Evaluating anthropogenic threats to the Northwestern Hawaiian Islands PDF

Preview

Summary

AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS Aquatic Conserv: Mar. Freshw. Ecosyst. (2008) Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/aqc.961 Evaluating anthropogenic threats to the Northwestern Hawaiian Islands KIMBERLY A. SELKOEa,b,*, BENJAMIN S. HAL...

Description

AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS

Aquatic Conserv: Mar. Freshw. Ecosyst. (2008) Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/aqc.961

Evaluating anthropogenic threats to the Northwestern Hawaiian Islands KIMBERLY A. SELKOEa,b,*, BENJAMIN S. HALPERNb and ROBERT J. TOONENa a

b

Hawai’i Institute of Marine Biology, University of Hawai’i, Kaneohe Bay, Hawai ’i, 97644, USA National Center for Ecological Analysis and Synthesis, 735 State Street, Santa Barbara, CA 93101, USA ABSTRACT

1. Every manager must assess and prioritize anthropogenic impacts on their management area from a long list of threats, but data which allow comparison of the relative ecological impacts of threats for decision-making is often unavailable. 2. This study employed a systematic and standardized method to collect and quantitatively synthesize expert opinion on the ecological effects of anthropogenic threats to the world’s largest marine protected area: the Papahanaumokuakea Marine National Monument (MNM) in the Northwestern Hawaiian Islands (NWHI). 3. In contrast to most threat ranking exercises, the method provides detail on the reasons behind a particular ranking of threats to an area and why the ranking may vary across the ecological landscape. 4. Survey results from 25 experts allowed the comparison of the vulnerability of eight NWHI ecozones to 24 anthropogenic threats in a quantitative manner. 5. Ecozones tended to have distinct top threats: sea-level rise was identified as the top threat to rocky intertidal, beach and terrestrial ecozones; sea temperature rise was the top threat to the coral reef ecozones; bottom fishing was the top threat to the deep reef/bank ecozone; and pelagic fishing in the wider Pacific was the top threat to the pelagic ecozone owing to impacts on bird, turtle and fish fauna that forage outside the Papahanaumokuakea MNM boundaries. 6. Many of these top threats are difficult for local managers to control because they arise from activities outside the Papahanaumokuakea MNM boundaries, indicating that additional work is needed to preserve the NWHI despite its highly protected status. The analysis indicates where inter-agency cooperation in removing and mitigating threats should be focused. Copyright # 2008 John Wiley & Sons, Ltd. Received 1 July 2007; Revised 26 November 2007; Accepted 8 January 2008 KEY WORDS:

vulnerability; coral reefs; ecosystem-based management; marine protected areas; expert opinion survey

*Correspondence to: Kimberly A. Selkoe, NCEAS, 735 State St., Santa Barbara, CA 93101. E-mail: [email protected]

Copyright # 2008 John Wiley & Sons, Ltd.

K.A. SELKOE ET AL.

INTRODUCTION The Northwestern Hawaiian Islands (NWHI) make up one of the most remote and isolated ecosystems in the world. The region is, nevertheless, subject to dozens of anthropogenic threats, from both past and ongoing activities. Populations of lobster, pearl oysters, pink coral, and sport fish are depressed or over-fished from recent and historical extraction (Friedlander et al., 2005; Heinemann et al., 2005). Likewise, ongoing bottom fishing, large amounts of marine debris, occasional shipwrecks and oil spills, established alien species, and a variety of other threats continue to affect the natural environment. Nonetheless, the NWHI are considered one of the world’s most pristine marine ecosystems (Pandolfi et al., 2005) that harbour one of the highest proportions of endemic species of any marine ecosystem (DeMartini and Friedlander, 2004), creating immeasurable value to the world’s heritage and a need for protection and careful management of the region. The large size and isolation of the NWHI ecosystem create particular challenges for managers tasked with overseeing its protection. Moreover, management of this region (shown in Figure 1) requires coordination of four distinct agencies with sometimes overlapping jurisdiction. The US Fish and Wildlife Service (USFWS) oversees two National Wildlife Refuges:

