ECOLOGICAL RISK ASSESSMENT

Transcription

1 ECOLOGICAL RISK ASSESSMENT as a Prioritization Too to Support Caifornia Fisheries Management Caifornia Ocean Science Trust and the Nationa Oceanic and Atmospheric Administration OCTOBER 2017

2 Contributors Authors Ocean Science Trust Errin Ramanujam, Hayey Carter NOAA Fisheries Jamea Samhouri, Joe Bizzarro About the Contributors Caifornia Ocean Science Trust Ocean Science Trust (OST) is an independent, non-profit organization that brings together governments, scientists, and citizens to buid trust and understanding in ocean and coasta science issues. We empower participation in the decisions that are shaping the future of our oceans. We were estabished by the Caifornia Ocean Resources Stewardship Act to support managers and poicymakers with sound science. For more information, visit our website at Nationa Oceanic and Atmospheric Administration The Nationa Oceanic and Atmospheric Administration (NOAA) is an agency that enriches ife through science. Our reach goes from the surface of the sun to the depths of the ocean foor as we work to keep citizens informed of the changing environment around them. From daiy weather forecasts, severe storm warnings, and cimate monitoring to fisheries management, coasta restoration and supporting marine commerce, NOAA s products and services support economic vitaity and affect more than one-third of America s gross domestic product. NOAA s dedicated scientists use cutting-edge research and high-tech instrumentation to provide citizens, panners, emergency managers and other decision makers with reiabe information they need when they need it. Project Team The project team incuded Caifornia Ocean Science Trust, the Caifornia Department of Fish and Widife (CDFW), NOAA Fisheries, and the Caifornia Ocean Protection Counci (OPC). Funding This project was funded by OPC with additiona support from Resources Legacy Fund and in kind support from OST and NOAA Fisheries. Acknowedgements OST woud aso ike to generousy thank the foowing CDFW fishery experts who participated in the piot project: Pau Reiy, Heather Giniak, Travis Buck, Marty Gingras, and Ryan Barting. We aso thank the foowing experts who were consuted during the project: Aistair Hobday, Rick Fetcher, Rebecca Martone, Phi Levin, John Fied, Chris Costeo, Steve Gaines, Mark Neson, Jim Hastie, Rod Fujita, and Wiow Bautista. Contact: Errin Ramanujam, Senior Scientist, Ocean Science Trust, errin.ramanujam@oceansciencetrust.org Suggested Citation: Ramanujam, E., Samhouri, J., Bizzarro, J., and Carter, H Ecoogica Risk Assessment as a Prioritization Too to Support Caifornia Fisheries Management. Oakand, Caifornia, USA. Cover image credit: Don DeBod i

3 Tabe of Contents About the Report... iv Highights and Key Findings...v Executive Summary...I 1. Introduction Overview Why Risk Assessment? Ecoogica Risk Assessment Approach Why These Two Risk Assessment Approaches? Approach 1: Appy an Existing ERA (NOAA PSA) Approach 2: Customize an ERA for Caifornia (Samhouri and Levin, 2012) Ecoogica Risk Assessment Piot Methods Customizing the ERA for Caifornia Fisheries Resuts Target Resuts Bycatch Resuts Habitat Resuts Next steps Key Decision Points for Too Refinement and impementation Key Decision Points for PSA vs. ERA for Target Fisheries Key Decision Points for Stakehoder Engagement Concusions References Appendix A: Stakehoder Workshops to Expore ERAs Appendix B: Attribute Tabes Appendix C: Draft ERA Scoring Spreadsheets Appendix D: Protoco for Too Appication & Anaytica Methods ii

4 List of Figures Figure 1. Overview of the piot ERA process steps based on Samhouri and Levin (2012)... 8 Figure 2. Nine fisheries or "units of anaysis" seected for the piot ERA anaysis Figure 3. Bycatch guids and habitat groups incuded in the piot ERA anaysis Figure 4. A generaized representation of how ERA resuts can be interpreted Figure 5. Target risk assessment resuts for the ERA piot Figure 6. Target risk assessment resuts for the ERA piot Figure 7. Habitat risk assessment for the ERA piot. Point size indicates the number of habitat groups affected by each fishery Figure 8. Resuts anaysis demonstration for Caifornia haibut Figure 9. Sampe resuts with data quaity scores integrated into the figure List of Tabes Tabe 1. Exposure and sensitivity attributes for the three ecosystem components of the ERA Tabe 2. Exampes of changed or deeted attributes from Piot ERA Scoresheet Tabe 3. Data quaity scoring descriptions Tabe 4. Comparison of considerations for the NOAA PSA and piot ERA for target species Tabe 5. Exposure attributes, definitions, and scoring categories for target, bycatch, and habitat Acronyms CDFW - Caifornia Department of Fish and Widife ERA - ecoogica risk assessment ERAF - ecoogica risk assessment for fishing ERAEF - ecoogica risk assessment for the effects of fishing FMP - fishery management pan GIS - geographic information system MPAs - marine protected areas MLMA - Marine Life Management Act NGO - non-governmenta organization NOAA - Nationa Oceanic and Atmospheric Administration OPC - Caifornia Ocean Protection Counci OST - Caifornia Ocean Science Trust OPC-SAT - Ocean Protection Counci Science Advisory Team PSA - productivity suceptibiity anaysis iii

5 About the Report Management Context The Marine Life Management Act (MLMA) was estabished in 1999 to ensure the conservation, restoration, and sustainabe use of Caifornia s marine iving resources. The Caifornia Department of Fish and Widife (CDFW) deveoped a Master Pan for Fishery Management Pans (Master Pan), which was adopted by the Fish and Game Commission in 2001, to hep achieve the objectives of the MLMA and focus management effort on the highest priority species. Through 2018, CDFW and its partners are amending the Master Pan to consider advancements in our understanding of the status of Caifornia s marine resources, as we as toos and approaches in the fied of fisheries science and ocean management. An important focus of the amendment is to deveop a more systematic, efficient, and transparent approach to prioritizing species for management action that refects the needs, risks, and vaue of fisheries, as we as CDFW s priorities and capacity. Project Overview This project to research, piot, and test ecoogica risk assessments (ERAs) was initiated by CDFW to inform the deveopment of a comprehensive, ecosystembased approach to fisheries prioritization for management action. CDFW fishery experts, incuding staff responsibe for overseeing the management of specific fisheries, were directy invoved in a aspects of this project, incuding deveoping, refining, and testing the ERA toos. CDFW fishery experts were aso invoved in stakehoder workshops to examine the ERA toos and soicit feedback for further revision and improvements. The project was funded by the Caifornia Ocean Protection Counci (OPC), with additiona support from the Resources Legacy Fund. The primary goa was to deveop a method to assess the potentia risk to seected fisheries from fishing. This method is intended for consideration by CDFW as part of the process to amend the MLMA Master Pan. This report has been provided to CDFW and may be integrated, in fu or in part, into a draft Master Pan Amendment. Additiona information about the Master Pan amendment process, incuding key resources and opportunities for stakehoder engagement, is avaiabe at ca.gov/conservation/marine/master-pan. This report and other supporting materias can be found on the OST project page at: iv Image credit: Dennis Jarvis

6 Highights and Key Findings A new ERA, taiored for Caifornia s fisheries management, provides a too that managers, fisheries scientists, and other stakehoders can use to understand the reative potentia risk posed by fishing activities to target species, bycatch species, and marine and estuarine habitats. The new ERA can be appied in both data poor and data rich contexts, and aows an evauation of the individua and cumuative risks posed to mutipe components of the ecosystem simutaneousy, giving breadth to its potentia appication. Based upon an existing appication that was customized to aign with the mandates and priorities of CDFW, the process to piot an ERA for seected Caifornia fisheries demonstrates the efficiency of drawing from an existing method and offers a too that can be used directy by others or customized further to meet ocaized needs. An incusive and transparent process has set the stage for broad buy-in and adoption. Co-deveopment of the ERA by scientists and CDFW fishery experts, and with stakehoder input during two in-person workshops, fostered incusion of mutipe perspectives and interpretations of existing data and information. This process deivers a scientificay rigorous too with broad utiity. The too can underpin scientific prioritization of the deveopment of management actions and screen for potentia impacts of fishing activity on an ecosystem. v Image credit: Dennis Jarvis

7 Executive Summary Image credit: Virginia Sea Grant The Caifornia Department of Fish and Widife (CDFW), with funding from the Ocean Protection Counci, requested that Caifornia Ocean Science Trust (OST) research, deveop and piot-test a risk-based decision framework that advances the goas of the Marine Life Management Act (MLMA) with regard to prioritization of fisheries for management attention, and is transparent, cost-effective to impement, and scientificay rigorous. CDFW sought an ecosystem-based management too that coud assess three ecosystem components: target species, bycatch species, and marine habitats. CDFW was aso interested in considering the potentia benefits of Caifornia s network of marine protected areas (MPAs). The goa of this project was to provide CDFW with a scientificayvetted piot test of an Ecoogica Risk Assessment (ERA) that coud inform the prioritization among fisheries as part of the MLMA Master Pan amendment process. ERAs can be vauabe quantitative toos to hep focus imited fishery management resources on species that are most at potentia risk from fishing activities and other stressors. In genera, ERA resuts are not mandates for management action, but serve to identify fisheries with potentiay greater risk. Once a set of fisheries has been scored, fisheries with potentiay higher risk shoud be further examined by users of the too, usuay fishery managers, to 1) better understand why they have been scored higher risk and, 2) use that information to decide if further assessment is required and/or a management action is needed. Buiding off previous research, OST worked with experts from NOAA Fisheries West Coast Region and MRAG Americas to test two different risk assessment toos, an ERA (adapted from Samhouri and Levin, 2012) and a NOAA Productivity and Susceptibiity Anaysis (PSA), respectivey. This project was one of thirteen MLMA Master Pan amendment information gathering projects intended to inform the State s process to amend the MLMA Master Pan for Fisheries. Customization and deveopment of the piot ERA too was iterative between the members of the project team (CDFW, OST, NOAA, OPC), and towards the end of the piot, with stakehoders incuding fishermen, environmenta NGOs, and other federa fisheries experts. The bycatch and habitat components were requested by CDFW. The target component, which is simiar to PSA, was added by NOAA and OST to demonstrate the functionaity of a singe ERA too and a customized approach taiored to Caifornia. Because Caifornia has a widespread and abundant MPA network, we aso considered the potentia benefits from MPAs to each ecosystem component (i.e., target, bycatch, habitat). In order to customize and test the piot ERA too, we seected nine fisheries that woud aow us to appy the too on a wide range of different fishery characteristics. A fishery was defined as a combination of a species, a fishing EXECUTIVE SUMMARY I

8 gear, and a sector (sport or commercia). The foowing nine fisheries were chosen from the ist of 45 fisheries for which a PSA was conducted: White Sturgeon - Hook and ine, sport Kep Bass - Hook and ine, sport Caifornia Haibut - Hook and ine, sport Caifornia Haibut - Traw, commercia Caifornia Haibut - Gi net, commercia Caifornia Haibut - Hook and ine, commercia Pacific Herring - Gi net, commercia Spiny Lobster - Hoop net, sport Spiny Lobster - Trap, commercia A ERAs create resuts by asking experts to score various attributes that may contribute to a species or habitat potentia risk to an identified stressor, in this case fishing activity. Sensitivity attributes are the expected response of a target species, bycatch guid, or habitat given that it is exposed to a stressor. Exposure attributes are those that represent the potentia for a stressor to infuence a target species, bycatch guid, or habitat based on attributes that incude spatia, tempora, and management factors. In this report they are sometimes referred to as the sensitivity axis or exposure axis. In this piot ERA, fishery experts were asked to compete scoring on a spreadsheet to be used to create resuts to hep us refine the piot ERA too. The fina spreadsheet has been evauated by CDFW with feedback from stakehoders. The spreadsheet is unique due to the highy coaborative and iterative nature of the process used to create it, its customization for Caifornia State fisheries, and they way in which it incorporated mutipe types of knowedge (e.g., academic, fishery experts, appied managers, fishermen). The spreadsheet represents a bridge between traditiona risk assessments, often created by scientists in isoation, and managers who need a too that meets their specific needs and the needs of stakehoders. The scores provided on the spreadsheet were anayzed, summarized, and presented in this report to hep readers understand the too, its appication, and how fisheries coud be prioritized if the piot ERA is used in the future. Resuts do not indicate a mandate for management action. Rather, they indicate a potentia risk that shoud be examined further by users of the too to determine whether and if a management action is needed. Target ERA - What is the potentia reative risk posed to target species from fishing activity? The target ERA has many simiarities to the NOAA PSA. The main difference between this ERA and other risk assessment approaches, incuding PSA, is that it was designed for Caifornia s fisheries with data that are readiy avaiabe to managers and with expicit consideration of Caifornia s MPA network (and is repicated for the bycatch and habitat ERAs). For each fishery in the target ERA, scores were added and averaged (i.e., arithmetic average method) independenty for sensitivity and exposure attributes. Of the nine fisheries assessed, reative risk to target species was greatest for white sturgeon in the sport hook-and-ine fishery, foowed by the four Caifornia haibut fisheries. For a five of these fisheries, reativey high risk scores resuted from reativey high scores for neary a exposure attributes as opposed to sensitivity attributes. Low sensitivity attribute scores for these target species impies that high exposure eves are not expected to jeopardize their persistence. EXECUTIVE SUMMARY II

