Biology of Marine Birds - Chapter 5 ppt

Seabird Demography and Its Relationship with the Marine Environment Henri Weimerskirch CONTENTS 5.1 Demography and Life History Strategies ...115 5.2 Seabirds and Other Birds ...117 5.3

Seabird Demography and Its Relationship with the Marine Environment

Henri Weimerskirch

CONTENTS

5.1 Demography and Life History Strategies 115

5.2 Seabirds and Other Birds 117

5.3 Demographic Parameters of Seabirds 118

5.4 Comparing the Demography of the Four Orders of Seabirds 120

5.5 Factors Responsible for Differences in Demographic Tactics 124

5.6 Intraspecific Variations in Demographic Traits 125

5.7 Population Regulation and Environmental Variability 129

5.8 Perspectives 131

Acknowledgments 132

Literature Cited 132

5.1 DEMOGRAPHY AND LIFE HISTORY STRATEGIES

Demography is the study of the size and structure of populations and of the process of replacing individuals constituting the population The study of demography was developed to forecast pop-ulation growth The rate at which a poppop-ulation increases or decreases depends basically on the fecundity (number of eggs laid) and survivorship of the individuals that belong to the population (Figure 5.1, bottom), but also to a lesser extent (especially for seabirds) on migration Because many organisms, and especially seabirds, breed several times in their lives, a population consists

of cohorts of individuals of different ages, born in different years Moreover, mortality and fecundity rates are generally age-specific; life tables represent these birth and death probabilities The rela-tionship between the rate of increase or decrease and demographic parameters can be translated into more or less complex equations The basic equation is the Euler–Lotka equation (Euler 1760, Lotka 1907) that specifies the relationships of age at maturity, age at last reproduction, probability

of survival to age classes, and number of offspring produced for each age class, to the rate of growth of the population (r)

The demography of organisms is a key to the evolution of life histories because it allows us

to examine the strength of selection on life history traits Although they can achieve similar population growth rates, i.e., being stable, increasing, or declining, each population living in a particular habitat has specific dynamics, with specific age-related survivals and fecundities The particular values of the demographic traits depend upon the adaptation of individuals and the attributes of the environment in which they live Therefore, comparing demographic traits of populations allows us to elucidate the ecological and evolutionary responses of populations to their

5

environments The comparison of demographic traits among taxa shows that demographic “tactics”exist; the concept of demographic tactic describes a complex co-adaptation of demographic param-eters (Stearns 1976) Basically these co-adaptations result in the existence of a gradient from taxawith high fecundity and a low survival, to species with a high survival and a low fecundity Thisfast–slow gradient (fast meaning fast turnover, and slow, slow turnover) or r/K gradient (Pianka1970) provides a convenient (although not perfect) summary of the patterns linking life historiesand habitats.

However, caution must be taken when life histories are compared First it is possible to comparetaxa from an ecological point of view as long as the allometric relationship linking them at a highertaxonomic level is known (Clutton-Brock and Harvey 1979) For example, within a taxa (a genus,for example), individuals of a particular species may live longer or produce fewer offspring thananother species, not because they rely on a different habitat, but only because they are larger.Because they are larger they have a lower metabolism and therefore could live longer; they mayproduce fewer offspring because their offspring are larger and therefore require more energy (Calder1984) The second constraint is phylogenetic (Harvey and Pagel 1991) Species are prisoners oftheir evolutionary past and can evolve to only a limited number of options The single egg clutch

of all Procellariiformes, and many other seabird species, has often been taken as an example forthis (Stearns 1992) The life histories and habitats of two albatrosses can reasonably be compared,but care has to be taken when an albatross is compared to a species belonging to a different order.Phylogeny sets limits on an organism’s life history and habitat but the ecological task of relatinglife histories to habitats is a fundamental challenge in ecology (Begon et al 1996) Comparingdemographic tactics within taxonomic levels that are closely related (ideally within the same species,see Lack 1947) to habitats or ecology remains a powerful tool to understand the influence of theenvironment on the evolution of life histories (Figure 5.1)

The aim of this chapter is first to describe the demographic traits of seabirds and comparethese traits between taxa to examine whether demographic tactics can be found between andwithin the four orders of seabirds Second, the variation in demographic traits will be examined

to see whether it can be related to differences in the marine environment or the way seabirds

FIGURE 5.1 Schematic representation of the relationships between demographic traits and the marine

envi-ronment.

