Click on the links below to skip ahead:
|Results of Prior NSF Support
GEOGRAPHIC STRUCTURE OF ADELIE PENGUIN POPULATIONS: DEMOGRAPHY OF COLONY GROWTH
Principal Investigators: D Ainley, G Ballard, P Wilson, K Dugger, P Lyver
OPP 9526865, 9814882, 0125608 (current study):
Factors regulating population size and colony distribution of Adélie Penguins
in the Ross Sea. $1,800,000. 9 years: 23 August 1996-31 July 2005.
Co-PIs: D. Ainley, G. Ballard, L. Ballance, K. Dugger, N. Nur, G. Rau, and
C. Ribic. Collaborators at LandCare Research NZ (P. Wilson, K Barton), funded
We have completed 8 of 9 NSF-funded field seasons in which we attempt to explain why 1) colonies in the southern portion of the Adélie Penguin range (e.g., Ross Sea) increased in size noticeably during the 1970-80s and then leveled off (Fig. 1); 2) the rate of increase and variability has been much greater at smallest colonies (and minimal at large colonies); and 3) adjacent colonies ranged greatly in size in the first place. To investigate these issues, we chose an isolated cluster of 4 colonies (a metapopulation) on or near Ross Island that together comprise 8% of the world population of this species (30% of which lies in the Ross Sea) and range three orders of magnitude from among the largest to the smallest for this species: Ross Island (Cape Crozier 130,000; Cape Bird 36,000; Cape Royds 4,000); and Beaufort Island 38,000 pairs. A 5th, incipient colony, Cape Barne (also on Ross Is), has been occupied irregularly in recent years, and at times throughout the Holocene, by 20 or more pairs. Only the colony at Beaufort Island is space limited. Our basic plan was to compare the reproductive and foraging ecology and effort of penguins nesting at colonies of different size, as well as to assess demographic parameters.
Fig. 1. Changes in breeding numbers of Adélie Penguins at three of five colonies in the Beaufort/Ross Is metapopulation, 1960-2000. Colonies differ in size by successive orders of magnitude: Royds smallest, Crozier largest. Beginning with the early 1980s, colony growth rate and variability differ by colony size.
Our first objective was to develop a computerized weighbridge (WB), from which, using implanted PIT tags (passively-interrogated transponder) to identify individual penguins, we could assess adult mass, meal size fed to chicks, and feeding frequency. We succeeded, and one reliable WB has been deployed at each of 3 colonies for 4-7 seasons (Crozier and Royds 7, Bird 4 seasons). We have since published results from the WB in 4 major publications (Ainley et al. 1998, 2004; Ballard et al. 2001, ms). We have found that penguins act like optimal foragers as long as feeding-trips are ≤ 2 d (food loads returned increase with trip length; Fig. 2).
Fig. 2. Amount of food fed to chicks (parent arrival mass minus departure mass) as a function of the duration of foraging trips (in hours); Ross Island, 1996-97. More food is provided as trips increase to 2 d; thereafter, less food is delivered as trip length increases.
However, as trip length increases (> 2 d), both meal size and adult body condition (mass) decreases as does chick growth. Further, trip times and meal sizes differ between sexes, but are closely related to parent mass at the start of chick feeding; heavy birds subsequently lose mass but lighter ones forage longer and gain mass. We have demonstrated a relationship to sea ice concentration as well, with smaller meal sizes and longer foraging trips in seasons of heavier sea ice (Fig. 3). Finally, trip times were consistently longest and adult body mass lightest at the largest colony, indicating negative factors involved in life at large colonies.
Fig. 3. Foraging trip duration (h) and food loads (g) related to sea-ice concentration within the foraging area of Adélie Penguins at Cape Crozier, 1997-2000. Trips were longer and food loads smaller in seasons of heavier ice cover while controlling for effects of sex, season, timing within breeding season, and individual (both P = 0.03). The effect on trip duration was stronger later in the chick-rearing period (P < 0.001).
During all summers, in addition to foraging effort as determined by the WB,
we also compared breeding effort among the colonies, measuring such factors
as chick growth, breeding success, numbers of breeding pairs, and survival
of breeders (Ainley et al. 1998, 2004; Ballard et al. 2001, ms). Chick growth
was consistently greatest at the small colony and least at the large one.
Breeding success did not vary. Numbers of breeding pairs was lowest at all
colonies during summers when spring sea ice was concentrated, thus hindering
In year 4, we used doubly-labeled water to gauge the degree to which energy expenditure was greater on longer foraging trips (Fig. 4). We also deployed time-depth-recorders to assess the structure of foraging trips (Fig. 5). The work is being prepared for publication (L. Ballance), but it appears that longest trips are the most costly, helping to explain the body mass loss of long-foraging penguins (Figs. 2-6). Penguins from the large Crozier colony are apparently spending much more effort in procuring food than are penguins from the smaller colonies.
Fig. 4. The amount of energy expended by Adélie Penguin parents as a function of foraging trip duration (in days) from colonies on Ross Island. Data collected January 2000 and 2002 using doubly-labeled water.
Fig. 5. Maximum and average diving depth of parents at Cape Crozier (solid line) and Cape Royds (dashed line) during 5-d periods from start of chick provisioning, 1999-2001. Shown here are predicted values from 2-way linear model that controls for seasonal variation (n = 47 individuals Crozier, 36 Royds).
We have quantified diet at each colony throughout the chick period every year
in order to determine, at the least, that diet quality does not vary geographically.
We have yet to directly assess prey availability, as the logistic support
has not been available. There is indirect evidence for equal availability
of foraging habitat, however, given that all the area offshore of colonies
is being exploited by the penguins (see below). We have found that diet composition
(and quality) is similar among colonies (Ainley et al. 1998, 2003a).
