Variability of cetacean distribution and habitat selection in the Alaskan Arctic, autumn 1982-91.

From: Arctic | Date: December 1, 2000| Author: | Copyright information

(Received 25 November 1999; accepted in revised form 5 July 2000) ABSTRACT. Ten years (1982-91) of autumn sighting data from aerial surveys offshore northern Alaska were analyzed to investigate variability in cetacean distribution and habitat selection. Habitat selection indices were calculated for bowhead, white, and gray whales in heavy, moderate, and light ice conditions; and for high, moderate, and low transport (inflow) conditions at Bering Strait. Bowhead whales selected shallow inner-shelf waters during moderate and light ice, and deeper slope habitat in heavy ice ... <0.05-0.001). White whales selected slope habitat ([chi square], p 0.001), and gray whales selected coastal/shoal and shelf/trough habitat ([chi square], p 0.025-0.001), in all ice conditions. In the Alaskan Beaufort Sea, bowheads selected shelf waters and white whales chose slope waters, without regard to transport conditions ([chi square], p 0.01-0.001). In the northern Chukchi Sea, gray whales selected coastal/shoal habitat in high transport conditions ([chi square], p 0.005), and shelf/trough habitat ([chi square], p 0.001) during moderate and low transport conditions. Variability in distribution and habitat selection among these species is likely linked to prey availability at dissimilar trophic levels, although this hypothesis has yet to be rigorously tested. Key words: Alaska, Arctic, Beaufort Sea, bowhead whale, Chukchi Sea, gray whale, habitat selection, white whale RESUME. Des donnees d'observation realisees en automne sur dix annees (1982-1991) grace a des releves aeriens au large de l'Alaska septenerional ont ete analysees dans le cadre de recherches sur la variabilite dans la distribution des cetaces et la selection de leur habitat. On a calcule les indices de selection de l'habitat pour la baleine boreale, la baleine blanche et la baleine grise de Californie dans des conditions de glace epaisse, moderee et mince; et pour des conditions de transport (courants de deversement) important, moyen et faible dans le detroit de Bering. La baleine boreale choisissait des eaux peu profondes de l'interieur du plateau continental durant les conditions de glace moderee et mince, et un habitat plus profond sur la pente durant des conditions de glace epaisse ([chi square], p 0,05- 0,001). La baleine blanche choisissait l'habitat sur la pente ([chi square], p 0,001) et la baleine grise choisissait l'habitat cotier/de hauts-fonds et celui du plateau/des fosses ([chi square], p 0,025-0,001), quelles que soient les conditions de glace. Dans la partie alaskienne de la mer de Beaufort, la baleine boreale choisissait les eaux du plateau et la baleine blanche celles de la pente, abstraction faite des conditions de transport ([chi square], p 0,01-0,001). Dans la partie septentrionale de la mer des Tchouktches, la baleine grise choisissait un habitat cotier/de hauts-fonds dans des conditions de transport important ([chi square], p 0,005) et un habitat de plateau/de fosses ([chi square], p 0,001) dans des conditions de transport allant de moyen a faible. La variabilite dans la distribution et la selection de l'habitat parmi ces especes est probablement liee a la disponibilite des proies a des niveaux trophiques dissemblables, bien que cette hypothese doive encore faire l'objet de tests approfondis. Mots cles: Alaska, Arctique, mer de Beaufort, baleine boreale, mer des Tchouktches, baleine grise de Californie, selection de l'habitat, baleine blanche Traduit pour la revue Arctic par Nesida Loyer. INTRODUCTION The Arctic Ocean has experienced a marked warming trend over the last century (Overpeck et al., 1997), with consistent, but not uniform, decreases in ice extent described for the last three decades (Maslanik et al., 1996; Martin et al., 1997). Marked interannual variation in salinity has also been reported (e.g., Bjornsson ee al., 1995; Roach et al., 1995), although without a clear trend toward, or against, freshening of Arctic waters. The effects of climate change are predicted to be amplified in the Arctic because of positive feedback mechanisms associated with deterioration of snow-ice albedo (Overpeck et al., 1997; Aagaard et al., 1999). Given this likely amplification, it would be useful to identify Arctic species that could serve as indicators of environmental change. Marine mammals are considered good candidates for this role because they are positioned as apex predators (Ainley and DeMaster 1990); the response of some species to changes in sea ice availability can be observed (DeMaster and Davis, 1995; Stirling, et al., 1999); and their importance as a subsistence resource encourages monitoring of population dynamics (e.g., Zeh et al., 1993). Specifically, bowhead whales (Balaena mysticetus), ringed seals (Phoca hispida), and white whales (Delphinapterus leucas) have been suggested as indicator species (Tynan and DeMaster, 1997). To this list I would add gray whales (Eschrichtius robustus), given their unique role as perturbators of the benthos in Arctic and Subarctic waters (Oliver and Slattery, 1985; Feder et al., 1994). Bowhead whales, white whales, and gray whales occupy dissimilar habitats offshore northern Alaska, which Moore and DeMaster (1997) provisionally described by average differences in bathymetry and ice cover. As a refinement to simply comparing differences in average measures of habitat features, Moore et al. (2000) calculated habitat selection ratios for the three species in summer and autumn. While descriptive in nature, the indices of habitat selection are based on quantification of survey effort and represent a key first step towards estimating resource selection by animals (Manly et al., 1993). These