INTRODUCTION
The census population size (N) is generally the only data used for assessing the status of most threatened species (Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). However, for evolutionary matters, the effective population size, not the census number, is an essential concern. The effective population size (N e) is expected as the size of an ideal population that has the same rate of increase in homozygosity or gene frequency change as the actual population under consideration (Wright, Reference Wright1931).
An important use of N e in conservation biology is the calculation of the minimum viable effective population size. An assessment of viability is necessary, in part, to determine whether the population is large enough to prevent inbreeding or if it has sufficient adaptive genetic variation (Vucetich & Waite, Reference Vucetich and Waite1998). A population with a high N e maintains a high level of genetic diversity, and therefore reduces the probability of inbreeding depression. However, a population with a very low N e is more susceptible to genetic drift and less able to respond to selection. This is because in small populations there is less genetic variation for natural selection to act upon, and there is a higher probability that beneficial alleles will not be maintained by selection and will instead be lost from the population because of random drift effects (Willi et al., Reference Willi, Van Buskirk, Schmid and Fischer2007).
Additionally, the estimate of N e reflects the number of individuals responsible for the maintenance of the genetic diversity of the species as well its evolutionary potential. Considering that the aim of conservation is maintaining species as dynamic entities capable of evolving to cope with environmental changes, it is important that the species' evolutionary potential is retained in order to respond to the current and unpredictable climate change scenario (Frankham et al., Reference Frankham, Balou and Briscoe2002).
The South American sea lion, Otaria flavescens, is the most abundant pinniped species along the Peruvian coast, with 118,220 individuals in 2006 found in breeding colonies and haul-out areas along the coast and nearshore islands (IMARPE, 2006). However, there is no information about its effective population size (N e). The only Ne estimate available is for the South American fur seal, Arctocephalus australis, for the Peruvian coast (N e = 2153: Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). The only Ne estimated for a Peruvian pinniped is that for the South American fur seal, Arctocephalus australis, for the Peruvian coast (Ne + 2153: Oliveira et al. 2006). This estimated N e is of critical importance for the Peruvian pinnipeds because their natural populations suffered frequent declines as a result of low food availability due to the replacement of cold and nutrient rich waters of the upwelling system with warm, poor nutrient and low productivity waters during El Niño events (Majluf & Trillmich, Reference Majluf and Trillmich1981). Indeed, the national census conducted by Instituto del Mar del Perú (IMARPE) along the Peruvian coast in 1996–1997 indicated that the O. flavescens population declined from 144, 087 (Arias-Schreiber & Rivas, Reference Arias-Schreiber and Rivas1998) to 27,991 individuals in December 1998 (IMARPE, 2006). Due to this drastic population decline of 80.57%, the Peruvian population of sea lions was considered vulnerable and in high danger of extinction along the Peruvian coast (Decreto Supremo No. 034-2004-AG).
It is important to mention that besides the El Niño effects, the Peruvian sea lions and fur seals also suffered local extinctions in the region due to indiscriminate commercial sealing for fur, meat and oil until 1946, when the hunting of both species (O. flavescens and A. australis) was prohibited (Piazza, Reference Piazza1969; Tovar & Fuentes, Reference Tovar and Fuentes1984). The sealing was totally banned only in 1959 (Grimwood, Reference Grimwood1969). Nevertheless at present, despite all the legal protection, seal poaching still remains, in order to supply the Asian Market with aphrodisiacs (Lama, Reference Lama2010).
It is important to mention that another conservation problem faced by Peruvian sea lions is the mortality due to interactions with fishery activity (Majluf et al., Reference Majluf, Babcock, Riveros, Schreiber and Alderete2002; Arias-Schreiber, unpublished results), which has been documented along the distribution of the species (Aguayo & Maturana, Reference Aguayo and Maturana1973; George-Nascimiento et al., Reference George-Nascimiento, Bustamante and Oyarzún1985; Koen Alonso et al., Reference Koen Alonso, Crespo, Pedraza, García and Coscarella2000; Szteren & Páez, Reference Szteren and Páez2002; Dans et al., Reference Dans, Alonso, Crespo, Pedraza, García, Gales, Nicholas and Hindell2003; Sepúlveda & Oliva, Reference Sepúlveda and Oliva2005). The South American sea lion is an opportunistic predator foraging on the most abundant prey, mainly on benthic and pelagic fish that are usually economically important (Vaz-Ferreira, Reference Vaz-Ferreira, Ridgeway and Harrison1981; Szteren & Páez, Reference Szteren and Páez2002; Sepúlveda & Oliva, Reference Sepúlveda and Oliva2005; Oliveira et al., Reference Oliveira, Ott, Malabarba, Reis, Peracci and Santos2008) mainly on benthic and pelagic fish that are usually economically important. As a result many sea lions were incidentally captured or even intentionally killed by fishery and fish farming operations throughout their range (see Crespo et al., Reference Crespo, Oliva, Dans and Sepúlveda2009 for a review).