Midway Atoll and the Hawaiian Islands National Wildlife Refuge, which contain all emergent land except Kure Atoll out to 10 fathoms. The State of Hawai’i Department of Land and Natural Resources (DLNR) oversees the 0–3 nautical mile (nm) zone and administers Kure Atoll as a State Wildlife Refuge. The National Marine Fisheries Service (NMFS) manages fishery resources and protected marine mammals and turtles in the Economic Exclusive Zone surrounding the NWHI. The National Marine Sanctuaries Program (NMSP) within the National Ocean Service (NOS) of the National Oceanic and Atmospheric Administration (NOAA) has jurisdiction over the remaining marine natural resources. In June 2006, Presidential Proclamation 8031 created the Papahanaumokuakea Marine National Monument (MNM) to provide the highest level of protection out to 50 nm around the NWHI (hawaiireef.noaa.gov/management/). The proclamation explicitly identifies the Secretary of Commerce (through NOAA), the Secretary of the Interior (through USFWS) and the State of Hawai’i (through DLNR) as cotrustees responsible for joint management of the natural resources of the Papahanaumokuakea MNM. Managers of the Papahanaumokuakea MNM, like those elsewhere throughout the world, must often prioritize limited funds for abating multiple anthropogenic threats to a

Figure 1. Map of the Hawaiian Archipelago, courtesy of NOAA. The Papahanaumokuakea Marine National Monument surrounds the emergent land of the NWHI out to 50 nm. Copyright # 2008 John Wiley & Sons, Ltd.

Aquatic Conserv: Mar. Freshw. Ecosyst. (2008) DOI: 10.1002/aqc

EVALUATING ANTHROPOGENIC THREATS TO THE NORTHWESTERN HAWAIIAN ISLANDS

protected area. Two obstacles typically arise in this situation: difficulty in comparing the impacts of diverse types of threats, and a lack of data on threat effects. Efforts to systematically evaluate the relative impacts of various threats are becoming more common. However, most approaches are developed for, and therefore tailored to, specific sites, threat datasets, or species groups, and so are difficult to adapt to a new situation (Cole and Landres, 1996; Jamieson and Levings, 2001; Zacharias and Gregr, 2005; Hiddink et al., 2007). The lack of data on the impacts of threats is usually addressed by relying on expert opinion (Reefs at Risk, Bryant et al., 1998; Kleypas and Eakin, 2007). Traditionally, only a few experts are haphazardly included. Because experts sometimes disagree, analyses based on the opinion of one or a few experts may not have good buy-in by other community members and stakeholders. Moreover, the process which led to experts’ recommendations is often undocumented and unclear, such that there is often no information on or standards for the criteria used to make the rankings. Thus, the rankings cannot be repeated or updated as new information becomes available or management goals evolve. In response to these obstacles, Halpern et al. (2007) developed an approach to assess the vulnerability of ecosystems to anthropogenic threats that can be applied to any ecological setting or list of threats, and it is applied here to the Papahanaumokuakea MNM. The approach builds on similar previous approaches that quantify or categorize threats by a combination of their intensity, extent, longevity, severity or duration of their direct and indirect effects (Cole and Landres, 1996; Bryant et al., 1998; Jamieson and Levings, 2001; Zacharias and Gregr, 2005; Ban and Alder, 2006). Expert opinion is assessed in a standardized, quantitative and

transparent way that allows issues of bias to be minimized or addressed and for results to be updated efficiently with new information (Halpern et al., 2007). The survey design allows fairly rapid data gathering (1 h to complete per respondent), so that large numbers of experts can easily be included and individual idiosyncrasies in opinion are subsumed by averaging a large sample. To apply the survey to the NWHI, the authors modified the Halpern et al. (2007) methodology to fit the unique list of threats and ecological communities of the NWHI. While most threats to the region are well known, the method here provides reasons why threats are ranked a certain way, and why their rankings might vary across the ecological landscape. With these data the study suggests management concerns for each ecozone, specific targets for inter-agency cooperation, and future research priorities. A second goal is to detail how the survey method can be adapted to a specific location, using the NWHI as a case study, for those interested in applying it to other regions of concern, and to identify needs for future improvements to the approach.

METHODS Approach The first step in assessing how the effects of threats vary across ecosystems is to build comprehensive lists of all relevant ecosystems and all possible threats (Halpern et al., 2007). Because the NWHI can be considered a single coral reef ecosystem, ecozones } distinct but interspersed benthic substrates or physical locations (Table 1) } were distinguished

Table 1. Descriptions and locations of ecozones used in the study Ecozone

Description

Distribution

Terrestrial Rocky Intertidal

Interior land distinct from the littoral zone. Solid substrate at intertidal depth composed of basalt rock.

Sandy Beach

Intertidal beach and adjacent shallows with soft benthos.

