.5 Model performance

3.38 Table 17 highlights the varying performance of the different models. Although the GLM performed very well in internal validation, its performance based on test data was among the worst of the three modelling approaches and, essentially, it is a model‟s performance on independent test data, and hence robustness, that will inspire confidence in the application of the model to environmental management.

3.39 The Maxent model outperformed the other models in all but one of the model performance indicators and on face value one might use this model in preference to the GLM output to support spatial management measures. This is particularly true in relation to monitoring and future survey effort in the Bay because its high specificity score means that it is able to accurately predict absences the majority of the time and the rate of false presences is much lower than for GLM. However, its performance with regard to sensitivity was much lower than for the GLM and this could be costly in terms of conservation effort (Fielding & Bell, 1997) for the pink sea fan. The GLM‟s high sensitivity score is mostly likely due to its optimistic and over simplified prediction based on rocky areas alone but in this regard it is less likely to „miss‟ potentially suitable habitat for the pink sea fan.

3.40 In addition, Fielding & Bell (1997) reported that prevalence in the test data (that is, the ratio of presence records to total number of test records) can influence the predictive power of the model based on the performance indictor selected. In this study, the test data for Maxent had a much lower prevalence than the GLM test data (as a result of the absence records which were not used in the Maxent modelling process). Fielding & Bell (1997) found that although specificity and sensitivity were not affected by prevalence, several other performance indictors were (with performance sometimes increasing as prevalence decreased). The article did not state whether CCR was affected by prevalence but by removing a 48

proportion of the absence records from the test data for Maxent, the relative performance of Maxent‟s CCR compared to GLM and the combined approach remained the same.

3.41 With regard to the relevance of the predictions to current protection measures, the fact that the Maxent model only predicts a small amount of suitable habitat outside the current closed area would suggest that as far as the pink sea fan is concerned, and any other species that it might provide a proxy distribution for, the current closed area provides adequate protection. In contrast, the GLM output would suggest that the current closure only accounts for the protection of approximately half of suitable habitat for the pink sea fan and that additional protection might be required.

3.42 The concept of a combined approach was developed to see if this approach could capture the high sensitivity of the GLM and combine it with the potentially more realistic specificity of the Maxent output. In reality, however, the combined approach merely represented a midpoint between the two models in terms of performance and did not outperform both models on any one indicator.

4 Conclusions

4.1 Most of the 62 long-lived sessile benthic taxa from the reefs of Lyme Bay are either fragile or of intermediate fragility. These taxa are likely to show recovery from the physical disturbance of demersal fishing (while robust taxa, which are all algae species, may not show recovery because they are insensitive to this pressure).

4.2 The different life histories of the sessile benthic organisms are likely to mediate recovery. Species that can regenerate from fragments are likely to recover faster than those with a solitary habit (sponges, bryozoans and corals), but only if they have not suffered local extirpation. This has been noted in a number of cases in the literature with rapid recovery of colonial species e.g. (Van Dolah et al., 1987, Asch & Collie, 2008).

4.3 Recovery of taxa that cannot regenerate is likely to take longer since they are reliant on external supply of larvae and subsequent recruitment processes, which can be very variable, and in addition many are slow growing. Complete recovery has rarely been established, in part due to a lack of reference areas, which has confounded many experimental approaches (Engel & Kvitek, 1998), although work on single species exploring demographic structure has resolved this in some cases (Linares et al., 2010).

4.4 Species interactions may mediate recovery and produce unexpected and dramatic responses and are hard to predict. Dramatic shifts have been documented in the literature from urchin barrens to macroalgal dominated communities, reversing trophic cascades e.g. (Salomon et al., 2008). These can occur over extended timescales. Nevertheless, analogous changes of a similarly dramatic nature are unlikely at Lyme Bay because there is no conspicuously dominant grazer equivalent to the urchins, nor is there any obviously grazer-dominated habitat. Secondly, this type of interaction is more likely to be seen in no-take reserves, where predators can recover, whereas Lyme Bay has only seafloor protection.

4.5 There is a paucity of quantitative, comparable studies in the literature on which to predict recovery of sessile benthic communities. This highlights the importance of the monitoring work in Lyme Bay, not just to quantify patterns and rates of recovery of a priority UK habitat, and associated community that includes many species of conservation importance, but this has profound importance for adding to the global body of knowledge of reef systems and their recoverability from physical disturbance.

4.6 From the modelling work it is clear that all three approaches predicted that a large proportion of the closed area is suitable for supporting the pink sea fan. In reality, many of these areas are associated with absence records (false positives).

4.7 Previous studies have indicated that the recovery of pink sea and other rocky reef communities is likely to take longer than the three years this current monitoring program extends to e.g. Hoskin (2003). The purpose of overlaying the predictions with a spatial map of scalloping effort (Figure 10) before the closure (2005-2006) was to see if there was any overlap between areas the model predicted as

suitable, absence records and past fishing activity because this might suggest that historic scalloping activity had disturbed suitable habitat making it unsuitable in some of these areas. The subsequent removal of this pressure following the establishment of the closure would mean that these areas could support pink sea fans in the future and that these areas would therefore be of priority to include in future long term monitoring of the area if the current list of sites was to be streamlined.

4.8 It is important to note, however, that the spatial association with scallop activity explained less than a third of the false positives within the closed area and that other factors not captured within the current model must play an important role in the distribution of pink sea fans in Lyme Bay. Future improvements to the model might include a longer time period for the remotely sensed variables, as opposed to a single year‟s worth, to reflect the long lived nature of the pink sea fan. Incorporating wave exposure dynamics into the model might also help to explain some of the absences on the rocky areas in the closure, especially in the northeast of the closed area where there is anecdotal evidence to suggest that turbidity and sedimentation rates can be high following long periods of strong south westerly winds. An updated and improved substrate map based on better resolution data would be a welcome improvement to both modelling and conservation efforts within the Bay in general.

4.9 The method developed as part of this work has demonstrated that species distribution modelling has the potential to support the development of monitoring programs. However, the Maxent and GLM approach appear to represent, in this example, a trade off between specificity and sensitivity. In some management scenarios it would be straightforward to identify which trait was preferable to another and in this case it would be easy to recommend Maxent purely on the basis that it outperformed GLMs on three of the performance indicators including, more importantly in terms of monitoring effort, specificity. However, it is important to remember that sensitivity and specificity are both important indicators in a conservation management and spatial planning context which is why the sites recommended for inclusion into future monitoring are those that are already part of the monitoring program and that both modelling techniques agree on.

