Dr. I. Poiner:

5.70. Sea turtles are very long lived animals that mature at a relatively late age (ca. 30 to 50 years). The interval between breeding events is also extended (ca. 5 to 15 years, depending on the species). While many eggs are produced, and egg predation is high, natural mortality of sub-adults and adults is probably relatively low. Because recruitment to the adult population is low, population modelling studies suggest even small increased mortality rates in adults and sub-adults could impact substantially on population numbers and viability (Crouse et. al., 1987; Crowder, et. al., 1994 Heppell et. al., 1995; Chaloupka and Musick 1997).

5.71. Most sea turtle populations in the world are severely depleted. All seven species are included in CITES Appendixes and the World Conservation Union's (IUCN) Red Data Book lists. Most species have been listed as endangered or threatened under various national legislation. For example, all five species found in the United States waters are listed under the Endangered Species Act of 1973 and the five species found in Australian waters are listed under the Commonwealth Endangered Species Protection Act 1992. Recovery from low population number (if non-natural sources of mortality have been removed) will be slow, and there are no clear documented cases of recovery in the world.

Table 1: Sea turtle species that have declined and current anthropogenic threat to sea turtle populations in Thailand (Monanunsap 1997; Limpus 1997; Settle 1995), Malaysia (Chark 1997; Limpus 1997; Liew 1995; Chan et. al., 1998) and United States (Lutcavage et. al., 1997).

5.72. There is data documenting declines in sea turtle populations and the causes of declines have included: habitat alteration and loss of nesting and foraging habitats, pollution and entanglement, and fishing and incidental capture (Table 1). However, apart from estimates of the incidental capture and mortality of sea-turtles in some shrimp trawl fisheries (United States and Australia) (Henwood and Stuntz, 1987; Poiner and Harris 1996; Robins 1992), estimates of mortalities from boat strike, oil pollution and explosive platform removal mortalities in the United States (Lutcavage et. al., 1997), most mortality factors are not well quantified and it is difficult to rank mortality sources either currently or over time. Furthermore, there is a paucity of information about total population size, age structure, age-specific growth and mortality rates of the turtle populations and turtle distributions (patch dynamics) (Chaloupka and Musick, 1996). Without this, and information on the size and age structure of the segment of the population impacted by the anthropogenic activity, it is difficult to rank the relative impact of the different sources of mortality on sea turtle species and populations.

1(b) Is it possible to rank the various sources of mortality according to their impact on sea turtle populations? In particular, is it possible to determine the relative role played by past practice of egg harvesting and direct catch as compared to more recent threats at sea (such as those related to modern fishing practices) on the depletion of sea turtle populations? If these determinations are possible, please explain the basis for them, in particular if the studies cited cover sea turtle populations in the countries parties to the dispute.

Dr. Scott Eckert:

5.73. Our perspective on the impacts of various types of mortality to sea turtle populations has evolved as sea turtle population models have become better refined. This process will likely continue; however, based on work by Frazer, Crouse, Crowder, and Heppell, the current perspective is quite different than that of 20 years ago (see review by Chaloupka and Musick, 1996). What has been determined is that it simply is not adequate to concentrate all efforts on protecting reproducing females and eggs as has been the traditional approach to restoring sea turtle populations. While it is obviously necessary to preserve the reproductive capacity of any sea turtle population, no population can be preserved by such methods alone. What both Frazer (1983) and Crouse et. al., (1987) pointed out in their population models is that it is vital to protect large juvenile and sub-adult turtles (so called "stage 3" turtles). Based on the reproductive value curves of Frazer (1983) for loggerheads, these larger turtles represent the highest reproductive value to the population, because significant reproductive investment has gone into their survival. There has not been any data presented to date to suggest that these value curves are not applicable to all species of sea turtle.

5.74. What is particularly critical to understand is that for many species (in particular those that have a neritic existence), stage 3 turtles are often the most subject to trawl fisheries (Crouse et. al., 1987). This is likely because this size class seems to pick foraging habitats that are most strongly correlated with shrimp fisheries. A number of possibilities have been proposed for this overlap. One is that this size class is more subject to chumming, i.e. the large quantity of bycatch discarded by the fishing boats attracts the turtle to scavenge. The other possibility is that this habitat is simply the developmental habitat for this size class of turtles. It is likely that these smaller size classes cannot dive as deeply nor as long as larger mature animals, and their ability to handle large prey is reduced. Thus, they forage in shallower waters with soft bottoms that characterize shrimp habitat.

