Fish assemblages inside and outside marine protected areas off northern Iceland: protection effects or environmental confounds?

Abstract

The density and mean size of demersal fish were analysed inside and outside three protected areas off the northern coasts of Iceland. One-way ANCOVA and mixed effects models, with depth, bottom temperature and tow duration as covariates, were used to determine differences in the two metrics between areas. In general, these differences were not statistically significant, except for one protected area in the northeast. In this area, small size classes of haddock (Melanogrammus aeglefinus) and long rough dab (Hippoglossoides platessoides) were found to be considerably more abundant and large size classes of cod (Gadus morhua) and haddock less abundant compared to the reference (fished) area. The mean sizes of haddock and long rough dab were smaller within this protected area. While these differences could not be confidently attributed to the closure, this area may be considered particularly suitable for the protection of juvenile fish. The large effect of covariates (mainly that of depth and temperature) on observed densities and mean size reasserts the necessity of their inclusion in models assessing the impact of area closures on fish communities. The study illustrates that a snapshot approach is unlikely to provide unambiguous evidence of the effectiveness of area closures. It can, however, provide useful information on various relationships between environmental gradients and fish distribution inside and outside a closure.

Introduction

It is widely recognised that marine protected areas (MPAs) are a potentially important regulatory measure to improve the management of human impacts on the marine environment (Horwood et al., 1998, Polunin, 2002, Gell and Roberts, 2003, Kaiser, 2005, Stefansson and Rosenberg, 2005, Jones, 2007, Jennings, 2009). Evaluation of their effectiveness, i.e. determining whether they actually meet their objectives, is a fundamental part of their implementation. It allows prompt modification of management and monitoring strategies (Gerber et al., 2005). More generally, such evaluation provides information to support management decisions within the framework of adaptive management (Agardy, 2000, Gerber et al., 2005).

Strict evaluation of the effectiveness of MPAs is difficult because in most cases, the objectives and criteria of success are poorly defined (Polunin, 2002, Willis et al., 2003). Moreover, availability of high-quality empirical data, such as time series, which would demonstrate the effectiveness of MPAs, is often limited. Empirical evidence for recovery in MPAs frequently suffers from a lack of rigour in the design of field surveys (Mosquera et al., 2000, Willis et al., 2003). Willis et al. (2003) identified the main problems associated with insufficient replication and a lack of control sites in evaluation studies. In the past, many sampling programmes for evaluation were initiated long after the establishment of MPAs (Pelletier et al., 2008), making it more difficult to ascertain whether observed differences between a given protected area and reference (fished) areas were an effect of the area closure or whether they existed before the closure. On the other hand, it may be difficult to distinguish changes in fish populations caused by natural processes in the environment from those induced by the reduction of fishing effort. As effects of area closures may be mitigated by a whole range of factors and processes (Polunin, 2002), detecting such effects is not guaranteed even with a valid survey design. Willis et al. (2003) stressed that little evidence exists to substantiate responses of fish populations to area closures. They noted, however, that this does not imply that MPAs fail in their objectives.

In Iceland, area closures are an important measure to protect fish stocks in addition to allocating quotas on species and mesh size regulations (Jaworski et al., 2006). Most closures (mainly off northwest, north and east Iceland) aim at protecting juvenile fish to increase the long-term yield to the fishery. Some areas are subject to temporary or seasonal closures, while others are closed permanently. In addition, some major spawning areas for the main commercial stocks are closed during spawning time. Despite the large number of closed areas in Iceland, their effect on fish populations has only recently been studied. Jaworski et al. (2006) investigated the impact of two area closures off the northeast and southeast coasts of Iceland on the demersal fish community. In that study, data from groundfish surveys carried out annually over a 20-year period were used to examine differences in density, mean size and fish species diversity between closed and fished areas, and between periods before and after the closure. Both closed areas had an effect on density and mean size of some commercially exploited fish species.

The present study explores, at relatively small spatial scales, differences between demersal fish assemblages in protected and fished sites off the north coast of Iceland. More specifically, its objective is to establish whether there were significant differences in fish density and mean size between protected and adjacent fished areas and whether any differences in these two metrics could be attributed to the protection status of the areas. The suitability and usefulness of such comparative analyses is discussed in the context of development, implementation and evaluation of MPAs.