The interactions with the fishery activity could be enhanced during El Niño events and increase the mortality of sea lions. For this reason the estimated N e combined with the current El Niño events as well as fishery interactions are matters of great concern for the survival of the species and should be taken into account in any future management plans to assure the conservation and protection of the species on the Peruvian coast.
The N e estimate is obtained by genetic (reviewed by Neigel, Reference Neigel, Smith and Wayne1996; Nunney, Reference Nunney2002) and demographic methods (reviewed by Husband & Barrett, Reference Husband and Barrett1992; Caballero, Reference Caballero1994; Nunney, Reference Nunney1995; Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006; Traill et al., Reference Traill, Bradshaw and Brook2007, Reference Traill, Brook, Frankham and Bradshaw2010). The most important factor influencing N e—the temporal oscillation in population size based on long-term census—is perhaps the most difficult to obtain (Vucetich et al., Reference Vucetich, Waite and Nunney1997).
Nunney & Elam (Reference Nunney and Elam1994) pointed out that estimates using data collected during a single season ignore the influence of temporal fluctuation and thus may represent gross overestimates. Vucetich & Waite (Reference Vucetich and Waite1998) highlighted how essential are long-term counts in order to improve the accuracy of the estimates of N e. Traill et al. (Reference Traill, Brook, Frankham and Bradshaw2010) also reinforce that conservationists working within developing nations rarely have the resources available to collect the long-term demographic and other data necessary to model viability for specific species or taxa. Fortunately due to the systematic efforts of IMARPE for more than 25 years, results from a long-term census size are available for fur seals and sea lions along the Peruvian coast. This paper presents the first estimate of N e of the Peruvian population of the South American sea lion, O. flavescens, based on the effects of the species' polygyny and oscillations in population size in different generations, which includes the fluctuations caused by the most severe El Niño of the 20th Century (1997–1998) (McPhaden, Reference McPhaden1999). We also discuss the importance of this value for the conservation of a population considered as vulnerable and which faces environmental changes like El Niño events.
MATERIALS AND METHODS
We used data from eight censuses conducted by IMARPE, between 1984 and 2006, which covered 71 breeding colonies and haulout areas from Los Órganos (04°10′S 81°07′W) to Morro Sama (18°00′S 70°53′W) and included the most important reproductive colonies of the species on the Peruvian coast (Isla Brava (11°22′S 77°45′W), Islas Chinchas (30°13′S 76°24′W) and Morro Quemado (14°20′S 76°07′W), Figure 1).
Fig. 1. The most important breeding sites and haul-out areas of South American sea lions, Otaria flavescens along the Peruvian coast, censused between 1984 and 2006 by IMARPE (see text).
In order to estimate N e we only used censuses that clearly identified adult breeding males and females (Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). Breeding males, also called ‘territorial males’, are the males that copulate on the beach and assemble harems, while other males, that are not in reproductive age as well as do not mate with females, are counted as ‘subadult’ males in the census (IMARPE, 2006).
The South American sea lion presents a polygynic breeding system with very few males being able to mate with many females (Capozzo & Perrin, Reference Capozzo, Perrin, Perrin, Würsig and Thewissen2008). Due to these differences in the number of breeding males and females, the effective population size is expected to be smaller than the census population size (Crow & Kimura, Reference Crow and Kimura1970).
We calculated the N e according to Oliveira et al. (Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006), using equations that took into account the effects of: (a) the mating system (and resulting skew in sex ratio variation); and (b) fluctuation in population size over generations (Hedrick, Reference Hedrick2000), from 1984 to 2006 (including the oscillations in before and after the most severe El Niño event of 1997–1998).
(a) The N e equation that accounts for the effects of unequal sex-ratio is:
$$N_e=\displaystyle{{4N_{ef} N_{em} } \over {N_{ef}+N_{em} }}$$where N ef is the number of breeding females and N em is the number of breeding males.