4.10 This work will be put into context with the findings from the first two years of monitoring in Lyme Bay to identify if the patterns reported here match those reported from the field work.

5 References

Asch R.G. & Collie J.S. (2008). Changes in a benthic megafaunal community due to disturbance from bottom fishing and the establishment of a fishery closure. Fishery Bulletin, 106, 438-456. Barham E.G. & Davies I.E. (1968). Gorgonians and water motion studies in Gulf of California. Underwater Naturalist, 5, 24-28. Behrens M.D. & Lafferty K.D. (2004). Effects of marine reserves and urchin disease on southern Californian rocky reef communities. Marine Ecology Progress Series, 279:, 129–139. Bevilacqua S., Terlizzi A., Fraschetti S., Russo G.F. & Boero F. (2006). Mitigating human disturbance: can protection influence trajectories of recovery in benthic assemblages? Journal of Animal Ecology, 75, 908-920. Black G. (2007). Lyme Bay pink sea fan survey 2006-07. Devon Biodiversity Records Centre, pp. Blyth R.E., Kaiser M.J., Edwards-Jones G. & Hart P.J.B. (2004). Implications of a zoned fishery management system for marine benthic communities. Ecology, 41, 951-961. Bradshaw C., Veale L.O., Hill A.S. & Brand A.R. (2000). The effects of scallop dredging on gravelly seabed communities. In M.J. Kaiser and S.J. De Groot (eds) Effects of fishing on non-target species and habitats, Oxford: Blackwell Science Limited, pp. 83-104. Bryan T.T. & Metaxas T.A. (2007). Predicting suitable habitat for deep-water gorgonian corals on the Atlantic and Pacific Continental Margins of North America. Marine Ecology Progress Series, 330, 113-126. Carpine C. & Grasshoff M. (1975). Les gorgonaires de la Mediterranee. Bulletin de l'Institute Oc\'eanographique, 71, 140. Castilla J.C. & Duran L.R. (1985). Human exclusion from the rocky intertidal zone of central Chile: The effects on Concholepas concholepas (Gastropoda). Oikos, 45, 391-399. Collie J.S., Hermsen J.M., Valentine P.C. & Almeida F.P. (2005). Effects of fishing on gravel habitats: assessment and recovery of benthic megafauna on Georges Bank. American Fisheries Society, [American Fisheries Society Symposium,325 pp. Connell J.H. & Keough M.J. (1985). Disturbance and patch dynamics of subtidal marine animals on hard substrata. In S.T.A. Pickett and P.S. White (eds) The Ecology of Natural Disturbance and Patch Dynamics, New York: Academic Press, pp. 125-151. Cramer J.S. (2003). Logit Models From Economics And Other Fields: Cambridge University Press. Edgar G.J. & Barrett N.S. (1999). Effects of the declaration of marine reserves on Tasmanian reef fishes, invertebrates and plants. Journal of Experimental Biology and Ecology, 242, 107-144. Engel J. & Kvitek R. (1998). Effects of otter trawling on a benthic community in Monterey Bay National Marine Sanctuary. Conservation Biology, 1204-1214. Fielding A.H. & Bell J.F. (1997). A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation, 24, 38-49. Grizzle R.E., Ward L.G., Mayer L.A., Malik M.A., Cooper A.B., Abeels H.A., Greene J.K., Brodeur M.A. & Rosenburg A.A. (2009). Effects of a large fishing closure on benthic communities in the western Gulf of Maine: recovery from the effects of gillnets and otter trawls. Fisheries Bulletin, 107, 308–317. 52

Guidetti P. (2007). Potential of marine reserves to cause community-wide changes beyond their boundaries. Conservation Biology, 21, 540-545. Hall-Spencer J.M. & Moore P.G. (2000a). Impact of scallop dredging on maerl grounds. In M.J.D.G. Kaiser, S.J. (ed) Effects of fishing on non-target species and habitats, Oxford: Blackwell Science Limited, pp. 105-117. Hall-Spencer J.M. & Moore P.G. (2000b). Scallop dredging has profound, long-term impacts on maerl habitats. ICES Journal of Marine Science, 57, 1407-1415. Halpern B.S. (2003). The impact of marine reserves: do reserves work and does reserve size matter? Ecological Applications, 13, S117-S137. Henry L.A., Kenchington E.L.R., Kenchington T.J., MacIsaac K.G., Bourbonnais- Boyce C. & Gordon D.C. (2006). Impacts of otter trawling on colonial epifaunal assemblages on a cobble bottom ecosystem on Western Bank (northwest Atlantic). Marine Ecology-Progress Series, 306, 63-78. Hermsen J., Collie J. & Valentine P. (2003). Mobile fishing gear reduces benthic megafaunal production on Georges Bank Marine Ecology Progress Series, 260, 97- 108. Hiddink J., Kaiser M.J., Hinz H. & Ridgeway A. (2007a). Quantification of epibenthic fauna in areas subjected to different regimes of scallop dredging activity in Lyme Bay, Devon. Bangor University, pp. Hiddink J.G., Jennings S. & Kaiser M.J. (2007b). Assessing and predicting the relative ecological impacts of disturbance on habitats with different sensitivities. Journal of Applied Ecology, 44, 405-413. Hoskin M. (2003). Effects of cessation of scallop-dredging on sessile macrofauna and scallops in Lyme Bay: Interim results and analyses for 2001 to 2003. MER Consultants Ltd, pp. Hoskin M.G., Coleman R.A. & von Carlshausen L. (2009). Ecological effects of the Lundy No-Take Zone: the first five years (2003-2007). Report to Natural England, DEFRA and WWF-UK. 160 pp. Jackson E.L., Langmead O., Barnes M., Tyler-Walters H. & Hiscock K. (2008). Identification of indicator species to represent the full range of benthic life history strategies for Lyme Bay and the consideration of the wider application for monitoring of Marine Protected Areas. Report to the Department of Environment, Food and Rural Affairs from the Marine Life Information Network (MarLIN). Defra Contract No. MB0101. Marine Biological Association of the UK Plymouth, 60 pp. Kaiser M.J., Clarke K.R., Hinz H., Austen M.C.V., Somerfield P.J. & Karakassis I. (2006). Global analysis of the response and recovery of benthic biota to fishing. Marine Ecology Progress Series, 3, 1-14. Knight E.P. (2005). The Effects of Trawling on Benthic Habitats: An Analysis of Recovery in the Western Gulf of Maine Closure. Master of Science, B.S. Salve Regina University. Linares C., Bianchimani O., Torrents O., Marschal C., Drap P. & Garrabou J. (2010). Marine Protected Areas and the conservation of long-lived marine invertebrates: the Mediterranean red coral. Marine Ecology Progress Series, 402, 69–79. Marshall, C.E., Glegg, G.A., Howell, K.L., Langston, B., Embling, C.B. & Stevens, T. (in prep). Using species distribution modelling and spatial fisheries data to inform monitoring of benthic recovery in a temperate marine protected area. Moreno C., Lunecke K. & Lépez M. (1986). The response of an intertidal Concholepas concholepas (Gastropoda) population to protection from Man in southern Chile and the effects on benthic sessile assemblages. Oikos, 46, 359-364.