Algal Beds

Primarily Halimeda beds in lagoons and deeper terraces, but also small stands of endemic seagrass at Midway. Refers to shallow, mostly protected reef areas (lagoonal, back, reticulated or patch reefs). Exposed seaward reefs from the crest down to the slope less than 30 m depth. Deep reef is designated >30 m depth. Banks are sites of high relief benthos in deep waters with rich fish communities, with or without reef builders. The entire water column, from surface to depth, outside of lagoon and shallow reef environments.

Kure, Midway, Lisianski, Lysan, Necker and Nihoa Gardner Pinnacles, La Perouse Pinnacle (FFS), Necker, and Nihoa Kure, Midway, Pearl and Hermes, FFS, Lisianski, Laysan, and Nihoa Kure, Midway, Pearl and Hermes, FFS, Lisianski, Laysan, Necker and Nihoa Kure, Midway, Pearl and Hermes, Lisianski, Neva Shoal, Laysan, Maro, and FFS Most NWHI locations where depth is 530 m

Inner reef Outer Reef Deep Reef/banks Pelagic Waters

There are approximately 30 deep banks in the NWHI. Deep reef is usually found adjacent to any shallow reef area. Makes up the majority of NWHI habitat.

French Frigate Shoals is abbreviated FFS. Copyright # 2008 John Wiley & Sons, Ltd.

Aquatic Conserv: Mar. Freshw. Ecosyst. (2008) DOI: 10.1002/aqc

K.A. SELKOE ET AL.

instead of ecosystems. The eight defined ecozones attempt to match the spatial divisions of the management agencies. Aside from biological and physical differences, ecozones also differ in human use or threat exposure (e.g. inner versus outer reef), and history of observation and research. For example, study of reef habitats in the NWHI has mainly been conducted by scuba diving at depths 530 m, so reefs and benthic habitat >30 m depth were considered separately with the expectation that information about the two areas might differ. However, many important taxa such as seabirds, turtles, seals, sharks and jacks use many or all of the ecozones. Other ways of subdividing the NWHI ecosystem are possible but are not considered here. The list of threats drew from previous efforts to catalogue threats to the NWHI (Maragos and Gulko, 2002), consultation with several NWHI experts at the start of this project, and the global list of threats in Halpern et al. (2007). In total, 24 potential threats were identified for use in the survey (see Appendix 1 for definitions of threats). Seven additional threats were added by single respondents during the interviews; these are also listed in Appendix 1 but were not fully evaluated. Because this study focused on contemporary conditions in the NWHI, past threats that have recently abated were only considered in light of their ongoing or lingering impacts (e.g. the lobster trap fishery, closed in 2000). Potential future threats were mostly avoided so that respondents could base their answers on observation and data rather than speculation. For example, on the topic of alien species invasions, experts were asked to base their answers only on their knowledge of the current record of the effects of alien species, not estimates of future potential effects. Similarly, respondents were asked to avoid extrapolating known impacts of recent climate change to future worst case scenarios. To assess the relative vulnerability of each ecozone to each threat, five attributes of how a threat affects a given ecozone were assessed, termed ‘vulnerability factors’: (1) the spatial scale at which a single threat event causes effects; (2) the temporal frequency of the threat, measured as the average time from one event to the next; (3) functional impact of the threat, measured as the number of trophic groups affected; (4)

resistance of the ecological community to the threat, measured qualitatively by the tendency for the community affected by the threat to remain in its ‘natural’ state and resist disturbance; and (5) recovery time needed for the affected aspects of the community to return to their ‘natural’ or previous state following removal or remediation of the threat (Halpern et al., 2007). Information on these five attributes was sought at an ‘order of magnitude’ level to capture gross differences in how threats act (Table 2). In addition, for each threat } ecozone combination the interviewers asked respondents to indicate the ‘certainty’ of their answers based on the source of knowledge they drew on to answer (Table 2). The certainty scores are an indication of where expertise is strong and conversely where gaps in knowledge exist to target with future research. Note that in the context of the survey, certainty was not defined as the certainty of future outcomes of threats (e.g. climate change factors have less certain effects than anchor damage).