9 Bycatch ERA - What is the potentia reative risk posed to bycatch from fishing activity? The approach taken to risk assessment for species caught as bycatch in the fisheries aows for users to score risk to a representative species within a guid (a group of reated species). For prioritization, the goa is to understand if there is a potentia risk and then ater work to understand the species and other detais of that potentia risk once a fishery has been prioritized. For the purposes of the ERA piot project and workshops, bycatch was defined as catch that is returned to the water. In the anaysis of each fishery in the bycatch ERA, the attribute scores for each guid (e.g., easmobranchs, samonids) are summed and averaged by axis, and those guids with no interaction with a fishery were removed from anaysis. In contrast to the Target ERA scoring, fisheries with mutipe bycatch guids had the score from each guid summed (i.e., cumuative method). Two of the bycatch ERA attributes were weighted to represent 50% of the tota score among a attributes within each axis to refect their reative importance. One of these weighted attributes reated to sensitivity attribute (reease mortaity) and the other reated to exposure (magnitude). The rank ordering of reative cumuative risk to bycatch guids differed from that of risk to target species. Specificay, the Caifornia haibut commercia gi net and traw fisheries posed the greatest reative cumuative risk. This resut emerged because both fisheries interact with six of the ten bycatch guids, the greatest number among the nine fisheries. Habitat ERA - What is the potentia reative risk posed to habitat from fishing activity? The approach to the habitat component foows a simiar structure to that of bycatch. One main difference between the bycatch and habitat ERA components is that for fisheries that occur in mutipe habitats, users are aso asked to estimate the percent of fishing activity that takes pace in each habitat (sum = 100%) for use in subsequent anaysis. In the anaysis of each fishery in the habitat ERA that has mutipe guids (e.g., kep, soft bottom) scores are summed and averaged by axis, and those guids no interaction with the fishery are removed from anaysis. The arithmetic average method was used to cacuate fina scores but was weighted by the percentage of fishing that occurs among impacted habitats. Simiar to the bycatch ERA, two habitat ERA attributes aso were weighted to refect their reative importance. One exposure attribute (gear footprint) and one sensitivity attribute (potentia damage to habitat from fishing gear) were weighted to represent 50% of the tota score for a attributes. This decision was made based on the recognition that it is the gear itsef which has the greatest potentia impact on the habitat among a attribute factors. Reative potentia risk to habitats was greatest for the Caifornia haibut commercia traw fishery, which affects nearshore soft bottom and habitat-forming marine invertebrates. Whie these two habitats were both highy exposed, neither was expected to be particuary sensitive. Compared with the risk assessment for bycatch guids, the remaining eight fisheries did not show as much variation in risk score for habitats. Data Quaity As part of this piot, data quaity was scored for every attribute and a written rationae for the attribute was provided. Data quaity was not integrated into any fina resuts because time and effort were focused on finaizing resuts for the three ecosystem components. However, this information is avaiabe in the appendices and coud be used to refine future ERAs and - as suggested at the workshops - to identify knowedge gaps reated to the nine fisheries. EXECUTIVE SUMMARY III

10 Next Steps and Concusions Going forward, it is within the purview of CDFW or other potentia users to consider whether and if this piot ERA meets their needs. This piot test invoved many key decisions. Severa others remain shoud CDFW choose to impement the ERA in whoe or in part. These incude: Key decision points for too refinement and impementation such as how to treat data quaity, and whether and how to identify actions and associated costs of owering risk Key decision points for PSA vs. ERA for target fisheries Key decision points for stakehoder engagement Consideration of the use of regiona ERAs The process we used to create the piot ERA was scientificay rigorous, efficient, highy iterative and coaborative among NOAA, OST, and CDFW, and incorporated stakehoder input. The resut is a too that, if impemented, can minimize capacity issues and can argey be conducted by CDFW. Additionay, shoud CDFW decide to impement the too, the five CDFW fishery experts who heped to piot the too are now in-house experts. CDFW can now assess whether and if the too meets current management needs to prioritize fisheries based on the three ecosystem components - target, bycatch, and habitat. Finay, the process we used demonstrated a way for CDFW to seect an existing ERA for customization to meet their specific management goas and ecosystem features. This piot ERA and process can be utiized by others to meet simiar goas or to seect and customize their own ERA efficienty. EXECUTIVE SUMMARY IV

11 1. Introduction Image credit: Ed Bierman 1.1 Overview CDFW requested that OST research, deveop, and piot test a risk-based decision framework that advances the goas of the MLMA with regard to prioritization of fisheries for management attention, and is transparent, cost-effective to impement, and scientificay rigorous. CDFW sought an ecosystembased management too that coud assess three ecosystem components: target species, bycatch species, and marine and estuarine habitats. CDFW was aso interested in considering the potentia benefits of Caifornia s network of MPAs. The goa of this project was to provide CDFW with a scientificay-vetted piot test of an ERA that coud inform the prioritization among fisheries as part of the MLMA Master Pan amendment process. REQUEST FROM CDFW Deveop risk-based too(s) to consider: Potentia risk of fishing activity to target species Potentia risk of a fishery s operation to bycatch species Potentia risk of a fishery s operation to habitats Consider impications of Caifornia s network of marine protected areas (MPAs) to the target, bycatch, and habitat ecosystem components 1.2 Why Risk Assessment? ERAs can be vauabe quantitative toos to hep focus imited fishery management resources on species that are most at risk from fishing activities and other stressors. ERAs for fishery management are frameworks for assessing the ikeihood that a fishery, species, or ecosystem component faces significant negative impacts from anthropogenic or natura sources (e.g., fishing activities, impaired water quaity, cimate change, ocean acidification). ERAs are one of the few readiy-avaiabe toos to hep a manager move away from singe species management and towards ecosystem based management (Link et a., 2002; FAO 2003). Caifornia began exporing the use of ERAs for state marine fishery management in 2013 through a series of OPC Science Advisory Team (OPC-SAT) meetings and severa exporatory projects (see Timeine of Events on the foowing page). INTRODUCTION 1

12 Timeine of Events Exporing Ecoogica Risk Assessments for Caifornia September 4, 2013 Ocean Protection Counci Science Advisory Team workshop: Advancing Science in Caifornia Fisheries Caifornia began exporing the use of ERAs for state marine fisheries management in 2013 at an OPC-SAT meeting Advancing Science in Caifornia Fisheries. The OPC-SAT came together with fishery decision makers to discuss the best ways for scientists to partner with CDFW to support sustainabe fisheries management. (OST 2013) Juy 29, 2015 Ocean Protection Counci provides funding to support deveopment of ERA toos for CDFW OPC funded OST to partner with CDFW to customize, test, and piot an ERA, and appy a PSA to specific fisheries. (OPC 2015) Piot ERA Deveopment OST worked with NOAA Fisheries and CDFW to expore risk to habitats and bycatch species by adapting an existing ERA framework for Caifornia state marine fisheries. ERA PILOT PROJECT September 2014 Ecoogica Risk Assessments: A Roadmap for Caifornia Fisheries December 2016 PSA Appied to 45 Caifornia Fisheries OST ed the deveopment of a OST contracted with MRAG report that examined mutipe ERA frameworks for understanding fishery Americas to conduct a NOAA PSA on 45 target fisheries (i.e., gear/ vunerabiities to various stressors, sector/species combinations), incuding fishing pressure and cimate change. This project identified essons earned from existing appications, representing 36 Caifornia statemanaged species. (MRAG Americas 2016) and considerations for adopting such methods for Caifornia fisheries. (OST 2014) February 25, 2015 June/Juy 2017 ERA Stakehoder Workshops OST hosted two stakehoder workshops to review the ERA draft too and scoring for nine piot fisheries. Ocean Protection Counci Science Advisory Team workshop: Readying Caifornia Fisheries for Cimate Change The roadmap report supported a second OPC-SAT meeting in 2015 where members of the scientific community were once again brought together with CDFW managers to discuss how the scientific community coud aid the State in the deveopment of the MLMA Master Pan amendment. (OST 2015) INTRODUCTION 2

13 There are many types of ERAs. The term refers to a cass of risk assessment frameworks, which in turn can address either absoute or reative risk (Box 1), aong with a range of goas. The particuar framework chosen, adapted as needed, and impemented depends on the management situation it is designed to support, but most contain eements that are considered key components (Box 2). It is important to create or adapt an ERA for ceary defined appications and management goas. ERAs that are buit for different projects wi therefore have different components and inputs, carefuy chosen and designed to address the specific management goas (e.g., maintain stock under fishing, minimize bycatch impacts). ERAs can support different types of managers information needs, which, in turn, inform different types of management decisions. ERAs can assess both data-poor and data-rich fisheries on the same scae. They each take uncertainty in the data or expert s knowedge into account with approaches that vary among toos. In addition to prioritization, ERAs can aso expose data gaps or identify specific concerns in a fishery to guide future management decisions. In genera, ERA resuts are not mandates for management action, but serve to identify fisheries of potentia concern for one or more ecosystem components. Once a set of fisheries has been scored, fisheries with potentia higher risk shoud be further examined by users of the too, usuay fishery managers, to 1. better understand why they have been scored higher risk, and, 2. use that information to decide if a management action is needed. ERAs err on the side of caution and are designed to produce fase positives (a fishery that is deemed high risk by the too, but in reaity is not at true risk) instead of fase negatives (a fishery that actuay has high risk for a negative outcome, but instead was scored as ow risk by the ERA). There are many reasons why a fishery may score a fase positive, and further investigation or interpretation of the resuts is needed to address these situations. Risk assessments can fag potentia issues and aow stakehoders to work together with managers to understand and potentiay avoid negative outcomes. For exampe, a fishery scored as higher potentia risk may have recenty deveoped a fishery management pan (FMP), so the State may consider the risk to be addressed (e.g. Caifornia spiny obster). Conversey, if a fishery is fasey estimated to have ower risk, it coud undergo a stock decine or some other avoidabe outcome whie managers are focused on other priorities. BOX 1. KEY DEFINITIONS Absoute Risk - the chance that a species, habitat, or ecosystem feature wi experience potentia decine on an absoute scae ranging from impossibe to certain. Cacuation of absoute risk scores requires vaidation and more time-consuming and data-intensive methodoogies than reative risk cacuations. Reative Risk - the chance that a species, habitat, or ecosystem feature wi experience potentia decine or degradation due to a particuar activity, in terms of higher or ower exposure and sensitivity scores, without reference to an absoute scae that defines the probabiity associated with these scores. (Adapted from Samhouri and Levin, 2012) INTRODUCTION 3

14 BOX 2. Key Piars of an ERA too and process There are tradeoffs with each of these key piars, and there is no one right ERA or approach. However, these are the key criteria to keep in mind when deveoping or customizing and ERA. Comprehensive Can assess mutipe or singe stressors (e.g., fishing activity, habitat degradation, cimate change) Fexibe Can assess data poor aongside data rich fisheries Can assess among or within fisheries (usuay not both) Appicabe to a types of fisheries, irrespective of size, fishing method, species, etc. Synthesizes knowedge Brings different types of expertise (e.g., managers, NGOs, academic scientists, fishermen, etc.) together to share knowedge Transparent and repeatabe Cear process, methods, data, and assumptions used in anayses Understandabe to different users (e.g., managers, stakehoders, scientists) Cear methods for how resuts wi be used by managers Cost effective Existing knowedge, information, and data within reaistic imits of time and resources Scientificay defensibe Undergoes independent peer review Usefu for management Inform risk management responses and decisions Takes a precautionary approach to uncertainty (Adapted from Hobday et a. 2011) INTRODUCTION 4

15 2. Ecoogica Risk Assessment Approach Image credit: NOAA Fisheries West Coast Buiding off previous research, OST worked with experts from NOAA Fisheries Science Centers and MRAG Americas to test two different risk assessment toos, a NOAA PSA and an adapted ERA (modeed after Samhouri and Levin, 2012), respectivey. This project was one of thirteen MLMA Master Pan amendment information gathering projects 1 intended to inform the State s process to amend the MLMA Master Pan for Fisheries 2. The goa of the project is to hep CDFW prioritize fisheries for management consideration (e.g., the deveopment of enhanced status reports, or enhanced status reports with ruemaking, or FMPs). By pioting two ERA methods aongside each other with varying eves of State fishery expert and stakehoder invovement, we provide CDFW with options to consider for future impementation. Based on the request from CDFW, OST was tasked with impementing two different risk assessment approaches: 1. Approach 1: Appy an existing ERA - Conduct an estabished risk assessment on a arge number of fisheries to assess risk to target species. The project team seected the NOAA productivity susceptibiity anaysis (PSA) to be conducted on 45 Caifornia marine fisheries. 2. Approach 2: Customize an ERA for Caifornia - Customize an ERA that can assess fishing risk to target, bycatch, and habitats that takes into account Caifornia s unique environment and assets (e.g., MPAs). The project team seected the Samhouri and Levin (2012) ERA too to adapt for Caifornia and piot on nine fisheries. For each approach, a fishery was defined as a fishing gear, sector (i.e., commercia or sport), and species combination (Box 3). We aso ca this unique fishery combination a unit of anaysis. This approach aows for differences between sport and commercia fisheries and among different gear types to be assessed separatey; however, the overa risk score can be assessed at the species eve, which is the way that the State currenty manages many of its fisheries. For exampe, in both of these information gathering risk assessment projects, spiny obster was BOX 3. KEY DEFINITION Fishery - for the purposes of the ERA piot, a fishery is defined as a unique gear type, sector (e.g., commercia or sport), and species combination. This is aso known as a "unit of anaysis." 1 For more information on the MLMA Master Pan projects, visit 2For more information on the MLMA Master Pan Amendment process, visit ERA APPROACH 5