Phillips S.J., Anderson R.P. & Schapire R.E. (2006). Maximum entropy modelling of species geographic distributions. Ecological Modelling, 190, 231-259. R Development Core Team (2009). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, pp. Robinson S.M.C., Bernier S. & MacIntyre A. (2001). The impact of scallop drags on sea urchin populations and benthos in the Bay of Fundy, Canada. Hydrobiologia, 465, 103-114. Salomon A.K., Shears N.T., Langlois T.J. & Babcock R.C. (2008). Cascading Effects of Fishing can Alter Carbon Flow Through a Temperate Coastal Ecosystem. Ecological Applications, 18, 1874-1887. Sewell J., Harris R., Hinz H., Votier S., & Hiscock K. (2007). An Assessment of the Impact of Selected Fishing Activities on European Marine Sites and a Review of Mitigation Measures. Report to the Seafish Industry Authority (Seafish). Marine Biological Association of the UK and the University of Plymouth, Plymouth, pp. http://www.seafish.org/pdf.pl?file=seafish/Documents/SR591.pdf Stevens T.F., Rodwell L., Beaumont K.L., Lewis T., Smith C. & Stehfest K.M. (2007). Surveys for Marine Spatial Planning in Lyme Bay. Marine Institute, University of Plymouth, pp. Turner S.J., Thrush S.F., Hewitt J.E., Cummings V.J. & Funnell G. (1999). Fishing impacts and the degradation or loss of habitat structure. Fisheries Management and Ecology, 6, 401-420. Van Dolah R.F., Wendt P.H. & Nicholson N. (1987). Effects of a research trawl on a hard-bottom assemblage of sponges and corals. Fisheries Research, 5, 39-54. Wood S.N. (2006). Generalized Additive Models. An Introduction with R, Boca Raton: Chapman & Hall. Wood S.N. (2008). Fast stable direct fitting and smoothness selection for generalized additive models. Journal of the Royal Statistical Society (B), 70, 495 - 518.

6 Appendices 6.1 Appendix 1: Lyme Bay benthic species list

Scientific Name Kingdom Phylum Class Order Family Clitellio arenarius Animalia Annelida Clitellata Haplotaxida Tubificidae Owenia fusiformis Animalia Annelida Polychaeta Sabellida Oweniidae Sabellaria alveolata Animalia Annelida Polychaeta Sabellida Sabellariidae Sabellaria spinulosa Animalia Annelida Polychaeta Sabellida Sabellariidae Bispira volutacornis Animalia Annelida Polychaeta Sabellida Sabellidae Bispira Animalia Annelida Polychaeta Sabellida Sabellidae Branchiomma bombyx Animalia Annelida Polychaeta Sabellida Sabellidae Chone duneri Animalia Annelida Polychaeta Sabellida Sabellidae Euchone rubrocincta Animalia Annelida Polychaeta Sabellida Sabellidae Megalomma vesiculosum Animalia Annelida Polychaeta Sabellida Sabellidae Megalomma Animalia Annelida Polychaeta Sabellida Sabellidae Myxicola infundibulum Animalia Annelida Polychaeta Sabellida Sabellidae Myxicola Animalia Annelida Polychaeta Sabellida Sabellidae Oriopsis armandi Animalia Annelida Polychaeta Sabellida Sabellidae Sabella pavonina Animalia Annelida Polychaeta Sabellida Sabellidae Sabellidae Animalia Annelida Polychaeta Sabellida Sabellidae Filograna implexa Animalia Annelida Polychaeta Sabellida Serpulidae Hydroides norvegicus Animalia Annelida Polychaeta Sabellida Serpulidae Pomatoceros lamarcki Animalia Annelida Polychaeta Sabellida Serpulidae Pomatoceros triqueter Animalia Annelida Polychaeta Sabellida Serpulidae Pomatoceros Animalia Annelida Polychaeta Sabellida Serpulidae Protula tubularia Animalia Annelida Polychaeta Sabellida Serpulidae Protula Animalia Annelida Polychaeta Sabellida Serpulidae Salmacina dysteri Animalia Annelida Polychaeta Sabellida Serpulidae Serpula vermicularis Animalia Annelida Polychaeta Sabellida Serpulidae Spirorbis (Spirorbis) tridentatus Animalia Annelida Polychaeta Sabellida Serpulidae Chaetopterus variopedatus Animalia Annelida Polychaeta Spionida Chaetopteridae Poecilochaetus serpens Animalia Annelida Polychaeta Spionida Poecilochaetidae Caulleriella alata Animalia Annelida Polychaeta Terebellida Cirratulidae Caulleriella zetlandica Animalia Annelida Polychaeta Terebellida Cirratulidae Chaetozone setosa Animalia Annelida Polychaeta Terebellida Cirratulidae Cirratulus cirratus Animalia Annelida Polychaeta Terebellida Cirratulidae Amphictene auricoma Animalia Annelida Polychaeta Terebellida Pectinariidae Lagis koreni Animalia Annelida Polychaeta Terebellida Pectinariidae 55