In all years, radio telemetry (combined with satellite transmitters in yr 5, and satellites exclusively since) provided information on foraging area size and degree of overlap among the 4 colonies (Fig. 6; Ainley et al. 2004). We found up to 50% overlap among the three smaller colonies (Beaufort, Bird, Royds), but minimal (if any) overlap with the large colony (Crozier). The foraging area of the 3 smaller colonies together was equivalent in size to that of the largest colony, but the density of foraging penguins in the common area was half that in the foraging area of the large colony. At the two largest colonies foraging by parents was at first close, but then progressively farther away as the chick-provisioning period advanced (larger chicks demanding more food). In the case of birds from Crozier, towards the end of the chick period they began to forage as far as the eastern shores of Beaufort Is, after which Beaufort birds no longer foraged there (Beaufort birds then foraged only to the west of their island). These results offered evidence that competition may be important in metapopulation structuring (i.e., the large colony affecting the size and location of foraging areas of smaller colonies located within foraging range).
Fig. 6. Penguin foraging areas for Cape Crozier (purple), Beaufort Is (blue), Cape Bird (red), and Cape Royds (green) prior to arrival of big icebergs; size of stars an index to colony size. Color intensity in grid related to triangulations per cell (dark, >10; medium, 5-9; light, <5). Cells with more than one color indicate overlap. Dark lines show fast ice boundaries; dotted lines indicate seasonal retreat of fast ice edge. Note that foraging area sizes proportional to colony size; little overlap with foraging of largest colony, but extensive overlap among small colonies.
During and after the 5th field season (2000), two very large icebergs
(C-16, 10 x 45 km; B-15, 40 x 160 km) grounded between Cape Crozier and Beaufort
Island (Arrigo et al. 2002; see below: The Study). The result was that Crozier
birds were prevented from foraging anywhere close to Beaufort Island, an area
from which Beaufort birds no longer retreated. This is experimental evidence
that the large population at Cape Crozier was displacing foraging birds from
Beaufort Island, indicating further that competition could be a factor involved
in geographic structuring of colonies (Ainley et al. 2004).
In the 7th field season (2002), we attached miniature GLS tags to penguins at Royds and Crozier in order to determine winter movements; in the 2003 season these archival tags were retrieved for data downloading (Fig. 7). Other tags were attached to be recovered in 2003-04. We were very pleased with results, as they will allow is to test some hypotheses generated in some retrospective analyses that relate colony growth to environmental factors. In these analyses, we used data on Adélie Penguin colony size at Cape Royds and Cape Bird, 1960 to the present (Fig. 1; thanks to our NZ collaborators). In the first, using path analysis, we found that 30% of annual change in population size is explained in a negative relationship to maximum sea-ice extent 5 yrs earlier (Wilson et al. 2001). Five years is the average age of recruitment in this species; and a sensitivity analysis showed that the measured population variability is most sensitive to changes in juvenile survivorship. We hypothesized that extensive ice moves the wintering penguins into unproductive waters north of the Antarctic Circumpolar Current southern boundary. In another study, we related colony size variation, as a function of the proportion of birds breeding, to the Antarctic Oscillation (affects wind patterns) and the size of the Ross Sea polynya (Ainley et al. ms). We continue to explore colony growth as related to sea ice characteristics for this pack ice-obligate species. We have submitted a proposal to NASA that would enlist the help of K Arrigo (productivity) and C Parkinson (ice characteristics; see below).
The natural experiment resulting from the icebergs has provided a gold mine of insights into the role of philopatry in structuring Adélie Penguin populations. One of the main features of our present request for renewal of this project concerns demography, and we will treat that subject below, including description of how the icebergs have helped.
Fig. 7. Uncorrected positions of penguins after being fitted with GLS tags in February 2003. Dark line is the southern boundary of the Antarctic Circumpolar Current (after Orsi et al. 1995). Dot colors are specific to months: March red, April orange, May yellow, June green, July blue, August purple, September pink, October brown. Penguin A, traveled to the eastern Ross Sea to molt. The ice it eventually encountered became entrained in a gyre. Penguin B, left later and encountered ice floes in the central Ross Sea that, not being part of the gyre, were advected north and west crossing into the unproductive waters of the ACC.
Besides hosting a Teacher Experience in Antarctic participant during 1999, and 5 pre-graduate-school interns in 2002-04, our contributions to the scientific and popular literature have been:
PUBLICATIONS (many of these papers can be acquired from our website)
Ainley, DG. 2002. Adélie Penguin: Bellwether of Climate Change. Columbia Univ. Press, NY.
Ainley, DG. 2002. The Ross Sea, Antarctica: Where all ecosystem processes still remain for study, but maybe not for long. Marine Ornithol. 31: 55-62.
Ainley, DG, G Ballard, SD Emslie, WR Fraser, PR Wilson & EJ Woehler. 2003. Adélie Penguins and environmental change. Science 300: 429.
Ainley, DG, G Ballard, KJ Barton, BJ Karl, GH Rau, CA Ribic & PR Wilson. 2003. Spatial and temporal variation of diet composition and quality within a presumed metapopulation of Adélie Penguins. Condor 105: 95-106.
Ainley, DG, ED Clarke, K Arrigo, WR Fraser, A Kato & PR Wilson. Ms. Decadal-scale changes in the climate and biota of the Pacific sector of the Southern Ocean, 1950s to the 1990s. Antarc. Sci., in press.
Ainley, DG & GJ Divoky. 2001. Seabirds: effects of climate change. Pp. 2669-2677 in Encyclopedia of Ocean Sciences (J. Steele, S. Thorpe & K. Tarekian, eds.). Academic Press, London.
Ainley, DG, CA Ribic, G Ballard, S Heath, I Gaffney, BJ Karl, KR Barton, PR Wilson & S Webb. 2004. Geographic structure of Adélie Penguin populations: size, overlap and use of adjacent colony-specific foraging areas. Ecol. Monogr. 74: 159-178.