16 anayzed separatey for the sport and commercia sectors to aow managers and stakehoders to understand the reative risk that each fishery poses to spiny obster and associated bycatch and habitats independenty. This approach aso aows managers to combine the scores of each fishery to understand potentia overa risk to spiny obster, bycatch, and habitats at the target species eve (i.e., a gear types and sectors combined into a singe score). 2.1 Why These Two Risk Assessment Approaches? Approach 1: Appy an Existing ERA (NOAA PSA) PSA is a widey used risk assessment too in fisheries management for assessing the vunerabiity of a fishery species or stock, using a set of predetermined measurabe attributes and score rankings. It was one of the toos used to prioritize management of fisheries in the previous 2001 CDFW Master Pan for Fisheries. PSA has evoved over time from its initia creation and has been adapted and appied across the gobe. Briefy, PSA assumes that the overa vunerabiity of a fished species to impacts from fishing depends on two characteristics: 1. the productivity of a species/stock based on ife history traits that determine whether a fished species coud sustain or recover from fishery-reated impacts. 2. the susceptibiity of the species/stock to impacts from fishery-specific activities. Productivity and susceptibiity attributes of each stock are examined and scored. There are severa different PSA methods, which vary sighty in their attributes and scoring methods, though most methods provide simiar resuts. See Box 4 for an overview of the 2016 PSA conducted by MRAG Americas on behaf of CDFW. BOX 4. PRODUCTIVITY AND SUSCEPTIBILITY ANALYSIS FOR SELECTED CALIFORNIA FISHERIES (MRAG AMERICAS, 2016) In 2016, a NOAA PSA was conducted on 45 units of anaysis as part of an information gathering process reated to Caifornia s MLMA Master Pan Amendment. The NOAA PSA methodoogy was seected for its incusion of attributes that evauate the management strategy and vaue of a fishery, aong with its abiity to generate and consider a data quaity score. The anayses were conducted by MRAG Americas, Inc. with input and review from CDFW experts. The ist of fisheries for anaysis represents a diversity of stocks that span commercia and sport sectors, gear types, coasta areas, and incude finfishes and invertebrates. It aso represents those state-managed fisheries with the highest commercia andings, recreationa catch, or commercia/recreationa participation, as evauated by CDFW. More information on the methods, interpretations, and resuts can be found in the PSA report onine here. ERA APPROACH 6

17 2.1.2 Approach 2: Customize an ERA for Caifornia (Samhouri and Levin, 2012) Seecting an ERA There are a arge number of existing ERAs and other fishery risk assessment toos utiized across the gobe. OST s first step was to whitte down this vast body of toos into severa for consideration for customization by the project team. Based upon our previous work, conversations and recommendations from ERA experts, and the scope of what the ERA too needed to do, we narrowed our fied of ERAs to the foowing: Ecoogica Risk Assessment for Fisheries (ERAF) - deveoped for Canadian fisheries (Martone et a. 2012) CARE and CARE ite too - deveoped by Environmenta Defense Fund (Battista et a. 2017) Ecoogica Risk Assessment for the Effects of Fishing (ERAEF) - deveoped for Austraian fisheries (Hobday et a. 2011) Mutipe Stressor Ecosystem-based Risk Assessment - deveoped as part of the NOAA integrated ecosystem assessment tookit (Samhouri and Levin, 2012) Quaitative Consequence Risk Assessment Anaysis - deveoped for Austraian fisheries (Fetcher 2005) Cumuative impacts assessment - a quantitative assessment for assessing mutipe human impacts (Stezenmüer et a. 2010, Hapern et a. 2009) When seecting an ERA for customization we considered many aspects of scientific rigor, capacity, efficiency, and the abiity to meet CDFW s needs. Some of the key considerations were if the too in its current state or, if easiy adapted, coud: Be appied at a state-wide scae whie accommodating regiona fisheries Be impemented efficienty Be scientificay rigorous (aready independenty peer-reviewed) Have previous rea-word appications Accommodate fisheries as Caifornia defines them (i.e., singe species eve) Address habitat, bycatch, and target species risk in the same too, on the same scae Consider Caifornia s assets and management structures (e.g., MPA network) Address stressors other than fishing activity Be customizabe in time frame aotted Be utiized by CDFW with minima outside expertise or support at the end of the piot The team utimatey chose to work with and customize the Samhouri and Levin (2012) methodoogy (Figure 1). The decision was primariy driven by the estabished scientific rigor of the too, its abiity to address fisheries at the species eve rapidy, and its abiity to be customized both to address bycatch and habitat but aso to take into account Caifornia s MPAs. ERA APPROACH 7

18 KEY PROCESS STEPS PILOT ERA FOR CALIFORNIA 1 Set Management Priorities Determine which management priorities are most important to address with the ERA Fishery sustainabiity* Top Leve Priorities Identified by CDFW Habitat conservation* *Marine Life Management Act goa Limit bycatch* Marine protected areas 2 Transate Priorities into ERA Goas Identify information needs, gaps, and targets to meet management priorities Project team transates priorities into goas for Caifornia piot ERA: Abiity to address habitat, bycatch, and target species Can be appied statewide, regionay Rapid and efficient Scientificay rigorous Rea-word appications Take into account MPA network Moderate time and effort 3 Deveop ERA Scope Create cear set of scientific questions and goas for the ERA to address Project team works with ERA experts from NOAA Fisheries to customize a piot ERA for Caifornia fisheries that can answer: What is the reative risk posed to a target species, bycatch species, or habitat from the prosecution of the fishery? 4 Conduct Risk Assessment Scientists, government, and stakehoders coectivey conduct ERA Piot too deveopment with NOAA Fisheries Experts Create process for scoring with CDFW Fishery experts Anayze draft scores with fu project team Stakehoder workshops Finaize draft resuts 5 Utiize Resuts Resuts of the ERA are pubished and inform previousy defined management outcomes CDFW to consider utiization of piot ERA in the MLMA Master Pan Amendment to assist with prioritization of species for management action Figure 1. Overview of the piot ERA process steps based on Samhouri and Levin (2012). ERA APPROACH 8

19 3. Ecoogica Risk Assessment Piot Image credit: Oregon Department of Fish and Widife The piot ERA, adapted from Samhouri and Levin (2012), was designed to ask the foowing question: What is the reative risk posed to a target species, bycatch species, or habitat from the prosecution of the fishery? CDFW was interested in exporing potentia risk posed to two ecosystem components based on their mandates in MLMA egisation: imiting bycatch and habitat conservation. Through conversations with CDFW, risk assessment experts from academia, NOAA, and some of the foremost risk assessment experts in Austraia, OST proposed expanding the piot to incude target species and the abiity to take into account our network of MPAs (which were not in pace during the creation of the previous Master Pan). 3.1 Methods Customizing the ERA for Caifornia Fisheries The Samhouri and Levin (2012) methodoogy was originay created to conduct an assessment that considered the overa heath of an ecosystem due to mutipe stressors (e.g., fishing activity, coasta deveopment, agricuture). The too did and continues to draw from ERAEF and PSA. The project team worked with the too s creator to adapt it to specificay address one stressor (fishing activity) on target species, bycatch species, and habitat. The bycatch and habitat components were requested by CDFW. The target component, which is simiar to PSA, was added by NOAA and OST to demonstrate the functionaity of a singe ERA too and demonstrate a customized approach taiored to Caifornia. Because Caifornia has a widespread and abundant MPA network, we aso considered how MPAs interact with each ecosystem component (i.e., target, bycatch, habitat). The Caifornia MPA network was estabished from 2007 to 2012, primariy to protect representative portions of ecosystems and those fished species which woud benefit from this protection. This network has ecoogica benefits to expoited species and their ecosystems by protecting the diversity and abundance of marine communities and ecosystems from environmenta and anthropogenic stressors. For each of the three ERA ecosystem components, an MPA attribute was created that expicity reduces their exposure fishing by taking into account the Caifornia MPA network (see Appendix B for attribute definitions for each component). METHODS 9

20 Each seected fishery receives an overa score for target, bycatch, and habitat based on Sensitivity and Exposure scoring attributes (Box 5). See Appendix B for a spreadsheet with compete attributes and definitions for sensitivity and exposure attributes for the piot ERA. Fishery Seection for the Piot In order to customize and test the piot ERA too, we seected nine fisheries that woud aow us to appy the too on a wide range of different fishery characteristics. We seected from the ist of 45 fisheries used for the PSA in order to faciitate easy comparisons between toos. Within that ist we sought to find a mix of fisheries for which we coud compare and test performance of the too based on the foowing comparisons and considerations: Based on these factors, the foowing nine fisheries (or units of anaysis ) were chosen (Figure 2): White Sturgeon - Hook and ine, sport Kep Bass - Hook and ine, sport Caifornia Haibut - Hook and ine, sport Caifornia Haibut - Traw, commercia Caifornia Haibut - Gi net, commercia Caifornia Haibut - Hook and ine, commercia Pacific Herring - Gi net, commercia Spiny Lobster - Hoop net, sport Spiny Lobster - Trap, commercia BOX 5. KEY DEFINITIONS Attribute - A factor or criterion which contributes to the overa risk to a target species, bycatch guid, or habitat in which the fishery occurs. Sensitivity - The expected response of a target species, bycatch guid, or habitat given that it is exposed to a stressor. Attributes incude resistance and recovery factors. It is sometimes referred to as the Sensitivity Axis. Exposure - The potentia for a stressor to infuence a target species, bycatch guid, or habitat based on attributes that incude spatia, tempora and management factors. Invertebrate and vertebrate species It is sometimes referred to as the Exposure Axis. Nearshore and peagic species Recreationa and commercia sectors within a fishery and fisheries with ony one sector Different gear types Varying economic vaue and/or cutura or regiona vaue Habitat or species that are considered ecosystem feature representatives High trophic eve species and forage species Geographic considerations: statewide and regiona; northern, centra, and southern Caifornia Data rich and data poor Fisheries with and without FMPs These fisheries were not previousy prioritized. They were seected to hep the team customize and refine the piot ERA too, based on stakehoder input, and to hep us understand how and if the too is working to meet its management goas and underpinning scientific rigor. METHODS 10

21 SPECIES Caifornia Haibut GEAR Hook and ine SECTOR Commercia SPECIES Caifornia Haibut GEAR Hook and ine SECTOR Sport SPECIES Caifornia Haibut GEAR Traw SECTOR Commercia SPECIES Caifornia Haibut GEAR Gi net SECTOR Commercia SPECIES White Sturgeon GEAR Hook and ine SECTOR Sport KEY GEAR SPECIES SECTOR SPECIES Spiny obster GEAR Hoop net SECTOR Sport SPECIES Spiny obster GEAR Trap SECTOR Commercia SPECIES Kep Bass GEAR Hook and ine SECTOR Sport SPECIES Pacific Herring GEAR Gi net SECTOR Commercia Figure 2. Nine fisheries or "units of anaysis" (representing a species, gear type, sector combination) seected for the piot ERA anaysis. Target Species Anaysis: What is the potentia reative risk posed to target species from fishing activity? We revisited the origina Samhouri and Levin (2012), PSA (Hobday et a., 2011; Patrick et a., 2010), and other risk assessment methodoogies to seect, customize, and define attributes to be scored as part of this piot ERA. The main difference between this ERA and other risk assessment approaches, incuding PSA, is that it was designed for Caifornia s fisheries with data that are readiy avaiabe to managers and with expicit consideration of Caifornia s MPA network. It aso has fewer attributes than the NOAA PSA. Bycatch Anaysis: What is the potentia reative risk posed to bycatch from fishing activity? For this component, the main chaenge was to find a way to assess risk to bycatch species without having to do an assessment on every species caught in a fishery. For prioritization, the goa is to understand if there is a potentia risk and then ater work to understand the species and other detais of that potentia risk once a fishery has been prioritized. In order to create this ERA component, NOAA and OST reviewed other methodoogies (Samhouri and Levin 2012; Zhou et a., 2008; Fetcher et a., 2005, Hobday et a., 2011). The approach taken aows for users to score risk to bycatch for a representative species within a guid (a group of reated species). Potentia benefits from Caifornia MPAs were aso a factor that was considered in the creation of this component by creating an expicit MPA attribute; this was scored reative to potentia benefit rather than potentia risk. Definition of Bycatch for Pioting ERA There are mutipe definitions of bycatch at both the state and federa eve, but this does not change the mechanics of the too itsef. Rather, it is more important to have a definition be consistenty appied within the bycatch ERA too appication. METHODS 11

