- Email: [email protected]
Short-term behavioural responses of Svalbard reindeer Rangifer tarandus platyrhynchus to direct provocation by a snowmobile
Biological Conservation 56 (1991) 179-194
?, \~t ' l
Short-term Behavioural Responses of Svalbard Reindeer Rangifer tarandus platyrhynchus to Direct Provocation by a Snowmobile N. J. C. T y l e r Department of Arctic Biology and Institute of Medical Biology, University of Tromso, Breivika, N-9000 Tromso, Norway (Received 15 February 1990; revised version received 20 June 1990; accepted 15 August 1990)
ABSTRACT Short-term behavioural responses of 101 groups of Svalbard reindeer Rangifer tarandus platyrhynchus to direct provocation by a snowmobile were recorded in April (late winter) 1987. The median size o f groups was 3"3 animals. The first visible responses of reindeer to an approaching snowmobile usually involved independent behaviour by different individuals in a group. Flight, by contrast, was a co-ordinated group response. Group median response distances were: minimum reaction distance, 640m; disturbance distance, 410 m; distance at initial flight, 80 m; and distance of flight, 160 m. Group median response times were: total running time, 22 s; total locomotion time, 38 s; maximum duration of disturbance, 193 s. Energy and time budget models indicate that one median flight response can cause an increase in a reindeer's daily energy expenditure o f approximately 0.4% and a loss of daily grazing time also of 0"4%. The study failed to detect any way in which the current level of snowmobile traffic might substantially reduce the physical wellbeing of Svalbard reindeer.
INTRODUCTION Disturbance to wildlife due to industrial development is an important environmental concern in the arctic (Shank, 1979; Martell & Russell, 1985; 179 Biol. Conserv. 0006-3207/91/$03"50© 1991 Elsevier Science Publishers Ltd, England. Printed in Great Britain
180
N . £ C Ty&r
Shideler, 1986; Shideler et al., 1986). Nevertheless, there have been few detailed studies of behavioural responses of reindeer and caribou Rangifer tarandus to specific stimuli, and none at all which investigate their reactions to snowmobiles. This paper describes the short-term behavioural responses of high arctic Svalbard reindeer to direct provocation by a snowmobile.
STUDY AREAS The study was conducted in Nordenskiold Land (78 ° N), Svalbard (Fig. 1) in April 1987. There are about 600 snowmobiles in Svalbard (Berg, 1987), most of which (c. 500) are owned privately by residents of Longyearbyen, the principal Norwegian settlement (population 1500) (Fig. 1). The two roads that lead out of the town have a combined length of approximately 20 km. Besides dog-sleds, snowmobiles are the only means of access to the interior and are important for both mineral exploration and recreation from February until snow melt, which usually occurs in mid-May. Nevertheless, one-third of traffic occurs within 10 km of the town (calculated from data in Persen, 1986). Beyond that, 96% of users normally follow ad hoc snowmobile routes along frozen river beds (Persen, 1986; Fig. 1).
80ON
~ ~'~"
~ J
/
I
"~ z~
I ....... l
~Sassendalen--
"~
'
"! -
L"I
.
.
.
road
- - - - ~-
....il -'~*-=-'-'Adven tdalan' ....
snowmobile route
,,~'...~
.
;
N
"
T I
I
I
I
Fig. 1. Map of Nordenskiold Land in Spitsbergen, largest island in the Svalbard Archipelago, showing the two study areas ('Adventdalen" and 'Sassendalen'), roads and principal snowmobile routes.
Responses of reindeer to provocation
181
There are approximately 4000 Svalbard reindeer R. t. platyrhynchus in Nordenskiold Land but the region is divided by mountains into distinct valley systems between which the animals normally do not move (Tyler & Oritsland, 1989). Each such system contains, in effect, a discrete subpopulation (Tyler, 1987a). Two subpopulations were selected to compare responses of reindeer in areas subjected to different levels of snowmobile traffic: 'Adventdalen' (c. 600 reindeer) and 'Sassendalen' (c. 700 reindeer) (Fig. 1). Adventdalen is the most heavily used area in Svalbard; there is considerably less traffic in Sassendalen. Both valleys are broad, fiatbottomed and bordered by plateaux that rise steeply to 400-500 m above sea level. No trees occur in Svalbard and reindeer (and snowmobiles) are often conspicuous over several kilometres, depending on the terrain. METHODS Provocation Two observers searched for reindeer using two 350 cc snowmobiles. When a reindeer group (with no two individuals > 30 m apart and forming a discrete social unit) was first seen the machines were parked together at a distance of 500-2500 m with engines stopped. The following information was recorded: (i) number of animals, their age (adult, > 24 months; yearling, 22-23 months; calf 10-11 months) and sex; (ii) their principal activity (feeding, lying down); (iii) type of terrain (open, broken). Groups were provoked by driving one snowmobile slowly (20 km/h) until reaching the spot where the animals had been. Reindeer were watched continuously during the approach and a stick was dropped to mark the position on the first occasion that any member of a group showed any visible reaction to the snowmobile. A second stick was dropped on the first occasion that any member of the group showed signs of unease or alarm. A third stick was dropped when the group fled. The 'attacker' then drove slowly back to the starting point and measured five distances to 80 m with the snowmobile odometer: (i) m i n i m u m reaction distance (distance between a group of reindeer and the 'attacker' at the first occasion that any member of the group reacted visibly to his presence); (ii) disturbance distance (distance between a group of reindeer and 'attacker' at the first occasion any member of the group showed any sign of unease or alarm); (iii) distance at initial flight (distance between a group of reindeer and the 'attacker' at the instant the animals first fled) and (iv) total distance (distance between the reindeer and the observer/'attacker' at the start of provocation). Total net distance of flight (v) was measured once the animals had settled. Distances < 80 m were estimated to 20 m by eye.