Ainley, DG, CT Tynan & I Stirling. 2003. Sea ice: a critical habitat for polar marine mammals and birds. Pp 240-266 in Sea Ice: An Introduction to its Physics, Biology, Chemistry and Geology (D.N. Thomas & G.S. Diekman, eds.). Blackwell Science, London.
Ainley, DG, PR Wilson, KR Barton, G Ballard, N Nur & BJ Karl. 1998. Diet and foraging effort of Adélie penguins in relation to pack-ice conditions in the southern Ross Sea. Polar Biol. 20: 311-319.
Ainley, DG, PR Wilson & WR Fraser. 2001. Effects of climate change on Antarctic sea ice and penguins. Pp 24-25 in Impacts of Climate Change on Wildlife (R.E. Green et al., eds). Royal Soc. Protection of Birds, Sandy. UK.
Arrigo, KR, GL van Dijken, DG Ainley, MA Fahnestock & T Markus. 2002. The impact of the B-15 iceberg on productivity and penguin breeding success in the Ross Sea, Antarctica. Geophys. Res. Letters 29(7), 10.1029/2001GLO14160.
Ballance, LT, DG Ainley & GL Hunt, Jr. 2001. Seabirds: foraging ecology. Pp 2636-2644 in Encyclopedia of Ocean Sciences (J. Steele, S. Thorpe & K. Tarekian, eds.). Academic Press, London.
Ballard, G, DG Ainley, N Nur, KJ Barton & PR Wilson. Ms. Seabird foraging decisions: Parental foraging effort of Adélie penguins in relation to gender, body condition, and environmental conditions. J. Animal Ecol., submitted February 2004.
Ballard, G, DG Ainley, CA Ribic & KR Barton. 2001. Effect of instrument attachment and other factors on foraging trip duration and nesting success of Adélie penguins. Condor103: 481-490.
Beigel, M, S Marcus & G Ballard. 2004. Exception management for RFID systems. Smart Labels Analysis 36:1-8.
Emslie, SD, PA Berkman, DG Ainley, L Coats & M Polito. 2003. Late-Holocene initiation of ice-free ecosystems in the southern Ross Sea, Antarctica. Mar. Ecol. Progr. Ser. 262: 19-25.
Smith, RC, DG Ainley, E Domack, S Emslie, WR Fraser, K Baker, J Kennett, A Leventer, E Mosley-Thompson, S Stammerjohn & M Vernet. 1999. Marine ecosystem sensitivity to historical climate change: Antarctic Peninsula. BioScience 49: 393-404.
Webb, S. 2000. My Season with Penguins. Houghton-Mifflin, Boston.
Wilson, PR, DG Ainley, N Nur, SS Jacobs, KJ Barton, G Ballard & JC Comiso. 2001. Adélie Penguin population change in the Pacific sector of Antarctica: relation to sea-ice extent and the Antarctic Circumpolar Current. Mar. Ecol. Progr. Ser. 213: 301-309.
During the course of this project, thus far, we have
presented oral papers at 6 and poster papers at 5 scientific meetings, including
SCAR Biology Symposium (1998), the 4th International Penguin Conference (2000),
Pacific Seabird Group Annual Meetings (1998, 2000, 2002), Gordon Conferences
(1999, 2001), and an International Sea-ice Symposium (2000).
We have submitted 4 papers for presentation at the 5th International Penguin Conference, September 2004, Ushuaia, Argentina: foraging trip energetics, effects of bands on survival and foraging, food loads vs foraging effort, diving behavior.
We have also been featured on CBS NEWS (Sunday Morning July 2000, The Early Show Feb 2001) and CNN (CNN Presents), as well as TIME MAGAZINE (Fall 2000: The Big Melt) and SCIENCE NEWS (May 2001: Big Bergs Ahoy! An armada of ice sets sail for new millennium).
Our project objectives are reviewed in the section Results of Prior NSF Support. The results reviewed there describe our study of the ecological factors that affect growth and geographic structuring of colonies in this and other colonial species (reviewed in Ainley et al. 2004). While that work has proceeded, we have been banding and applying PIT tags in order to establish adequate samples of known-age penguins, from which our goal is to answer questions about the demography of population change and metapopulation dynamics in a coastal, polar top-trophic species. This demographic aspect, quite labor intensive (band searching), has become the prime focus during the current segment of our project (2002-04) as it would in the segment proposed here. Obtaining adequate demographic samples is a process that takes several years, as Adélie Penguins in this population, based on data gathered during decline in the 1960-70s (Ainley et al. 1983, Ainley 2002), do not first breed until they are 4-5 years old (average). Only in the most recent effort have we extended our field season from 2.0 mo (early Dec – early Feb, 1996-2001) to 3.5 mo (early Nov – early Feb), required to assess breeding propensity. In the renewed project, we also propose international collaboration with Italian scientists who conduct complementary efforts at nearby Terra Nova Bay, where several Ross Island-banded penguins have been sighted.
Fig. 8. The icebergs B-15 and C-16 as they have been positioned since February 2001, arriving just after the breeding season. During 2003, B-15 broke into three pieces, which have remained just slightly separated in the same configuration as the original. Nevertheless, the full breakup may have begun. McMurdo Sound, which lies to the west of Ross Island, is covered by fast ice in this image. This fast ice, which normally breaks out to leave all of the Sound north of Cape Royds open, has been present to varying degrees since the icebergs grounded. This ice, if extensive, forces the penguins to exert increased effort and changes the propensity of individuals to occupy, visit or recruit to and among the 3 western colonies depending on relative access to open water.