Seabird Demography and Its Relationship with the Marine Environment 117

exploit it, when comparing species within the same order, but also by comparing populationswithin the same species

5.2 SEABIRDS AND OTHER BIRDS

In this study, a seabird is considered the species breeding along the seashore and relying on marineresources during the breeding season Therefore several species of Pelecanidae, Laridae, Sternidae,and Phlacrocoracidae breeding inland or relying on freshwater resources are excluded, althoughthey often winter in marine habitats The data set used here includes 177 species of seabirds, withinformation on fecundity for 103 species, on age at first breeding for 111 species, and onsurvival/life expectancy for 76 species All three parameters were simultaneously available for 62species, and fecundity and age at first breeding for 84 species Data were taken from Cramp(1978), Jouventin and Mougin (1981), Cramp and Simmons (1983), Marchant and Higgins (1990),Del Hoyo et al (1992, 1996), Gaston and Jones (1998), unpublished data from a long-term database for southern seabirds (CEBC-IFRTP), and unpublished data provided by E A Schreiber fortropical Pelecaniformes

When compared with other birds, seabirds have lower fecundity; they breed at an older age andhave higher adult survival Since age at first breeding, survival, and to a lesser extent clutch size,are explained in part by mass (relationship between log body mass and log of demographicparameters: clutch y = –0.081x – 1.33, r2 = 0.028, p < 0.01, n = 362 species of seabirds and other

birds; age at first breeding y = 0.215x – 0.545, r2 = 0.313, p < 0.001, n = 261, survival y = 0.249x

+ 0.4859, r2 = 0.394, n = 127, p < 0.001), it is important to remove the effect of size Indeed it

could be argued that on average, seabirds are larger than land birds To remove the variation ofdemographic traits related to body mass, they were transformed as log (parameter) – 0.25 log (mass)(Stearns 1983, Gaillard et al 1989) Once the effect of body mass has been removed for clutch sizeand age at first breeding, seabirds still appear to stand at the extreme slow end of the fast–slowgradient that exists for bird species (Figure 5.2), underlying the low reproductive rate of seabirds

FIGURE 5.2 Relationship between the clutch size and minimum age at first breeding (both corrected for

body mass) in 175 species of seabirds (black dots, y = 0.571x + 0.72, r 2 = 0.541, p < 0.001) and 187 species

of other birds (white dots, y = 0.306x – 0.118, r 2= 0.216, p < 0.001) belonging to all the existing orders of

birds for which data are available.

5.3 DEMOGRAPHIC PARAMETERS OF SEABIRDS

When examining the demographic parameters of seabirds, extensive differences exist between andwithin orders, families, and species (Table 5.1) Fecundity is the product of clutch size, breedingfrequency, and breeding success Fecundity of seabirds is generally low, with all Procellariiformes,Phaethontidae, Fregatidae, and several species of Sulidae, Alcidae, Sternidae, Spheniscidae, andeven some Laridae having a clutch of one (Table 5.1; see also Appendix 2) Several species of

Diomedeidae, one of Procellaridae (the White-headed petrel Pterodroma lessonii; Figure 5.3), andprobably most Fregatidae (at least females) breed only every second year when successful On theother hand, some species of Phalacrocoracidae can have clutch sizes reaching five to seven eggsand many species of Laridae have clutch sizes of three and are able to lay a replacement clutchwhen failing early in the season (Figure 5.4)

The reasons for the low fecundity of seabirds have been much debated, and David Lack usedseabirds, especially pelagic seabirds with a very low fecundity, to illustrate his general theory onclutch size (Lack 1948, 1968) Basically, Lack suggested that altricial birds should lay the clutchthat fledges the most offspring The ability to provide enough food to offspring would therefore bethe main reason for the low reproductive rate of some seabirds The development of life historytheory and especially the concept of cost of reproduction and residual reproductive value (Williams1966) later sophisticated this view The basic idea is that, because resources are assumed to belimited, reproduction can have a negative influence on the probability of survival to the nextreproduction, and therefore individuals should balance present and future reproduction (allocation;see Figure 5.1) For a long-lived species, the risk taken, especially during the first years of life,should be limited in order to enhance future reproductive success Long-lived animals wouldtherefore behave as “prudent parents,” trying to limit risks of increased mortality when reproducing.Therefore the single clutch of albatrosses and many other seabirds may have evolved as theresult of the low provisioning rate of chick due to distant foraging zones (Lack 1968), but also ofthe “prudent” behavior of the parents that would limit energetic investment because of their highreproductive value However, whether a clutch of one is the best option for other seabirds with adifferent ecological specialization is not clear (Ricklefs 1990) Indeed the low fecundity of seabirds

is generally attributed to the marine environment on which they rely, an environment that is assumed

to be poor, patchy, and unpredictable (Ashmole 1971) However, obviously the marine environment

is very diverse and heterogeneous, with localized rich feeding areas or areas of low productivity.Therefore we might expect differences in demographic tactics within taxa according to the envi-ronment exploited, or to the foraging technique used, or diet Conversely, convergence might beexpected between taxa exploiting the same resources or environment, and divergence within taxawhen environments exploited are different