Scientific Name Kingdom Phylum Class Order Family Eupolymnia nebulosa Animalia Annelida Polychaeta Terebellida Terebellidae Lanice conchilega Animalia Annelida Polychaeta Terebellida Terebellidae Lanice Animalia Annelida Polychaeta Terebellida Terebellidae Pista cristata Animalia Annelida Polychaeta Terebellida Terebellidae Pista lornensis Animalia Annelida Polychaeta Terebellida Terebellidae Terebellides stroemi Animalia Annelida Polychaeta Terebellida Trichobranchidae Heteromastus filiformis Animalia Annelida Polychaeta Capitellidae Mediomastus fragilis Animalia Annelida Polychaeta Capitellidae Euclymene lumbricoides Animalia Annelida Polychaeta Maldanidae Euclymene oerstedi Animalia Annelida Polychaeta Maldanidae Praxillella affinis Animalia Annelida Polychaeta Maldanidae Aricidea (Acmira) catherinae Animalia Annelida Polychaeta Paraonidae Paradoneis lyra Animalia Annelida Polychaeta Paraonidae Scalibregma inflatum Animalia Annelida Polychaeta Scalibregmidae Elminius modestus Animalia Arthropoda Maxillopoda Sessilia Archaeobalanidae Semibalanus balanoides Animalia Arthropoda Maxillopoda Sessilia Archaeobalanidae Balanus crenatus Animalia Arthropoda Maxillopoda Sessilia Balanidae Balanus perforatus Animalia Arthropoda Maxillopoda Sessilia Balanidae Balanus Animalia Arthropoda Maxillopoda Sessilia Balanidae Chthamalus montagui Animalia Arthropoda Maxillopoda Sessilia Chthamalidae Chthamalus stellatus Animalia Arthropoda Maxillopoda Sessilia Chthamalidae Megatrema anglicum Animalia Arthropoda Maxillopoda Sessilia Pyrgomatidae Verruca stroemia Animalia Arthropoda Maxillopoda Sessilia Verrucidae Bicellariella ciliata Animalia Bryozoa Gymnolaemata Cheilostomatida Bicellariellidae Hippoporina pertusa Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Pentapora fascialis Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Pentapora Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Schizomavella auriculata Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Schizomavella linearis Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Schizomavella Animalia Bryozoa Gymnolaemata Cheilostomatida Bitectiporidae Porella concinna Animalia Bryozoa Gymnolaemata Cheilostomatida Bryocryptellidae Bugula angustiloba Animalia Bryozoa Gymnolaemata Cheilostomatida Bugulidae Bugula plumosa Animalia Bryozoa Gymnolaemata Cheilostomatida Bugulidae Bugula turbinata Animalia Bryozoa Gymnolaemata Cheilostomatida Bugulidae Bugula Animalia Bryozoa Gymnolaemata Cheilostomatida Bugulidae Caberea boryi Animalia Bryozoa Gymnolaemata Cheilostomatida Candidae Scrupocellaria scruposa Animalia Bryozoa Gymnolaemata Cheilostomatida Candidae Cellaria fistulosa Animalia Bryozoa Gymnolaemata Cheilostomatida Cellariidae Cellaria sinuosa Animalia Bryozoa Gymnolaemata Cheilostomatida Cellariidae

Scientific Name Kingdom Phylum Class Order Family Cellaria Animalia Bryozoa Gymnolaemata Cheilostomatida Cellariidae Cellepora pumicosa Animalia Bryozoa Gymnolaemata Cheilostomatida Celleporidae Cellepora Animalia Bryozoa Gymnolaemata Cheilostomatida Celleporidae Turbicellepora avicularis Animalia Bryozoa Gymnolaemata Cheilostomatida Celleporidae Cryptosula pallasiana Animalia Bryozoa Gymnolaemata Cheilostomatida Cryptosulidae Electra pilosa Animalia Bryozoa Gymnolaemata Cheilostomatida Electridae Chartella papyracea Animalia Bryozoa Gymnolaemata Cheilostomatida Flustridae Flustra foliacea Animalia Bryozoa Gymnolaemata Cheilostomatida Flustridae Securiflustra securifrons Animalia Bryozoa Gymnolaemata Cheilostomatida Flustridae Securiflustra Animalia Bryozoa Gymnolaemata Cheilostomatida Flustridae Celleporella hyalina Animalia Bryozoa Gymnolaemata Cheilostomatida Hippothoidae Membranipora membranacea Animalia Bryozoa Gymnolaemata Cheilostomatida Membraniporidae Escharella immersa Animalia Bryozoa Gymnolaemata Cheilostomatida Romancheinidae Escharella ventricosa Animalia Bryozoa Gymnolaemata Cheilostomatida Romancheinidae Parasmittina trispinosa Animalia Bryozoa Gymnolaemata Cheilostomatida Smittinidae Parasmittina Animalia Bryozoa Gymnolaemata Cheilostomatida Smittinidae Oshurkovia littoralis Animalia Bryozoa Gymnolaemata Cheilostomatida Umbonulidae Alcyonidium diaphanum Animalia Bryozoa Gymnolaemata Ctenostomatida Alcyonidiidae Alcyonidium gelatinosum Animalia Bryozoa Gymnolaemata Ctenostomatida Alcyonidiidae Alcyonidium Animalia Bryozoa Gymnolaemata Ctenostomatida Alcyonidiidae Flustrellidra hispida Animalia Bryozoa Gymnolaemata Ctenostomatida Flustrellidridae Vesicularia spinosa Animalia Bryozoa Gymnolaemata Ctenostomatida Vesiculariidae Crisia denticulata Animalia Bryozoa Stenolaemata Cyclostomatida Crisiidae Crisiidae Animalia Bryozoa Stenolaemata Cyclostomatida Crisiidae Disporella hispida Animalia Bryozoa Stenolaemata Cyclostomatida Lichenoporidae Disporella Animalia Bryozoa Stenolaemata Cyclostomatida Lichenoporidae Ciona intestinalis Animalia Chordata Ascidiacea Aplousobranchia Cionidae Ciona Animalia Chordata Ascidiacea Aplousobranchia Cionidae Clavelina lepadiformis Animalia Chordata Ascidiacea Aplousobranchia Clavelinidae Clavelina Animalia Chordata Ascidiacea Aplousobranchia Clavelinidae Didemnum coriaceum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Didemnum maculatum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Didemnum maculosum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Diplosoma listerianum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Diplosoma spongiforme Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Diplosoma Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Lissoclinum perforatum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Lissoclinum Animalia Chordata Ascidiacea Aplousobranchia Didemnidae