When we wrote the proposal that covers our present activities, in essence, we had a chance to study the demographics of population increase to counter the 1960-70s study of population decrease (see Fig 1). Soon after our project was renewed, however, some large icebergs (B15, C16) grounded to almost entirely occupy the feeding area of the large Crozier colony and essentially erected a wall separating it from the other colonies (cf Figs 6, 8). This wall was 165 km long (~1.5 o latitude) thus abruptly stemming the flow of potential emigrants from Crozier to the smaller colonies (artificially increased philopatry), and increasing the flow of alien-recruits among the latter (artificially decreased philopatry; Fig. 9). We now have an unparalleled natural experiment that seemingly will provide the insights to explain the previously problematic genetic homogeneity detected among Adélie Penguins, including work on the evolution of DNA >10,000 years old (Roeder et al. 2001, Lambert et al. 2002). In other words, the frequency by which such large icebergs appear is high with respect to a genetic or geologic time scale (twice per millennium; D. MacAyeal, pers. comm.). Therefore, we wish to continue the present study, but in addition to our current emphasis on age-, sex-, colony-, and time-dependent survival, emigration, and breeding propensity, we propose to include a shift toward understanding the short- and long-term demographic consequences of the natural experiment associated with the arrival (and retreat?) of icebergs B-15 and C-16.
Fig. 9. Changes in emigration patterns as a function of the arrival of the large icebergs (dotted line) and extent of fast ice in McMurdo Sound (red = most of McMurdo Sound covered). Capes Bird and Royds are on the west side and Crozier on the east side of the icebergs.
In this amazing, ongoing natural experiment, depending on the extent of fast
ice in McMurdo Sound (which has varied subsequently from open to entirely
covered; Figs 8, 9), we have already documented to varying degrees depending
on year and colony, the short-term demographic effects of increased immigration,
poor reproductive success, decreased breeding population size and decreased
breeding propensities at capes Royds, Bird and Crozier. Before the bergs,
appreciable numbers of Crozier birds were sighted elsewhere, a flow that has
since been reduced (Fig. 9, lower right). On the other hand, the iceberg wall
seemingly has forced more McMurdo Sound residents to reside at Crozier, having
chosen a route around the bergs that did not get them to their intended destination
(Fig. 9, lower left, upper right). The extensive fast ice that required 20-50
km over-ice treks by Royds birds, has encouraged them to recruit to Bird,
closer to open water (Fig. 9, upper left). While these patterns have annihilated
the conception that this species is highly philopatric (Ainley et al. 1983),
the long-term demographic effects, particularly those associated with nearly
complete failure of recent age cohorts, will not be fully realized until 4
or 5 years hence, at minimum, when affected age-classes recruit into the breeding
population. With continued assessment of reproductive success, breeding population
size, and our banded known-age birds we can begin to understand how environmental
disturbance events impact both short-term and long-term demographic processes
in this species.
Band-resighting data for 8 years of known-age birds are currently available to us, but at least 5 more years of resighting data is necessary to fully understand the impacts of the iceberg experiment. The average life span of this species is ~ 15 years (up to 20 years; Ainley et al. 1983) and it would be useful to continue at least through one penguin generation. In addition, continued collection of resighting data beginning in the laying/early incubation stage each season will allow us to model age-specific breeding probabilities (Lebreton et al. 2003) and survival differences (if any) between breeders and non-breeders (Spendelow et al. 2002) in relation to environmental factors (i.e., icebergs, spring Ross Sea ice conditions, winter ice extent, Antarctic Oscillation, etc.). Long-term known-age data sets are invaluable for understanding population demographics, particularly for long-lived species with delayed maturation.
In the continuation of this study we will test these and other hypotheses:
1) The demography of the Cape Crozier colony will no longer affect growth
of the McMurdo Sound colonies, which will act as a separate unit; demographic
rates will diverge.
2) Flipper bands can adversely affect survival of Adélie Penguins, especially during years of poor food availability or greater winter ice extent, by decreasing foraging efficiency.
3) Wintering areas of Adélie Penguins in the Ross Island metapopulation vary according to ice conditions and the timing of post-breeding migration. Further, annual variation in adult survival is dependent on the wintering area of the population, which because of B-15, now differs between Cape Crozier and the McMurdo Sound colonies.
4) Among colonies in which reproductive success does not vary, foraging effort, as a result of interference competition for prey, rather than a breeding performance based assessment, is the major factor that affects colony choice among recruits: to be philopatric or not.
Demography. The effects of
icebergs B-15/C-16 on annual survival, philopatry, breeding
propensity, and population size are of particular importance. Because Adélie
Penguins do not recruit until 3-8 years after hatching, current efforts to
band chicks and collect resighting data on known-age birds must be continued
in order to fully understand the dynamic demography of this metapopulation.
We hypothesize a large decline in populations 4 or 5 years hence because entire
cohort classes were lost in years 2001-02 (all colonies), 2002-03 (Royds,
to a lesser extent Bird) and 2003-04 (Crozier). These low fledging rates will
result in low recruitment rates, and subsequently low breeding populations.
Continued banding efforts will allow us to explore breeding propensity, age
at first breeding, survival differences between breeders and non-breeders,
and the degree to which these rates can vary annually and geographically.
For instance, during the extensive ice years of 2002-03 and 2003-04, no 3-
or 4-year olds attempted to breed at Royds, a pattern not consistent with
age of recruitment as determined by Ainley et al. (1983). In addition, when
the severe ice conditions relaxed in 2003-04 (Fig. 9), of the 341 banded known-age
penguins seen at Royds, fully 14% were originally hatched at other colonies.
Obviously, philopatry had relaxed!
We will continue to arrive at capes Royds, Bird and Crozier early in the breeding season (egg-laying/early incubation) in order to definitively document breeding status of banded known-age birds. When banded birds are found, their territories will be marked (GPS location) and revisited regularly to note status and success. We will continue with band resighting throughout the breeding season and a sample of chicks will be banded just prior to fledging at each colony, each year.
Estimates of time-, age-, sex-related breeding probabilities and breeding-status-related survival at each colony, in addition to movement rates among colonies will be generated using Cormack-Jolly-Seber mark-recapture models and Program MARK (White & Burnham 1999; Fig. 10). In addition, specific hypotheses regarding variation in survival in relation to time, age, environmental covariates and breeding status will be tested using mark-recapture models and an Information Theoretic Approach (Burnham & Anderson 2002).
