Population decline in the Delta caribou herd with reference to other Alaskan herds

After growing continuously for nearly 15 years, the Delta caribou herd began to decline i n 1989. M o s t other Interior Alaskan herds also began declining. In the Delta herd, and i n other herds, the declines were caused primarily by high summer mortality o f calves and increased natural mortality of adult females. Other minor causes included increased winter mortality of calves, and reduced parturition rates o f 3-year-old and older females. T h e decline i n the Delta herd also coincided w i t h increased w o l f (Canis lupus) numbers, winters w i t h deeper than normal snow, and w a r m summers. M e a n body weight of annual samples of 10-month-o ld female calves was consistently l o w during the decline. Except i n some of the smallest Interior Alaskan herds, we conclude that evidence for population regulation i n Alaskan caribou is weak, and that herds are l ikely to fluctuate w i t h i n a wide range o f densities due to complex interactions o f predation and weather. Unless w o l f numbers are influenced by man, the size of a caribou herd i n a given year is l ikely to be largely a function of its size during the previous population l o w and the number of years of favorable weather i n the interim.


Introduction
Caribou herds in Alaska and elsewhere have fluctuated in size over time, and the factors involved in these fluctuations have been widely debated (Leopold & Darling, 1953;Skoog, 1968;Van Ballenberghe, 1985;Messier et ah, 1988;Bergerud & Ballard, 1989;Seip, 1991;Eberhardt & Pitcher, 1992;Bergerud, 1993).Several caribou herds in Alaska were intensively studied during late 1970s and 1980s when herds were generally increasing (Davis et ah, 1991;Cameron et ah, 1993;Adams et ah, 1994;Whitten, 1994).This paper reports results of a continuing study of limiting and regulating factors in the Delta caribou herd, during the period of population decline from 1979 to 1993 and compares more limited data from other Alaskan herds.

Study area and population
The Delta caribou herd is one of 31 herds composing a total population of about 880,000 caribou in Alaska (Table 1, ADF&G files).About 750,000 of these caribou occur in the 3 largest herds: Western Arctic (29), Porcupine (22),and Mulchatna (19) Rangifer, Special Issue No. 9, 53-62 (Numbers correspond to herd numbers in Fig. 1 and Table 1).Most of the other herds occur in the mountainous areas of Interior Alaska and range in size from a few hundred to about 40,000 caribou.
The Delta herd (9) occupies an area of about  Initial studies were to determine the causes of low calf production and/or survival prevalent in the herd from 1971 to 1974.However, after a wolf control program primarily to benefit moose (Alces alas), the Delta herd increased rapidly (Gasaway et al., 1983), and data collected through 1989 was representative of a growing population.
From 1979 to 1989 the Delta herd grew continuously from 4,191 to 10,690 (Davis et al, 1991).
From 1979 to 1982 the herd grew rapidly (k = 1.20), because harvest was light, adult female mortality was low, and natality and calf survival were high (Davis & Valkenburg, 1985).

Methods
We annually estimated population size, recruitment of calves to autumn, and age-specific natality rates of females in the Delta herd.In most years we also collected data on weights of 10-month-old female calves, and mortality rates and causes of death of radiocollared females older than 10 months.Starting in 1991 we also began weighing and collaring 4month-old females.Movements and distribution of radiocollared caribou in the Delta herd and surrounding herds were monitored to detect immigration or emigration.
Population censuses (total counts) were conducted during mid June to mid July each year and followed techniques described by Davis et al. (1979) and Valkenburg et al (1985).We estimated calf recruitment to September/October and April with helicopter surveys.Allocation of sampling effort was based on the distribution of radiocollared females.
Natality rates of radiocollared females were estimated by looking for distended udders, hard antlers, or calves at heel from a Piper Super Cub or Bellanca Scout aircraft during the calving period (15 May 1 Jun) (Bergerud, 1964;Davis et al, 1991).Weights Rangifer, Special Issue No. 9, 1996 These weather variables were plotted and compared with data on caribou parturition (natality) rates and September/October calfxow ratios.

Results and discussion
Immediate causes of the decline in the Delta herd The proximate or immediate causes of decline of the Delta herd from 1989 to 1993 are clear.In order of importance they were: 1) high natural mortality of calves from birth to late September during 1990-1993, 2) high natural mortality of females older than calves primarily from wolf preda¬ tion, 3) high mortality of radiocollared calves from September/October to April during 1991-1993, and 4) relatively low natality rates of adult females during 1990, 1991, and 1993 (Table 2).