Scientific Name Kingdom Phylum Class Order Family Didemnidae Animalia Chordata Ascidiacea Aplousobranchia Didemnidae Distaplia rosea Animalia Chordata Ascidiacea Aplousobranchia Holozoidae Aplidium elegans Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Aplidium glabrum Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Aplidium proliferum Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Aplidium punctum Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Aplidium Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Morchellium argus Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Morchellium Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Polyclinum aurantium Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Aplidium turbinatum Animalia Chordata Ascidiacea Aplousobranchia Polyclinidae Pycnoclavella aurilucens Animalia Chordata Ascidiacea Aplousobranchia Pycnoclavellidae Pycnoclavella producta Animalia Chordata Ascidiacea Aplousobranchia Pycnoclavellidae Ascidia conchilega Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Ascidia mentula Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Ascidia virginea Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Ascidia Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Ascidiella aspersa Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Ascidiella scabra Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Phallusia mammillata Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Phallusia Animalia Chordata Ascidiacea Phlebobranchia Ascidiidae Corella parallelogramma Animalia Chordata Ascidiacea Phlebobranchia Corellidae Corella Animalia Chordata Ascidiacea Phlebobranchia Corellidae Molgula manhattensis Animalia Chordata Ascidiacea Stolidobranchia Molgulidae Molgula occulta Animalia Chordata Ascidiacea Stolidobranchia Molgulidae Molgula Animalia Chordata Ascidiacea Stolidobranchia Molgulidae Pyura microcosmus Animalia Chordata Ascidiacea Stolidobranchia Pyuridae Botrylloides leachii Animalia Chordata Ascidiacea Stolidobranchia Styelidae Botryllus schlosseri Animalia Chordata Ascidiacea Stolidobranchia Styelidae Botryllus Animalia Chordata Ascidiacea Stolidobranchia Styelidae Dendrodoa grossularia Animalia Chordata Ascidiacea Stolidobranchia Styelidae Distomus variolosus Animalia Chordata Ascidiacea Stolidobranchia Styelidae Polycarpa pomaria Animalia Chordata Ascidiacea Stolidobranchia Styelidae Polycarpa scuba Animalia Chordata Ascidiacea Stolidobranchia Styelidae Polycarpa Animalia Chordata Ascidiacea Stolidobranchia Styelidae Stolonica socialis Animalia Chordata Ascidiacea Stolidobranchia Styelidae Stolonica Animalia Chordata Ascidiacea Stolidobranchia Styelidae Styela clava Animalia Chordata Ascidiacea Stolidobranchia Styelidae Styela coriacea Animalia Chordata Ascidiacea Stolidobranchia Styelidae

Scientific Name Kingdom Phylum Class Order Family Branchiostoma lanceolatum Animalia Chordata Leptocardii Branchiostomidae Actinia equina Animalia Cnidaria Anthozoa Actiniaria Actiniidae Actinia fragacea Animalia Cnidaria Anthozoa Actiniaria Actiniidae Anemonia viridis Animalia Cnidaria Anthozoa Actiniaria Actiniidae Aulactinia verrucosa Animalia Cnidaria Anthozoa Actiniaria Actiniidae Urticina eques Animalia Cnidaria Anthozoa Actiniaria Actiniidae Urticina felina Animalia Cnidaria Anthozoa Actiniaria Actiniidae Aiptasia mutabilis Animalia Cnidaria Anthozoa Actiniaria Aiptasiidae Aiptasia Animalia Cnidaria Anthozoa Actiniaria Aiptasiidae Capnea sanguinea Animalia Cnidaria Anthozoa Actiniaria Aurelianiidae Scolanthus callimorphus Animalia Cnidaria Anthozoa Actiniaria Edwardsiidae Peachia cylindrica Animalia Cnidaria Anthozoa Actiniaria Haloclavidae Adamsia carciniopados Animalia Cnidaria Anthozoa Actiniaria Hormathiidae Amphianthus dohrnii Animalia Cnidaria Anthozoa Actiniaria Hormathiidae Calliactis parasitica Animalia Cnidaria Anthozoa Actiniaria Hormathiidae Calliactis Animalia Cnidaria Anthozoa Actiniaria Hormathiidae Metridium senile Animalia Cnidaria Anthozoa Actiniaria Metridiidae Actinothoe sphyrodeta Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Cereus pedunculatus Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Sagartia elegans Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Sagartia troglodytes Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Sagartia Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Sagartiogeton laceratus Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Sagartiogeton undatus Animalia Cnidaria Anthozoa Actiniaria Sagartiidae Alcyonium digitatum Animalia Cnidaria Anthozoa Alcyonacea Alcyoniidae Alcyonium glomeratum Animalia Cnidaria Anthozoa Alcyonacea Alcyoniidae Alcyonium Animalia Cnidaria Anthozoa Alcyonacea Alcyoniidae Parerythropodium coralloides Animalia Cnidaria Anthozoa Alcyonacea Alcyoniidae Sarcodictyon roseum Animalia Cnidaria Anthozoa Alcyonacea Clavulariidae Sarcodictyon Animalia Cnidaria Anthozoa Alcyonacea Clavulariidae Eunicella verrucosa Animalia Cnidaria Anthozoa Alcyonacea Gorgoniidae Eunicella Animalia Cnidaria Anthozoa Alcyonacea Gorgoniidae Cerianthus lloydii Animalia Cnidaria Anthozoa Ceriantharia Cerianthidae Cerianthus Animalia Cnidaria Anthozoa Ceriantharia Cerianthidae Corynactis viridis Animalia Cnidaria Anthozoa Corallimorpharia Corallimorphidae Caryophyllia (Caryophyllia) inornata Animalia Cnidaria Anthozoa Scleractinia Caryophylliidae Caryophyllia (Caryophyllia) Animalia Cnidaria Anthozoa Scleractinia Caryophylliidae