The minimum age at first breeding ranges from 2 to 4–5 years in most species of seabirds,except for Diomedeidae and Fregatidae and some species of Procellaridae that start breeding later(Table 5.1) Late age at first breeding is generally assumed to be necessary for long-lived species

to attain similar foraging skills to those of adults, either because skills are complex to attain (e.g.,Orians 1969, Burger 1987) and/or because of the high reproductive value of young birds.Like age at first breeding, but even more importantly, survival is a parameter that is difficult

to estimate accurately because it requires the marking of birds and their recapture over severalyears Estimates of adult survival are available for a limited number of species (Table 5.1) and have

to be treated with caution Indeed, the statistical methods to estimate survival are in constantrefinement, resulting in an overall increase of the estimates of survival rates within a species astechniques improve (Clobert and Lebreton 1991) Therefore, comparisons of survival are oftendifficult to perform unless the same method has been used Average longevity is generally used toillustrate survival but cannot be compared to longevity records that only give maximum age based

on isolated recaptures Most Procellariidae and Diomedeidae have high survival and life expectancy,but also several species within the other orders, for example, several species of Alcidae and one

in Figures

Number of Species with

at Least One Parameter

Average Clutch

Frequency

of Breeding

Age at First Breeding

Adult Life Expectancy (number of species with an estimate

of survival)

Relationship between Age at First Breeding – 1/Fecundity (both corrected for body mass)

Relationship between 1/Fecundity – Life Expectancy (both corrected for body mass)

Sphenisciformes Spheniscidae Cross 15 1–2 0.7–1 2–5 6.4–20.5 (10) y = 0.219x – 0.648, r 2 =

Laridae Unfortunately, no estimate is available for Fregatidae, nor for most tropical Procellariidae,Laridae, and Sternidae, limiting the scope of a general comparison The low fecundity and late age

at first breeding of Fregatidae suggest high survival rate (maximum age recorded 34 years [E.A.Schreiber personal communication]), probably similar to Diomedeidae One reason for the highsurvival of seabirds, especially those breeding on oceanic islands, is the absence of terrestrialpredators; this is probably true for most large species, but not for the smaller species that can sufferheavy mortality from avian predators Estimates of survival between fledging and recruitment intothe breeding population are more difficult to obtain logistically because of the delayed age at firstbreeding, and are rare in the literature, limiting the scope for meaningful comparisons between groups

5.4 COMPARING THE DEMOGRAPHY OF THE FOUR ORDERS OF

SEABIRDS

Within seabirds, minimum age at first breeding and life expectancy (log transformed) are somewhatrelated to the log of mass (y = 0.092x + 0.666, r2= 0.0788, p < 0.01 and y = 0.1148x + 1.675, r2

= 0.1532, p < 0.001) These relationships express the allometric component of demographic pattern

and indicate that body mass is a significant, but not fundamental, determinant of the variation indemographic traits in seabirds They represent a first-order tactic which expresses the biomechanicalconstraints of body mass (Western 1979, Gaillard et al 1989) When parameters are corrected forthe effect of body mass, the relationships between demographic traits are still very significant(Figure 5.5), representing a second-order tactic (Western 1979) It indicates that demographicparameters of seabirds covary after correction for the effect of body mass, which suggests theexistence of demographic tactics among seabirds The relationship between fecundity and lifeexpectancy is very significant (Figure 5.5) and highlights the classical balance between clutch sizeand survival rates The relationship between fecundity and age at first breeding, and that betweenage at first breeding and life expectancy, are also highly significant (Figure 5.5) The regressionlines for the three relationships each describe a similar gradient within seabirds going from specieswith a fast turnover (high fecundity, early age at first breeding, and short life expectancy) to specieswith a slow turnover

When examining the species within each order, they appear not to be distributed evenly alongthis fast–slow gradient Spheniciformes appear to be distributed at the left-hand size of the gradient

FIGURE 5.3 A White-headed Petrel They breed only every other year, incubating their egg for 60 days and

spending 112 days raising their single chick (Photo by H Weimerskirch.)