22 For the purposes of the ERA piot project and workshops, bycatch was defined as catch that is returned to the water (Box 6), with the foowing additiona guidance: For seven of the 10 guids, a bycatch guid is scored if there is significant bycatch (greater than 1% of the catch of the target species by either weight or number) for any species within that guid. We use the most frequenty caught species within that guid and score it as appropriate. If a species within the guid have what we consider to be non-significant bycatch, we do not score that guid. For the other three guids (marine mammas, marine birds, and threatened and endangered species and/or overfished rockfish), if there is any bycatch of these guids, we score the most common species within the guid. Sub- and supra-egas of the target species are scored as bycatch. Defining Bycatch Guids Ten guids were initiay suggested by CDFW to contain the fu spectrum of species possibe as bycatch in Caifornia s marine fisheries, but to imit the number of guids so that scoring was not an unreasonabe task. The guids were refined by the project team and stakehoders during workshops. Bycatch guids were as foows (Figure 3a): Marine mammas Marine birds Threatened or endangered species, and/or overfished rockfish Easmobranchs Samonids BOX 6. KEY DEFINITION Bycatch - for the purposes of the ERA piot, the project team defined bycatch as catch which is returned to the water. Fatfish Other rockfish Peagic finfish Non-peagic finfish Marine invertebrates Habitat Anaysis: What is the potentia reative risk posed to habitat from fishing activity? The approach to the habitat component foows a simiar structure to that of bycatch. When a fishery is being scored, the user is asked to ony score the habitat types in which the fishery operates. In order to create this ERA component, NOAA and OST reviewed other methodoogies (Wiiams et a., 2011; Arkema et a., 2014) to propose options to the project team for their input and feedback. One main difference between the bycatch and habitat ERA components is that for fisheries that occur in mutipe habitats, users are aso asked to estimate the percent of fishing activity that takes pace in each habitat (sum = 100%) for use in subsequent anaysis. Caifornia MPAs were aso a factor that was considered in the creation of this component by creating an expicit MPA attribute; this was scored reative to potentia benefit rather than potentia risk. Defining Habitat Groups Ten habitat groups were seected based on knowedge of Caifornia coasta and oceanic ecosystems, management definitions of habitat utiized by CDFW, and avaiabiity of GIS mapping data, using the smaest number of groups feasibe for efficiency (Figure 3b). One of the options for this component in the future woud be for some of the attributes to be scored using GIS spatia anaysis rather than expert scores. For exampe, spatia overap with MPAs coud be automated to anayze percent cover of tota habitat inside MPAs and outside MPAs in state waters. METHODS 12

23 Habitat types incuded: Habitat-forming marine vegetation Habitat-forming marine invertebrates EstuariesNearshore hard bottom (0-200m) Nearshore soft bottom (0-200m) Offshore hard bottom (>200m) Offshore soft bottom (>200m) Peagic Hard bottom intertida Soft bottom intertida A. BYCATCH GUILDS Marine mammas Marine birds Threatened or endangered species, and/or overfished rockfish Easmobranchs Samonids Fatfish Other rockfish Peagic finfish Non-peagic finfish Marine invertebrates B. HABITAT GROUPS Habitat-forming marine vegetation Habitat-forming marine invertebrates Estuaries Nearshore hard bottom (0-200m) Nearshore soft bottom (0-200m) Offshore hard bottom (>200m) Offshore soft bottom (>200m) Peagic Hard bottom intertida Soft bottom intertida Figure 3. (A.) Bycatch guids and (B.) habitat groups incuded in the piot ERA anaysis. Photo credits: Marine Appied Research and Exporation; other creative commons. METHODS 13

24 Creating a Process and Structure for Scoring the ERA Customization and deveopment of the piot ERA was iterative among the project team, and with stakehoders incuding fishermen, environmenta NGOs, and other federa fisheries experts. NOAA and OST worked to deveop options for an approach to scoring and anaysis to present to the project team. Once an approach was chosen, NOAA and OST began deveopment of a scorabe spreadsheet. Deveopment of the spreadsheet was highy coaborative and iterative, initiay invoving mutipe rounds of review by CDFW and suggested changes, foowed by CDFW experts testing the spreadsheet by scoring the nine fisheries. Prior to the stakehoder workshops, CDFW experts competed at east two rounds of scoring and the project team made changes to the spreadsheet after each scoring exercise based on feedback from CDFW. The project team made further changes after receiving input from each of the two stakehoder workshops and the CDFW fishery experts rescored the spreadsheets after each workshop. The spreadsheet and resuts in this report represent the fina iteration of the piot ERA too that resuted from this process. CDFW may decide to pursue further changes and refinements shoud they decide to use the piot ERA in the future. Finaizing the Scoring Spreadsheets The fina draft spreadsheet is incuded in Appendix B aong with scored spreadsheets and scoring instructions. Many changes were made during its deveopment based upon: 1) knowedge of users and stakehoders, 2) data that most users coud access and appy, and 3) knowedge of how Caifornia fisheries and fishermen operate. At first gance, the scoring bin descriptions and cut-offs may seem arbitrary. However, the project team made efforts to base them in science, when possibe, and then refine them utiizing knowedge about management in Caifornia and data avaiabe. This approach heped to carify, standardize, and quantify the different score bin descriptors to decrease the ikeihood of users scoring or interpreting attributes differenty. The fina draft spreadsheet has been tested by CDFW with some feedback from stakehoders. The spreadsheet is unique due to the highy coaborative and iterative nature of the process used to create it, its customization for Caifornia State fisheries, and they way in which it incorporated mutipe types of knowedge (e.g., academic, fishery experts, appied managers, fishermen).. The spreadsheet represents a bridge between traditiona risk assessments, often created by scientists in isoation, and managers who need a too that meets their specific needs. In creating the fina draft spreadsheet the team had severa goas in mind: Continue to ensure strong scientific grounding and rigor. Create a spreadsheet that is easiy understood by mutipe types of users and stakehoders. Since mutipe users may compete the scoring, efforts were made to reduce subjectivity in attribute descriptions and scoring bins. For exampe, we moved away from descriptions ike reduce impacts somewhat to concrete or quantitative exampes of reduced impacts as part of the definition and scoring bin descriptors. Seect a reasonabe number of attributes for each ERA component to create an efficient scoring process for the user as we as fina scores that vary among Units of Anaysis. (Azose and Samhouri, unpubished manuscript; Samhouri and Levin (2012) sensitivity anaysis in suppementa materia; Hobday et a. 2007). Additionay, studies have shown that the number of attributes affects the sensitivity of the attributes themseves. With too few attributes, each attribute affects the overa score disproportionatey. With too many attributes, none of them change the score enough, possiby resuting in simiar overa scores that coud hamper efforts by managers to make decisions. Seecting a Scae The project team decided to extend the scae of risk scores from a range of 1-3 (which is the scae used in the NOAA PSA) to range of 1-4 (where 1 represents the owest potentia risk and 4 represents the highest). This METHODS 14

25 decision was made for two key reasons: 1) to create a broader range and potentiay greater differences in resuts among fisheries and 2) to create a scoring system where there is no haf-way, or midde score (i.e. on a scae of 1-3, 2 is the exact midde, on a scae of 1-4, users are forced to pick a score on either side of the midde). This second reason draws upon previous risk assessment iterature and essons earned (Morrison et a., 2015; OST 2014) and ensures the user wi seect a score either above or beow the midde mark. We aso instructed users to score every attribute. Seecting Attributes The fina draft spreadsheet represents the current decisions made by the project team with input from stakehoders. Tabe 1 shows the attributes for target, bycatch, and habitat. See Appendix B for a fu spreadsheet, which incudes each attribute, attribute definitions, and the scoring bin descriptions for each attribute. However, other attributes were considered over the course of the piot and there are sti others that were never considered but are represented in other fisheries risk assessment toos across the gobe. In the future, CDFW or other users of the too may choose to incorporate these attributes, change definitions, or change scoring bin descriptions. Tabe 2 provides exampes of changes or rationae for excuding origina attributes (and their definitions). Data Quaity Scoring and Rationae Accounting for the quaity of the data used to score attributes is a key component of any modern ERA. Because ERAs ask a user to generate scores for fisheries that are both data poor and data rich, the data used can range from expert opinion to forma stock assessments. Because it is important to understand how the attributes were scored, experts are asked to provide a rationae about why they chose a particuar score. An expert can expain if the score was based on peer-reviewed iterature, andings data, persona observation, or some other source. The rationae used in scoring is very important for interpreting resuts, for ensuring standardization in scoring rationae among experts, and for transparency with stakehoders. Because of the above, the experts are asked to assign each attribute score an additiona data quaity score (Tabe 3). The data quaity scores were adapted from the Monterey Bay Nationa Marine Sanctuary appication of the Samhouri and Levin (2012) methodoogy. Testing the Scoresheet and Generating Draft Resuts for Too Refinement For this piot, one CDFW fishery expert was responsibe for scoring each of the five chosen species, and competed the scoring for each fishery mutipe times. One of the fishery experts, the CDFW project ead, worked with the other fishery experts to review scores to ensure that a experts were scoring the attributes consistenty. Some of the CDFW fishery experts reached out to fishermen to assist in the scoring process. The process aowed for us refine the spreadsheet and to standardize the scoring rationae among attributes. Additionay, we received feedback on the spreadsheet from participants at the stakehoder workshops. This heped us to further understand each fishery and resuted in more spreadsheet and scoring changes. Whie this process was beneficia for deveoping and pioting the too itsef, it is not necessariy the process for future use. This is discussed further in Section 4: Next Steps. Anayzing Scores Approach to Assessment of the Three Ecosystem Components: Target, Bycatch, and Habitat Mutipe approaches for methodoogy and anaysis were expored as part of this piot. Efforts were made to not choose methodoogies that fit our preconceived notions of the resuts. When making choices the project team reied on peer-reviewed scientific iterature to draw from tested methodoogies, ruing out methodoogies that produced resuts that were highy unreasonabe, and accepting methodoogies even if they produced surprising, but sti beievabe resuts. Additionay, at times the project team saw certain decisions as being more poicydriven than science-driven. For exampe, in some management contexts it might make sense to weight certain bycatch guids higher than others. METHODS 15

26 Tabe 1. Exposure and sensitivity attributes for the three ecosystem components of the ERA. Exposure attributes Sensitivity Attributes Target Bycatch Habitat Current anding trends and management strategy MPA coverage and/or other permanent spatia cosure (MPAs, RCAs, etc.) Gear Seectivity Spatia intensity Tempora intensity Vaue of expoited species Age at maturity Behaviora response Fecundity Breeding strategy Fishing mortaity Management effectiveness MPA coverage and/or permanent spatia cosures (RCA, gear cosures) Current status Spatia intensity Tempora intensity Magnitude Age at maturity Behaviora response Fecundity Reease mortaity Management effectiveness MPA coverage and/or permanent spatia cosures (RCA, gear cosures) Spatia overap Tempora (# months /yr) and other fishery cosures Gear footprint Current status Recovery time Popuation connectivity Potentia damage to habitat from fishing gear Tabe 2. Exampes of changed or deeted attributes from Piot ERA Scoresheet. Origina Attribute Origina Attribute Description Change or Rationae for excuding Management effectiveness and current stock status The track record of current management approaches used to mitigate impacts of fishing activity, taking into account stock status when known For target ERA, changed to current anding trends and management strategy to refect both data and metrics used by CA fishery managers and to make the attribute score bin descriptions more concrete and ess subjective. Simiar changes that mirror this decision were made to the bycatch and habitat ERAs as we. Life Stage Intensity Life stage(s) affected by a stressor; if stressor affects individuas before they have the opportunity to reproduce, recovery is ikey to be inhibited The intensity of the fishing pressure (amount of catch) in the area where a habitat is known to occur This attribute was removed from both the target ERA and the bycatch ERA. Athough, this attribute has precedence in other ERAs there was not strong support from stakehoders or from State managers for its incusion. The scientists on the team did not think the overa ERA approach or score ERA was weakened by its remova. This was removed from the habitat ERA due to concern and feedback from the first stakehoder workshop that it was dupicative of spatia intensity. There shoud be consideration given in the future to whether this attribute shoud be added or combined with spatia intensity, especiay if a GIS anaysis component is pursued. METHODS 16