Density-dependent resource limitation
Evidence for density-dependent resource limitation in the Delta and other Interior Alaskan herds was ambiguous.Although there was a weak relationship between density and population growth rate between 1989 and 1993 (Fig. 2).Some low density herds (e.g., Denali and Mentasta) declined, while others that had as high or even higher densities than the Delta herd did not decline (Table 3).However, the greatest decline occurred in the Delta herd which also had the highest density of caribou.Rangifer, Special Issue No. 9, 1996

Nutrition as a limiting factor
Nutrition in the Delta herd apparently was poorer from 1990 to 1993 than in most prior years.
However, it is not clear how decreased nutrition may have contributed to reduced population performance.Body weights of 10-month-old Delta calves have generally been lower since 1989 (Fig. 3), and the parturition rate of females was significantly lower in 1990 and 1991 than from 1984 through 1989 {% 2 = 9.99, P < 0.01) (Table 2).In 1993 natality was very low.Since 1979 body weight of 10-month-old calves (which presumably reflects overall body condition in the Delta herd) has been a reasonably good predictor of calf survival to autumn (Fig. 4).This correlation may reflect increased vulnerability of calves to mortality factors during their first summer of life in years when overall herd nutrition is suboptimal prior to calving (Adams et al, 1994).Poor survival of offspring in populations of animals with suboptimal nutrition has been widely reported (Skogland, 1985).
It is tempting to conclude that the reduced natality rate in the Delta herd in 1993, and perhaps in 1990 and 1991, contributed to the caribou decline.However, in 1993 natality rates in the adjacent Denali herd and in the Chisana herd were at least twice as high as in the Delta herd, and autumn calfxow ratios were similar (6:100 in the Denali, 4:100 in the Delta, and 2:100 in the Chisana) (Adams, pers. commun.;Valkenburg, 1993).In addition, in 1992 natality in the Delta herd was the highest recorded, and the autumn calfxow ratio was among the lowest recorded ( Rangifer, Special Issue No. 9, 1996 where most of the Denali herd and about half of the Delta herd wintered.

Predation
There is compelling evidence that predation by wolves has been a major influence on the Delta herd over time.After wolf control in the mid-1970s the Delta herd became the most rapidly growing caribou herd in the state (Davis et al, 1983;Gasaway et al, 1983).In the mid-1980s, as wolf density approached precontrol levels, recruitment of caribou calves decreased and mortality of adults increased (Table 2).During 1983-1993 in the Delta herd, wolves were implicated in 23 out of 26 cases where the cause of death of adult radiocollared female caribou could be determined.In 26 additional cases, the precise cause of death could not be determined primarily because many of these caribou died during summer.In summer it was difficult to find evidence of hemorrhaging, so even if the kill site had been visited by bears or wolves, it was not possible to determine if the predators were scavenging or whether they killed the caribou.Long bones were recovered at about 50% of the winter kill sites, and in no case was malnutrition (as indicated by marrow fat content of less than 20%) documented as a possible contributing or direct cause of death.
In the recent decline, we did not determine causes of neonatal calf mortality.However, in the adjacent Denali herd (10) and the nearby Mentasta herd (18), wolf and grizzly bear predation were the major causes of high calf mortality (Adams, pers. comm.; Jenkins, pers. comm.).Grizzly bear densities are lower and wolf densities are higher in the range of the Delta herd than in the Denali herd (Dean, 1987;Mech et al, 1991;Boertje, 1993;Reynolds, 1993) and it therefore appears probable that wolves are more important as predators of calves in the Delta herd than in the Denali herd.
Prior to summer 1989 our data suggests that moose were the primary prey of wolves, but shortly thereafter, wolves switched to eating caribou.In  1990;Boertje, 1993).Subsequently, coincident with severe winter weather, wolves increased, caribou declined and moose continued to increase until 1992 (Boertje, 1993).Comparative data from radiocesium (CS-137) concentrations in wolves corroborated this behavioral switch in prey selection over time (Boertje et al, 1992).In addition, Mech et al. (1994) presented evidence that wolves included a higher proportion of caribou in their diet after 1989 in the range of the Denali herd.
Because much of the evidence for wolf predation as the main limiting factor in the Delta herd is circumstantial, we cannot be absolutely certain that the Delta herd would not have declined if wolf numbers had been substantially lower.If wolves are effectively removed from the calving and summer ranges of the Delta herd during the ongoing control program (winters 1993-1994 and 1994-1995) and the herd fails to recover, we will accept this as strong evidence that wolves were not the cause of the high calf mortality.
Another important question is whether wolves could have caused a decline in caribou without the presence of adverse weather.From 1985 on, wolves were an important limiting factor.The April 1988 and 1989 calfxow ratios suggested increased overwinter mortality of calves prior to the onset of severe winters.However this probable decrease in recruitment, prior to the onset of severe weather, was insufficient to cause the herd to decline.In addition, wolves did not show a concurrent numerical response-wolf numbers remained relatively stable from 1985 to 1989 prior to the onset of bad weather (Boertje, 1993).
Wolf predation as a density dependent limiting factor Bergerud (1993) proposed a conceptual model of population regulation in woodland caribou where wolf predation acts in a density dependent way and maintains caribou density at low levels (i.e., <0.1/km 2 ) because caribou lose the ability to effectively 'space out' from wolves at higher density.
Although this model may fit some of the smallest Alaskan herds, clearly there are many herds which survive for long periods at moderate densities and neither 'space out' nor 'space away' from wolves.
Furthermore, we found no clear relationship between caribou:wolf ratio or caribou equivalents:wolf ratio and caribou growth rate (  1 Population estimates during the period ranged from 2,393 to 2,697 but no trend was apparent. f Assuming a fall population of 80 wolves within the range of the herd (data from Tobey, 1991).
ce (Picea qlauca) forest.Adjacent herds include the Macomb herd (17) to the east, Denali herd (10) to the west, White Mountains herd (31) to the north, and Nelchina herd (20) to the south.Until the early 1970s the Delta herd was considered one of many rather insignificant groups of caribou in Alaska.It was relatively small in size, had an inaccessible range, and management and research efforts were concentrated on the larger, road-accessible Fortymile and Nelchina herds.However, after the decline of the Fortymile and Nelchina herds in the early 1970s, the Delta herd received more attention from hunters, and consequently, from the Alaska Department of Fish and Game (ADF&G).Efforts to determine population identity and recruitment had begun in the late 1960s, but the first sys-Rangifer, Special Issue No. 9, 1996 tematic census was not done until 1973.In 1979 ADF&G identified the need for a long-term population dynamics study of an Interior caribou herd, and began intensive research on the Delta herd.