182
N.J.C. Tyler
The observer watched each group continuously during the provocation to corroborate marking of each distance. He recorded each phase of the animals' flight, including total running time, total locomotion (running and walking) time and maximum duration of disturbance (time from the instant that a group bolted until all members of the group had subsequently settled) to 1 s with a stop watch.
Frequency of disturbance The frequency with which reindeer in Adventdalen were overtly disturbed by normal snowmobile traffic was determined by watching groups of between three and eight reindeer continuously for periods of between 24 and 29h. Individual records were kept for each animal in each group. Observations were made in April 1981 and 1982 during the Easter holidays when traffic is heaviest. Daylight is continuous in Svalbard at that time.
Analysis of data Seven measured responses were compared with respect to six variables: group size, group composition by age and sex, location (Adventdalen vs Sassendalen), activity (feeding vs lying down), terrain (open vs broken) and the animals' view of the 'attack' approach (restricted vs unrestricted). Four types of groups were recognised in the analysis: M groups (adult and yearling males only), F groups (adult and yearling females only), FFC groups (groups of adult females and one or more calves), MX groups (groups including at least one adult male and one adult female). Results for adults and yearlings were combined by sex. Skew was tested using the t statistic (Sokal & Rohlf, 1969). Time or distance parameters were compared between groups or areas using Mann-Whitney U tests (twosample comparisons) or Kruskal-Wallis one way analysis of variance (k sample comparisons) (Siegel, 1956). Association between parameters was measured using Spearman's rank correlation (rs). H o was rejected at p (twotailed) < 0-05 in all tests.
RESULTS Data were recorded from 101 groups of reindeer (336 animals: 71 adult males, 206 females, 2 male yearlings, 1 female yearling, 2 yearlings of unknown sex and 54 calves), 64 groups in Adventdalen and 37 in Sassendalen. The median size of groups was 3.31 reindeer (range = 1-11).
Responses of reindeer to provocation
183
First visible reactions A reindeer's first visible reaction to provocation was normally a m o m e n t a r y glance up at the approaching snowmobile, after which it resumed whatever it had been doing (e.g. feeding, lying). Reindeer normally betrayed unease or alarm either by 'alerting' or by bunching up. Alerting involved one or more animals quickly looking up, ears pricked and rotated forward, and often urinating at the same time. Bunching involved peripheral animals quietly breaking off feeding and walking slowly in towards the group. Those that moved did not usually 'alert', and usually avoided disturbing the group.
Flushing and flight Groups almost invariably fled as a single unit but animals which had alerted or bunched were usually the first to run. The median duration of the first run was 15 s ( r a n g e = 3-110s, n = 67). In most cases (79"3%, n = 69) the group then suddenly stopped, turned a r o u n d and stood staring at the snowmobile (Fig. 2). Animals that had not already urinated often did so at this stage. The median duration of the first pause was 16 s (range = 1-500 s, n = 64). After it, most groups (73-5%, n = 47) either started running again (42.2%) or began to feed (31"3%). The median duration o f the second run was 15 s (range = 3-120 s, n = 27), after which most groups (n = 21) again stopped abruptly, turned a r o u n d and stared at the snowmobile. There was a strong positive correlation between the duration o f the first and second runs (r s = 0"478, n = 27, p < 0"02). After fleeing all reindeer in a group soon resumed feeding or lay down.
Minimum reaction distance (MRD), disturbance distance (DD) The median m i n i m u m reaction distance was 640m and the median disturbance distance was 4 1 0 m (Table 1). Neither M R D nor D D differed significantly with respect to size, composition or location of groups, or with the nature of the terrain or the animals' maintenance activity immediately prior to provocation.
Distance at initial flight (DIF) The median distance between reindeer and the snowmobile when the animals first fled was 8 0 m (Table 1). Reindeer tended to flee at a greater distance in Sassendalen (mean D I F = 148 m) compared to in Adventdalen (mean D I F = l19m). The median D I F (all groups) was the same (80m) in both areas (Table 1) but the rank variance of the data differed between them
N. J. C. Tyler
184
GROUPS PROVOKED
NEXT ACTIVITY: (N = 87)
NEXT ACTIVITY: (N = 6 4 )
I
NEXT ACTIVITY : (N
=
&
START R U N N I N G ~
FEED WALK STAND STOP, TURN & STAND
2.3% 5.7% 6.9%
79.3%
ILIE DOWN ~[WALK IFEED ISTART R U N N I N G AGAIN
42.2%
[WALK ~ISTOP,
TURN & STAND
19.2% 80.8%
[FEED ~[WALK START R U N N I N G AGAIN
n = 4 n = 8 n = 9
1.6%
25.0% 31.3%
26)
NEXT ACTIVITY: (N = 21)
1
NEXT ACTIVITY: (N = 8)
]FEED ~[WALK [STOP, TURN & STAND
n = 1 n : 3 n = 4
Fig. 2. Sequence of responses of groups of reindeer after being provoked into running. The chart shows the proportion (%) of groups ( n = 87) that exhibited one of four different activities (feeding; walking; standing; stop, turn and stand) after first being provoked into running. Thereafter, the chart shows the next activity of as many as possible of those groups that exhibited the most common behaviour at each stage.