27 Tabe 3. Data quaity scoring descriptions. Score Description Exampe Very imited data. Information based on expert opinion surveys or on genera iterature reviews from a wide range of habitats or species. Limited data. Estimates with high variation and imited confidence, or based on studies of simiar habitats/species or of the foca habitat/species in other regions. Adequate data. Information is based on imited spatia or tempora coverage, moderatey strong or indirect statistica reationships, or for some other reason is deemed not sufficienty reiabe to be designated as best data Best data. Substantia information exists to support the score and is based on data coected for the habitat or species in the study region. No empirica iterature exists to justify scoring for a foca habitat or species in reation to a particuar activity/pressure but reasonabe inference can be made by the person conducting the risk assessment. Scoring based on a study of a simiar habitat or species outside of the study region. Use of presence-absence data from ad hoc samping efforts; use of reativey od information; etc. Data-rich assessment of habitat or species status, with reference to historica extent and current trends. For the purposes of this piot study, we defined a fishery (or unit of anaysis ) as each unique target/gear/sector combination, simiar to the PSA approach. The piot ERA produces a reative risk score for impacts to target, bycatch, and habitat separatey. When mutipe units of anaysis exist for a target species, they can be averaged to produce an overa target-specific score. For exampe, Caifornia haibut was scored as four separate fisheries, to produce four separate scores, but can aso be averaged to create one score for the species across a fisheries. A resuts were anayzed using customized code to be utiized in an open source appication, R. Compementary anayses were conducted to enabe the presentation of different options for consideration by scientists and fisheries managers at CDFW. For a anayses, possibe scores for each attribute ranged from 1-4. This approach varies sighty from that used for the PSA, which ranged from 1-3 and did not specificay define a non-interaction score. Target ERA Anaysis For each target unit of anaysis in the target ERA, scores were added and averaged (i.e., arithmetic average method) independenty for Sensitivity and Exposure attributes. For a fished species with mutipe units of anaysis (e.g. Caifornia haibut), exposure and sensitivity resuts were further averaged to provide an overa target-specific score. Bycatch ERA Anaysis Each unit of anaysis in the bycatch ERA that has mutipe guids (e.g., easmobranchs, samonids) had scores for each guid summed and averaged by axis. Average vaues from each guid were further summed ( cumuative method ) to provide fina scores for each target or unit of anaysis. The arithmetic average method (as described above for target anaysis) does not incorporate sampe size (i.e., number of guids that interact with the fishery) and gives a guids in the fishery equa weight, though the weighting of individua bycatch guids coud be METHODS 17

28 modified if so desired by CDFW. By contrast, fisheries with more bycatch guids score higher using the cumuative method. The cumuative method aso gives greater emphasis to guids that score reativey high and are therefore reativey more at risk from the fishery. Two of the bycatch ERA attributes were weighted to refect their reative importance. One sensitivity attribute (reease mortaity) and one exposure attribute (magnitude) were each weighted to represent 50% of the tota score among a attributes within each axis. This decision was made based primariy on the feedback of stakehoders and CDFW, who fet that the risk associated with these attributes best matched their understanding of how fisheries affected bycatch. The reative weighting of bycatch attributes can be further modified as needed (e.g., by choosing to weight reease mortaity at 50%, magnitude at 25%, or choosing different attributes atogether). Reative weighting was not impemented among bycatch guids because doing so woud require a subjective vaue judgement. Options exist for dispaying the resuts that can highight fisheries that interact with specia status species or which interact with higher numbers of bycatch guids (see section 3.2 Resuts for exampes of this). Graphica Dispay of Resuts Resuts are represented most simpy by ecosystem component-specific (target, bycatch, habitat) figures that dispay reative risk among targets and units of anaysis, or among bycatch or habitat groups within a target or unit of anaysis (Figure 4). Sensitivity and Exposure scores combine to create an overa risk score, with absoute and reative risk scores easiy interpreted using contour ines of equivaent risk drawn at intervas ranging from 1-4. For bycatch resuts, the size of the dots on the figure indicates how many guids with protected status (maximum = 3) contributed to the overa score for a target or unit of anaysis. For habitat resuts, the size of the dots on the figure indicates how many groups contributed to the overa score for a target or unit of anaysis. Data Quaity Figure 4. A generaized representation of As part of this piot, data quaity was scored for every how ERA resuts can be interpreted. attribute and a written rationae for the attribute was provided. Data quaity was not, however, integrated into any fina resuts because imited time and effort were focused on finaizing resuts for the three ecosystem components. The project team discussed options for assessing data quaity and stakehoders provided feedback as we. The treatment of data quaity remains as one of the key decision points for the State or other users to consider in the future. There are we-estabished methodoogies and the piot too can easiy incorporate any of the options, since the data are readiy avaiabe, once a decision has been made. The options for this key decision are discussed in Section 4: Next Steps. Sensitivity Lower potentia risk Exposure Higher potentia risk METHODS 18

29 3.2 Resuts The resuts presented in this report represent the fina piot resuts for the pioting and testing of this ERA too. We produced and updated draft resuts throughout the iterative too deveopment process to hep us understand if the too was meeting its intended goa - to meet the State s prioritization needs with scientific rigor. The fina piot resuts are presented to hep understand how the too works and how resuts shoud be interpreted if the too is impemented. These resuts do not indicate a mandate for management action. Rather, they indicate a potentia risk that shoud be examined further by users of the too to determine whether and if a management action is needed. We pioted nine units of anaysis, representing five fished species. This ERA too coud be used to score the remaining 36 fisheries for which PSAs were competed. Aong with the PSA resuts, the 45 ERAs coud then be used to hep CDFW prioritize these fisheries for management actions Target Resuts Of the nine fisheries assessed, reative risk to target species was greatest for white sturgeon in the sport hook and ine fishery, foowed by Caifornia haibut in a four fisheries (Figure 5). For a five of these fisheries, reativey high risk scores resuted from high scores for neary a exposure attributes. Reative risk to Caifornia spiny obster from trap and sport/hoop fisheries and to kep bass from the hook and ine fishery was simiar to reative risk to Caifornia haibut from its four fisheries, but due to somewhat higher sensitivity scores rather than exposure scores. For obster, high fishing mortaity, behaviora response, and age at maturity scores ed to high sensitivity scores, whereas for kep bass sensitivity scores were reativey high because of popuation connectivity and breeding strategy considerations, in addition to their behaviora response to the fishery. Pacific herring exhibited the owest reative risk due to the commercia gi net fishery, with ow scores for most exposure and sensitivity attributes. 4 3 Fishery Key CGN - commercia, gi net CHL - commercia, hook & ine CTP - commercia, trap CTR - commercia, traw SH - sport, hoop net SHL - sport, hook & ine Sensitivity 2 Herring CGN Lobster CTP Lobster SH Bass SHL Haibut CHL Haibut CGN Haibut CTR Haibut SHL Sturgeon SHL Reative Risk Risk Higher Lower Exposure Figure 5. Target risk assessment resuts for the ERA piot. RESULTS 19

30 3.2.2 Bycatch Resuts The rank ordering of reative cumuative risk to bycatch guids differed from that of risk to target species (Figure 6). Specificay, the Caifornia haibut commercia gi net and traw fisheries posed the greatest reative cumuative risk. This resut emerged because both fisheries interact with six of the ten bycatch guids, more than in the other fisheries. In the case of the commercia gi net fishery, reative cumuative risk to bycatch guids was caused by the high sensitivity scores for other fatfishes, peagic finfishes, non-peagic finfishes, and threatened and endangered species and/or overfished rockfishes. The cause of high reative cumuative risk for bycatch guids due to the Caifornia haibut commercia traw fishery was more variabe, with some guids ike marine mammas that had high sensitivity scores and others ike fatfishes with high exposure scores. Intermediate reative risk scores for bycatch associated with the sport hook and ine fisheries for white sturgeon and kep bass resuted from their interactions with five bycatch guids each. For both fisheries, non-peagic finfishes were highy exposed, but not very sensitive, whereas other bycatch guids were expected to be moderatey exposed and sensitive. The remaining five fisheries exhibited much ower reative cumuative risk to bycatch guids, argey because those fisheries interact ony with one or two bycatch guids. It is important to note that since the Bycatch ERA utiized the cumuative approach, the fina resuts were more heaviy infuenced by the number of bycatch guids that interacted with a fishery than by individua attribute scores. Sensitivity Lobster CTP Haibut CHL Herring CGN Lobster SH Haibut CGN Sturgeon SHL Bass SHL Haibut SHL Haibut CTR Exposure Fishery Key CGN - commercia, gi net CHL - commercia, hook & ine CTP - commercia, trap CTR - commercia, traw SH - sport, hoop net SHL - sport, hook & ine Number of affected protected groups Reative Risk Reativ Higher Lower Figure 6. Target risk assessment resuts for the ERA piot. Point size indicates the number of bycatch guids with protected status affected by each fishery (maximum possibe was three). RESULTS 20

31 3.2.3 Habitat Resuts Reative risk to habitats was greatest for those affected by the Caifornia haibut commercia traw fishery, which incuded nearshore soft bottom and habitat-forming marine invertebrates (Figure 7). Whie these two habitats were both highy exposed, neither was expected to be particuary sensitive. Compared with the risk assessment for bycatch groups, the remaining eight fisheries did not show as much variation in risk score for habitats. Habitats infuenced by the commercia gi net fishery for Caifornia haibut showed the greatest reative risk of those in an intermediate risk group, which aso incuded the two Caifornia spiny obster fisheries, the commercia gi net fishery for Pacific herring, and the sport hook-and-ine fishery for white sturgeon. Notaby, the two Caifornia spiny obster fisheries affect the most habitats, incuding nearshore hard and soft bottom, marine vegetation, and invertebrates. The owest reative risk to habitats emerged for three hook-and-ine fisheries: the sport fishery for kep bass and both the commercia and sport fisheries for Caifornia haibut. Of the three habitats infuenced by the kep bass fishery, ony marine vegetation exhibited somewhat high exposure and sensitivity scores. 4 3 Fishery Key CGN - commercia, gi net CHL - commercia, hook & ine CTP - commercia, trap CTR - commercia, traw SH - sport, hoop net SHL - sport, hook & ine Sensitivity 2 1 Lobster CTP Lobster SH Haibut CGN Bass SHL Herring CGN Sturgeon SHL Haibut SHL Haibut CHL Haibut CTR Exposure Number of affected habitat groups Reative Risk 4 Higher Lower Figure 7. Habitat risk assessment for the ERA piot. Point size indicates the number of habitat groups affected by each fishery (out of 10). RESULTS 21

32 What is the reative risk posed by fishing activities to TARGET species, BYCATCH guids and HABITATS? Target ERA Bycatch ERA Habitat ERA Sensitivity Sensitivity Sensitivity Exposure Exposure Exposure What is the reative risk to TARGET SPECIES among haibut fisheries? HIGHER RISK What is the reative risk to BYCATCH GUILDS among haibut fisheries? HIGHER RISK Which is the reative risk to HABITATS among haibut fisheries? HIGHER RISK LOWER RISK LOWER RISK LOWER RISK What attributes are driving higher reative risk? Which BYCATCH GUILDS exihibit higher reative risk? Which HABITATS exhibit higher reative risk? Spatia intensity Vaue Gear seectivity HIGHER RISK TEP/overfished Fatfish Peagic finfish Non-peagic finfish HIGHER RISK LOWER RISK Habitat-forming marine invertebrates Nearshore soft bottom KEY COMMERCIAL LOWER RISK Invertebrates Easmobranchs SPORT TRAWL GILL NET HOOK & LINE What attributes are driving higher reative risk? Reease mortaity Age at maturity Current status Tempora intensity What attributes are driving higher reative risk? Fishing gear damage Gear footprint Spatia overap Tempora & other cosures POTENTIAL ACTIONS Prioritize for management action Monitor or coect more information No action needed Figure 8. Resuts anaysis demonstration for Caifornia haibut. 22

33 4. Next steps Image credit: Mooseaope Going forward, it is within the purview of CDFW or other potentia users to consider whether and if this piot ERA meets their needs. This piot test invoved many key decisions. Severa others remain shoud CDFW choose to impement the ERA in whoe or in part. This section outines those key decisions, considerations and options for future use of the too. 4.1 Key Decision Points for Too Refinement and impementation Many decisions regarding too deveopment and impementation were agreed upon by the project team with feedback from stakehoders. Data Quaity consideration (discussed in Section 3: Methods). Before impementation, we recommend deciding on a preferred method for incorporating data quaity information. This information is hepfu for interpreting resuts, understanding data gaps, and mobiizing pubic and private priorities for research and monitoring. We have summarized the three main options (athough variations of each exist) and considerations for data quaity anaysis beow. However severa remain and are discussed beow. 1. Keep data quaity scores independent from resuts. In this option, data quaity scores are anayzed but do not impact the overa score of the fishery itsef. This separate anaysis aows for an understanding of the reative eve of data quaity in each fishery and can aso be used when interpreting resuts. Furthermore, the independent data quaity scores can be used to identify knowedge gaps reated to certain attributes, fisheries, and/or ecosystem components. The option to keep data quaity scores independent was used in the NOAA PSA scoring. However, we recommend that the data quaity scores be integrated into the resuts figure for easier interpretation instead of creating a separate graph of data quaity scores as was presented in the NOAA PSA MRAG report (MRAG 2016) (Figure 9). 2. Low data quaity fisheries are weighted higher. In this approach, fisheries that receive ower data quaity scores are weighted more heaviy overa. This approach was deveoped by the creators of PSA and is considered to be more precautionary (Hobday et a., 2007; Hobday et a., 2011). Whie it has been adopted by others, the NOAA PSA departed from this approach in its appication. NEXT STEPS 23

