Inferring shark population trends from generalized linear mixed models of pelagic longline catch and effort data

Abstract

We estimate recent (1992–2005) trends in relative abundance for Northwest Atlantic oceanic and large coastal sharks, using generalized linear mixed models to standardize catch rates of eight species groups as recorded by U.S. pelagic longline fishery observers. Models suggest precipitous (76%) declines in hammerhead (Sphyrna species) and large coastal (dusky, night, and silky shark, genus Carcharhinus) species, and moderate declines (53%) in blue and oceanic whitetip sharks over this period. In contrast, mako and thresher sharks appear to have stabilized, and the tiger shark population appears to be increasing. A comparison of nominal shark catch rates from this fleet's observer and logbook data (to evaluate the veracity of trends previously estimated from the latter) showed a high degree of concordance for each species group, both in individual sub-areas and overall. Models of these two datasets for the common time period (1992–2000) show that compared to the observer data the logbook data indicate greater declines for some species, but lesser declines for others. Signs of recovery for some shark species are encouraging, but must also be set in the context of the significant declines that occurred in previous decades.

Introduction

Concern about increased exploitation of sharks, coupled with the inherent vulnerability to overexploitation of many of these species, has brought this group of fishes to the forefront of marine conservation in recent years (FAO, 1998, FAO, 2000, Musick et al., 2000, ICCAT, 2004, CITES, 2006, Anon, 2009). Large pelagic sharks are circumglobally distributed top predators and among the most heavily exploited sharks (Camhi et al., 2008a, Dulvy et al., 2008). Species in this group, which includes wide-ranging oceanic sharks such as blue (Prionace glauca) and mako (Isurus species) and more coastal tiger (Galeocerdo cuvier) and hammerhead (genus Sphyrna) species, comprise the majority of those traded in Asia's shark fin trade (Clarke et al., 2006) and are also increasingly sought after for their meat (Hareide et al., 2007).

Quantifying the impacts of exploitation remains a challenge for most shark populations because of a paucity of data (Camhi et al., 2008a). Few stock assessments have been conducted for sharks, and results for many of those that have been were uncertain (e.g. ICCAT, 2008). Indices of abundance are key components of the complex population dynamics models used in stock assessments (Maunder and Punt, 2004), and also important indicators of the direction and magnitude of changes in abundance for the many shark species for which there are inadequate catch records and biological information to conduct stock assessments.

Estimating unbiased indices of abundance for large pelagic sharks is, however, complicated by several factors (Camhi et al., 2008b). Distributed in epipelagic and upper mesopelagic waters, these species are rarely caught in fishery-independent research surveys. Surveys that have sampled sharks often are limited by low sample size to provide estimates only for the most frequently caught coastal species. Conversely, fisheries sample intensely over large regions closer in size to the geographic ranges of shark populations, but are much more variable than designed research surveys making standardization of the catch rates a challenge (Maunder and Punt, 2004, Bishop, 2006). What is more, there is a dearth of long-term fishery-dependent data for sharks: most commercial fisheries began recording shark catches at the species level only in the 1990s, and reliable species identification remains a challenge. There also is a tradeoff between logbook data, which are self-reported by fishermen, and scientific observer data, which should be more accurate but often monitor only a small proportion of commercial fleets. The situation is exacerbated for oceanic sharks because much of their exploitation occurs on the high seas, where their catches are unrestricted and often un- or under-reported (Camhi et al., 2008b).

In the Northwest Atlantic Ocean, one of the most data-rich regions for sharks, many large pelagic shark species appear to have declined significantly (Musick et al., 1993, Simpfendorfer et al., 2002, Baum et al., 2003, Ha, 2006, Myers et al., 2007, Aires-da-Silva et al., 2008). For example, two dedicated shark-targeted longline surveys conducted annually on the U.S. east coast since 1972 and 1974 respectively, have provided valuable multi-decadal records for many large coastal shark species; analyses of these data indicate substantial declines in dusky, tiger, blacktip and sandbar sharks (Ha, 2006, Myers et al., 2007). Examination of fisheries logbooks from 1986 to 2000 also suggested significant changes in large pelagic shark population abundance in this region, ranging from 40% declines for two mako shark species up to 89% declines for three hammerhead species (Baum et al., 2003). In those analyses, generalized linear models (GLM) were fitted to the non-zero catches with the truncated negative binomial distribution to avoid the potential bias of any change in fishermen's tendency to record shark catches over time (Baum et al., 2003). Six additional analyses using different statistical distributions and subsets of the data (based on the tendency of sharks to be recorded on different vessels) led to some quantitative differences in trends, but similar conclusions of significant declines in abundance (Baum et al., 2003, Supplementary Online Material). That research has, however, been criticized for inferring trends in abundance from a single data source, particularly since the data were from logbooks (Burgess et al., 2005, but see rebuttal in Baum et al., 2005, and analyses of additional data sources in Myers et al., 2007).

To address these concerns and to examine more recent changes, here we build upon this earlier research by using the U.S. Atlantic pelagic longline fishery's observer monitoring program data: (i) to describe the spatial distribution and concentrations of large pelagic sharks in the Northwest Atlantic Ocean, (ii) to estimate trends in their relative abundance using the most recent available observer data (1992–2005), (iii) to compare these data and estimates to those from the same fleet's logbook data, and (iv) to suggest improvements for future observer data collection and models.