185
Responses o f reindeer to provocation
(all g r o u p s , z = 1.95, n~ = 62, n 2 = 37, p = 0.05). A l l - m a l e g r o u p s in S a s s e n d a l e n b o l t e d at a s i g n i f i c a n t l y g r e a t e r d i s t a n c e ( m e d i a n = 240 m) t h a n allm a l e g r o u p s in A d v e n t d a l e n ( m e d i a n = 50 m : z = 2-01, nl = 7, n 2 = 6, p < 0"05). R e i n d e e r t h a t w e r e l y i n g d o w n t e n d e d t o b o l t at a g r e a t e r d i s t a n c e ( m e d i a n D I F = 160 m) t h a n g r o u p s t h a t w e r e f e e d i n g w h e n p r o v o k e d ( m e d i a n D I F = 8 0 m ; all g r o u p s c o m b i n e d , z = 2.03, n~ = 74, n 2 = 18, p < 0"05). R e i n d e e r t h a t h a d a n u n b r o k e n v i e w o f the a p p r o a c h i n g s n o w m o b i l e t e n d e d t o b o l t at a g r e a t e r d i s t a n c e ( m e d i a n D I F = 160 m) t h a n g r o u p s t h a t h a d a r e s t r i c t e d v i e w ( m e d i a n D I F = 80 m ; all g r o u p s c o m b i n e d , z = - 2-35, nl = 56, n z = 31, p < 0-02). T h e difference w a s s i g n i f i c a n t in a l l - m a l e g r o u p s ( u n r e s t r i c t e d view, m e d i a n D I F = 2 0 0 m ; r e s t r i c t e d view, m e d i a n D I F = 5 0 m ; z = - 2 . 1 5 , n 1 = 8 , n 2 = 4 , p < 0 - 0 5 ). TABLE 1 Data Used in an Evaluation of Seven Measured Responses by Svalbard Reindeer to Direct Provocation by a Snowmobile Data are given for all of groups combined with respect to location, except when significant differences by group type occur. FFC groups, see Methods Variables by measured response
Adventdalen
Sassendalen
Both areas
Minimum reaction distance (m) All groups a Median 600 Minimum 50 Maximum 1 840 N 48
640 80 1 520 35
640 50 1 840 83
Disturbance distance (m) All groups b Median Minimum Maximum N
480 80 1 280 35
410 50 1 280 82
400 50 1 280 47
Distance at initial flight (m) All groups b Median 80 Minimum - 10c Maximum 480 N 62 All-male groups b Median 50 Minimum 15 Maximum 240 N 7
80 10 480 37 240 80 480 6
80 -10
480 99 8O 15 480 13 (~ ontbnted~
TABLE 1--contd. Variables by measured response Distance of flight (m) All groups b Median Minimum Maximum N F F C groups b Median Minimum Maximum N Total running time (s) All groups b Median Minimum Maximum N F F C groups b Median Minimum Maximum N Total locomotion time (s) All groups b Median Minimum Maximum N
Adventdalen
Sassendalen
Both areas
160 10 1 300 57
*
160 50 1 440 34
160 10 1440 91
80 10 400 11
*
200 50 1 040 12
80 10 1 040 23
25 9 180 33
22 0 210 94
33 10 90 13
23 5 90 26
45 10 640 29
38 5 890 84
290 65 765 27
193 16 2130 84
275 65 765 12
190 35 2 130 24
21 0 210 61 13 5 60 13
35 5 890 55
M a x i m u m duration of disturbance (s) All groups b Median 185 Minimum 16 Maximum 2 130 N 57 F F C groups b Median 116 Minimum 35 Maximum 2 130 N 12 a Frequency distribution skewed b Frequency distribution skewed c One group (adult female and a snowmobile had driven between * p _< 0"05; ** p < 0"02.
*
to the right, p < 0"01. to the right, p < 0-001. calf) did not start to run until after the and continued 10m beyond them.
Responses of reindeer to provocation
187
Distance of flight ( D F )
Reindeer in Sassendalen tended to flee further (mean D F = 303 m) than reindeer in Adventdalen (mean D F -- 210 m, Table 1). Median D F was the same (160m) in both areas but the rank variance of the data differed significantly between them (all groups, z = 2"05, nl = 57, n 2 = 34, p < 0-05). Groups o f adult females which included one or more calves (FFC) in Sassendalen fled further (median D F = 200m) than FFC groups in Adventdalen (median D F = 8 0 m : z=2.30, n ~ = l l , n2=12, p<0"05). Reindeer that bolted early tended to run farther: distance of flight was weakly but significantly (p < 0.002) correlated with distance at initial flight (Fig.
3).
Total locomotion time, total running time (TRT)
Median total locomotion time was 38 s (Table 1). Reindeer in Sassendalen tended to spend more time running (median TRT = 25 s) than reindeer in Adventdalen (median TRT = 21 s, Table 1) but the difference was significant only in F F C g r o u p s ( z = 1.93, nl--- 13, n 2 13, p = 0"05). ----
M a x i m u m duration o f disturbance ( M D D )
Maximum duration of disturbance was significantly shorter in Adventdalen (median M D D = 185 s) c o m p a r e d t o in Sassendalen (median M D D ----2 9 0 s; all g r o u p s , z = 2.00, n 1 = 57, n 2 = 2 7 , p = 0 - 0 4 6 ) and in FFC g r o u p s (z = 2.34, n I = 12, n 2 = 12, p < 0-02; T a b l e 1). M D D was significantly shorter in o p e n terrain (median = 185 s) c o m p a r e d t o i n b r o k e n terrain (median = 2 9 6 s; all 1500 "E 1200 ,_m ,4-
900
~,
soo
c
•C~ ~
m O0
300
•
• •
•
,
0 0
1 O0
•
,O
• 200
'~ 300
• 400
500
Distance at initial flight (m)
Fig. 3. Distance (m) g r o u p s o f reindeer fled after provocation plotted against the distance between them and the snowmobile at the instant that they fled. Spearman's rank correlation coefficient r s = 0-0335, n = 89, p < 0"002. The correlation remains significant even after the u p p e r right-hand two points are excluded. Overlapping data points are not s h o w n .
188
N. J. C. Tyler
groups, z = 2.28, nl = 51, n2 = 24, p = 0"02) but none of the differences by group type were significant.