indices indicate that, in summer, bowhead whales selected continental slope waters and moderate ice conditions; white whales selected slope and basin waters and moderate to heavy ice conditions; and gray whales selected coastal/ shoal waters that were usually ice-free. In autumn, bowheads shifted to shallow inner-shelf waters and light to ice-free conditions; white whales selected outer shelf and slope waters in moderate to heavy ice cover; and gray whales selected coastal/shoal and trough habitats in light to ice-free conditions. Habitat differences among species were significant in both summer and autumn (ANOVA F><0.0001). This habitat partitioning is likely associated with distinct differences in preferred prey and foraging strategies among the three species (see Moore et al., 2000), although this hypothesis has not been tested. The calculation of autumn habitat selection ratios in Moore et al. (2000) did not address variability in cetacean distribution in relation to changes in environmental conditions. However, for whales to be useful as indicators of environmental change in the Arctic, as suggested in Tynan and DeMaster (1997), some baseline for modeling this variability is required. As a first step toward this type of modeling, this paper describes variable distribution and habitat selection among bowhead, white, and gray whales calculated for different conditions of sea ice cover and inflow (hereafter, transport) at Bering Strait. As in Moore et al. (2000), cetacean distribution and habitat selection indices were derived from ten years (1982-91) of sighting data from aerial surveys conducted offshore northern Alaska and funded by the Minerals Management Service (MMS) of the U.S. Department of the Interior. Oceanographic Variability An overview of physical oceanography offshore northern Alaska is provided in Moore and DeMaster (1997) and Moore et al. (2000). Ice cover and transport (i.e., in-flow) at Bering Strait are two aspects of Alaskan Arctic physical oceanography that may affect cetacean distribution and habitat selection. Sea ice affects productivity in the Arctic (Smith and Nelson, 1985; Smith and Sakshaug, 1990; Smith and Schnack-Schiel, 1990) and can also act as a physical barrier to migrating animals. Transport at Bering Strait provides an advective pathway for nutrients and zooplankton between the northern Bering, Chukchi, and western Beaufort Seas (Grebmeier and Barry, 1991; Niebauer and Schell, 1993) and, via the Beaufort Undercurrent, possibly as far as the eastern Alaskan Beaufort Sea (Aagaard, 1984; Fissel et al., 1987). Because cetaceans are apex predators in the short food webs common to the Arctic (Ainley and DeMaster, 1990), physical factors that influence productivity and prey availability will likely influence habitat selection. Alternatively, bathymetrically channeled currents may provide migratory cues to animals. Over the course of the study, sea ice cover varied from years when the Alaskan Beaufort and northern Chukchi Seas remained ice covered to years when the ice retreated nearly 170 km offshore. If ice affects cetacean distribution, either as an affiliate of prey or as an impediment to migration, differences in distribution among years ranked as heavy, moderate, and light in ice cover should be apparent. Transport at Bering Strait also exhibited strong interannual variation during the study period, with mean transport ranging from 0.61 to 0.96 Sv (Roach et al., 1995). If transport affects cetacean distribution, either by influencing foraging opportunities via advection of prey or enhancement of productivity, or by providing migration cues via distinctive water mass qualities, differences in distribution among years ranked as high, moderate, and low in transport should be identifiable. Because both ice cover and transport are wind forced (Niebauer, 1988; Roach et al., 1995), they are not independent variables. However, when study years were ranked into heavy/high vs. moderate vs. light/low categories for analyses, correspondence between data sets was incomplete, so distribution and habitat selection were examined separately for years pooled by (1) ranks of sea ice cover and (2) transport at Bering Strait. METHODS The study area, survey protocol, and statistical approach are described in Moore et al. (2000). Specifically, habitat selection was tested by species and oceanographic parameter via chi-square analysis and the calculation of habitat selection ratios (Manly et al., 1993). One note of caution is needed: the chi-square test generally requires that expected frequencies be five or more, and this was rarely the case for the northern Chukchi Sea analyses (see Tables 4 and 6). Manly et al. (1993) suggest that the test may still be valid when the condition is not met, but results must be treated with some reservation. Because much of the Chukchi Sea falls into the 36-50 m depth regime, most of the survey effort and sightings occur there. However, the Chukchi shelf is not a featureless plain; in fact, its topography plays a major role in channeling currents. Thus, the chi-square analysis is included here for completeness, to investigate cetacean use of shoal and trough features, although habitat selection results should indeed be viewed cautiously. Heavy 0 0 0 0.10 2.500 0.40 0.30 0.405 0.06 0.60 3.333 0.54 Total 1.0 6.238 1.0 Moderate 0 0 0 0.35 5.000 0.72 0.14 0.226 0.03 0.51 1.700 0.25 Total 1.0 6.926 1.0 Light 0 0 0 0.08 2.000 0.41 0.40 0.571 0.12 0.52 2.261 0.47 TABLE 4. Variability of bowhead whale (BH), white whale (WW), and gray whale (GW) transect sightings (t-SI) with ice condition rank, by depth regime, in the northern Chukchi Sea. Ice conditions ranked from navy/NOAA (1992). [Part 1 of 3] Ice Condition [years] Depth (m) Effort (t-km) Heavy><= 35 4130 Total 22767 Moderate><= 35 5412 Total 17822 Light>