We estimated effective population sizes before and after the El Niño event in order to calculate the effect of changes in population size over time. To calculate the effective population size prior to the El Niño event we used data collected during the year 1993 by IMARPE (2006), with a census size of 76,349 individuals, among which 6435 were reproductive males and 45,080 were reproductive females. To calculate the effective population size after the El Niño event we used data collected for the year 2000 by IMARPE (2006), with a census size of 48,088 individuals, among which 558 reproductive males and 12,323 were reproductive females. The overall effective population size through different generations was calculated using this estimate of N e for population before and after the El Niño of 1997–1998.
(b) The N e for a population that varies in size over generations is given by the harmonic mean of the N e in each generation (Hedrick, Reference Hedrick2000):
$$N_e=\displaystyle{t \over {\sum {{1 / {N_i }}} }}$$where N i is the effective population size in the ith generation and t is the number of generations considered.
RESULTS AND DISCUSSION
The estimate N e that accounts for the effects of unequal sex-ratio in 1993 was 22,525 and 2,135 for the year 2000, which means that the N e prior to the 1997–1998 El Niño was 10 times bigger than the N e for the year 2000. However, the general effective population size given by the harmonic mean of the N e in each generation was 7715 (SD = 6403), that comprises the influence of mating system and variation in population size. This very high value of standard deviation is probably due to the oscillation in population size caused by the 1997–1998 El Niño event (see Table 1), when the population declined by 80.57%.
Table 1. Census size of Otaria flavescens, between 1984 and 2006 (before and after the strongest El Niño event from 1996–1998). N ef, number of breeding females; N em, number of breeding males; N, census size; N e, effective population size per year.
Very little information is available prior to the period used in this analysis (IMARPE, 2006). In the 1950s Piazza (Reference Piazza1969) presented the earliest records about haul-out and breeding areas for the Peruvian coast. The first national censuses were conducted between 1968 and 1979 and, estimated the population size in roughly 20,000 specimens, with no suggestion about population trends due to insufficient data (Majluf & Trillmich, Reference Majluf and Trillmich1981).
For pinniped species, it is important to mention that 7715 could be an underestimated N e, particularly due to the variation in reproductive success among individuals and to the possibility that peripheral males are reproductively active as well (Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). Unfortunately it is very difficult to obtain information on variation in male reproductive success, including from the peripheral males. Nevertheless, we can discuss the probable impacts of these factors upon our results. If individual variance in reproductive success exists, its effect will reduce the estimate of N e to a more conservative value (Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). Even so, Campagna et al. (Reference Campagna, Le Boeuf and Cappozzo1988) observed that 39% of the peripheral (non-territorial) males of South American sea lions were able to mate with harem females in the Argentinean colonies. Unfortunately, there is no information about non-territorial males in the census data set for the Peruvian coast (IMARPE, 2006), which makes it impossible to calculate the contribution of the peripheral males to the N e. These males are specimens that remain outside of the breeding colony in an area close to the sea, with a difficult access to the breeding sites. We believe that the impact of this fact on our estimates could be very low, presuming that they have very few chances to breed and leave offspring.
The most important factor reducing the N e in general is fluctuation in the population size, followed by variation in family size, with variation in sex-ratio having a smaller effect (Frankham, Reference Frankham1995). The N e estimate presented here takes into account a combination of population fluctuations caused by current El Niño events affecting the Peruvian fur seals and also an unequal sex-ratio, characteristic of many species of pinnipeds, including O. flavescens (Riedman, Reference Riedman1990). The oscillation of the population size due to El Niño effects is clearly observed on the partial values of N e for the years before and after the 1997–1998 El Niño (Table 1). The estimate of N e prior this event was 10 times bigger than the N e for the year 2000. We believe that this population suffered a demographic bottleneck due to the decline of roughly 80.57% of the original numbers of sea lions as a result of El Niño effects.
According to Glantz (Reference Glantz1996) the El Niño is a climatological event characterized by anomalous conditions in the atmosphere and ocean, mainly related to the warming of the sea-surface temperature (SST) from 2oC to 9oC along the coast of Ecuador and Peru. It reappears at intervals ranging between two and seven years (Cane, Reference Cane1983). The Humboldt Current upwelling system, which is the richest in the world (Idyll, Reference Idyll1973; Cushing, Reference Cushing1982), is affected by El Niño, with increased SST and decreased primary productivity directly influencing the depth distribution and abundance of the Peruvian anchovy, Engraulis ringens, the most important prey of O. flavescens and A. australis in Peru (Idyll, Reference Idyll1973; Majluf, Reference Majluf1992; Arias-Schreiber, Reference Arias-Schreiber2003).
