Studies of no-take marine reserves: Methods for differentiating reserve and habitat effects

Highlights

• We examined how marine reserve studies differentiated reserve from habitat effects.

• Few studies accounted sufficiently for potentially confounding habitat effects.

• Partitioning of habitat effects has not greatly improved over time worldwide.

• Without BACI study designs, unequivocal reserve effects are difficult to detect.

• Few studies included baseline data and temporal monitoring, reducing robustness.

Abstract

No-take marine reserves (NTMRs) are promoted extensively as conservation and fisheries management tools in response to marine ecosystem degradation and fisheries decline, but their use remains controversial. Numerous studies indicate that NTMRs result in higher abundances of fishery-targeted marine species, but few studies utilised robust sampling designs. Comparisons of NTMRs and fished sites must account for potential confounding of habitat type and quality in heterogeneous marine environments. In a review of 164 NTMR studies, many studies failed to adequately account for habitat effects, with no apparent improvement in the literature through time (1983–2013). Five methods employed to partition reserve and habitat effects were used in conjunction with three basic sampling designs: single-point-in-time comparisons, temporal monitoring, and before-after-control-impact (BACI) studies. Researchers evaluating NTMRs: a) ignored habitat or described sites as broadly similar (28.7%); b) constrained sampling to similar habitats (43.9%); c) explicitly tested for habitat differences among sites (12.8%); d) used statistical methods to partition habitat and reserve effects (26.2%); and e) examined changes in habitat caused by reserve protection (17.1%). Over half of the studies (54.3%) made no statistical attempt to account for habitat effects. Robust studies, utilising methods (c-e) within BACI designs, are needed to accurately differentiate NTMR from habitat effects and permit managers to more accurately assess the benefits of NTMRs.

Introduction

Worldwide, fisheries stocks and marine environments are severely affected by anthropogenic impacts. Excessive declines of fish stocks and catch-per-unit-effort are indicative of collapsing fisheries (Froese et al., 2012, Jackson et al., 2001, Myers and Worm, 2003, Pauly et al., 2002, Watson et al., 2012) and many iconic marine habitats are declining in quality and geographic extent (Burrows et al., 2011, De’ath et al., 2012, Wilkinson, 2004). No-take marine reserves (NTMRs) are often promoted as a tool to manage ecosystems more holistically and fisheries more conservatively. In this review, NTMRs are defined as marine protected areas in which all extractive activities, including fishing or resource extraction, are prohibited (Sobel and Dahlgren, 2004).

No-take marine reserves aim to spatially restrict impacts from fishing effort and fishing mortality (Allison et al., 1998, Lubchenco et al., 2003). They are usually employed for one or both of two broad objectives: to protect marine biodiversity and ecosystem health and services (Bohnsack, 1998, Lubchenco et al., 2003, Russ, 2002), and to improve fisheries by protecting a portion of stocks (Gell and Roberts, 2003, Pauly et al., 2002, Russ, 2002, Sale et al., 2005). Many studies worldwide of NTMRs have demonstrated positive reserve effects on targeted species (Babcock et al., 1999, Edgar and Barrett, 1999, Polunin and Roberts, 1993, Russ and Alcala, 2010). Meta-analyses generally conclude that NTMRs enhance density, biomass, body size, diversity, and fecundity of targeted species within their boundaries (Côté et al., 2001, Halpern, 2003, Lester et al., 2009, Molloy et al., 2009, Mosquera et al., 2000). However, the use and efficacy of NTMRs are controversial with vocal critiques (Agardy et al., 2003, Rife et al., 2012, Willis et al., 2003) and support (Lubchenco et al., 2003, Pauly et al., 2002, Roberts et al., 2005).