Variations in age at first breeding, age-, sex-, and breeding-status fecundity, and annual breeding propensity and breeding population size will also be estimated for each colony.
Fig. 10. Resighting rates of Adélie Penguins within the Ross Island metapopulation. Survival (resighting) varies annually, with distinct sex-related differences in most but not all years. Resighting rate has decreased and genders have become more disparate in the latest, most environmentally severe years.
Flipper bands.The effects of flipper bands on penguins appear to be variable by species and years of study, with negative effects on resighting rates (or annual survival) reported for Adélie (Ainley et al. 1983, Trivelpiece & Trivelpiece 1994), Gentoo P. papua (Trivelpiece & Trivelpiece 1994), Chinstrap P. antarctica (Trivelpiece & Trivelpiece 1994), and King Aptenodytes patagonicus (Froget et al. 1998) penguins, but no effects reported for Royal Eudytes chrysolophus (Hindell et al. 1996), Adélie (Clarke & Kerry 1998), nor Magellanic penguins Spheniscus magellanicus (Boersma, pers. comm.). A more recent study of King penguins revealed no significant difference in survival rates between banded and unbanded individuals, but implicated impacts to foraging efficiency and related reproductive output (Gauthier-Clerc et al. 2004). While improvements in band design and application have decreased problems with wounding (Sallaberry & Valencia 1985; we have observed none), possible increased energy costs associated with flipper bands may still negatively impact banded individuals (Bannasch et al. 1994, Culik et al. 1993, Gauthier-Clerc et al. 2004). However, flipper bands continue to be the only mark that is easily seen and read, economical to produce, and easy to apply and have provided the only information on penguin demography. We have found that PIT tags can be used effectively to monitor colonies or sub-colonies that are small enough to funnel through a weighbridge or similar reading device (see Results of Previous NSF Support; Clarke & Kerry 1998), but small colonies are not representative of large colonies (Ainley et al. 2004) and no technology has yet been invented that allows fast, efficient scanning of large numbers of birds with minimal disturbance. Measuring dispersal among colonies is an important aspect of our work that would be very difficult and expensive to do without bands. In addition, very few studies have truly quantified effects of flipper bands on penguin survival or reproductive success; at best, most have indexed survival based on band return rates (Trivelpiece & Trivelpiece 1994, Froget et al. 1998, Clarke & Kerry 1998, Gauthier-Clerc et al. 2004) which, without detection probabilities, can be negatively biased (but see Ainley et al. 1983, Ainley 2002).
Fig. 11. Preliminary results comparing resighting rates of flipper-banded and PIT-tagged adults. During the three years when sufficient data were available, PIT-tagged-only individuals exhibited slightly higher rates, but ones that lay below the highest values exhibited by banded birds in some years. In the years of comparison, breeding propensity and philopatry were severely affected by the iceberg experiment (Figs. 7, 8).
Our work employing advanced mark-recapture methodologies, in conjunction with
PIT tags, will be the first to model band effects in regard to sighting probability,
sex, time or environmental variation (Fig. 11). Using other comparisons between
banded and unbanded pit-tagged birds in the weighbridge subcolonies, we have
also begun to measure effects of bands on foraging trip duration (detectable,
but with significant annual variation) and meal sizes (no detectable effect).
We are aware of ethical concerns related to experiments that may increase
mortality of study animals, and we believe a scientific understanding of what
factors actually contribute to mortality and other potential impacts of bands
is necessary for the correct interpretation of existing data as well as for
the continued improvement of marking techniques. Therefore, in addition to
continued effort to apply rapidly changing transponder technology (e.g., Beigel
et al. 2004), it is important to fully understand band effects on penguins,
especially regarding any variation associated with gender and resource changes.
All that we currently know about penguin demography is based on studies of
flipper-banded penguins, studies that are ongoing on various species in Australia,
New Zealand and South Africa, as well as the Antarctic. In a separate proposal,
in conjunction with the engineers who originally perfected RFID technology
(and who designed our weighbridges), we will work to apply new transponder
technology with the intent to eliminate the need for further banding of penguins.
Wintering area and survival. In the 2002-03 and 2003-04 seasons, for the first time ever, we deployed micro-GLS tags, with results providing, also for the first time ever, a complete picture of the winter journeys of a migratory penguin (see Results of Prior NSF Support; Fig. 7). Some of these penguins traveled > 7,000 km between breeding seasons, which surprised us. We can now test an hypothesis generated in one of our retrospective analyses (Wilson et al. 2001): Adélie Penguins when wintering north of the southern boundary of the Antarctic Circumpolar Current (ACC), as a result of northward sea-ice extent, exhibit lower survival than when wintering to the south (years of lesser ice extent). Waters to the south of the ACC are more productive than those of the ACC itself (Tynan 1998, Nicol et al. 2000). Our preliminary results are very encouraging (Fig. 12).
Fig. 12. Resighting rates of male Adélie Penguins from Ross Island regressed against sea-ice extent during July. Extent equals the amount of ice in the Ross Sea sector of the Southern Ocean x 106 (data from NOAA).
Obviously, using the GLS tags alone, in conjunction with our demographic studies, we
will learn a great deal about effects of sea ice extent on this population
of Adélie Penguins. However, to learn even more, we have submitted a proposal
to NASA, responding to announcement NRA-04-OES-02
(Oceans and Ice), with colleagues Kevin Arrigo (Stanford University) and Claire Parkinson (NASA Goddard Space Flight
Center). If funded, we would be able to relate penguin demographics and
winter journeys to characteristics of sea ice and ocean productivity using
satellite remote sensing. Analysis would be coincident with our field work.
The means by which the movement and extent of ice may lead to the location
where Adélie Penguin wintering potentially may have central influence on
how this species responds to climate change (Fig 13).