Arias-Schreiber & Rivas (Reference Arias-Schreiber and Rivas1998) and Arias-Schreiber (2000) described that during the severe El Niño in 1997–1998, the Peruvian population of both, sea lions and fur seals declined as a result of low food availability due to the replacement of cold and nutrient rich waters of the upwelling system with warm, poor nutrient and low productivity waters (Majluf & Trillmich, Reference Majluf and Trillmich1981). This population decline caused a genetic bottleneck in the Peruvian fur seals (Oliveira et al., Reference Oliveira, Meyer, Hoffman, Majluf and Morgante2009), with loss of genetic diversity in seven microsatellite loci and presumably some loss of evolutionary potential for the species. Nevertheless, there is no published information about the genetic consequences for the demographic bottleneck suffered by the Peruvian sea lions in 1997.
A direct consequence of short periods of small population sizes (bottlenecks) or continued small population sizes is typically the loss of genetic diversity (Hedrick, Reference Hedrick2000). In this context, we can only speculate that Peruvian sea lions could have suffered some loss of genetic diversity based on the remaining N e of 2135 individuals from the year 2000 (see Table 1). After the 1997 drastic population decline, all remaining individuals are descendants of the bottleneck survivors.
As a measure of comparison we estimate the N e for the Falkland/Malvinas population of O. flavescens based on the census information from 1990 and 1995 published by Thompson et al. (Reference Thompson, Strange, Riddy and Duck2005). The Falkland/Malvinas population corresponds to the N e of 826 individuals that belong to a slow recovering population. There is also the estimate of N e of 2153 individuals for the Peruvian population of the South American fur seal, Arctocephalus australis (Oliveira et al., Reference Oliveira, Arias-Schreiber, Meyer and Morgante2006). In all cases the calculated N e for the sea lions seems extremely large when compared with the Falklands/Malvinas population as well with the fur seals probably due to a larger population size of the Peruvian sea lions as well differences in population dynamics or natural histories. An explanation for this fact could be the more available and adequate niches to sea lions requirements, like abundant sandy beaches along the Peruvian coast as well as flexible and opportunistic feeding behaviour, preying the most abundant species, which is virtually advantageous during El Niño periods. It has been reported that O. flavescens rather prefers sandy beaches than rocky areas (Vaz-Ferreira, Reference Vaz-Ferreira1965; Tuñez et al., 2008) which are more used by fur seals in this region (Vaz-Ferreira, Reference Vaz-Ferreira1982; Stevens & Boness, Reference Stevens and Boness2003). Moreover, O. flavescens is considered an opportunistic top predator, usually eating the most abundant fish and squid along its entire distribution (George-Nascimento et al., 1985; Crespo et al., Reference Crespo, Nepomnaschy, Alonso, García and Oporto1990; Szteren & Páez, Reference Szteren and Páez2002; Soto et al., Reference Soto, Trites and Arias-Schreiber2006; Oliveira et al., Reference Oliveira, Ott, Malabarba, Reis, Peracci and Santos2008). This behaviour could improve the species survival during El Niño events and also its maintenance in an unpredictable ecosystem such as the Peruvian coast (Soto et al., Reference Soto, Trites and Arias-Schreiber2006).
The estimated N e of 7715 for Peruvian sea lions is even higher than the median estimates of minimum viable population (MVP) size of 5816 breeding adults, value estimated for 102 species of vertebrates. This estimate is suggested for conservation programmes in order to ensure long-term persistence for wild populations of vertebrates (Reed et al., Reference Reed, Grady, Brook, Ballou and Frankham2003). Traill et al. (Reference Traill, Brook, Frankham and Bradshaw2010) critically reviewed the theoretical estimates of minimum population size made over the past few decades and concluded that for both evolutionary and demographic constraints, the populations require sizes to be at least 5000 adult individuals. This seems to be a large requirement, but a number of studies across taxonomic groups have made similar findings: 4169 individuals from a meta-analysis of 212 species (Traill et al., Reference Traill, Bradshaw and Brook2007) and the median MVP from population viability analyses of vertebrates was 5816 individuals (Reed et al., Reference Reed, Grady, Brook, Ballou and Frankham2003). The census-based MVP of 5500 reported by Thomas (Reference Thomas1990) is also remarkably congruent; all these findings are similar to the recommended census N of 5000 individuals (Frankham, Reference Frankham1995).