Survey methods To gather information on the five attributes of each threat, the authors surveyed scientists (i.e. individuals with a Master’s degree or Ph.D. in some field of science) who (1) had conducted field research personally in the NWHI and (2) had expertise in some aspect of the ecology of the NWHI. Scientists were contacted by email and asked to set up a meeting to complete the survey. They were given reminder emails or phone calls if no response was received after initial contact. The majority of respondents completed the survey during the meeting in 1–2 h, but four started the survey during the meeting and completed the survey on their own, and four completed the survey entirely on their own after reading detailed instructions because a meeting could not be arranged. This study focused on one-on-one interviews as opposed to a web-based survey as in the global analysis (Halpern et al., 2007) in hopes of increasing the response rate and minimizing any potential confusion about survey questions that could arise from improper understanding of the survey instructions and potentially add error to results.

Table 2. Ranking system for each vulnerability factor Scale

Frequency

Functional impact

Resistance

Recovery time

Certainty

No impact Rare Occasional Annual or regular

No impact Species level Single trophic group Multiple trophic groups

Totally resistant Highly Moderately Weakly

No Impact 51 year 1–10 years 10–100 years

Low Medium High

4

No impact 5100 m2 100m2–1 km2 1–10 km2 (small or partial atoll) 10–100 km2 (large atoll)

Constant

Entire community and/or habitat forming structure

>100 years

Very high

5

Entire NWHI

0 1 2 3

Copyright # 2008 John Wiley & Sons, Ltd.

Aquatic Conserv: Mar. Freshw. Ecosyst. (2008) DOI: 10.1002/aqc

EVALUATING ANTHROPOGENIC THREATS TO THE NORTHWESTERN HAWAIIAN ISLANDS

Interviews were conducted by one of two authors (KAS, BSH) between January and April 2006, using a standardized approach. The interviewers explained the goal of the project, the structure of the survey, the definitions of the ecozones and the vulnerability factors with examples of appropriate ranks for several threats. Then the respondents were shown the list of 24 threats and asked to choose the top three threats with the worst ecological impacts currently affecting the NWHI. This initial ranking of threats is called the ‘stated’ response. Finally, respondents filled in ‘vulnerability factors’ and a ‘certainty score’ for each threat in each ecozone (explained below), with the option of skipping threats and/or ecozones for which they felt they were unqualified to answer. Interviewers asked respondents to indicate mechanisms and relevant literature on the threats where possible.

all ecozones were used as replicates in an ANOVA and tested for statistical difference among three affiliation categories: government researchers (Pacific Islands Fisheries Center of NMFS, NOAA, Sea Grant and USGS affiliations), academic scientists (University of Hawai ’i and Bishop Museum affiliations) and managers (USFWS, DLNR and Monument affiliations). Similarly, difference in vulnerability scores and certainty scores by gender were assessed with t tests. To assess bias by number of days spent in the NWHI, ordinary least squares regression of number of days against mean vulnerability or certainty score of each respondent was used.

RESULTS Survey statistics

Data analysis To determine the ranking of threats, values for the five vulnerability factors were combined to give each threat a single ‘vulnerability score,’ which represents how vulnerable each ecozone is to each threat under current conditions in the NWHI. This analysis uses a simple average with equal weights, to be conservative. However, if the five factors were thought to have unequal importance in determining the impact of threats, a weighting scheme representing their relative importance could be applied. In all cases, rank values for scale and resistance were first rescaled from 0–4 so that all factors had the same range of values. Vulnerability scores for each threat in each ecozone were averaged across replicate survey responses, and threats were ranked by their vulnerability scores. To assess variance in responses, the standard deviation of a respondent’s vulnerability scores for each threat/ecozone combination was calculated. To evaluate whether sample size affected the mean or standard deviation of vulnerability scores, ordinary least squares linear regression with all threat–ecozone combinations as replicates (N=192) was used. Two approaches were used to determine whether the top three threats identified by the vulnerability analysis matched the ‘stated’ top three rankings. First, each respondent’s top three stated choices were compared with the top three identified from analysis of their vulnerability scores, for those respondents who completed the survey for all threats. Also, the stated rankings were averaged across all respondents and compared with the top three threats calculated from the mean vulnerability scores across all ecozones and all respondents. To assess possible respondent bias in both the vulnerability scores and certainty scores differences in responses among institutional affiliation, gender and number of days spent working in the NWHI were assessed. Each respondent’s mean vulnerability score or mean certainty score across all threats in Copyright # 2008 John Wiley & Sons, Ltd.

Forty-one people were contacted and verified as meeting the two criteria for participation in the survey. From this group, 25 (61%) filled in the survey and had a variety of affiliations (Table 3). Twenty-two ...