34 Reative Risk 4 Higher Lower Data quaity score Limited data Best quaity data Figure 9. Sampe resuts with data quaity scores integrated into the figure. 3. Higher data quaity attributes are weighted higher in overa score. In the previous appication of the Samhouri and Levin (2012) ERA method, attributes that were scored with high data quaity were given a higher weight in the overa risk score for a Unit of Anaysis. Create open access scoresheet and automated resuts anaysis. Once key refinements and decisions have been made, we recommend that the too be structured into a format that is user-friendy and automates the anayses and resuts. This coud be accompished through either an automated spreadsheet and/or webbased app utiizing the methods and R code contained in Appendix XX. Timeine for updates and addition of fisheries. We have deveoped an ERA that is based on best practices, and is transparent and repeatabe. This means that as new information becomes avaiabe, new fisheries need assessment, or as time progresses, the State can add or reassess fisheries reativey quicky. Over time, this too becomes much faster to utiize, especiay for fisheries previousy assessed where the user is just re-evauating changes in attribute scores. Stakehoders emphasized that any prioritization too shoud be updated periodicay, rather than conducted as a one-time anaysis. We recommend consideration be made as to whether and how frequenty to re-assess or add new fisheries to the anaysis as part of their ERA impementation pan, and note that some attributes wi require more frequent updating (e.g., current andings trend and management strategy for target ERA) than others (e.g., age at maturity for target ERA). Definition of bycatch. The bycatch ERA can accommodate different definitions of bycatch. Whatever definition is chosen must be adhered to and impemented consistenty by the fishery experts conducting the scoring. For the purposes of the piot ERA, bycatch was defined as catch that is returned to the water. Under this definition, sub-ega and supra-ega sized individuas of the target species are considered and NEXT STEPS 24

35 scored as bycatch, as are individuas that are intentionay reeased even though they can egay be retained (catch-and-reease fishing). This scoring decision coud potentiay infate bycatch risk for a fishery that may not typicay be considered to have high rates of bycatch. Process for scoring. There is a wide array of iterature and methods for utiizing expert judgment to conduct risk assessments. For the piot, OST foowed some of the best practices as described in Burgman (2005), but not others due to the preiminary nature of the work. Many ERAs use the modified dephi version as described in Burgman (2005). In this method, users are trained and briefed on the too, then score the assessment, discuss their scores with others, foow-up with independent experts, then choose or combine to make fina scores. We recommend that users review the Dephi method and make informed decisions about how scoring is competed beyond the piot. We aso recommend that mutipe experts score each Unit of Anaysis to incorporate different opinions that may ead to variabiity among attribute scores, or have experts convene to discuss and either agree upon or make changes to scores based on discussion. Regiona assessment of fisheries. For the ERA piot, the project team assessed state marine fisheries, i.e. the scae of our assessment was for fisheries operating within state waters with no geographica partitioning. However, during the stakehoder workshops, the project team received feedback that a regiona approach for some fisheries may be more appropriate. For exampe, some fisheries, within the same sector and using the same gear, may operate at different scaes and have different bycatch species north and south of Point Conception. This practice coud resut in different eves of reative risk between ocations. The ERA piot too can accommodate different definitions of fisheries (or Units of Anaysis) without changing the too itsef. Therefore, it can anayze fisheries any way the user chooses to define them. 4.2 Key Decision Points for PSA vs. ERA for Target Fisheries For the information gathering phase of the MLMA Amendment, there were two options pioted for prioritizing target species for impacts from fishing activity. The PSA target species anaysis was requested by CDFW and the target ERA piot component was pursued by OST and NOAA to provide another option aongside the bycatch and habitat ERAs. Both of these toos represent scientificay sound options for the State to consider. As a next step, CDFW wi evauate the usefuness of both of these toos for consideration as part of the amended Master Pan. Whie both toos have simiar structure, methodoogies, and goas, there are differences to consider when seecting a too for future use (Tabe 4). There are aso options for combining the PSA and ERA into a singe too and for transitioning between toos over time (Box 6). 4.3 Key Decision Points for Stakehoder Engagement To increase the probabiity of success, CDFW is encouraged to consider mutipe approaches to stakehoder engagement and decide upon the best approach given avaiabe resources, capacity, and priorities, before the impementation of any risk assessment. Best practices for ERA deveopment and impementation invove stakehoders (Hobday et a. 2011, Fetcher 2005). In genera, stakehoder engagement can ead to better data and knowedge than academic scientists or managers can generate aone. Fishermen and other stakehoders often have distinctive insights and information that can greaty inform the resuts of an ERA. Their invovement in the scoring process of an ERA aso buids trust and buy-in to the resuts of the too. Typicay, the eve of effort invoved in engaging stakehoders in ERA processes decreases over time, as has been the case in many instances of stakehoder participation in ERAs, particuary in Austraia. Over time, ess effort is required for stakehoders to earn and understand the too itsef, and often times ony a few attributes (ones with new information or that are contentious) require ongoing or reguar updates or discussion as fisheries are revisited. NEXT STEPS 25

36 Tabe 4. Comparison of considerations for the NOAA PSA and piot ERA for target species. Consideration NOAA PSA Piot ERA (adapted from Samhouri and Levin, 2012) Scientificay rigorous Yes Yes Independenty peer-reviewed Yes Previous versions of the too were peerreviewed Estabished methodoogy Yes This version of too is not yet estabished Number of fisheries competed for CA 45 9 Spread of fina scores Less More Abiity to be further customized to consider stressors other than fishing No Considers MPA network Not expicity Yes Yes Customized for Caifornia Participation Pubicy accessibe spreadsheet with automated resuts No, customized AUS methodoogy for federa U.S. fisheries Deveoped in consut with NOAA fishery scientists Yes Yes Deveoped in consut with NOAA and CDFW fishery scientists and stakehoders No, woud need to be created As part of this piot project, the project team convened two stakehoder workshops to introduce the ERA too whie it was sti under deveopment (Appendix A). The primary goa for seeking input from stakehoders at this eary stage was to ensure the too was deveoped to refect stakehoder priorities and needs. Engagement aso offered an opportunity to share information about the ERA process and its potentia roe in supporting fisheries management in Caifornia. Visit the OST webpage (here) for the project for a summary of key themes and highights that incudes an overview of discussion topics, key questions, and identified next steps that emerged from both workshops, as we as input received during informa discussions with participants. During the stakehoder workshops, participants expressed interest in heping to score fisheries in the future if CDFW were to adopt an ERA too for prioritizing fisheries and recommended severa approaches to stakehoder invovement (here). There are different approaches to stakehoder engagement, which vary in the eve and timing of invovement by stakehoders. Each of these approaches must be examined in ight of the current needs and capacity of CDFW and it s partners. Not a of the options may be feasibe. Based on feedback from the workshops, exampes and essons earned from stakehoder engagement for other ERAs, and considerations about efficiency and capacity of CDFW, the foowing are severa approaches to stakehoder engagement for CDFW s discussion and consideration: Draft ERA scores and spreadsheets coud be made pubicy avaiabe onine and stakehoders coud provide feedback through written comments or submit a competed score sheet to be considered CDFW. Scoring panes coud be created for each fishery and those CDFW fishery experts in charge of scoring a NEXT STEPS 26

37 BOX 6. OPTIONS FOR TARGET SPECIES PRIORITIZATION THAT COMBINE OR TRANSITION THE NOAA PSA AND PILOT ERA. Combine both methods: Take unique eements of the target ERA and insert into NOAA PSA. Unique attributes (such as the MPA attribute) from the customized Target ERA that are not part of the NOAA PSA woud be added to the anaysis of the PSA. Considerations: Creates a customized PSA approach that considers aspects of stakehoder feedback from workshops as we as key state assets such as the MPA network. Creates an anaysis with more attributes and therefore information about potentia risk from fishing impacts. No onger has a readiy avaiabe open access nature of the too itsef (NOAA PSA avaiabe onine as exce workbook). This combined version was not tested, however each of its sub-components (NOAA PSA and Target ERA) were tested. Care needs to be taken in the cross-wak from ERA Exposure and Sensitivity attributes to PSA Productivity and Susceptibiity attributes. Need to re-score and reanayze resuts, which may not be a time-consuming undertaking since much of the structure is aready in pace Phased: Transition from NOAA PSA to ERA Can utiize the ERA, if that is the preferred too, over time to imit any cumbersome front end of impementation issues. Considerations: Utiize resuts of PSA right now, but moving forward use the ERA instead (assuming the State pans to continue to re-evauate fisheries and evauate new ones). Effort to conduct Target ERA woud be more time consuming than updating PSA scores. PSA and ERA scores coud be made to be equivaent (i.e. a score of on PSA is the same as the ow risk score of 1-2 on ERA) to simpify this transition. Aows the State to utiize the PSA scores now, with no further modification. If using PSA in the short term, target, bycatch, and habitat cannot be scored on the same scae immediatey (however, they can sti be paced into high, medium, and ow risk score bins). NEXT STEPS 27

38 fishery woud be required to either conduct or review their scores with a diverse group of stakehoders (e.g., fishermen, academic scientists, and environmenta NGOs). CDFW fishery experts woud be responsibe for considering stakehoders recommendations for scoring changes and documenting whether a scoring request was approved or denied with rationae. CDFW coud choose to hod briefings and/or webinars with stakehoders to share information and fina resuts without soiciting feedback or making changes to the scores based on stakehoder knowedge. Simiar to the ERA workshops hed during the piot, CDFW coud host workshops to share, discuss, and refine scores for fisheries with stakehoders. For exampe, CDFW coud present draft spreadsheets and work with stakehoders to compete fina scoring, or coud score fisheries in rea time with stakehoders. The former approach coud be more suitabe for an ERA too that incudes attributes that are not based on vaues or information avaiabe in peer-reviewed iterature vaues (e.g., fecundity is a iterature-based attribute, but spatia intensity is not and coud therefore benefit from mutipe types of knowedge to determine a more accurate score). 5. Concusions This piot ERA process resuted in a too that addresses potentia management needs and can be conducted reativey efficienty whie remaining scientificay rigorous. The process we used to create the piot ERA was highy iterative and coaborative between NOAA, OST, CDFW, and incorporated stakehoder input. The resut is a too that can argey be conducted by CDFW shoud sufficient resources and capacity be avaiabe. CDFW can now assess if the too meets their current management needs to prioritize fisheries based on the three ecosystem components - target, bycatch, and habitat. Key decision points for potentia future use of the too pertain mosty to too impementation and stakehoder engagement, rather than too deveopment. This process demonstrated a way for CDFW to seect an existing ERA for customization to meet their specific management goas and ecosystem features. This piot ERA and process can be utiized by others to meet simiar goas or to seect and customize their own ERA efficienty. CONCLUSION 28

39 References Arkema, K.K., Verutes, G., Bernhardt, J.R., Carke, C., Rosado, S., Canto, M., et a Assessing habitat risk from human activities to inform coasta and marine spatia panning: a demonstration in Beize. Environmenta Research Letters 9: Battista, W., Narr, K., Sarto, N., and Fujita, R Comprehensive assessment of risk to ecosystems. Fisheries Research 185: Burgman, M., Risks and decisions for conservation and environmenta management. Cambridge University Press, Cambridge, UK. FAO The ecosystem approach to marine capture fisheries. FAO Technica Guideines for Responsibe Fisheries, No. 4 (Supp. 2): 112 pp. Fetcher, W.J., The appication of quaitative risk assessment methodoogy to prioritize issues for fisheries management. ICES Journa of Marine Science 62: Hapern, B.S., Wabridge, S., Sekoe, K.A., Kappe, C.V., Michei, F., D'Agrosa, C., Bruno, J.F., et a A goba map of human impact on marine ecosystems. Science 319: Hobday, A.J., Smith, A., Webb, H., Daey, R., Wayte, S., Buman, C., Dowdney, J., Ecoogica risk assessment for effects of fishing: Methodoogy. Report R04/1072 for the Austraian Fisheries Management Authority, Canberra. Hobday, A.J., Smith, A.D.M., Stobutzki, I.C., Buman, C., Daey, R., Dambacher, J.M., Deng, R.A. et a Ecoogica risk assessment for the effects of fishing. Fisheries Research 108: Link J. S., Brodziak J., Edwards S., Overhotz W., Mountain D., Jossi J., Smith T., et a Marine ecosystem assessment in a fisheries management context. Canadian Journa of Fisheries and Aquatic Sciences 59: O.M., Martone, R., Hannah, L., Grieg, L., Boutiier, J. and Patton, S. A risk-based framework for ecosystem-based oceans management. CSAP Working Paper 2012/P46. Morrison, J.P, Sharma, A.K., Rao, D., Pardo, I.D., Garman, R.H., Kaufmann, W., and Boon, B Fundamentas of transationa neuroscience in toxicoogy pathoogy. Toxicoogy Pathoogy 43: MRAG Americas Productivity and susceptibiity anaysis for seected Caifornia fisheries. Report to Caifornia Ocean Science Trust and Caifornia Department of Fish and Widife. 55 pp. Caifornia Ocean Science Trust Ecoogica Risk Assessments: A Roadmap for Caifornia Fisheries. Avaiabe onine FINAL.pdf Patrick, W.S., Spencer, P., Link, J., Cope, J., Fied, J., Kobayashi, D., Lawson, P., et a., Using productivity and susceptibiity indices to assess the vunerabiity of United States fish stocks to overfishing. Fishery Buetin 108: Samhouri, J.F., and Levin, P.S Linking and- and sea-based activities to risk in coasta ecosystems. Bioogica Conservation 145: REFERENCES 29