Scientific Name Kingdom Phylum Class Order Family smithii Caryophyllia Animalia Cnidaria Anthozoa Scleractinia Caryophylliidae Hoplangia durotrix Animalia Cnidaria Anthozoa Scleractinia Caryophylliidae Leptopsammia pruvoti Animalia Cnidaria Anthozoa Scleractinia Dendrophylliidae Epizoanthus couchii Animalia Cnidaria Anthozoa Zoanthidea Epizoanthidae Epizoanthus Animalia Cnidaria Anthozoa Zoanthidea Epizoanthidae Isozoanthus sulcatus Animalia Cnidaria Anthozoa Zoanthidea Parazoanthidae Isozoanthus Animalia Cnidaria Anthozoa Zoanthidea Parazoanthidae Parazoanthus axinellae Animalia Cnidaria Anthozoa Zoanthidea Parazoanthidae Bougainvillia muscus Animalia Cnidaria Hydrozoa Anthoathecata Bougainvilliidae Corymorpha nutans Animalia Cnidaria Hydrozoa Anthoathecata Corymorphidae Hydractinia echinata Animalia Cnidaria Hydrozoa Anthoathecata Hydractiniidae Tubularia indivisa Animalia Cnidaria Hydrozoa Anthoathecata Tubulariidae Aglaophenia parvula Animalia Cnidaria Hydrozoa Leptothecata Aglaopheniidae Aglaophenia pluma Animalia Cnidaria Hydrozoa Leptothecata Aglaopheniidae Aglaophenia tubulifera Animalia Cnidaria Hydrozoa Leptothecata Aglaopheniidae Gymnangium montagui Animalia Cnidaria Hydrozoa Leptothecata Aglaopheniidae Campanularia hincksii Animalia Cnidaria Hydrozoa Leptothecata Campanulariidae Obelia geniculata Animalia Cnidaria Hydrozoa Leptothecata Campanulariidae Obelia longissima Animalia Cnidaria Hydrozoa Leptothecata Campanulariidae Halecium beanii Animalia Cnidaria Hydrozoa Leptothecata Haleciidae Halecium halecinum Animalia Cnidaria Hydrozoa Leptothecata Haleciidae Halecium Animalia Cnidaria Hydrozoa Leptothecata Haleciidae Antennella secundaria Animalia Cnidaria Hydrozoa Leptothecata Halopterididae Kirchenpaueria pinnata Animalia Cnidaria Hydrozoa Leptothecata Kirchenpaueriidae Kirchenpaueria Animalia Cnidaria Hydrozoa Leptothecata Kirchenpaueriidae Lafoea dumosa Animalia Cnidaria Hydrozoa Leptothecata Lafoeidae Nemertesia antennina Animalia Cnidaria Hydrozoa Leptothecata Plumulariidae Nemertesia ramosa Animalia Cnidaria Hydrozoa Leptothecata Plumulariidae Plumularia setacea Animalia Cnidaria Hydrozoa Leptothecata Plumulariidae Abietinaria abietina Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Abietinaria Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Amphisbetia operculata Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Diphasia attenuata Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Dynamena pumila Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Hydrallmania falcata Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Hydrallmania Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Sertularella mediterranea Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Sertularella gayi Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae

Scientific Name Kingdom Phylum Class Order Family Sertularella polyzonias Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Sertularella Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Sertularia argentea Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Sertularia cupressina Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Thuiaria thuja Animalia Cnidaria Hydrozoa Leptothecata Sertulariidae Gastrochaena dubia Animalia Mollusca Bivalvia [unassigned] Gastrochaenidae Eufistulana Animalia Mollusca Bivalvia [unassigned] Gastrochaenidae Hiatella arctica Animalia Mollusca Bivalvia [unassigned] Hiatellidae Modiolarca subpicta Animalia Mollusca Bivalvia Mytiloida Mytilidae Modiolus modiolus Animalia Mollusca Bivalvia Mytiloida Mytilidae Musculus discors Animalia Mollusca Bivalvia Mytiloida Mytilidae Mytilus edulis Animalia Mollusca Bivalvia Mytiloida Mytilidae Anomia ephippium Animalia Mollusca Bivalvia Pectinoida Anomiidae Pododesmus patelliformis Animalia Mollusca Bivalvia Pectinoida Anomiidae Anomiidae Animalia Mollusca Bivalvia Pectinoida Anomiidae Mimachlamys varia Animalia Mollusca Bivalvia Pectinoida Pectinidae Hemilepton nitidum Animalia Mollusca Bivalvia Veneroida Lasaeidae Kurtiella bidentata Animalia Mollusca Bivalvia Veneroida Montacutidae Clausinella fasciata Animalia Mollusca Bivalvia Veneroida Veneridae Timoclea ovata Animalia Mollusca Bivalvia Veneroida Veneridae Venerupis senegalensis Animalia Mollusca Bivalvia Veneroida Veneridae Veneridae Animalia Mollusca Bivalvia Veneroida Veneridae Antalis vulgaris Animalia Mollusca Scaphopoda Dentaliida Dentaliidae Phoronis hippocrepia Animalia Phoronida Phoronida Leuconia nivea Animalia Porifera Calcarea Baerida Baeriidae Clathrina clathrus Animalia Porifera Calcarea Clathrinida Clathrinidae Clathrina coriacea Animalia Porifera Calcarea Clathrinida Clathrinidae Clathrina Animalia Porifera Calcarea Clathrinida Clathrinidae Grantia compressa Animalia Porifera Calcarea Leucosolenida Grantiidae Leucosolenia botryoides Animalia Porifera Calcarea Leucosolenida Leucosoleniidae Leucosolenia complicata Animalia Porifera Calcarea Leucosolenida Leucosoleniidae Sycon ciliatum Animalia Porifera Calcarea Leucosolenida Sycettidae Sycon Animalia Porifera Calcarea Leucosolenida Sycettidae Pachymatisma johnstonia Animalia Porifera Demospongiae Astrophorida Geodiidae Pachymatisma Animalia Porifera Demospongiae Astrophorida Geodiidae Dercitus bucklandi Animalia Porifera Demospongiae Astrophorida Pachastrellidae Thymosia guernei Animalia Porifera Demospongiae Chondrosida Chondrillidae Aplysilla rosea Animalia Porifera Demospongiae Dendroceratida Darwinellidae Aplysilla sulfurea Animalia Porifera Demospongiae Dendroceratida Darwinellidae