Energy cost of disturbance Median total locomotion time after provocation was 38 s, including 22 s running (Table 1). Assuming that reindeer ran at 40 km h - 1, walked at 4 km h-1 and did not break through the snow-crust (which they rarely do in April), the gross cost of locomotion for a 50-kg reindeer during a flight of median duration is 46 kJ (from Nilssen et al., 1984). If the total daily energy expenditure (DEE) of a 50-kg animal in April is 11-7 MJ (Tyler, 1987c), one median flight response represents an increment in DEE of 0"4%. Corresponding values for one maximum and one minimum flight response are 4.7% and 0.01% of DEE, respectively.
Frequency of disturbance During observation of six groups of reindeer (n = 32 animals) over 817 h, in which no continuous record for any animal lasted < 24 h, reindeer were overtly disturbed on only four occasions: (1) three snowmobiles drove by 100 m from four animals which had been lying down for 1"5 h; they stood up and walked slowly away for 20 min before starting to feed for 4.5 h; (2) a single snowscooter drove by 200 m from an adult female, which had been lying down for 30 min; she stood up and began to feed for 40 min before lying down again; (3) three skiers walked 50 m past three reindeer lying on the snow; the animals stood up, walked slowly away for 5 min before starting to feed for 1.5 h before lying down again; (4) thirty minutes later these animals leapt up and stood staring at something beyond the observer's view for 5 min before lying down for 3-5 h.
DISCUSSION Groups of Svalbard reindeer often bunched when provoked. Bunching is an 'anti-predator behaviour' displayed by many species including ungulates (Curio, 1976). Its occurrence is unexpected in Svalbard reindeer because there are no large predators in Svalbard (except for polar bears Ursus maritimus which do not prey on reindeer there). Moreover Svalbard reindeer have been fully protected by Norwegian law since 1925. Residents of the archipelago, however, were granted dispensation to hunt them by the Royal Ministry of the Environment, Oslo, in 1983 and approximately 100 animals have been shot each year since then. Hunting remains forbidden in
Responses of reindeer to provocation
189
Adventdalen. Nevertheless, despite this protection, the reindeer display typical alarm and flight behaviours when disturbed, including bunching and the 'excitation jump' or 'stott' (unpublished data). Altering and bunching in response to provocation by a snowmobile contrasts with the behaviour of barren-ground caribou at the approach of a three-quarter ton truck: these normally fled on sight (Horejsi, 1981). The difference supports Horejsi's (1981) suggestion that it is the rate at which an object approaches, not its size or movement per se, which causes reindeer/caribou to flee: his truck travelled between three and four times faster than the snowmobile in the present study. In further contrast, unlike Horejsi (1981) I recorded no 'no-flight' responses, a reflection of the fact that the snowmobile was driven straight at reindeer until reaching the spot where they had been to start with. This aspect of the experimental design has important consequences for interpretation of results. Svalbard reindeer, like caribou (Horejsi, 1981), evidently distinguish and respond according to the direction at which a vehicle approaches: I was frequently able to drive close ( < 3 0 m ) but obliquely past animals without putting them to flight. Driving straight at them, therefore, must represent a 'worst-case' situation and distance parameters recorded here ought to be regarded as potential maxima, at least with respect to a single snowmobile travelling at moderate speed. So far as assessing potential short-term effects of ordinary snowmobile traffic on the Svalbard reindeer, however, the bias will be conservative. This is corroborated by the four 'real-life' examples of disturbance (see Results), none of which provoked reindeer into running. Disturbance by motor vehicles can have overt effects on animals, such as trauma, and covert effects, such as causing them to avoid otherwise favoured habitat in which regular traffic occurs. This study was concerned only with overt consequences of disturbance. I observed no injuries nor circumstances likely to produce them. Svalbard reindeer are not nomadic; individual animals occupy small, traditional home ranges for weeks or even months at a time in winter (Tyler & Qritsland, 1989). Each animal, therefore, is presumably familiar with its local topography; flight, consequently, is probably rarely a random, stumbling scramble. On the contrary, animals that live in small home ranges are probably capable of instantly selecting the surest route to the safest ground. This may very largely explain how Svalbard reindeer avoid injury when galloping tightly bunched and at full speed over rough terrain, although the small size of groups may also be significant. Miller and Gunn (1979), for example, never observed Peary caribou, which normally live in groups of less than five, injure themselves when fleeing from a helicopter. By contrast, in wildebeest Connochaetes taurinus--a notably gregarious species--several hundreds of animals can be
190
N. J. C. Tyler
drowned or trampled during stampedes (Sinclair, 1979). Some Svalbard reindeer evidently become completely habituated to snowmobiles and other forms of human activity. Four adult males currently live in Longyearbyen for most of the year. One (marked) is known to have been there for nine years. Other males habitually establish summer home ranges around the airstrip where commercial jets land three times each week and where, in summer, there is helicopter activity around the clock. Despite this, and despite regular and increasing snowmobile traffic each winter, the density of reindeer in Adventdalen remains higher than in any other natural subpopulation in Svalbard (Oritsland & Alendal, 1985). Evidence for habituation in reindeer beyond the settlement is less clear. Animals in Sassendalen tended to be more shy than animals in Adventdalen (Table 1). While this reflects different levels of snowmobile traffic in the two areas it need not have been due to different levels of habituation. Herbivore activity is strongly affected by the abundance of food. Foraging time, in particular, increases with decreasing food availability (Jarman & Jarman, 1979; Owen-Smith, 