Fig. 2 .
Fig. 2. Linear regression of crude caribou density versus annual average population growth rate (k).

Fig. 3 .
Fig. 3. Mean weight and standard error bars of samples of female calves weighed in April 1979-1993.

February-
March 1989, just prior to the caribou decline, we investigated prey selection in 4 wolf packs in the Delta herd's range-by tracking collared individuals 2-3 times daily.Caribou and moose were abundant within the ranges of all packs.By weight, moose comprised two-thirds and caribou one-third of the wolves' diet (assuming 1 average moose = 3 average caribou).During the 30-day period, the 4 wolf packs studied killed 16 moose, 23 caribou, and 2 sheep.The small wolf packs killed as many caribou as the larger ones.At that time caribou and moose were both increasing (McNay,

Fig. 5 .
Fig. 5. Scatter plot of annual average population growth rate (A.) versus caribou:wolf ratio for 6 Interior Alaska caribou herds.
plot of annual average population growth rate (À.) versus caribou equivalents:wolf ratio for 6 Interior Alaska caribou herds.therut) and decreased calf survival the following summer (through reduced body condition of calves at birth).The caribou decline in the Delta herd was coincident with 4 of the most severe winters since 1972 and followed 3 warm summers (Fig.7).This was probably also true for the Macomb and Denali herds where weather was similar.However, in east central Alaska on the winter ranges of the Chisana, Mentasta, and Nelchina herds only the winter of 1989-1990 was severe, and snow depth barely exceeded 70 cm (snow data from Northway).The Chisana and Mentasta herds declined rapidly ( Fig. 7. Plot of parturition rate and autumn calfxow ratio in relation to snow depth, summer precipitation, and summer temperature in the range of the Delta herd.

Table 1 .
Estimated size and crude density of Alaskan caribou herds.
b Not shown on fig. 1. c 1992 estimate.d Stablilized through harvest.
* Fixed wing count only.b Count probably not representative of herd.

Table 2
).The cause of the low natality in the Delta herd in 1993 is unknown, however, weather in May and September 1992 was unusual and the growing season was short.Persistent cold and snow in May resulted in the latest leaf out ever recorded in Fairbanks (25 May), and Eriophorum T FEMALE CALF WEIGHT (KG) Fig. 4. Linear regression of female calf weight in April onSeptember-October calfxow ratio (data from Table2).

Table 4
ted by short periods of rapid decline.Superficially, this may appear to be density-dependent predation, but growth rate of caribou may be more sensitive to influence of stochastic environmental factors rather than caribou density.

Table 4 .
Characteristics of 6 Interior Alaska caribou herds, 1980-1993.This area was formerly considered part of the range of the Fortymile herd, the herd was first recognized in the late 1970s.
bThe population peak actually occurred in 1990.c d Growth rate is approximate because the 1980 population estimate was poor.