In addition, we hope to deploy a remote camera at Cape Royds to document the time that chicks go to sea, and the number of adults who return to molt. We depart Royds before this time owing to the end of helicopter support in the US program. In 1959-60, R. Taylor (1962) noted that a large percentage of adults molted at Royds, contrary to what happens at capes Crozier and Bird. If adults depart Royds several weeks later than they do at the latter colonies, they will encounter pack ice much farther west in the Ross Sea on their journey north, as the autumn-winter ice field grows from east to west. In turn, this means that Royds birds could winter in the area of Penguin B (in Fig 7), with most Crozier and Bird adults wintering to the east (Penguin A in Fig. 7).
Fig. 13. Ice drift derived from AMSR-E passive-microwave data using wavelet analysis (courtesy Y. Zhao, T. Liu & J. Comiso). Sea ice shown in blue; Antarctic continent (to the south) and ice-free ocean (to the north) both shown in black. Compare ice trajectories to the disparate sets of penguin tracks in Fig. 6. Where the penguin undertook its post-breeding molt (on an ice floe) may have determined where it wintered.
competition within the Ross/Beaufort Island metapopulation.
Investigators have described nest-site prospecting by subadult kittiwakes
Rissa tridactyla late in the chick-rearing period at various colonies
and subsequent recruitment the following year (e.g., Porter 1990, Coulson
& Neve de Mevergnies 1992, Boulinier et al. 1996). Danchin et al. (1998),
with supporting evidence from Suryan & Irons (2001), hypothesized that
subsequent colony of recruitment depended on the perception of success (number
of near-to-fledging chicks) among conspecifics at a given colony, called performance-based
attraction. In our study before the icebergs, large numbers of immigrant Adélie
Penguins were showing up at Cape Royds but we found no evidence of higher
reproductive success there (Ainley et al. 2004). However, we did find evidence
for trophic interference competition, the intensity of which was proportional
to colony size (least at tiny Royds). It is possible, therefore, that this
negative aspect of large colonies had contributed greatly to the observed
relaxation in philopatry.
To test this hypothesis, we propose to use a central-place foraging model similar to that developed for the Black-legged Kittiwake Rissa tridactyla of Prince Wm Sound (Ainley et al. 2003b, Ford et al. ms), but one specific to the Ross/Beaufort Island Adélie Penguin metapopulation. The model uses cellular automata coupled with a spatial model of food availability (Aassine & El Jai 2002) to simulate animals moving through a grid-based environment, searching and feeding until returning to colonies. The automata behave optimally, maximizing food uptake rate and minimizing trip time (as noted in Ainley et al. 1998), but their ability to optimize is limited by imperfect knowledge of food availability. By combining the activities of many automata from different colonies, a picture of the foraging areas of interacting colonies can be assembled in the manner suggested by Wootton (2001). By dynamically coupling food availability to the number of birds feeding in a given cell, the model can simulate prey depletion resulting from individuals from multiple colonies competing for food (as per Ainley et al. 2004).
The model requires estimates of relative food availability in each grid cell in the model universe. Estimating this will require a model for assessing fish and krill density and predictability. Available to us are the results of acoustic surveys of Euphausia crystallorophias, a major penguin prey (Ainley et al. 2003a), conducted by Assali & Kalinowski (2000) and Ackley et al. (2003) in the Ross Sea. We will test alternative methodologies to find the best model and predictor variables for prey abundance. Candidate predictors include ice concentration, depth, time of year, and chlorophyll (Ainley et al. 1998; Arrigo et al. 2002, 2003; Ballard et al. 2001, ms). We will also test the method described by Marin & Delgado (2001) developed for the Shetland Island area as a means of refining these estimates.
In addition to cell-based estimates of food density, model parameters not easily measured directly are: 1) uncertainty regarding food availability, and 2) the number of cells that a foraging bird looks into the future (i.e. its prior knowledge). However, dive bouts from TDRs and inter-bout movements from satellite tags will offer clues (Fig. 14). We will estimate these parameters using a fitting procedure that finds combinations, which minimize the difference between observed and predicted foraging distributions. Our work with kittiwakes showed that a consistent set of model parameters was capable of accounting for 59-66% of variation in the spatial distribution of foraging, radio-tracked kittiwakes.
Once we have calibrated the model for the Adélie Penguin population, we will use it to examine several questions that relate to the general hypothesis (see above):
1) Does the location of colonies minimize the amount of overlap among foraging penguins? We will generate multiple scenarios in which colonies are placed in randomized locations around the study area. For each scenario, the model will determine the foraging areas of each colony and calculate the corresponding degree of overlap. The null hypothesis is that the amount of overlap among colonies sited in their actual locations does not differ from the overlap expected if the colonies were located randomly.
2) Is the pattern of decreasing mass gain in trips >48 h evidence of non-optimal foraging? The observed negative relationship between mass gain and trip length for trips exceeding 48 h (Ainley et al. 1998, Watanuki et al. 2002) could occur for several reasons. Food availability (or luck) may be low on some foraging trips, and increased trip time results from poor foraging success. Alternatively, the relationship could result from the need to balance adult body mass with food delivered to the chicks (Ballard et al. ms). Or, it is possible that the birds are simply behaving sub-optimally. We will use the model to determine whether any of these factors can account for the observed relationship.
3) Is the observed shift in foraging activity away from the colonies as the season progresses consistent with local prey depletion? We will configure the model to adjust prey levels so that prey availability decreases proportional to the number of birds foraging in a given cell. The null hypothesis would be that seasonal changes in food availability will result in the observed shift of foraging activity away from the larger colonies without invoking prey depletion.
Fig. 14. Foraging areas determined by satellite tags in the 2003 season. Ross Sea is shown with shallower banks in greener colors. Similar to patterns before the iceberg arrivals, the foraging of parents from Royds, Bird and Beaufort overlapped extensively (see Fig. 6); in contrast, the icebergs barred foraging by Crozier parents in the vicinity of Beaufort and Bird. By the end of the season, Crozier birds were making trips of 5 days duration to the east.