5.75. It is my belief that nothing is as destructive to any turtle population as incidental mortality caused by fishing operations. Beside the issue of how some fisheries focus mortality on critically important size classes, fishery impacts can cause population declines far more rapidly than mortality associated with beaches. Good examples of this are the loggerhead populations in North Carolina, South Carolina and Georgia. All have declined approximately 80 per cent in 26 years, due primarily to shrimp fishing (NRC, 1990). The crises of the Pacific leatherback, has undoubtably been due to high mortality in the high seas driftnet fishery and the South American swordfish driftnet and longline fisheries. In this latter case we have seen the world's largest leatherback nesting population (estimated by Pritchard, 1982, to support 75,000 females in 1980) decline more than 95 per cent to less than 1,000 females by 1997 (Sarti et. al., 1996). The rate of decline caused by these impacts are often too fast for us to respond until it is too late. This latter situation is near to my own experience, as I have been working with colleagues from Mexico on one of the primary nesting beaches for the leatherback in Mexico since 1986. We should have seen this terrible decline, but I have described how long a nesting beach must be monitored before a trend will become apparent. We were not concerned that this decline was real until about 5 years ago, and it took 3 more years to confirm our suspicions. Finally it took until this year to determine that the problem lay with gillnetting in South America, and it may be too late to reverse this trend. The rate of decline caused by incidental fisheries mortalities is simply too rapid to respond with mitigation actions.

5.76. With declines associated with egg mortality such problems take substantially longer and they tend to be far more gradual, as was the case in Terengannu, Malaysia. Thus, our ability to detect these declines is enhanced, and while such perturbations to the population may require many years to turn around, there are many techniques available to mitigate (beach protection programmes, in situ beach hatcheries, enforcement of egg harvest regulation etc.). In the case of Terengannu, the problem was that when the population decline was identified, too little was understood about population dynamics of turtles to realize that preserving approximately 10 per cent of the harvested eggs was not enough.

Dr. J. Frazier:

5.77. Turtles that reproduce contribute to the maintenance of the population; hence, those animals are critical to sustaining the population: without reproduction, there will be extinction, sooner or later. Individuals that are not yet reproducing are not yet contributing to the maintenance of the population. Therefore, turtles that have survived the many and diverse risks over a period of decades, growing to reproductive maturity, are essential to the maintenance of the population, because they are capable of reproducing for many more years (decades, apparently). These adult animals are the immediate key to the future of the population. Animals that are not yet mature, still have to survive more years before they can contribute to the maintenance of the population. The younger a sea turtle is, the more time will have to pass before it will begin breeding; during this time, it will be liable to different sources of mortality, and in the end it may not survive to reproduce.

5.78. Therefore, removing a reproductive female from a population will have an immediate impact on the population, by reducing the amount of reproduction in the population. As there is a high probability that a breeding female will nest for many seasons, removing her will eliminate her reproductive contribution, not only in the short term but also in the long term. In comparison, removing a recently hatched turtle ("hatchling") from a population will have no immediate impact on reproduction. This animal would need to evade different sources of mortality for decades, before it began to reproduce. If we assume - for the sake of argument - that the chances of a hatchling surviving to maturity are 1 in 1,000, then on average the removal of 1,000 hatchlings would have an impact comparable to removing one animal that has just reached maturity, but the reproductive contribution of the hatchling that survived to maturity would not be manifested for more than a decade after it had hatched. Clearly, a healthy population needs individuals in all stages of development and maturity; there must be constant recruitment of young animals into the population to gradually replace the older animals as they die or cease reproducing. Thus, even although removing a hatchling may have no immediate effect on the reproduction of a population, continual removal of hatchlings will produce a situation in which the population would "die of old age", that is, there would be no new animals to replace the old ones.