1979; Trudell & White, 1981; Skogland, 1984). Adult females in Adventdalen feed 30 % more per 24-h day in winter than females introduced onto ungrazed pasture at Ny ,~lesund, 100 km further north (Tyler, 1987a, and unpublished data). The density of reindeer in Adventdalen was 50% higher than in Sassendalen when this study was carried out (Tyler, 1987b; N. A. Oritsland, pers. comm.). There is no reason to assume that the per capita supply of food in winter was the same in each area. Consequently, it is not possible to say to what extent differences in responses to a snowmobile between reindeer in the two areas were due to different levels of habituation or differences in nutritional status. Time and energy costs of flight from a snowmobile are surprisingly small. The figures (increments in DEE of 4.7 %, 0.4 % and 0.01% for one maximum, median and minimum flight response, respectively) considerably overestimate the energy costs of 'normal' encounters between snowmobiles and Svalbard reindeer. They are based on responses to highly artificial, direct provocation and, moreover, represent the gross, or instantaneous, energy costs of flight. What matters, so far as the animals' fitness is concerned, is the net cost. Gross and net costs are equal only when animals are unable to compensate for involuntary expenditure of energy by appropriate adjustments in food intake, or are prevented from doing so. Moen et al. (1982) pointed out that changes in animals' behaviour are not a dependable indicator of the effects of disturbance on heart rate, which itself is known to correlate closely with energy expenditure both in Rangifer and other ungulate genera (Nilssen et al., 1984; Fancy & White, 1986). This is a potential limitation of studies of harassment which quantify only sensible responses to disturbance (MacArthur et al., 1979, 1982a, b; Moen et al.,
Responses of reindeer to provocation
191
1982). However, the significance of Moen et al.'s (1982) observation is not clear. Increased heart rate may be a useful indicator of psychological activation but it is not a reliable indicator of oxygen consumption in emotionally excited subjects (Blix et al., 1974; Str~mme et al., 1978). Shideler et al. (1986) suggested that 'the effects of linear facilities and/or off-road vehicles may have important consequences to foraging success of some caribou herds...'. Svalbard reindeer spend approximately 12.9h grazing per 24h during fine weather in late winter: there is no difference between the sexes or between adults and calves in this respect (Tyler, 1987a). The median maximum-duration of disturbance (MDD) following a single provocation by a snowmobile was 193s (Table 1). Thus, assuming all disturbance time represents lost grazing time, then the loss from a single provocation with minimum, median and maximum M D D represents 0.03, 0.4 and 4.6% of daily grazing time, respectively. These values are also likely to overestimate the amount of grazing time that Svalbard reindeer typically lose following an encounter with a snowmobile. The figures are calculated on the basis of the 'maximum duration of disturbance' of each group. They represent, therefore, not the average but the extreme response in each group provoked in this study. Time and energy costs of flight following deliberate provocation are small but, in a marginal environment, the cumulative effect of small costs may have substantial ecological significance. The potential cumulative effect of disturbance can be appreciated by considering the frequency with which reindeer are normally disturbed in late winter in Svalbard. Six groups of reindeer watched continuously for periods > 24 h were disturbed altogether four times, i.e. rather less than one disturbance per group per day at the height of the holiday season. This low frequency reflects the fact that Svalbard reindeer normally inhabit mountain sides and high plateaux in late winter whereas snowmobile routes follow river beds (Tyler, 1987d). Given both the low frequency of disturbance and the low cost of each disturbance, it seems unlikely that snowmobiles may be responsible for substantial increases in energy expenditure or losses of grazing time in Svalbard while the intensity of traffic remains at its present level. This conclusion is contrary to interpretations of results from many previous studies of effects of motor traffic (including aircraft) on reindeer and caribou (see reviews by Shideler (1986) and Shideler et aL (1986)). There are several potential reasons for this. This study considered only immediate, overt responses of Svalbard reindeer to provocation. It ignored entirely all psychological and physiological aspects. But so did all the studies which Shideler (1986) and Shideler et al. (1986) reviewed, so the difference is unlikely to lie here. Alternatively, animals often respond vigorously to motor traffic and it sometimes may be difficult to avoid conclusions being
192
N.J. C. Tyler
influenced by the visual drama of the event. Obviously, galloping is energetically costly. However, the fact that animals gallop when directly provoked by a vehicle tells us little about how normal traffic affects their wellbeing. At the very least we must calculate the gross energy cost of each flight and record how frequently they are put to flight. Perhaps the simplest explanation is that there is no reason to expect Svalbard reindeer to respond to provocation in the same way as other subspecies of reindeer and caribou. Svalbard reindeer are notable for their docile, sedentary behaviour. They normally spend less than 4 % of each 24 h walking and they rarely run (Nyholm, 1976; Kastnes, 1979; Tyler, 1987a): without predators they have no need to. In fact, their activity budget in winter is not substantially different from that of other subspecies (e.g. Skogland, 1984; Boertje, 1985) but caribou and continental wild reindeer live under constant threat of suddenly having to gallop off from biting flies or wolves Canis lupus. Having to escape in a hurry is part of their daily life; for Svalbard reindeer this is not the case.
ACKNOWLEDGMENTS This study was paid for by Store Norske Spitsbergen Kulkompani A/S and Norsk Hydro A/S. I thank Hhkon Stensrud--the provocateur--for assistance in the field and S. D. Albon, R. T. Bowyer, D. R. Klein, J. P. Kelsall, N. Leader-Williams and F. L. Miller for their comments on the manuscript.
