A number of scientists have in recent years used US longline data to provide information on the abundance and status of oceanic species. The estimated number of hooks set by the longline fleet within the Caribbean area has been decreasing since the late 1990s and was less than 200 000 hooks in 1999. The fleet has been moving away from its fishing grounds in the Caribbean and the Florida east coast and increasing its activities in the Gulf of Mexico (Figure 11). Unfortunately, with the exception of Baum et al. (2001), who report area-specific trends for oceanic sharks, most studies do not report the trends for individual area strata, such as the Caribbean, because the emphasis of their analysis is on developing population indices.
Cramer and Ortiz (2000) developed indices for yellowfin tuna for the period 1982-1999, but the data for the Caribbean were so sparse that the indices were developed based on data for the Gulf of Mexico and the northwest Atlantic only. The Florida East Coast area, the southernmost area included in their analysis, includes waters around the Bahamas, Turks and Caicos Islands and northeast of Cuba.
Ortiz and Scott (2000) developed indices for blue and white marlin for the period 1986-1999. Indices suggest a steady decrease in abundance for both species during that period. These authors did consider the Caribbean data in their standardization. Similarly Ortiz and Brown (2001) developed indices of abundance for sailfish for the period 1986-2000 and provided a map of the cumulative spatial distribution of longline effort (in number of hooks). The index, although uncertain, suggests that sailfish abundance grew from 1985 to 1993, then decreased by half from 1994 to 1995 and has remained low since.
Number of longline hooks in three areas where the US longline fleet operates: Caribbean (CAR), Florida East Coast (FEC) and Gulf of Mexico (GOM)
Cramer and Ortiz (2001) developed abundance indices for three-year-old and older swordfish for the period 1982-2000. The index suggests that abundance dropped in the 1980s and has remained relatively unchanged since.
Prager (2000) used US longline logbook data to develop an index of abundance for dolphinfish for the Gulf of Mexico, the US Atlantic coast and the Caribbean, excluding areas in the Caribbean south of Puerto Rico. He developed this index for the period 1986-1997 but expressed little confidence in its usefulness, because it fluctuates by a factor of 15 in the period considered and by a factor of four in the period 1994-1997. In the same period, estimates of fishing mortality changed from 1.8 in 1985-1986 to 0.3 in the 1990s. Prager doubts that such fluctuations reflect abundance and fishing mortality changes, and acknowledges that they suggest that the indices are highly uncertain, partially due to how infrequently dolphinfish is reported in the catch. His analysis suggests that the stock was at very low indices of abundance in the mid-1980s and increased eightfold during the early 1990s. Prager suggests that such large changes in abundance are plausible, given the life-history characteristics of dolphinfish, but also says the trend could be an artefact. He notes that the indices may be biased because of the problem of properly defining dolphinfish stocks and the incomplete coverage of the dataset.
There is another large fleet of longliners operating in the southern Caribbean area and offshore of the Guyana area, the Venezuelan industrial longline fleet, and abundance indices of oceanic species have recently been computed for it. Most of the operations of this fleet occur in the Caribbean area of Venezuela, although the fleet extends its operations east of Barbados and all the way down to 5°S. Arocha and Ortiz (2000) developed indices of abundance for yellowfin tuna, for the period 1991-1999, that showed no increasing or decreasing trend but considerable variation from year to year. ICCAT (2001b) developed abundance estimates for blue and white marlin for this fleet for the period 1991-1999, and Arocha and Ortiz (2001) developed indices of abundance for sailfish for the period 1991-2000. All three indices suggest abundance being rather unchanged for most of that period.
In recent times, the Japanese and Taiwanese longline vessels have operated much less frequently in the waters of the Caribbean than they did prior to the 1990s. The number of hooks deployed by Japanese vessels in the Caribbean region was 1.5 million in 1999 and 1 million in 2000 (ICCAT, 2001b). During the 1990s the activities of Taiwanese vessels in the Caribbean were confined to the waters of the Lesser Antilles and east of that, whereas in the 1970s and 1980s they operated around the Large Antilles as well (Hsu and Liu Amorim, 2000). In recent standardizations of Japanese catch per unit effort (CPUE) that were made to develop abundance indices for yellowfin tuna (Matsumoto and Miyabe, 2001) and marlins (Yokawa and Takeuchi, 2001), longline data from the western areas of the Atlantic, including the Caribbean, were not used because of poor data coverage. Although Taiwanese scientists do use data from the western Atlantic for the estimation of yellowfin abundance indices (Hsu and Liu Amorim, 2000), the area stratum that includes the Caribbean encompasses the whole of the northwestern Atlantic, west of 40°W and north of 10°N. Therefore neither the Japanese nor the Taiwanese scientists consider longline data for the Caribbean appropriate for the estimation of a relative abundance measure for the region.
In summary, neither of the two longline fisheries for which data within the Caribbean are available - US and Venezuelan - covers a large portion of the fished area for any offshore oceanic stocks, and the US fleet has been reducing its presence. This means that although it may be possible to compare the relative abundance of a stock in an area of the Caribbean with that in an area of the North Atlantic (e.g. the Gulf of Mexico), it is not possible to estimate how that relates to the abundance of the entire stock. In the case of the two other large longline fleets that operate Atlantic-wide - Japanese and Taiwanese - over the last 10-20 years their operations within the Caribbean have been greatly reduced. In recent years, the effort of these two fleets in the Caribbean has probably not been representative because of the restricted areas in which they are operating. This precludes the use of data from these two fisheries in comparing recent abundance indices for the Caribbean.
There are at least two different options for tackling this problem, and both would need the cooperation of scientists from all the major longlining countries. One option would be to construct a spatially explicit, relative estimation model of abundance that integrates all the different longline datasets. Given a number of assumptions about the distributional patterns of abundance of oceanic fish, it may be possible to design area strata that would provide enough information to determine relative abundance for the Caribbean area or smaller subsets of the area.
Another option would be to develop an integrated assessment and spatially explicit model such as the one proposed by the ICCAT Bigeye Tuna Program (Maury and Restrepo, 2001). The most complex conceptual model for the stock being considered by these authors has 12 areas for the entire Atlantic, two of which cover parts of the Caribbean and the Gulf of Mexico.
The success of the above options depends on whether there is enough time and space overlap between fleets to be able to estimate the fleet-specific factors that relate CPUE across fleets. This could be evaluated using the ICCAT database, with which an appropriate space-time stratification could then be developed for analysis of the full trip-by-trip datasets. However, as the proposed bigeye tuna models indicate (Maury and Restrepo, 2001), it is highly unlikely that such a model would ever be able to estimate abundance indices at the level of a single countrys EEZ.
The conclusion of these analyses is that there is no simple way to use data from a single longline fishery to estimate spatially structured, relative indices of abundance for offshore oceanic stocks in the Caribbean in the context of the overall biomass of the stock.