40 Stezenmüer, V., Lee, J., South, A., Rogers, S., Quantifying cumuative impacts of human pressures on the marine environment: a geospatia modeing framework. Marine Ecoogy Progress Series 398: Wiiams, A., Dowdney, J., Smith, A.D.M., Hobday, A.J., and Fuer, M Evauating impacts of fishing on benthic habitats: a risk assessment framework appied to Austraian fisheries Fisheries Research 112: Winemier, K.O Patterns of variation in ife history among South American fishes in seasona environments. Oecoogia 81: Zhou, S. and Griffiths, S Sustainabiity Assessment for Fishing Effects (SAFE): A new quantitative ecoogica risk assessment method and its appication to easmobranch bycatch in an Austraian traw fishery. Fisheries Research REFERENCES 30

41 Appendix A: Stakehoder Workshops to Expore ERAs as a Potentia Prioritization Too to Support Fisheries Management in Caifornia Ocean Science Trust hosted two stakehoder workshops, one in northern Caifornia and one in southern Caifornia, were hed in June and Juy 2017 to expore the roe that ERAs may have in prioritizing and informing the management of Caifornia fisheries. Fisheries experts were invited to: Review draft ERA scoring for nine piot fisheries as exampes for considering target species, aong with habitat and bycatch risks, and provide feedback on the toos reated to CDFW and stakehoder priorities. Expore PSA, another type of risk assessment, which focuses on the risk of fishing activity to fisheries, and its preiminary resuts. Learn about the Marine Life Management Act (MLMA) Master Pan amendment process, incuding how ERA may support CDFW s broader prioritization goas. Workshops were hed in the foowing ocations: Southern Caifornia: Long Beach Date: Thursday, June 15 Time: 9:00 am 3:00 pm Location: Veteran s Park Community Center Northern Caifornia: Santa Rosa Date: Thursday, Juy 27 Time: 9:00 am 3:00 pm Location: Justice Joseph A. Rattigan Buiding Workshop resources Handouts Agenda 1 (here) Productivity Susceptibiity Assessment Overview 2 (here) Summary of Workshop Key Themes and Discussion 3 (here) Draft piot ERA score sheet 4 (here) Powerpoint presentations Introduction sides 5 (here) Productivity Susceptibiity Anaysis sides (45 fisheries) 6 (here) Ecoogica Risk Assessment sides (9 piot fisheries) 7 (here) 1 2http:// 3http:// 4http:// 5http:// http:// APPENDIX A 31

42 Appendix B: Attribute Tabes Tabe 5. Exposure attributes, definitions, and scoring categories for target (T), bycatch (B), and habitat (H). * indicates that attribute does not vary across fisheries. Low resistance factors Sensitivity Description 4 (High) (Low) Behaviora response T,B Popuation-wide behaviora effect of the fishery (or fishing gear) on a species Behaviora response significanty increases impact (e.g., baited hook/pot/trap, ighted squid boats - attracted into it) Behaviora response increases impact somewhat (e.g., schooing behavior - herring) Behaviora response does not change impact (e.g., sedentary species) Behaviora response reduces impact (e.g., buit in inefficiency or abiity of a species to get out of a gear) Current status H The regiona status of the habitat; increasingy critica status signifies a decrease in the abiity of the habitat to recover from the impacts of the pressure High concern (endangered or threatened status or thought to be imperied); unrecognizabe or substantiay degraded compared to historica status Habitat is highy degraded, but either has no officia status or is undergoing significant or successfu management efforts to rebuid or restore Moderate to ow concern (e.g., impact studies exist but do not revea major probems); somewhat degraded compared to historica; efforts underway to rebuid the habitat No concern; negigibe difference from historica Fishing mortaity T The proportion of the tota popuation ethay removed from the fish stock by fishing activities. > <0.20 Reease mortaity B Sow recovery factors Fish surviva after capture and reease varies by species, region, and gear type or even market conditions, and thus can affect the susceptibiity of the stock. Probabiity of surviva < 25% Probabiity of surviva 26-50% Probabiity of surviva 51-75% Probabiity of surviva > 75% Age at maturity T,B,* Age at maturity - popuation-wide average age at maturity; greater age at maturity corresponds to onger generation times and ower productivity >10 years 6-10 years 2-5 years <2 years Breeding strategy T,B,* The breeding strategy of a stock provides an indication of the eve of mortaity that may be expected for the offspring in the first stages of ife. Additiona information in Winemier or Externa fertiization and no parenta care with known ow successfu reproduction rates 2 or Externa fertiization and no parenta care 3 or Interna fertiization or parenta care but not both 4 or Interna fertiization and parenta care Fecundity T,B,* The popuation-wide average number of offspring produced by a femae each year <10e1 10e1-10e2 10e2 10e3 >10e3 Popuation connectivity H For biotic habitats, reaized exchange with other popuations based on spatia patchiness of distribution, degree of isoation, and potentia dispersa capabiity; based on monitoring surveys, and popuation genetic or direct tracking estimates. Abiotic habitats shoud be scored as 1. There is a recognized biogeographicay boundary for a or most individuas within the state (e.g. Point Conception); the habitat or some of the organisms that form it are isted species or have protected status. There is a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Point Conception); the habitat or some of the organisms that form it are NOT isted species or have protected status. There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has either an egg or arva dispersa period ess than 1 month or has no egg and arva dispersa period There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has an egg or arva dispersa period of 1 month or greater. Popuation connectivity T,B,* Reaized exchange with other popuations based on spatia patchiness of distribution, degree of isoation, and potentia dispersa capabiity; based on monitoring surveys, and popuation genetic or direct tracking estimates There is a recognized biogeographicay boundary for a or most individuas within the state (e.g. Point Conception); one or more popuation(s) is identified as DPS or EU There is a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Point Conception); no popuations are identified as DPS or EU. There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has either an egg or arva dispersa period ess than 1 month or has no egg and arva dispersa period There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has an egg or arva dispersa period of 1 month or greater. Potentia damage to habitat from fishing gear H Within the footprint of the fishery: The potentia modification of habitat when exposed to a fishery (gear, chains, anchors, boats, etc.). Potentia modification to habitat structure is caused by fishing activity using bottom traw or beam traw or new gear with unstudied effects Potentia modification to habitat structure is caused by fishing activity using traps on strings with ground ines and weights (e.g. spot prawn, hagfish fisheries), gi nets, or purse seine Potentia modification to habitat structure is caused by trap and hoop nets with individuas ines and foats (e.g. obster, Dungeness crab), or occasionay by anchor or chain damage from vesses not drifting when hook-and-ine or hand-coection type gear. No or insignificant potentia modifications to habitat structure caused by any of these gears used: hook-and-ine, cam fork, abaone iron, urchin rake, hand coection (e.g. sea cucumber fishery), A-frame, midwater traw Recovery time H For biotic habitats, we refer to recovery time of the habitat as a whoe (e.g., a mature kep forest) rather than recovery time of individuas. For abiotic habitats, shorter recovery times for habitats such as mudfats decrease the sensitivity of exposure to human activities, whereas for habitats made of bedrock, recovery wi occur on geoogica time scaes. Recovery time >100 years Recovery time >10 years Recovery time 1-10 years Recovery time <1 year APPENDIX B 32

43 Spatia and tempora factors Sensitivity Description 4 (High) (Low) Spatia intensity T,B The overap between the regiona abundance of the species and the reative intensity of the target fishery throughout the region (consider both area and vertica overap, but not MPA coverage) Very High overap (>40%) between species and fishery High overap (>20-40%) between species and fishery Moderate overap (10-20%) between species and fishery Low overap (<10%) between species and fishery Spatia overap H The regiona status of the habitat; increasingy critica status signifies a decrease in the abiity of the habitat to recover from the impacts of the pressure High concern (endangered or threatened status or thought to be imperied); unrecognizabe or substantiay degraded compared to historica status Habitat is highy degraded, but either has no officia status or is undergoing significant or successfu management efforts to rebuid or restore Moderate to ow concern (e.g., impact studies exist but do not revea major probems); somewhat degraded compared to historica; efforts underway to rebuid the habitat No concern; negigibe difference from historica Tempora intensity T, B The proportion of the tota popuation ethay removed from the fish stock by fishing activities. > <0.20 Reease mortaity B Sow recovery factors Fish surviva after capture and reease varies by species, region, and gear type or even market conditions, and thus can affect the susceptibiity of the stock. Probabiity of surviva < 25% Probabiity of surviva 26-50% Probabiity of surviva 51-75% Probabiity of surviva > 75% Age at maturity T,B,* Age at maturity - popuation-wide average age at maturity; greater age at maturity corresponds to onger generation times and ower productivity >10 years 6-10 years 2-5 years <2 years Breeding strategy T,B,* The breeding strategy of a stock provides an indication of the eve of mortaity that may be expected for the offspring in the first stages of ife. Additiona information in Winemier or Externa fertiization and no parenta care with known ow successfu reproduction rates 2 or Externa fertiization and no parenta care 3 or Interna fertiization or parenta care but not both 4 or Interna fertiization and parenta care Fecundity T,B,* The popuation-wide average number of offspring produced by a femae each year <10e1 10e1-10e2 10e2 10e3 >10e3 Popuation connectivity H For biotic habitats, reaized exchange with other popuations based on spatia patchiness of distribution, degree of isoation, and potentia dispersa capabiity; based on monitoring surveys, and popuation genetic or direct tracking estimates. Abiotic habitats shoud be scored as 1. There is a recognized biogeographicay boundary for a or most individuas within the state (e.g. Point Conception); the habitat or some of the organisms that form it are isted species or have protected status. There is a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Point Conception); the habitat or some of the organisms that form it are NOT isted species or have protected status. There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has either an egg or arva dispersa period ess than 1 month or has no egg and arva dispersa period There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has an egg or arva dispersa period of 1 month or greater. Popuation connectivity T,B,* Reaized exchange with other popuations based on spatia patchiness of distribution, degree of isoation, and potentia dispersa capabiity; based on monitoring surveys, and popuation genetic or direct tracking estimates There is a recognized biogeographicay boundary for a or most individuas within the state (e.g. Point Conception); one or more popuation(s) is identified as DPS or EU There is a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Point Conception); no popuations are identified as DPS or EU. There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has either an egg or arva dispersa period ess than 1 month or has no egg and arva dispersa period There is not a recognized biogeographicay boundary for a of or most individuas within the state (e.g. Pt. Conception), and the species has an egg or arva dispersa period of 1 month or greater. Potentia damage to habitat from fishing gear H Within the footprint of the fishery: The potentia modification of habitat when exposed to a fishery (gear, chains, anchors, boats, etc.). Potentia modification to habitat structure is caused by fishing activity using bottom traw or beam traw or new gear with unstudied effects Potentia modification to habitat structure is caused by fishing activity using traps on strings with ground ines and weights (e.g. spot prawn, hagfish fisheries), gi nets, or purse seine Potentia modification to habitat structure is caused by trap and hoop nets with individuas ines and foats (e.g. obster, Dungeness crab), or occasionay by anchor or chain damage from vesses not drifting when hook-and-ine or hand-coection type gear. No or insignificant potentia modifications to habitat structure caused by any of these gears used: hook-and-ine, cam fork, abaone iron, urchin rake, hand coection (e.g. sea cucumber fishery), A-frame, midwater traw APPENDIX B 33

44 Appendix C: Draft ERA Scoring Spreadsheets Direct downoad inks to a of the individua draft ERA scoring spreadsheets by fishery can be found beow and on the Ocean Science Trust website project page here: A (.zip): Bank: Caifornia haibut Commercia, hook and ine: com_hl_era-scoresheet-2017.xsx Sport, hook and ine: ERA-scoresheet-2017.xsx Commercia, gi net: net-_era-scoresheet-2017.xsx Commercia, traw: ERA-scoresheet-2017.xsx Kep bass Sport, hook and ine: HL_ERA-scoresheet-2017.xsx Pacific herring Commercia, gi net: ERA-scoresheet-2017.xsx Spiny obster Commercia, trap: trap_era-scoresheet-2017.xsx Sport, hoop net: hoop_era-scoresheet-2017.xsx White sturgeon Sport, hook and ine: rec_era-scoresheet-2017.xsx APPENDIX C 34