Scientific Name Kingdom Phylum Class Order Family Dysidea fragilis Animalia Porifera Demospongiae Dictyoceratida Dysideidae Dysidea pallescens Animalia Porifera Demospongiae Dictyoceratida Dysideidae Dysidea Animalia Porifera Demospongiae Dictyoceratida Dysideidae Cliona celata Animalia Porifera Demospongiae Hadromerida Clionaidae Adreus fascicularis Animalia Porifera Demospongiae Hadromerida Hemiasterellidae Stelligera rigida Animalia Porifera Demospongiae Hadromerida Hemiasterellidae Stelligera stuposa Animalia Porifera Demospongiae Hadromerida Hemiasterellidae Stelligera Animalia Porifera Demospongiae Hadromerida Hemiasterellidae Polymastia boletiformis Animalia Porifera Demospongiae Hadromerida Polymastiidae Polymastia mamillaris Animalia Porifera Demospongiae Hadromerida Polymastiidae Polymastia penicillus Animalia Porifera Demospongiae Hadromerida Polymastiidae Polymastia Animalia Porifera Demospongiae Hadromerida Polymastiidae Pseudosuberites sulphureus Animalia Porifera Demospongiae Hadromerida Suberitidae Suberites carnosus Animalia Porifera Demospongiae Hadromerida Suberitidae Suberites domuncula Animalia Porifera Demospongiae Hadromerida Suberitidae Suberites ficus Animalia Porifera Demospongiae Hadromerida Suberitidae Suberites pagurorum Animalia Porifera Demospongiae Hadromerida Suberitidae Suberites Animalia Porifera Demospongiae Hadromerida Suberitidae Terpios fugax Animalia Porifera Demospongiae Hadromerida Suberitidae Terpios gelatinosa Animalia Porifera Demospongiae Hadromerida Suberitidae Tethya aurantium Animalia Porifera Demospongiae Hadromerida Tethyidae Tethya Animalia Porifera Demospongiae Hadromerida Tethyidae Axinella damicornis Animalia Porifera Demospongiae Halichondrida Axinellidae Axinella dissimilis Animalia Porifera Demospongiae Halichondrida Axinellidae Tethyspira spinosa Animalia Porifera Demospongiae Halichondrida Dictyonellidae Ciocalypta penicillus Animalia Porifera Demospongiae Halichondrida Halichondriidae Ciocalypta Animalia Porifera Demospongiae Halichondrida Halichondriidae Halichondria (Halichondria) bowerbanki Animalia Porifera Demospongiae Halichondrida Halichondriidae Halichondria (Halichondria) panicea Animalia Porifera Demospongiae Halichondrida Halichondriidae Hymeniacidon perlevis Animalia Porifera Demospongiae Halichondrida Halichondriidae Hymeniacidon kitchingi Animalia Porifera Demospongiae Halichondrida Halichondriidae Halisarca dujardini Animalia Porifera Demospongiae Halisarcida Halisarcidae Haliclona (Reniera) cinerea Animalia Porifera Demospongiae Haplosclerida Chalinidae Haliclona (Halichoclona) fistulosa Animalia Porifera Demospongiae Haplosclerida Chalinidae Haliclona (Haliclona) oculata Animalia Porifera Demospongiae Haplosclerida Chalinidae Haliclona (Haliclona) Animalia Porifera Demospongiae Haplosclerida Chalinidae

Scientific Name Kingdom Phylum Class Order Family simulans Haliclona (Haliclona) urceolus Animalia Porifera Demospongiae Haplosclerida Chalinidae Haliclona (Rhizoniera) viscosa Animalia Porifera Demospongiae Haplosclerida Chalinidae Haliclona Animalia Porifera Demospongiae Haplosclerida Chalinidae Oscarella lobularis Animalia Porifera Demospongiae Homosclerophorida Plakinidae Iophon hyndmani Animalia Porifera Demospongiae Poecilosclerida Acarnidae Iophon Animalia Porifera Demospongiae Poecilosclerida Acarnidae Iophon nigricans Animalia Porifera Demospongiae Poecilosclerida Acarnidae Amphilectus fucorum Animalia Porifera Demospongiae Poecilosclerida Esperiopsidae Amphilectus Animalia Porifera Demospongiae Poecilosclerida Esperiopsidae Esperiopsis Animalia Porifera Demospongiae Poecilosclerida Esperiopsidae Ulosa stuposa Animalia Porifera Demospongiae Poecilosclerida Esperiopsidae Hemimycale columella Animalia Porifera Demospongiae Poecilosclerida Hymedesmiidae Hemimycale Animalia Porifera Demospongiae Poecilosclerida Hymedesmiidae Hymedesmia (Hymedesmia) paupertas Animalia Porifera Demospongiae Poecilosclerida Hymedesmiidae Phorbas fictitius Animalia Porifera Demospongiae Poecilosclerida Hymedesmiidae Phorbas plumosus Animalia Porifera Demospongiae Poecilosclerida Hymedesmiidae Clathria (Microciona) atrasanguinea Animalia Porifera Demospongiae Poecilosclerida Microcionidae Ophlitaspongia papilla Animalia Porifera Demospongiae Poecilosclerida Microcionidae Myxilla (Myxilla) fimbriata Animalia Porifera Demospongiae Poecilosclerida Myxillidae Myxilla (Myxilla) incrustans Animalia Porifera Demospongiae Poecilosclerida Myxillidae Raspailia (Clathriodendron) hispida Animalia Porifera Demospongiae Poecilosclerida Raspailiidae Raspailia (Raspailia) ramosa Animalia Porifera Demospongiae Poecilosclerida Raspailiidae Raspailia Animalia Porifera Demospongiae Poecilosclerida Raspailiidae Phascolion (Phascolion) strombus Animalia Sipuncula Sipunculidea Golfingiida Phascolionidae Desmarestia aculeata Chromista Heterokontophyta Phaeophyceae Desmarestiales Desmarestiaceae Desmarestia ligulata Chromista Heterokontophyta Phaeophyceae Desmarestiales Desmarestiaceae Desmarestia viridis Chromista Heterokontophyta Phaeophyceae Desmarestiales Desmarestiaceae Dictyopteris polypodioides Chromista Heterokontophyta Phaeophyceae Dictyotales Dictyotaceae Dictyota dichotoma Chromista Heterokontophyta Phaeophyceae Dictyotales Dictyotaceae Taonia atomaria Chromista Heterokontophyta Phaeophyceae Dictyotales Dictyotaceae Fucus serratus Chromista Heterokontophyta Phaeophyceae Fucales Fucaceae Fucus spiralis Chromista Heterokontophyta Phaeophyceae Fucales Fucaceae