5.79. The purpose of the above simplified example is to clarify the immediate impacts of different sources of mortality. A very clear explanation of the issues of mortality, survivorship and life stages is given in National Research Council (1990: Chapter 5). In more precise, scientific terms population models provide quantitative ways to evaluate how different sources of recruitment or mortality are likely to impact a population. The studies of Crouse et. al. (1987), Crowder et. al. (1994; 1995), and Heppell (1996a; 1996b) have used population models to predict the relative effects of increasing recruitment or increasing mortality on different phases of the life cycle, and how these relate to conservation priorities. Because basic information is lacking, and only two populations of loggerheads have sufficient data to even begin constructing population models, the work has been limited to loggerheads. While some details of life history parameters will be different with other species and other populations, these models are the best predictive tools that we have at the moment. Furthermore, it is not likely that general conclusions will change, given the general similarity in life history parameters between the different species of marine turtles.

5.80. A concept that is used to integrate the above ideas is that of "reproductive value". According to the work of Crouse et. al. (1987), if the reproductive value of an egg is 1, then the reproductive value of a subadult would be 116 and that of a breeding animal would be 584. These numbers are indices of the relative "value" of an individual, at different stages during its life cycle, for reproduction and maintenance of the population.

5.81. Given the above paradigm, a population would be resistant to mortality concentrated on early life stages (e.g., egg harvesting), but mortality of animals that are reproducing, or just about to reproduce, would have an immediate effect on the level of reproduction in the population, and these deficits would continue for many years. The closer to maturity the turtles get, the more they are worth to the population, and the less it can afford to lose them. On the other hand, continual egg harvesting will in the end result in the collapse of a population, but a few years of total loss of eggs, or moderate levels of egg loss over a period of years would have less impact on the status of a population than would several years of removing breeders - this would be true both in the short term and in the long term.

5.82. Hence, sources of mortality that affect animals that are mature, or nearly mature, have far greater instantaneous impact on the status of the population than taking the same number of eggs or young animals, for they reduce levels of reproduction very quickly. Harvesting of breeding animals, or incidental capture in fishing gear, are examples of these very "costly" sources of mortality. Modern fishing practices have been repeatedly documented to cause mortality; and more specifically, large-sized, especially adult, turtles are known to be caught and drown in shrimp trawls in Australia (Poiner and Harris, 1994; Robins, 1995); Pacific Costa Rica (Arauz, 1990; 1996b); and the United States (National Research Council, 1990).

5.83. Systematic data on present population size, trends in population size, rates and sources of mortality, population structure, and indeed geographic distribution are incomplete for many areas. Hence, many decisions related to conservation and management of sea turtles are made with imperfect knowledge, "splicing" together the best information available, from wherever it can be obtained. While this clearly has drawbacks and limitations, the logic is to adopt a conservative approach so that mistakes in decision making will be "conservative", i.e., least likely to be detrimental to the population. In terms of the present discussion, this means paying special attention to those animals that are of greatest value to the maintenance of the population: in other words, making sure that the animals that are breeding, or close to breeding, are protected. Clearly, all stages of a population are important for its long-term continuity, but where risks are focused on individuals with the greatest reproductive value, an immediate priority is to reduce these risks. This is particularly true of populations that are under risk, because they have been decimated.

5.84. A study was carried out using a population model to evaluate the effect of mortality on different stages of the life cycle of loggerheads from eastern Australia. They found that even if hatchling emergence success could be elevated to 90 per cent (well above the natural average), with the present rate of mortality of adult and subadult females, the population may be headed for extinction in what are the equivalent of 3 turtle generations (Heppell et. al., 1996b). Since a major source of mortality for these turtles is incidental capture in prawn trawls (Poiner and Harris, 1994; Robins, 1995), a priority for the survival of these populations is significantly reducing this risk; Heppell et. al. concluded that the use of TEDs, together with other conservation measures, would be instrumental in the survival of these populations. Similar analyses, giving comparable results, have been carried out for the east coast of the United States, where once again it was concluded that eliminating, or significantly reducing, mortality of breeders and large juveniles in shrimp trawls was critical to the long term survival of these populations (Crouse et. al., 1987; Crowder et. al., 1994; 1995).