45 Appendix D: Protoco for Too Appication & Anaytica Methods Protoco for Too Appication - A User s Guide to Cacuating ERA Scores for Caifornia Fisheries *the R coding packages referred to in this section are avaiabe for downoad on the project site at: oceansciencetrust.org/projects/era/ Background Prior to data compiation and anaysis, you wi need to: 1) compete expert scoring for each fishery and unit of anaysis (e.g., target, bycatch, habitat) of interest, 2) save each corresponding fishery/unit of anaysis scoring spreadsheet as a.csv fie, and 3) downoad and insta R ( which wi serve as the patform for anaysis. Aternativey, you can downoad and insta R-Studio ( which is an interface for R that incudes a consoe, syntax-highighting editor that supports direct code execution, and toos for potting, history, debugging, and workspace management. Because the ERA for Caifornia fisheries contained nine separate fisheries, each with three units of anaysis (target species, bycatch groups, and habitat groups), 27.csv fies were generated. Each of these fies shoud be ocated in the same foder, caed Input fies*, as the R-code wi navigate to this foder during scoring compiation and anaysis. Four R-scripts are required: Compie scores, Risk code for bycatch, Risk code for habitats, and Risk code for target groups. Pace each script in a foder abeed Code*. Data Compiation The first step in processing scores is to compie a fishery data for each unit of anaysis. Open the Compie scores script in R or R-Studio. You wi need to ca severa ibraries (which perform specific compiation, anaytica, or graphing functions in the script) before you start. Most of these ibraries wi require additiona packages that aso must be instaed (instructions: Once a ibraries and packages are successfuy instaed, you can start to run the script to compie scores. The initia step in this process requires some text editing in ine 13 to change the path to match the ocation of your input fies. The script contains hepfu suppementa information and instructions throughout, a of which are indicated by a # symbo at the start of a ine of code. You can incude these statements when you run the script or omit them. R understands that these statements are text strings (not code) and wi ignore this information. Rather than running the entire script at once, it is usefu to run it in the indicated subdivisions to better foow and understand the procedure, and to pinpoint any errors as they arise. At ine 247, you wi need to identify the ocation of an Output Fies* foder where the compied scores.csv fie wi be created. Otherwise, you can run the code directy as written. Once this process is compete, you can begin anaysis of each fishery unit. Data Anaysis Each unit of anaysis is anayzed separatey using a distinct R script. Each script requires that you oad ibraries (and associated packages) and set the working directory to match that of your oca computer. The first step in the process is to read in compied scores and fiter them to match the unit of anaysis that you are anayzing. Run the script in the indicated sub-sections so that you can best understand the process and pinpoint any errors that may arise. For the target script, you wi need to edit ines 87, 129, and 149 to indicate the intended ocation of the output fies. In the bycatch script, edit ines 148, 227, 241, 261, and 278, and in the habitat script, edit APPENDIX D 35

46 ines 135, 187, 202, and 228. For each unit of anaysis, resuts are output as.csv fies, and aso used to generate figures. These figures incude: 1. Target risk by fishery, 2. Bycatch cumuative risk to a bycatch groups in each fishery, average risk to a bycatch groups in each fishery, risk panes by bycatch group for each fishery, and risk panes by fishery for each bycatch group, and 3. Habitat - average risk to a habitat groups in each fishery, risk panes by habitat group for each fishery, and risk panes by fishery for each habitat group. A figures are saved as.pdf fies, which aows for editing in Adobe Iustrator or simiar vector graphics creation and editing software. Short version: 1. Run the Compie scores script. a.) For target species, run the Risk code for target species script. b.) For bycatch groups, run the Risk code for bycatch script. c.) For habitats, run the Risk code for habitats script Anaytica Methods Defining Units of Anaysis and Assessing Risk We conducted scoring and anaysis for five target species, incuding Caifornia haibut, Caifornia spiny obster, kep bass, Pacific herring, and white sturgeon. These species were seected by CDFW because of the diversity of habitat types, fishery types (commercia or recreationa), and gear types that they represent. Mutipe fisheries were anayzed for two of the target species, resuting in nine units of anaysis as foows: four Caifornia haibut fisheries, two spiny obster fisheries, and singe fisheries targeting Pacific herring, white sturgeon, and kep bass. For each unit of anaysis, we assessed the risk posed by each fishery to the target species, 10 bycatch groups, and 10 habitat groups. The risk assessment for each unit of anaysis was based on the exposure and sensitivity of each target, bycatch, or habitat group (hereafter, foca component). Foca components that were reativey more exposed and more sensitive were considered to be at higher risk. CDFW experts quantified exposure and sensitivity based on the sets of attributes described beow. We use the term baseine to refer to attributes that did not vary among fisheries. A singe expert scored target, bycatch, and habitat for each unit of anaysis, and competed the scoring for each fishery mutipe times foowing discussions as a group to ensure the scoring categories were standardized. One fishery expert worked with the other fishery experts to review scores and ensure that the attributes were scored consistenty. The 10 bycatch groups and 10 habitat groups were chosen to be maximay representative whie not overy encumbering expert scorers with an exhaustive ist. Risk to a bycatch or habitat group was assessed for a representative species or feature within that group. Initia group definitions and orientation between axes were modified and finaized foowing interactive workshops with stakehoders. Bycatch and habitat groups not affected by a particuar unit of anaysis (fishery) were not scored by experts and were not incuded in anaysis. For bycatch, ten groups were initiay suggested by CDFW to contain the fu spectrum of species considered bycatch in Caifornia s marine fisheries, but to imit the number of groups so that scoring was not an unreasonabe task. The groups were refined by the project team and stakehoders during workshops. We used the most frequenty caught species within a group and scored it as appropriate. For seven of the 10 guids, a bycatch group APPENDIX D 36

47 was scored if it represented greater than 1% of the catch of the target species by either weight or number for any species within that group. For the other three groups (Marine Mammas, Marine Birds, and Threatened and Endangered species and/or Overfished Rockfish), if there was any bycatch of these groups, we scored the most common species within the group. Sub- and supra-egas of the target species were scored as bycatch. Ten habitat groups were seected based on knowedge of Caifornia coasta and oceanic ecosystems, management definitions of habitat utiized by CDFW, and avaiabiity of GIS mapping data, using the smaest number of groups feasibe for efficiency. For fisheries that occur in mutipe habitats, we estimated the percent of fishing activity that occurs in each habitat (sum = 100%) for use in subsequent anaysis. Exposure Attributes Definitions and scoring categories for each exposure attribute are incuded in Appendix B. Target Experts assessed exposure of each target species to each fishery based on two baseine attributes, the vaue of the expoited species and MPA coverage (and/or other permanent spatia cosure) in pace to protect the species. In addition, experts determined exposure of each target species based on four attributes that did vary among fisheries, incuding: spatia intensity, tempora intensity, gear seectivity, and current andings trend and management strategy. Bycatch Two baseine attribute and four additiona attributes were scored for bycatch groups. Current status and MPA coverage were considered baseine attributes. Non-baseine attributes incuded: magnitude, management effectiveness, spatia intensity, and tempora intensity. Habitat One baseine attribute (MPA coverage) and three additiona attributes were scored for habitat groups. Nonbaseine attributes incuded: management effectiveness, spatia overap, and tempora cosures. Sensitivity Attributes Definitions and scoring categories for each sensitivity attribute are incuded in Appendix B. Target Sensitivity of each target species to each fishery was scored by CDFW experts based on four baseine attributes and two attributes that varied among fisheries (Tabe 3). Baseine attributes incuded: age at maturity, breeding strategy, fecundity, and popuation connectivity. Behaviora response and fishing mortaity were the non-baseine attributes. Bycatch Four baseine attributes and two additiona attributes were scored for bycatch groups. Age at maturity, breeding strategy, fecundity, and popuation connectivity were baseine attributes. Behaviora response and reease mortaity were considered non-baseine attributes. Habitat Two baseine attributes and two non-baseine attributes were scored for habitat groups. Popuation connectivity and current status were the baseine attributes whereas non-baseine attributes incuded recovery time and potentia damage to habitat from fishing gear. APPENDIX D 37

48 Data Quaity Accounting for the quaity of the data used to score attributes is a key component of any modern ERA. Because ERAs ask a user to generate scores for fisheries that are both data poor and data rich, the data used can range from expert opinion to forma stock assessments. Because it is important to understand how the attributes were scored, experts are asked to provide a rationae about why they chose a particuar score. An expert can expain if the score was based on peer-reviewed iterature, andings data, persona observation, or some other source. The rationae used in scoring is very important for interpreting resuts, for ensuring standardization in scoring rationae among experts, and for transparency with stakehoders. Because of the above, the experts are asked to assign each attribute score an additiona data quaity score. The data quaity scores were adapted from the Monterey Bay Nationa Marine Sanctuary appication of the Samhouri and Levin (2012) methodoogy. Anaysis of Expert Scores We deveoped a framework for evauating the risk that fisheries wi ead to negative effects on marine species or habitats over the next ca. 20 years (assuming management practices continue unchanged) based on two axes of information. The first axis was reated to the exposure E of a species or habitat to a fishery, and the other axis was a conditiona probabiity reated to the sensitivity S of the species or habitat to the fishery, given its exposure. We define a negative effect as an unwanted outcome, here assumed to be the decine in abundance of a species or habitat. Anayses were conducted to enabe the presentation of different options for consideration by scientists and fisheries managers at CDFW and were utimatey taiored to best meet their needs. For a anayses, scores ranged from 1-4, with a 0 score indicating no interaction with a fishery or a very minor interaction (for habitats and bycatch species that do not interact with a fishery). This approach varies sighty from that used for the PSA, which ranged from 1-3 and did not specificay define a non-interaction score. The reative risk R i to species or group i was cacuated as the Eucidean distance of the species or group from the origin in a space defined by exposure and sensitivity indices, or Under this framework, the risk to a species or group increased with distance from the origin and each axis received equivaent weight in estimating risk. Vaues for each exposure attribute a e,i and sensitivity attribute as,i were determined by assigning a score ranging from one to four for a standardized set of A e or A s attributes (for the exposure and sensitivity axes, respectivey). These scores were used to cacuate an exposure or sensitivity index with each attribute weighted by a factor w i (ranging from [0,1] with ) reated to its importance, as and (1) (2) (3) Target For each target unit of anaysis, scores were added and averaged (i.e., Arithmetic Average Method) independenty for Sensitivity and Exposure attributes. Because a target species interact with their fisheries, no 0 vaues are associated with these units of anaysis. For target species with mutipe Units of Anaysis, Exposure and Sensitivity resuts were further averaged to provide an overa target-specific score. A attributes were assigned an equa weight w i =1/A e or w i =1/A s. APPENDIX D 38

49 Bycatch Bycatch units contained mutipe groups (e.g., easmobranchs, samonids) for each target or Unit of Anaysis; therefore, scores for each group were summed and averaged by axis, and those groups with 0 scores were removed from anaysis. Average vaues among categories were then further summed ( Cumuative Method ) to provide fina scores for each target or unit of anaysis. That is, after using Equation 1 to cacuate risk R b for each individua Bycatch group b, cumuative risk to a bycatch groups C B was cacuated as (4) The Arithmetic Average Method (as described above for target anaysis) does not incorporate sampe size (i.e., number of groups that interact with the fishery) and gives a groups in the fishery equa weight. By contrast, target and target/fisheries units interacting with more bycatch groups score higher using the Cumuative Method. The Cumuative Method aso gives greater emphasis to groups that score reativey high and are therefore more ikey to be heaviy affected by the fishery. Bycatch attributes were weighted to refect their reative importance. One Sensitivity attribute (Reease Mortaity) and one Exposure attribute (Magnitude) were weighted to represent 50% of the tota score for each axis. That is, w magnitude = 0.5 and w i = 0.5/(A e -1) for the other exposure attributes, whie w reease mortaity = 0.5 and w i = 0.5/(A s -1) for the other sensitivity attributes. This decision was made based primariy on the feedback of stakehoders and CDFW, who fet that these attributes were the main drivers of each category and provided resuts that better matched their understanding of how fisheries impacted bycatch. The reative weighting of bycatch attributes can be further modified as needed (for exampe, coud choose to weight Reease Mortaity at 50%, Magnitude at 25%, or choose different attributes atogether). Reative weighting was not impemented among bycatch groups because doing so woud require a subjective vaue-judgment. Options exists for dispaying the resuts that can highight fisheries that interact with specia status species or which interact with higher numbers of bycatch guids (see Resuts Section XX) for exampes of this. Habitat The Habitat foca component contained mutipe groups (e.g., kep, soft bottom) for each target or unit of anaysis; therefore, (as with bycatch units) scores for each group were summed and averaged by axis. A Weighted Arithmetic Average Method was used to cacuate fina scores. This decision was made argey based on stakehoder feedback and can be modified as needed. Instead of utiizing the cumuative method, each fished habitat h was weighted by a factor w h based on the reative amount of fishing effort. For exampe, if a fishery mosty interacts with estuaries (95%), and nominay with soft bottom (5%), the fina score wi be weighted using 95% of the estuaries score and weighted 5% with the soft bottom scores. That is, risk to a habitats R H was cacuated based on risk to individua habitats R h (from Equation 1) as (5) Like bycatch attributes, habitat attributes were weighted to refect their reative importance. One Sensitivity attribute (Potentia Damage to Habitat from Gear Type) and one Exposure attribute (Gear Footprint) were weighted to represent 50% of the tota score for a attributes on each axis. That is, w gear footprint = 0.5 and w i = 0.5/ (A e -1) for the other exposure attributes, whie w damage = 0.5 and w i = 0.5/(A s -1) for the other sensitivity attributes. This decision was made based primariy on the feedback of CDFW and stakehoders, who aso considered other options such as a gear mutipier and weighting of additiona attributes. The weighting of habitat attributes can be further adapted as needed. APPENDIX D 39

50 Image credit: Kaus Stiefe