Scientific Name Kingdom Phylum Class Order Family Fucus vesiculosus Chromista Heterokontophyta Phaeophyceae Fucales Fucaceae Halidrys siliquosa Chromista Heterokontophyta Phaeophyceae Fucales Sargassaceae Laminaria digitata Chromista Heterokontophyta Phaeophyceae Laminariales Laminariaceae Laminaria hyperborea Chromista Heterokontophyta Phaeophyceae Laminariales Laminariaceae Saccharina latissima Chromista Heterokontophyta Phaeophyceae Laminariales Laminariaceae Scytosiphon lomentaria Chromista Heterokontophyta Phaeophyceae Scytosiphonales Scytosiphonaceae Cladostephus spongiosus Chromista Heterokontophyta Phaeophyceae Sphacelariales Sphacelariaceae Saccorhiza polyschides Chromista Heterokontophyta Phaeophyceae Tilopteridales Phyllariaceae Chaetopteris Chromista Ochrophyta Ochrophyta Bryopsis plumosa Plantae Chlorophyta Bryopsidophyceae Bryopsidales Bryopsidaceae Cladophora pellucida Plantae Chlorophyta Ulvophyceae Cladophorales Cladophoraceae Ulva linza Plantae Chlorophyta Ulvophyceae Ulvales Ulvaceae Porphyra leucosticta Plantae Rhodophyta Bangiophyceae Bangiales Bangiaceae Schmitziella endophloea Plantae Rhodophyta Florideophyceae Acrochaetiales Acrochaetiaceae Asparagopsis armata Plantae Rhodophyta Florideophyceae Bonnemaisoniales Bonnemaisoniaceae Bonnemaisonia asparagoides Plantae Rhodophyta Florideophyceae Bonnemaisoniales Bonnemaisoniaceae Aglaothamnion byssoides Plantae Rhodophyta Florideophyceae Ceramiales Callithamniaceae Aglaothamnion hookeri Plantae Rhodophyta Florideophyceae Ceramiales Callithamniaceae Callithamnion tetricum Plantae Rhodophyta Florideophyceae Ceramiales Callithamniaceae Antithamnionella ternifolia Plantae Rhodophyta Florideophyceae Ceramiales Ceramiaceae Ceramium ciliatum Plantae Rhodophyta Florideophyceae Ceramiales Ceramiaceae Ceramium gaditanum Plantae Rhodophyta Florideophyceae Ceramiales Ceramiaceae Ceramium shuttleworthianum Plantae Rhodophyta Florideophyceae Ceramiales Ceramiaceae Pterothamnion plumula Plantae Rhodophyta Florideophyceae Ceramiales Ceramiaceae Heterosiphonia plumosa Plantae Rhodophyta Florideophyceae Ceramiales Dasyaceae Acrosorium venulosum Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Apoglossum ruscifolium Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Cryptopleura ramosa Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Delesseria sanguinea Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Drachiella heterocarpa Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Drachiella spectabilis Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Erythroglossum laciniatum Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Hypoglossum hypoglossoides Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Membranoptera alata Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Phycodrys rubens Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae Polyneura bonnemaisonii Plantae Rhodophyta Florideophyceae Ceramiales Delesseriaceae

Scientific Name Kingdom Phylum Class Order Family Brongniartella byssoides Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Chondria dasyphylla Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Osmundea hybrida Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Osmundea pinnatifida Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Polysiphonia elongata Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Polysiphonia nigra Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Polysiphonia stricta Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Polysiphonia Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Rhodomela confervoides Plantae Rhodophyta Florideophyceae Ceramiales Rhodomelaceae Compsothamnion thuyoides Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Griffithsia corallinoides Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Halurus flosculosus Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Pleonosporium borreri Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Plumaria plumosa Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Ptilota gunneri Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Sphondylothamnion multifidum Plantae Rhodophyta Florideophyceae Ceramiales Wrangeliaceae Corallina officinalis Plantae Rhodophyta Florideophyceae Corallinales Corallinaceae Lithophyllum incrustans Plantae Rhodophyta Florideophyceae Corallinales Corallinaceae Lithophyllum Plantae Rhodophyta Florideophyceae Corallinales Corallinaceae Phymatolithon calcareum Plantae Rhodophyta Florideophyceae Corallinales Hapalidiaceae Gelidium spinosum Plantae Rhodophyta Florideophyceae Gelidiales Gelidiaceae Gelidium pusillum Plantae Rhodophyta Florideophyceae Gelidiales Gelidiaceae Catenella caespitosa Plantae Rhodophyta Florideophyceae Gigartinales Caulacanthaceae Calliblepharis ciliata Plantae Rhodophyta Florideophyceae Gigartinales Cystocloniaceae Calliblepharis jubata Plantae Rhodophyta Florideophyceae Gigartinales Cystocloniaceae Cystoclonium purpureum Plantae Rhodophyta Florideophyceae Gigartinales Cystocloniaceae Rhodophyllis divaricata Plantae Rhodophyta Florideophyceae Gigartinales Cystocloniaceae Dilsea carnosa Plantae Rhodophyta Florideophyceae Gigartinales Dumontiaceae Dumontia contorta Plantae Rhodophyta Florideophyceae Gigartinales Dumontiaceae Chondracanthus acicularis Plantae Rhodophyta Florideophyceae Gigartinales Gigartinaceae Chondrus crispus Plantae Rhodophyta Florideophyceae Gigartinales Gigartinaceae Callophyllis laciniata Plantae Rhodophyta Florideophyceae Gigartinales Kallymeniaceae Kallymenia reniformis Plantae Rhodophyta Florideophyceae Gigartinales Kallymeniaceae Meredithia microphylla Plantae Rhodophyta Florideophyceae Gigartinales Kallymeniaceae Erythrodermis traillii Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae Mastocarpus stellatus Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae Phyllophora crispa Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae Phyllophora Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae

Scientific Name Kingdom Phylum Class Order Family pseudoceranoides Schottera nicaeensis Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae Stenogramme interrupta Plantae Rhodophyta Florideophyceae Gigartinales Phyllophoraceae Polyides rotundus Plantae Rhodophyta Florideophyceae Gigartinales Polyidaceae Hildenbrandia rubra Plantae Rhodophyta Florideophyceae Hildenbrandiales Hildenbrandiaceae Scinaia furcellata Plantae Rhodophyta Florideophyceae Nemaliales Scinaiaceae Schizymenia dubyi Plantae Rhodophyta Florideophyceae Nemastomatales Schizymeniaceae Palmaria palmata Plantae Rhodophyta Florideophyceae Palmariales Palmariaceae Plocamium cartilagineum Plantae Rhodophyta Florideophyceae Plocamiales Plocamiaceae Lomentaria articulata Plantae Rhodophyta Florideophyceae Rhodymeniales Lomentariaceae Lomentaria clavellosa Plantae Rhodophyta Florideophyceae Rhodymeniales Lomentariaceae Lomentaria orcadensis Plantae Rhodophyta Florideophyceae Rhodymeniales Lomentariaceae Cordylecladia erecta Plantae Rhodophyta Florideophyceae Rhodymeniales Rhodymeniaceae Rhodymenia ardissonei Plantae Rhodophyta Florideophyceae Rhodymeniales Rhodymeniaceae Rhodymenia delicatula Plantae Rhodophyta Florideophyceae Rhodymeniales Rhodymeniaceae Rhodymenia holmesii Plantae Rhodophyta Florideophyceae Rhodymeniales Rhodymeniaceae Rhodymenia pseudopalmata Plantae Rhodophyta Florideophyceae Rhodymeniales Rhodymeniaceae