Read all of this article with a FREE trial to HighBeam Research

Related newspaper, magazine, and trade journal articles from HighBeam Research

(Including press releases, facts, information, and biographies)

Test of a measurement model for investigating the youth physical activity promotion model. (Measurement).(Brief Article)
; ...12 iterations, and overall model fit statistics ranged from excellent (PSW [[chi square].sup.2.sub.(9)] = 7.15; LVE [[chi square].sup.2.sub.(9)] = 4.93; PA[[chi square].sup.2.sub.(5)] = 7.78) to moderate (PPC[[chi square].sup.2.sub.(14)] = 31...
Four electronic tables for probability distributions.
; ...distribution. The case of the noncentral [chi square] distribution serves as an example. Suppose...percentage points x of the noncentral [chi square] distribution with given degrees of freedom...the random variable X has a noncentral [chi square] distribution with parameters v and...
Survival Characteristics of Corn Seed during Storage: I. Normal Distribution of Seed Survival.(Statistical Data Included)
; ...there are several reports in which [Chi square] tests for normality indicated that seed-surviv...germination always resulted in significant [Chi square]. They suggested that adjustment for...responsible for a significantly large [Chi square] (Finney, 1971). First, individual seeds...
A causal modelling test of the relationship between chief executive officer experience and corporate strategy.
; ...for analysing categorical data has been chi square analysis, but this approach has several...Loglinear models provide information that chi square test and non-parametric tests are not...the variables, if necessary. Thus the chi square approach is not capable of handling the...
Cetacean habitat selection in the Alaskan Arctic during summer and autumn.
; ...selection de l'habitat des cetaces. On a decrit des habitats distincts pour la baleine boreale (Balaena mysticetus), la baleine blanche (Delphinapterus leucas) et la baleine grise de Californie (Eschrichtius robustus) en se fondant sur les taux...
The radiologic work-up in thyroid surgery: Fine-needle biopsy versus scintigraphy and ultrasound.(Brief Article)
; ...analyzed according to the chi-squared ([chi square]) test. Of 189 thyroid scintigraphy scans...specificity = 19%; accuracy = 38%; [chi square] = 7.67; p = 0.006). Of 66 ultrasounds...specificity = 14%; accuracy = 32%; [chi square] = 2.47; p = 0.116 [not statistically...
Assessment of effects of pre- and post-training programme for healthcare professionals about breastfeeding.(SHORT REPORT)(Report)
; ...in the before and the after group ([chi square]=29.31, p<0.001). 16.7% and 36.5% gave exclusive breastfeeding for six months ([chi square]=72.85, p<0.001), and 28.5% and 23...breastfeeding within the first five months ([chi square]=17.20, p=0.002). Ninety-four percent...
4: f-Divergence and contingency tables.(Information Theory and Statistics: A Tutorial)
; ...also often used in statistics. They are called [chi square]-divergence, Hellinger distance, and variational...f-divergences is the following approximation by the [chi square]-divergence [chi square][(P,Q) = [summation](P(a) - Q(a)).sup.2]/Q(a...
Daylight models--Singapore sky.
; ...gradation function for the zenith f ([chi square] the indicatrix function L sky patch...cd m-2) LZ Zenith luminance(cd m-2) [chi square] scattering angle(rad) a,b,c,d,e constants...luminance (LZ) as given below: (1) L/LZ =f ([chi square]) [phi] (Z)/f (Zs) [phi] (0[degress...
The relationship between amphetamine use, crime and psychiatric disorder among prisoners in New South Wales.
; ...analysis was carried out in the form of chi-square frequency tables and logistic regression...to be for robbery or property crimes ([chi square = 22. 185, p = .014). Amphetamine users...admissions to psychiatric hospitals ([chi square] = 13.740, p < .001) than non-users...