Intellectual Merit. No other program in Antarctica is studying the demography
of the Adélie Penguin, deemed by the CCAMLR Ecosystem Monitoring group to
be a so-called indicator species of environmental and fishery conditions.
In that aspect alone, our project has been making important contributions
toward understanding the polar marine ecosystem. Moreover, the iceberg natural
experiment is an unparalleled opportunity to study the demographics of a polar
(or any) seabird and to respond to the challenges identified by Lebreton et
al. (2003) who, in setting the standard for metapopulation studies of birds,
have called for a far better grasp of emigration rates. The latter requires
a very labor-intensive effort, as we know full well, and for that reason few
other tagging studies effectively gather such information. This will be the
first study of Adélie Penguin demography using state-of-the art analytical
procedures with mark-recapture data. Moreover, thanks to electronic wonders,
we are supplying an ecological context to the demographic patterns observed,
which also is rare in this field of research.
As far as we know, no other field research program is working to develop marking technologies to replace current penguin bands. This effort would facilitate the study of penguin demographics throughout the Southern Hemisphere.
Broader Impacts. Many researchers now study the effects of global climate change on coastal and polar ecosystems. Given climate projections for the coming century, this is not surprising, and we have been contributing directly and indirectly to this effort (cf. Croxall et al. 2002, Ainley et al. 2003). Important to this endeavor is the close collaboration we enjoy with New Zealand scientists, with plans to expand collaboration to include Italian scientists at Terra Nova Bay, immediately to the north.
In regard to education, and looking to the future, we have involved and will be involving the public and persons who would not normally participate in this type of scientific exploration: one high school teacher, 6 masters-level assistants (3 have gone on for PhDs), and 4 pre-masters interns (2 have since enrolled in graduate programs). We have constructed a popular website to highlight our results, as well as contributing to the popular literature (see Results of Prior NSF Support). Finally, we plan to insert a camera at Cape Royds to beam penguin nesting activities to the support personnel of McMurdo Station and to the browsers of the internet web.
Components of the project will be placed under the management of
the principal collaborators as follows: David Ainley is overall PI, with close help from Grant Ballard.
Ballard is the data manager and with Katie Dugger handles data analysis. Dugger
is a demographer. Glenn Ford, a modeler will participate for first three years.
Peter Wilson and Kerry Barton head the NZ component (Cape Bird colony)
and Silvia Olmastroni the Italian component (Edmonson Point colony, Terra Nova Bay). The NZ
component is responsible for aerial photography, and at times this will include
the Terra Nova Bay colonies as well.
Each year the US and NZ components have been meeting at McMurdo/Scott Base, and we would like to visit the Italian base as well beginning with this phase of the project. We usually meet in NZ as well after the season to discuss results and make plans. Otherwise, Ballard has relatives in Italy, and has met (and will continue to meet) with Olmastroni there.
We envision at least one major publication coming from the modeling
to be conducted by Ford, perhaps to be published in Ecology or Ecological
Monographs. We should be well on the way to publishing a paper on the
foraging effort of penguins, using WB, TDR and satellite data, perhaps to
be published in Journal of Animal Ecology. During the present grant,
we have computerized all the demographic data from the 1960-70s study at Cape Crozier and plan
to compare patterns evident then with those in the current population using
modern statistical analyses. This might be another paper destined for Ecology.
We are working toward a metapopulation model that would include all of the
colonies in the southwestern Ross Sea, including
perhaps those at Terra Nova Bay; how the
big icebergs behave may determine the timing for our analyses. During the
first year of the proposed project, we should be well along in the process
of publishing the four papers we intend to be presenting at the International
Penguin Symposium in September 2004 (see Results of Prior NSF Support). There
will be several other papers that result from our effort during the 5 years
of its span.
We plan to be upgrading the weighbridges and anticipate publishing a paper, with the engineers responsible for the weighbridge, describing this technology and how we achieved success with its application in our project.
Eventually our data, including those from the 1960-70s, will be made available on our website for pre-approved use by others.
Ainley DG. 2002. The Adelie Penguin: Bellwether of Climate Change. NY:
Columbia Univ. Press. In press.
Ainley DG & RE LeResche. 1973. The effect of weather and ice conditions on breeding in Adélie Penguins. Condor 75:235-239.
Ainley DG, RE LeResche & WJL Sladen. 1983. Breeding Biology of the Adelie Penguin. Univ. Calif. Press, Los Angeles. 240 pp.
Ainley DG, EF O'Connor & RJ Boekelheide. 1984. Ecology of Seabirds in the Ross Sea, Antarctica. A.O.U. Monograph No. 32. 79 pp.
Ainley DG, CA Ribic & WR Fraser. 1994. Ecological structure among migrant and resident seabirds of the Scotia-Weddell Confluence region. J. Anim. Ecol. 63:347-364.
Ainley DG, PR Wilson, KR Barton, G Ballard, N Nur & BJ Karl. 1998. Diet and foraging effort of Adélie penguins in relation to pack-ice conditions in the southern Ross Sea. Polar Biology 20: 311-319.
Ainley DG, RC Wood & WJL Sladen. 1978. Bird life at Cape Crozier, Ross Island. Wilson Bull. 90:492-510.
Ballard G, DG Ainley, CA Ribic & KR Barton. 2001. Effect of instrument attachment and other factors on foraging trip duration and nesting success of Adélie penguins. Condor. In press.
Barbraud C & H Weimirskirch. 2001. Emperor penguins affected by Southern Ocean anomalies. Nature, in press.
Blackburn N, RH Taylor & PR Wilson. 1991. An interpretation of the growth of the Adélie penguin rookery at Cape Royds, 1955-1990. NZ J Ecol 15: 117–121.
Brownie C. JD Nichols, KH Pollock & JB Hestbeck. 1993. Capture-recapture studies for multiple strata including non-Markovian transitions. Biometrics 49:1173-1187.