Seabird Demography and Its Relationship with the Marine Environment 121

or fast turnover end of the gradient: penguins breed relatively early, have a short life expectancy,and a high fecundity relative to their size Conversely, many Procellariiformes species are found

at the slow turnover extreme (Figure 5.5) Since the relationship considers all seabirds, i.e., fourdifferent orders, it is important to examine whether the relationships are a result of taxonomicdifferences in demography Controlling for phylogeny (Harvey and Pagel 1991) was not possiblebecause of the lack of a complete phylogeny covering all species of seabirds, and was out of thescope of this study When investigating the existence of a gradient within orders, it appears thatsignificant relationships persist within Procellariiformes and Pelecaniformes, whereas there is atendency, yet nonsignificant for Charadriiformes, and no relationship for Spheniciformes (Table5.1) This suggests the existence of different demographic tactics within Procellariiformes andPelecaniformes, and perhaps Charadriiformes We will now examine whether these tactics amongtaxa tending to show a fast or a slow turnover can be related to different environmental conditions

or foraging strategies

(a)

(b)

FIGURE 5.4 Seabird species exhibit a range of fecundities (a) Some gulls, such as this Herring Gull, may

raise three chicks in a year, spending 45 to 50 days feeding them before they fledge (b) Giant Petrels raise one chick a year and spend 100 to 120 days feeding it before it fledges (Photos by J Burger.)

FIGURE 5.5 Relationships between 1/fecundity, age at first breeding, and life expectancy (corrected by body

mass) in the four orders of seabirds (Sphenisciformes, crosses; Procellariiformes, symbols filled in black; Pelecaniformes in gray; and Charadriiformes in white) The inverse of fecundity is used for clarity, so that the three variables are positively linked Correspondences of symbols for families are given in Table 5.1 Fecundity is estimated as the number of young produced per female per year It is the product of the average clutch size per year by the overall breeding success Because data on the average age at first breeding are scarce, minimum age at first breeding is used Adult life expectancy is directly derived from adult survival and is measured as (0.5 + 1/(1 – s)) (Seber 1973) When parameters are available for several populations, average values are used.

Seabird Demography and Its Relationship with the Marine Environment 123

To allow an easier representation of the ranking of species along this gradient, the species havebeen plotted along the first component of a principal component analysis (PCA) performed on thedemographic parameters When the three parameters are used, the first principal component explains71.1% of the total variance (Figure 5.6a) One extreme, the left-hand side, is characterized by ahigh fecundity, short life expectancy, and early age at first breeding, while the other extreme presentsthe opposite characteristics Because of the low number of species for which life expectancy isknown, with an absence of data for some families like Fregatidae (see Table 5.1), a PCA was alsoperformed on the fecundity and age at first breeding only, to be able to plot a larger number ofspecies The first principle component then explains 74.3% of the total variance (Figure 5.6b).Because the two analyses provide very similar ranking (compare Figure 5.6a and 5.6b, Factor 1 (2parameters) = 0.924 × Factor 1 (3 parameters) + 0.043, r = 0.956, p < 0.001) We use the ranking

obtained from the PCA performed on fecundity and age at first breeding only, with the largernumber of species (Figure 5.6b)

(a)

(b)

FIGURE 5.6 Ranking of the four orders of seabirds along a slow–fast gradient described by the first principal

component of the PCA analyses (see symbols for families in Table 5.1): (a) PCA performed on 1/fecundity, life expectancy, and age at first reproduction, all corrected for body weight (eigenvalues 2.133, 0.546, and 0.321); and (b) PCA performed on 1/fecundity and age at first reproduction, both corrected for body weight (eigenvalues 1.487 and 0.513).

Spheniciformes and Procellariiformes almost do not overlap on the gradient, whereas iformes extend throughout the gradient, and Charadriiformes are intermediate (Figure 5.6b).Whereas the species within the four families of Procellariiformes are scattered throughout thegradient, in Pelecaniformes the four families appear to be clearly separated from one another:Phalacrocoracidae, Sulidae, Phaethontidae, and Fregatidae ranking separately on the fast–slowgradient This ranking probably reflects a strong phylogenetic effect on demographic tactics withinthis order, with each family having a distinct morphology and feeding specialization Conversely,within Procellariiformes, Diomedeidae and Procellaridae are very similar in terms of morphologyand feeding technique and are ranked similarly Similarities in demographic traits between somefamilies belonging to different orders suggest convergence Phalacrocoracidae appear to haveequivalent demographic tactics to those of Spheniciformes, having a fast turnover Diving petrels,Pelecanoididae, also appear have faster turnover than most other petrels This tendency to be at thefast extreme of the gradient in these three families could be associated with the constraints of divingthat make birds poor fliers and therefore reduce foraging range Convergence in demographic tacticsmay also be found between Fregatidae and the longest lived albatrosses and petrels These birdshave in common a pelagic life but especially economic flight In Charadriiformes, a ranking ofdemographic tactics by families is also apparent, although less clear-cut than in Pelecaniformes,