REFERENCES Berg, A. (1987). Response from the snowmobile club. Svalbardposten, Longyearbyen, 11, 11. Blix, A. S., Str~mme, S. B. & Ursin, H. (1974). Additional heart rate--an indicator of psychological activation. Aerospace Med., 45, 1219-22. Boertje, R. D. (1985). An energy model for adult female caribou of the Denali herd, Alaska. J. Range Manage., 38, 468-73. Curio, E. (1976). The Ethology of Predation. Springer-Verlag, Berlin. Fancy, S. G. & White, R. G. (1986). Predicting energy expenditures for activities of caribou from heart rates. Rangifer, Special Issue, 1, 123-30. Horejsi, B. L. (1981). Behavioral response of barren ground caribou to a moving vehicle. Arctic, 24, 180-5. Jarman, P. J. & Jarman, M. V. (1979). The dynamics of ungulate social organization. In Serengeti: Dynamics of an Ecosystem, ed. A. R. E. Sinclair & M. NortonGriffiths. University of Chicago Press, Chicago, pp. 185-220.
Response of reindeer to provocation
193
Kastnes, K. (1979). Svalbardreinens aktivitetsm~nster gjennom ~ret (Activity patterns of Svalbard reindeer throughout the year). Cand. real. thesis, University of Oslo. MacArthur, R. A., Johnston, R. H. & Geist, V. (1979). Factors influencing heart rate in free-ranging bighorn sheep: a physiological approach to the study of wildlife harassment. Can. J. Zool., 57, 2010-21. MacArthur, R. A., Geist, V. & Johnston, R. H. (1982a). Cardiac and behavioral responses of mountain sheep to human disturbance. J. Wildl. Manage., 46, 351-8. MacArthur, R. A., Geist, V. & Johnston, R. H. (1982b). Physiological correlates of social behaviour in bighorn sheep: a field study using electrocardiogram telemetry. J. ZooL Lond., 196, 401-15. Martell, A. M. & Russell, D. E. (eds). (1985). Caribou and human activity. Proc. N. Am. Caribou Workshop, 1st, Whitehorse, Yukon, September 1983. Miller, F. L. & Gunn, A. (1979). Responses of Peary caribou and muskoxen to helicopter harassment. Can. Wild. Serv. Occ. Pap., No. 40, 1-89. Moen, A.N., Whitemore, S. & Buxton, B. (1982). Effects of disturbance by snowmobiles on heart rate of captive white-tailed deer. N. Y. Fish & Game J., 29, 177-83. Nilssen, K. J., Johnsen, H. K., Rognmo, A. & Blix, A. S. (1984). Heart rate and energy expenditure in resting and running Svalbard and Norwegian reindeer. Am. J. Physiol., 246, R963-7. Nyholm, E.S. (1976). Monthly general rhythm pattern variations in the Nordenskiold Land-Sabine Land population of the Svalbard reindeer (Rangifer tarandus platyrhynchus). Norsk Polarinst. ,~rbok, 1974, 129-38. Oritsland, N. A. & Alendal, E. (1985). Bestandens storrelse og livshistorie (Population size and life history). In Svalbardreinen og dens Livsgrunnlag (Ecology of Svalbard Reindeer), ed. N.A. IDritsland. Universitetetsforlaget, Oslo, pp. 52-60. Owen-Smith, N. (1979). Assessing the foraging efficiency of a large herbivore, the kudu. S. Afr. J. Wildl. Res., 9, 102-9. Persen, E. (1986). Snascooteren og naturmilj~et (The snowmobile and the natural environment). Typed report to the Governor of Svalbard. (Available from Sysselmannen p~ Svalbard, N-9170 Longyearbyen, Norway). Shank, C.C. (1979). Human-related behavioral disturbance to northern large mammals: a bibliography and review. Report prepared for Foothills Pipe Lines (South Yukon) Ltd, Calgary. Shideler, R. T. (1986). Impacts of human developments and land use on caribou: a literature review, II. Impacts of oil and gas development on the central Arctic herd. Alaska Dep. Fish & Game, Div. Habitat, Tech. Rep., 86-3. Shideler, R. T., Robus, H. H., Winters, J. J. & Kuwada, M. (1986). Impacts of human developments and land use on caribou: a literature review, I. A worldwide perspective. Alaska Dep. Fish & Game, Div. Habitat, Tech. Rep., 86-2. Siegel, S. (1956). Nonparametric Statistics for the Behavioural Sciences. McGrawHill Kogakshua, Tokyo. Sinclair, A. R. E. (1979). The eruption of the ruminants. In Serengeti: Dynamics of an Ecosystem, ed. A. R. E. Sinclair & M. Norton-Griffiths. University of Chicago Press, Chicago, pp. 82-103.
194
N.J. C. Tyler
Skogland, T. (1984). Wild reindeer foraging niche organization. Holarctic Ecol., 7, 345-79. Sokal, R. R. & Rohlf, F. J. (1969). Biometry: The Principles and Practice of Statistics in Biological Research. W. H. Freeman, San Francisco. Str~mme, S. B., Wikeby, P. C., Blix, A. S. & Ursin, H. (1978). Additional heart rate. In Psychobiology of Stress: A Study of Coping Men, ed. H. Ursin, E. Baade & S. Levine, Academic Press, New York, pp. 83-9. Trudell, J. & White, R. G. (1981). The effect of forage structure and availability on food intake, biting rate, bite size and daily eating time of reindeer. J. appl. Ecol., 18, 63-81. Tyler, N. J. C. (1987a). Natural limitation of the abundance of the high arctic Svalbard reindeer. PhD thesis, Cambridge University. Tyler, N. J. C. (1987b). Environmental studies related to petroleum activities on Svalbard. Report from the reindeer project 1987. Report prepared for Store Norske Spitsbergen Kulkompani A/S & Norsk Hydro A/S, Longyearbyen, Norway. Tyler, N. J. C. (1987c). Body composition and energy balance of pregnant and nonpregnant Svalbard reindeer during winter. Syrup. zool. Soc. Lond., 57, 203-29. Tyler, N. J. C. (1987d). Environmental studies related to petroleum activities on Svalbard. Status and distribution of reindeer in Reindalen. Report from the reindeer project 1986. Report prepared for Store Norske Spitsbergen Kulkompani A/S & Norsk Hydro A/S, Longyearbyen, Norway. Tyler, N. J. C. & Oritsland, N. A. (1989). Why don't Svalbard reindeer migrate? Holarctic Ecol., 12, 369-76.