We believe that the estimated N e for the Peruvian population is not a critical value from a conservation standpoint, because it is higher than the mean MVP for vertebrates (roughly 5000 breeding age adults: Reed et al., Reference Reed, Grady, Brook, Ballou and Frankham2003). Nevertheless, it is important to mention that this number does not take into account the special current and future selective pressures that Peruvian sea lions are under, such as strong El Niño, competition with fisheries and future possible culling, which means this number may not be so reliable after all. Moreover the viability of O. flavescens on the Peruvian coast may depend primarily on local availability of food resources and its effects on pup growth and survival (Soto & Trites, Reference Soto and Trites2004). The sea lions face the productive but unpredictable Peruvian upwelling ecosystem (Ryther, Reference Ryther1969), and are directly exposed to inter-annual and highly stochastic fluctuations in the distribution and abundance of their principal prey, E. ringens (Arias-Schreiber, Reference Arias-Schreiber2003).
Based on Soto et al. (Reference Soto, Trites and Arias-Schreiber2006) the maternal attendance patterns would be strictly related to abundance of prey and oceanographic features close to the rookeries. Major prey shortage during El Niño resulted in females increasing the amount of days their foraging trips and decreasing the time they spent onshore with their pups, which died due to starvation (Soto & Trites, Reference Soto and Trites2004). Thus, stochastic fluctuations in the marine environment should directly affect the maternal behaviour and possibly also the reproductive success of this species. Climatic change models predict stronger and more frequent El Niños along the Peruvian coast (NCDC–NOOA, 2004). In this sense, the estimated N e of 7715 should be considered as a ‘target value’ to be maintained in order to keep the population large enough to avoid inbreeding or to retain adaptive genetic variation to survive future and more frequent El Niño events.
The essential flaw of the estimate of N e is that it requires a data set from long-term censuses that include temporal fluctuations in population size (~20 years) (Vucetich et al., Reference Vucetich, Waite and Nunney1997). Nunney & Elam (Reference Nunney and Elam1994) stated that estimates based on data collected during a single season ignores the influence of temporal fluctuation and thus may represent gross overestimates of N e. This is why the data presented here are so important and unique in comparison to the information available on pinniped populations from other countries throughout South America. The long-term census size available for fur seals and sea lions along the Peruvian coast are results of the systematic efforts of IMARPE for more than 25 years. Vucetich & Waite (Reference Vucetich and Waite1998) emphasized the importance of long-term counts in order to improve the accuracy of the estimates of N e. Traill et al. (Reference Traill, Brook, Frankham and Bradshaw2010) also strengthened the view that conservationists working within developing nations rarely have the resources available to collect the long-term demographic and other data necessary to model viability for specific species or taxa.
This N e result has an important application related to management decisions for the conservation of sea lions in Peruvian waters. The N e of 7715 must be taken into account mainly when discussions of resuming culling activities resurface, based on the statements of increasing competition between fishery activities and sea lions during El Niño events. In 1997 Peruvian fishermen called for a cull of sea lions to protect fisheries and the Peruvian Fisheries Ministry was considering a pilot programme to kill up to 60 sea lions. However, as a result of the 1997–1998 El Niño event the numbers of sea lions onshore were drastically reduced and the programme was abandoned (Lama, Reference Lama2010; Seal Conservation Society, 2010). Therefore, we recommend that the conservation managers consider this estimated N e in future management strategies to ensure the conservation of the species and maintenance of its evolutionary potential on the Peruvian coast.
ACKNOWLEDGEMENTS
The authors would like to thank the Instituto del Mar del Perú (IMARPE), who kindly provided information related to Peruvian studies; Enrique Alberto Crespo for conservation insights; Rocío Loizaga de Castro for supplying important references; Adriano Duarte who kindly prepared Figure 1; Fernando Lopes for your general computer skills; Emily Toriane and Robert Brownell Jr. for reviewing our English; and Francisco Aquino for about El Niño events and climatic change. We also thank the anonymous referees for their comments and suggestions on this manuscript. This study was Gemars contribution no. 35. This work was partially supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (MCT/CNPq 14/2010-479199/2010-8).