Pelecan-with Laridae (Pelecan-with the exception of one species, the Swallow-tailed Gull Creagrus furcatus) and

Stercoraridae toward the fast extreme Conversely, Sternidae and Alcidae are distributed over awider range, rather at the slow end, suggesting convergence in demographic tactics with speciesthat are well known to be long-lived like Procellariiformes

5.5 FACTORS RESPONSIBLE FOR DIFFERENCES IN DEMOGRAPHIC

TACTICS

Some demographic traits are phylogenetically conservative and fixed at high taxonomic levels Forexample, all Procellariiformes have a clutch size of one Others, like minimum age at first breedingand maximum life expectancy, probably do not vary within populations of a species because theyare likely not to be adapted to local environmental conditions Maximum life expectancy is probablymainly related to allometric pressures or phylogeny Small birds have a higher energy expenditureand therefore shorter life span than larger birds (Lindstedt and Calder 1976; see Chapter 11) Thereare negative correlations between survival and vigorous, energy-expensive activity such as flight(Bryant 1999); consequently, birds with a low-energy flight such as albatrosses may live longercompared to birds with a highly expensive flight such as shags On the other hand, breeding success,breeding frequency, average age at first breeding, and adult and juvenile survival express theinteractions between phenotype and environment and are influenced by the environment (Figure5.1) These demographic traits are likely to be different between closely related species exploitingdifferent marine environments, or even within the same species exploiting different environments.Therefore, families covering a wide range over the fast–slow gradient suggest a broad range ofdemographic tactics due, for example, to a group of species exploiting a diversity of habitats.Conversely, families with a restricted range along the fast–slow gradient suggest that all speciesbelonging to this group probably face similar environmental conditions For example, Sulidae rankover a relatively restricted range, but they breed from tropical to sub-Arctic waters

Seabirds have been classically separated into inshore, offshore, and oceanic or pelagic (Ashmole1971), and it is generally assumed that pelagic species are the most long-lived, whereas inshorespecies are shorter lived (Lack 1968) Therefore, we might expect that pelagic species should befound at the slow turnover extreme of the fast–slow gradient When considering the four orderssimultaneously, there is indeed a tendency for oceanic families to stand at the slow end of thegradient (e.g., most Procellaridae, Hydrobatidae, Diomedeidae, or Fregatidae), whereas moreinshore families are found at the other extreme However, this is mainly due to the fact that many

Seabird Demography and Its Relationship with the Marine Environment 125

Procellariiformes are pelagic and stand at the slow extreme of the gradient Examining the bution of inshore, offshore, or oceanic species within families does not lead to the clustering ofinshore or oceanic species at one or the other extreme of the gradient (Figure 5.7a) This suggeststhat the assumption that pelagic species are more long-lived than inshore species only exists whengroups are compared at high taxonomic levels (for example, when comparing Procellariiformesand Charadriiformes) But when the effect of size is controlled and each order examined separately,the data available today do not allow us to conclude that pelagic species have a slower turnoverthan inshore species

distri-Polar waters are generally more productive than tropical waters, which may have influencedthe evolution of demographic traits Therefore, we might expect that tropical species might belocated at the slow end of the gradient compared to polar species This tendency is not apparentwith the data available (Figure 5.7b) For most families, either some breed only over a narrowrange of climates (Alcidae or Fregatidae, for example) or data are not available (tropical Procel-laridae, for example), limiting the possibility of making generalizations

These first examinations indicate that the role of the marine environment in shaping graphic tactics is difficult to determine, and that the conjunction of several factors has probablybeen involved in shaping the demographic traits of marine birds Because data are lacking for manygroups, comparisons at lower taxonomic levels are impossible at this time

demo-5.6 INTRASPECIFIC VARIATIONS IN DEMOGRAPHIC TRAITS

Some species that are separated geographically show very homogenous demographic traits between

populations For example, Wandering Albatrosses (Diomedia exulans) breeding in the Atlantic,

Indian, or Pacific Ocean have very similar demographic traits (Weimerskirch and Jouventin 1987,Croxall et al 1990, Weimerskirch et al 1997, DeLamare and Kerry 1992; Figure 5.8), suggestingthat each population is relying on similar resources in the three regions Indeed both in the Atlantic

(a)

(b)

FIGURE 5.7 Ranking of the four orders of seabirds along a slow–fast gradient described by the first principal

component of the PCA on 1/fecundity and age at first breeding: (a) Symbols are inshore (white), offshore (gray), and oceanic (black) species (b) Symbols are polar (white), temperate (gray), and tropical (black) species.