Possibly the greatest source of controversy regarding NTMR science is the quality of empirical evidence supporting their benefits, which is a function of study design (Jones et al., 1992, Russ, 2002, Willis et al., 2003). The most robust methods for evaluating environmental and ecological impacts, including NTMR implementation, are before-after-control-impact (BACI) (Underwood, 1992, Underwood, 1994) and before-after-control-impact-pair (BACIP) designs (Jones et al., 1992, Russ, 2002). However, if a study begins after reserve implementation, or resources for monitoring are limited, rigorous temporal or spatial sampling with baseline (“before”) data may not be possible (Claudet and Guidetti, 2010, Mosquera et al., 2000), leaving spatial comparisons as the only study option (Evans and Russ, 2004, Guidetti et al., 2005, Kamukuru et al., 2004, Miller et al., 2012, Polunin and Roberts, 1993). The majority of studies on NTMR effects lack baseline data and thus may confound reserve effects with spatial heterogeneity of the environment, particularly the benthic environment. A recent meta-analysis of reserve studies found that only 23 of 221 studies (10.4%) were of a BACI design (Lester et al., 2009), suggesting the large potential for confounding NTMR effects with habitat effects.

Few, if any, NTMRs are placed randomly. Reserves may be located deliberately in areas of high habitat quality, species richness, or fish biomass for conservation or tourism objectives (Roberts et al., 2003, Willis et al., 2003), e.g. Apo Island NTMR in the Philippines (Russ and Alcala, 2010). Other sites with lower targeted fish biomass may be designated as reserves for opportunism and socio-economic priorities (Pressey et al., 1993, Roberts, 2000), e.g. the Galapagos Islands (Edgar et al., 2004). If NTMR and control sites are initially dissimilar in factors such as habitat, larval supply, or history, their comparison is inevitably confounded by inherent site differences.

Habitat differences between NTMR and fished sites may confound analyses of NTMR effects, even when attempts were made to ensure sites were as similar as possible. Many NTMR studies have not accounted for potential confounding effects of habitat on marine assemblages, resulting in ambiguous conclusions about NTMR effectiveness and utility (Osenberg et al., 2006, Osenberg et al., 2011). Proponents of NTMRs may unintentionally overstate reserve benefits, and critics may claim that benefits are uncertain due to poor study design. Differentiating NTMR and habitat effects is a priority for NTMR assessments (Claudet and Guidetti, 2010).

Variations in quality, diversity, and extent of habitats have long been identified as primary contributors to distribution and abundance of fishes and invertebrates (Almany, 2004, Bellwood and Hughes, 2001, Carpenter et al., 1981, Chittaro, 2004, Luckhurst and Luckhurst, 1978). In this review, habitat typically refers to the benthic habitat in which the organisms reside, particularly coral and rocky reef habitat. Few studies of NTMRs quantitatively describe and account for heterogeneity of habitat (Côté et al., 2001, Osenberg et al., 2006), even though confounding effects of habitat in NTMR studies have been acknowledged for over a decade (Chapman and Kramer, 1999, Edgar and Barrett, 1997, Garćla-Charton and Pérez-Ruzafa, 1999). Including measures of habitat can improve assessments of NTMR effects (Claudet and Guidetti, 2010).

Given the importance of habitat in determining fish density and assemblage structure and the non-random placement of NTMRs, with the concomitant high potential for control sites having dissimilar habitats, it is essential to determine how potential habitat variation may affect assessment of NTMR effects. No studies to date have examined how habitat has been incorporated into NTMR assessments worldwide and how these methods influence conclusions about NTMR effects. This review examines how spatial heterogeneity of habitat has been accounted for in NTMR studies globally. For the purposes of this review, “reserve effects” (=NTMR effects) are defined as significantly higher density, biomass, species richness or significantly different assemblage composition of marine organisms between NTMR and control (fished) sites caused by NTMR protection. Specifically, this review aims to critique different methodologies, statistical and sampling, for differentiating NTMR and habitat effects. These methodologies may be applied with different sampling designs: single-point-in-time, temporal sampling without baseline data, or before-after-control-impact. Thus, the review attempts to identify the combinations of methods used to account for habitat effects in NTMR studies and how these methods are combined with different sampling designs. It also suggests the best combinations of methods and sampling designs for differentiating reserve and habitat effects and describes the implications for science and management.