?, \~t ' l
Short-term Behavioural Responses of Svalbard Reindeer Rangifer tarandus platyrhynchus to Direct Provocation by a Snowmobile N. J. C. T y l e r Department of Arctic Biology and Institute of Medical Biology, University of Tromso, Breivika, N-9000 Tromso, Norway (Received 15 February 1990; revised version received 20 June 1990; accepted 15 August 1990)
ABSTRACT Short-term behavioural responses of 101 groups of Svalbard reindeer Rangifer tarandus platyrhynchus to direct provocation by a snowmobile were recorded in April (late winter) 1987. The median size o f groups was 3"3 animals. The first visible responses of reindeer to an approaching snowmobile usually involved independent behaviour by different individuals in a group. Flight, by contrast, was a co-ordinated group response. Group median response distances were: minimum reaction distance, 640m; disturbance distance, 410 m; distance at initial flight, 80 m; and distance of flight, 160 m. Group median response times were: total running time, 22 s; total locomotion time, 38 s; maximum duration of disturbance, 193 s. Energy and time budget models indicate that one median flight response can cause an increase in a reindeer's daily energy expenditure o f approximately 0.4% and a loss of daily grazing time also of 0"4%. The study failed to detect any way in which the current level of snowmobile traffic might substantially reduce the physical wellbeing of Svalbard reindeer.
INTRODUCTION Disturbance to wildlife due to industrial development is an important environmental concern in the arctic (Shank, 1979; Martell & Russell, 1985; 179 Biol. Conserv. 0006-3207/91/$03"50© 1991 Elsevier Science Publishers Ltd, England. Printed in Great Britain
180
N . £ C Ty&r
Shideler, 1986; Shideler et al., 1986). Nevertheless, there have been few detailed studies of behavioural responses of reindeer and caribou Rangifer tarandus to specific stimuli, and none at all which investigate their reactions to snowmobiles. This paper describes the short-term behavioural responses of high arctic Svalbard reindeer to direct provocation by a snowmobile.
STUDY AREAS The study was conducted in Nordenskiold Land (78 ° N), Svalbard (Fig. 1) in April 1987. There are about 600 snowmobiles in Svalbard (Berg, 1987), most of which (c. 500) are owned privately by residents of Longyearbyen, the principal Norwegian settlement (population 1500) (Fig. 1). The two roads that lead out of the town have a combined length of approximately 20 km. Besides dog-sleds, snowmobiles are the only means of access to the interior and are important for both mineral exploration and recreation from February until snow melt, which usually occurs in mid-May. Nevertheless, one-third of traffic occurs within 10 km of the town (calculated from data in Persen, 1986). Beyond that, 96% of users normally follow ad hoc snowmobile routes along frozen river beds (Persen, 1986; Fig. 1).
80ON
~ ~'~"
~ J
/
I
"~ z~
I ....... l
~Sassendalen--
"~
'
"! -
L"I
.
.
.
road
- - - - ~-
....il -'~*-=-'-'Adven tdalan' ....
snowmobile route
,,~'...~
.
;
N
"
T I
I
I
I
Fig. 1. Map of Nordenskiold Land in Spitsbergen, largest island in the Svalbard Archipelago, showing the two study areas ('Adventdalen" and 'Sassendalen'), roads and principal snowmobile routes.
Responses of reindeer to provocation
181
There are approximately 4000 Svalbard reindeer R. t. platyrhynchus in Nordenskiold Land but the region is divided by mountains into distinct valley systems between which the animals normally do not move (Tyler & Oritsland, 1989). Each such system contains, in effect, a discrete subpopulation (Tyler, 1987a). Two subpopulations were selected to compare responses of reindeer in areas subjected to different levels of snowmobile traffic: 'Adventdalen' (c. 600 reindeer) and 'Sassendalen' (c. 700 reindeer) (Fig. 1). Adventdalen is the most heavily used area in Svalbard; there is considerably less traffic in Sassendalen. Both valleys are broad, fiatbottomed and bordered by plateaux that rise steeply to 400-500 m above sea level. No trees occur in Svalbard and reindeer (and snowmobiles) are often conspicuous over several kilometres, depending on the terrain. METHODS Provocation Two observers searched for reindeer using two 350 cc snowmobiles. When a reindeer group (with no two individuals > 30 m apart and forming a discrete social unit) was first seen the machines were parked together at a distance of 500-2500 m with engines stopped. The following information was recorded: (i) number of animals, their age (adult, > 24 months; yearling, 22-23 months; calf 10-11 months) and sex; (ii) their principal activity (feeding, lying down); (iii) type of terrain (open, broken). Groups were provoked by driving one snowmobile slowly (20 km/h) until reaching the spot where the animals had been. Reindeer were watched continuously during the approach and a stick was dropped to mark the position on the first occasion that any member of a group showed any visible reaction to the snowmobile. A second stick was dropped on the first occasion that any member of the group showed signs of unease or alarm. A third stick was dropped when the group fled. The 'attacker' then drove slowly back to the starting point and measured five distances to 80 m with the snowmobile odometer: (i) m i n i m u m reaction distance (distance between a group of reindeer and the 'attacker' at the first occasion that any member of the group reacted visibly to his presence); (ii) disturbance distance (distance between a group of reindeer and 'attacker' at the first occasion any member of the group showed any sign of unease or alarm); (iii) distance at initial flight (distance between a group of reindeer and the 'attacker' at the instant the animals first fled) and (iv) total distance (distance between the reindeer and the observer/'attacker' at the start of provocation). Total net distance of flight (v) was measured once the animals had settled. Distances < 80 m were estimated to 20 m by eye.