5.85. Along the Pacific coast of Central America it is estimated that some 60,000 turtles are caught annually in shrimp trawls; and in Costa Rica, which contributes a third of the total, there may be 24 to 60 per cent mortality (Arauz, 1996a). Systematic data are not available for other populations (neither for the countries involved in this dispute, nor for any others). Thus, these sorts of determinations can only be carried out for other populations by analogies based on existing knowledge of the basic similarities in life history parameters. Until systematic information is available to refute these assumptions, it is generally considered that the most conservative approach is to use the results of the population models as indicators for conservation priorities.

Mr. M. Guinea:

5.86. It is difficult to rank the various threats to sea turtles on a global scale. Hence the paradigm of breeding units becomes an essential tool to assess, for that unit, the relative impacts of human pressures. The major threats to each breeding unit must be independently assessed and managed. Sea turtle populations have declined in some countries which have had a long period of intensive egg harvest. As have the populations in countries that have focused on the exploitation of tortoise shell or meat. This has happened regardless of modern fishing practices such as trawling. Other countries with intensive trawling activities have also experienced a decline in their sea turtle numbers. It is the breeding unit of each species that has to be examined. If nesting beaches have been destroyed by commercial or industrial development, then abatement measures should be directed to halt, modify or remove that development. If introduced predators have reduced the productivity of rookeries, then their reduction becomes the target of abatement measures. If adults are being killed while nesting and by set nets off shore, then legislation to protect rookeries and their off shore refuges needs to be enacted or enforced. Should trawling be responsible for the deaths of sea turtles of any age, then management regulations involving reduction of fishing effort, by exclusion areas, closed seasons, vessel and gear size restrictions, limits to tow duration, adoption and enforcement of bycatch reduction devices such as TEDs, should be adopted.

5.87. As indicated, the sources of mortality should be examined in relation to the breeding units. Malaysia and Thailand because of their proximity may share breeding units of some species. Malaysia may share breeding units with the Philippines and Indonesia. India and Pakistan could share breeding units of some species. The United States and Mexico may share breeding units as well. It is speculative to suggest that southeastern United States shares a breeding unit with any of the other countries in the dispute.

Mr. H.-C. Liew:

5.88. The truth and accuracy of whatever ranks produced are only as good as the information that is available. There will always be shortcomings of such reports and their reliability will vary from country to country depending on how accurate and extensive the information is made available. Scientific studies are still being conducted to improve on the information but gaps still exist especially on mortalities at sea of hatchlings, juveniles and adults. For example, we still do not know what degree of mortalities is affecting our hatchlings at sea. How many are killed by natural predation or by consuming floating debris like tar balls and styrofoam beads, etc. Attempts have been made to estimate these unknowns by modelling, but these estimates are hinged on assumptions made of what is unknown. Turtle landings, egg production, or turtle catch statistics produced by governments or NGOs are sometimes extrapolated figures, misreported, biased or even falsified. With the absence of any other data, they are often assumed to represent the true situation. The degree of error does vary from country to country depending on expertise available to collect the data and various other limitations. Bearing these limitations in mind, attempts can still be made to rank them when such requests are made.

5.89. In developed countries like the United States where the human population is generally rich, educated, with cheap protein available, they could afford to have strict conservation management policies and effective enforcement. Hence turtle mortalities caused by egg harvesting or killing for meat is negligible (ranked low). With mortalities by these causes removed, mortalities caused by their high technologies like shrimp trawling becomes more prominent and overshadows the other causes (ranked high). In developing countries like India, Pakistan, Malaysia, Thailand, Indonesia, etc., conservation management policies and their enforcement are at a different level. Egg harvesting for consumption is still legal or poorly enforced in many of these countries. Turtles are still being slaughtered for meat in some of them. Fishing technologies like shrimp trawling may not be as developed or still artisanal to be of prime impact on turtle populations. They may even be using other fishing methods which may have a greater impact on sea turtles than shrimp trawling, like the sunken set-net or "pukat pari" in Malaysia.

Dr. I. Poiner:

5.90. It is not possible to rank the various sources of mortality according to their impact on sea turtle populations especially with respect to relative role of past practices.

1(c) A survey of current anthropogenic threats to sea turtle populations in the five countries involved in this dispute would be appreciated. In particular, are anthropogenic threats currently more important at sea or on the nesting grounds? What is the relative impact on sea turtle populations of egg harvests and direct harvesting of sea turtles vs. incidental capture of sea turtles in fishing operations, in particular shrimp trawling? Is this situation similar in different parts of the world? Are different species of sea turtles affected differently?