Cooch E, R Pradel & N Nur. 1996. A Practical Guide to Capture/Recapture Analysis using SURGE. Centre d’Ecologie Fonctionelle et Evolitive - CNRS, Montpellier, France. 125 pp.
Clarke JR & KR Kerry. 1994. The effects of monitoring procedures on Adélie Penguins. CCAMLR Science 1:155-164.
Crawford RJM, LJ Shannon & PA Whittington. 1999. Population dynamcis of the African Penguin Spheniscus demersus at Robben Island, South Africa. Mar. Ornithol. 27: 139-147.
Dann P & JM Cullen. 1990. Survival, patterns of reproduction, and lifetime reproductive output in Little Blue Penguins (Eudyptula minor) on Phillip Island, Victoria, Australia. In LS Davis & JT Darby, eds., Penguin Biology, pp. 63-84. Orlando: Academic Press.
Darby JT & PJ Seddon. 1990. Breeding biology of Yellow-eyed Penguins (Megadyptes antipodes). In LS Davis & JT Darby, eds., Penguin Biology, pp. 45-62. Orlando: Academic Press.
Falla RA.1937. Birds. B. A. N. Z. Antarctic Res. Exped. 1929-1931. Repts., Series B, vol. II. Adelaide: B. A. N. Z. A. R. Expedition Committee Adelaide: Hassell Press.
Fraser WR, WZ Trivelpiece, DG Ainley & SG Trivelpiece.1992. Increases in Antarctic penguin populations: Reduced competition with whales or a loss of sea ice due to environmental warming? Polar Biol. 11:525-531.
Gloersen PW, J Campbell, DJ Cavalieri, JC Comiso, CL Parkinson & HJ Zwally. 1992. Arctic and Antarctic sea ice, 1978-1987: Satellite Passive-Microwave Observations and Analysis. NASA SP-511. Washington, DC: U.S. Natl. Aeron. Space Admin.
Lebreton J-D, KP Burnham, J Clobert & DR Anderson. 1992. Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecol. Monogr. 62:67-118.
Levick GM. 1915. Natural history of the Adélie Penguin. British Antarc. (“Terra Nova”) Exped., 1910-1913, Zool. 1:55-84.
Nicol S, T Pauly, NL Bindoff, S Wright, D Thiele, GW Hosle, PG Strutton & E Woehler. 2000. Ocean circulation off east Antarctica affects ecosystem structure and sea-ice extent. Nature 406: 504-507.
Pollock KH, JD Nichols, C Brownie, and JE Hines. 1990, Statistical inference for capture-recapture experiments. Wildl. Mongr. 107:1-97.
Richdale LE. 1957. A population study of penguins. Oxford: Oxford Univ. Press, Oxford UK.
Roeder AD, RK Marshall, AJ Mitchelson, T Visagathlagar, PA Ritchie, DR Love, TJ Pakai, HC McPartlan, ND Murray, NA Robinson, KR Kerry & DM Lambert. 2001. Extraordinary genetic homogeneity among Adélie penguin colonies around Antarctica. Molecular Genetics: in press.
SCAR. 1993. Report of Workshop on Researcher-Seabird Interactions (W. Fraser, compiler). Monticello, NN.
Shannon LJ & RJM Crawford. 1999. Management of the African Penguin Spheniscus demersus — Insights from modeling. Mar. Ornithol. 27: 119-128.
Smith RC, D Ainley, K Baker, E Domack, S Emslie, B Fraser, J Kennett, S Leventer, E Mosley-Thompson, S Stammerjohn & M Vernet. 1999. Marine ecosystem sensitivity to climate change. BioScience 49(5): 393-404.
Sokal RR & FJ Rohlf. 1995. Biometry: Principles and Practice of Statistics in Biological Research, 3rd ed., WH Freeman, San Francisco.
Spurr EB. 1975. Breeding of the Adélie Penguin Pygoscelis adeliae at Cape Bird. Ibis 117:324-338.
Stammerjohn SE & RC Smith .1997. Opposing Southern Ocean climate patterns as revealed by trends in regional sea ice coverage. Climate Change 37: 617-639.
Taylor RH. 1962. The Adélie Penguin Pygoscelis adeliae at Cape Royds. Ibis 104:176-204.
Taylor RH & PR Wilson. 1990. Recent increase and southern expansion of Adélie Penguin populations in the Ross Sea, Antarctica, related to climatic warming. NZ J Ecol 14: 25-29.
Taylor RH, PR Wilson & BW Thomas. 1990. Status and trends of Adélie penguin populations in the Ross Sea region. Polar Rec 26 (159): 293-304.
Tynan CT. 1998. Ecological importance of the southern boundary of the Antarctic Circumpolar Current. Nature 392: 708-710.
White GC & KP Burnham. 1999. Program Mark: survival estimation from populations of marked animals. Bird Study 46 Supple:120-138.
Wilson PR, DG Ainley, N Nur, SS Jacobs, KJ Barton, G Ballard & JC Comiso. 2001. Adélie Penguin Population Change in the Pacific Sector of Antarctica: Relation to Sea-Ice Extent and the Antarctic Circumpolar Current. Mar. Ecol. Progr. Ser. 213: 301-309.
Wilson RP & G Peters. 1999. Foraging behavior of the Chinstrap Penguin Pygoscelis antarctica at Ardley Island, Antarctica. Mar. Ornithol. 27:85-95.
Wilson RP, K Peutz, CA Bost, BM Culik, R Bannasch, T Reins & D Adelung. 1993. Diel dive depth in penguins in relation to diel vertical migration of prey: Whose dinner by candlelight? Mar. Ecol. Progr. Ser. 94:101-104.
Yeates GW. 1968. Studies on the Adélie Penguin at Cape Royds 1964-65 and 1965-66. N.Z. J. Mar. Freshwater Res. 2:472-496.