Dr. Scott Eckert:

5.91. Identifying primary sources of mortality within these countries is quite challenging, primarily because most of these countries (with the exception of the United States) are simply not putting much effort into studying the problem. This seems particularly true for measures of fishery bycatch. The purpose of most government fisheries agencies is to support fisheries with research and technology advances. Generally, measuring bycatch is not a high priority. Hence, bycatch data is rarely gathered. Furthermore with fisheries such as shrimp fishing, the boats are relatively small and numerous, making the application of an observer programme difficult and very expensive. However, without an independent observer programme any data (such as logbook data or even port sampled data) must be suspect. In my experience captains uniformly under-report bycatch data, sometimes unintentionally but often out of concern for what reporting turtle mortalities will mean to their livelihood. It is not surprising to me, therefore, to find a great paucity of rigorous study on incidental take in shrimp fishing in Malaysia, Thailand, Pakistan or India. In the United States there is a large body of information (see NRC, 1990, Crouse et. al., 1992, Murphy and Murphy, 1989), driven primarily by the requirements of the Endangered Species Act. All that is generally reported in these other 4 countries are anecdotal or very limited reports, most gathered during interviews with fishermen.

5.92. In the United States threats to sea turtles are somewhat species and regionally dependent; however, a few generalizations can be made. For green, loggerhead and Kemp's ridley turtles in the Atlantic, the most serious threat was shrimp trawling (NMFS and USFWS 1992, 1991a, 1991b, NRC, 1990). This threat has been well documented and is probably indisputable (Maley et. al., 1994, NRC, 1990). The requirement that TEDs be utilized in all waters at all times has reduced this threat. At this time the most significant threat has to be enforcement of existing regulations, and the scope of this problem is minor when compared to the previous unlimited incidental take. Also, of significant importance to green turtles and hawksbills occurring within the US insular Pacific Ocean (except in Hawaii) and Caribbean is the direct killing of turtles. (NMFS and USFWS 1996a, 1996e). However, the scale of this problem is substantially less than in other countries of the region. The primary threats to olive ridleys in the United States is the due to incidental take of turtles in the Hawaiian-based longline fishery (NMFS and USFWS, 1996f). Threats to the leatherback in US waters include the shrimp fishery on the Atlantic seaboard, the Hawaii longline fishery and gillnet fishing in Northern California.

5.93. In Malaysia, it is apparent that egg harvest is still a serious problem for green, hawksbill and possibly leatherback turtles, despite regulations designed to limit such harvest (Eckert, 1993, and paragraphs 5.34 and 5.75). During research I conducted in 1989 in Terengannu, Malaysia, leatherback and green turtle eggs were openly sold in the local markets despite their protected status. Trawling has also been described as a threat to turtles residing off Terengannu, though this report may be out-of-date to the current situation (Chan, et. al., 1988). However, in 1991, Chan reported that incidental capture in fishing gear "is now recognized at one of the most serious threats to the survival of the remaining sea turtles in Malaysia." (Chan, 1991). I have no other information on the situation in Malaysia, but based on my general experience with trawl fisheries and sea turtles, I would not be surprised that any area in southeast Asia which support trawl fishing also has incidental catch of sea turtles.

5.94. In Thailand there appear to be a number of threats to sea turtle populations, the most serious appear to be shrimp trawling, killing of turtles and taking of eggs on nesting beaches (Eckert, 1993, Hill, 1991, Hill, 1992, Chantrapornsyl 1997). There also seems be problems with enforcement of trawling regulations (Hill, 1991, Hill, 1992). Both green and hawksbill populations in Thailand are severely depleted (Eckert, 1993, Chantrapornsyl, 1997).

5.95. As described earlier, there are a host of anthropogenic threats to sea turtles in India, including the killing of nesting females, harvest of eggs and incidental mortality associated with shrimp fishing. However, the large numbers of olive ridley killed by legal and illegal trawling operations is extraordinary and must represent the single largest threat to sea turtle populations in India.

To Continue With Chapter 5.96