182
N.J.C. Tyler
The observer watched each group continuously during the provocation to corroborate marking of each distance. He recorded each phase of the animals' flight, including total running time, total locomotion (running and walking) time and maximum duration of disturbance (time from the instant that a group bolted until all members of the group had subsequently settled) to 1 s with a stop watch.
Frequency of disturbance The frequency with which reindeer in Adventdalen were overtly disturbed by normal snowmobile traffic was determined by watching groups of between three and eight reindeer continuously for periods of between 24 and 29h. Individual records were kept for each animal in each group. Observations were made in April 1981 and 1982 during the Easter holidays when traffic is heaviest. Daylight is continuous in Svalbard at that time.
Analysis of data Seven measured responses were compared with respect to six variables: group size, group composition by age and sex, location (Adventdalen vs Sassendalen), activity (feeding vs lying down), terrain (open vs broken) and the animals' view of the 'attack' approach (restricted vs unrestricted). Four types of groups were recognised in the analysis: M groups (adult and yearling males only), F groups (adult and yearling females only), FFC groups (groups of adult females and one or more calves), MX groups (groups including at least one adult male and one adult female). Results for adults and yearlings were combined by sex. Skew was tested using the t statistic (Sokal & Rohlf, 1969). Time or distance parameters were compared between groups or areas using Mann-Whitney U tests (twosample comparisons) or Kruskal-Wallis one way analysis of variance (k sample comparisons) (Siegel, 1956). Association between parameters was measured using Spearman's rank correlation (rs). H o was rejected at p (twotailed) < 0-05 in all tests.
RESULTS Data were recorded from 101 groups of reindeer (336 animals: 71 adult males, 206 females, 2 male yearlings, 1 female yearling, 2 yearlings of unknown sex and 54 calves), 64 groups in Adventdalen and 37 in Sassendalen. The median size of groups was 3.31 reindeer (range = 1-11).
Responses of reindeer to provocation
183
First visible reactions A reindeer's first visible reaction to provocation was normally a m o m e n t a r y glance up at the approaching snowmobile, after which it resumed whatever it had been doing (e.g. feeding, lying). Reindeer normally betrayed unease or alarm either by 'alerting' or by bunching up. Alerting involved one or more animals quickly looking up, ears pricked and rotated forward, and often urinating at the same time. Bunching involved peripheral animals quietly breaking off feeding and walking slowly in towards the group. Those that moved did not usually 'alert', and usually avoided disturbing the group.
Flushing and flight Groups almost invariably fled as a single unit but animals which had alerted or bunched were usually the first to run. The median duration of the first run was 15 s ( r a n g e = 3-110s, n = 67). In most cases (79"3%, n = 69) the group then suddenly stopped, turned a r o u n d and stood staring at the snowmobile (Fig. 2). Animals that had not already urinated often did so at this stage. The median duration of the first pause was 16 s (range = 1-500 s, n = 64). After it, most groups (73-5%, n = 47) either started running again (42.2%) or began to feed (31"3%). The median duration o f the second run was 15 s (range = 3-120 s, n = 27), after which most groups (n = 21) again stopped abruptly, turned a r o u n d and stared at the snowmobile. There was a strong positive correlation between the duration o f the first and second runs (r s = 0"478, n = 27, p < 0"02). After fleeing all reindeer in a group soon resumed feeding or lay down.
Minimum reaction distance (MRD), disturbance distance (DD) The median m i n i m u m reaction distance was 640m and the median disturbance distance was 4 1 0 m (Table 1). Neither M R D nor D D differed significantly with respect to size, composition or location of groups, or with the nature of the terrain or the animals' maintenance activity immediately prior to provocation.
Distance at initial flight (DIF) The median distance between reindeer and the snowmobile when the animals first fled was 8 0 m (Table 1). Reindeer tended to flee at a greater distance in Sassendalen (mean D I F = 148 m) compared to in Adventdalen (mean D I F = l19m). The median D I F (all groups) was the same (80m) in both areas (Table 1) but the rank variance of the data differed between them
N. J. C. Tyler
184
GROUPS PROVOKED
NEXT ACTIVITY: (N = 87)
NEXT ACTIVITY: (N = 6 4 )
I
NEXT ACTIVITY : (N
=
&
START R U N N I N G ~
FEED WALK STAND STOP, TURN & STAND
2.3% 5.7% 6.9%
79.3%
ILIE DOWN ~[WALK IFEED ISTART R U N N I N G AGAIN
42.2%
[WALK ~ISTOP,
TURN & STAND
19.2% 80.8%
[FEED ~[WALK START R U N N I N G AGAIN
n = 4 n = 8 n = 9
1.6%
25.0% 31.3%
26)
NEXT ACTIVITY: (N = 21)
1
NEXT ACTIVITY: (N = 8)
]FEED ~[WALK [STOP, TURN & STAND
n = 1 n : 3 n = 4
Fig. 2. Sequence of responses of groups of reindeer after being provoked into running. The chart shows the proportion (%) of groups ( n = 87) that exhibited one of four different activities (feeding; walking; standing; stop, turn and stand) after first being provoked into running. Thereafter, the chart shows the next activity of as many as possible of those groups that exhibited the most common behaviour at each stage.
185
Responses o f reindeer to provocation
(all g r o u p s , z = 1.95, n~ = 62, n 2 = 37, p = 0.05). A l l - m a l e g r o u p s in S a s s e n d a l e n b o l t e d at a s i g n i f i c a n t l y g r e a t e r d i s t a n c e ( m e d i a n = 240 m) t h a n allm a l e g r o u p s in A d v e n t d a l e n ( m e d i a n = 50 m : z = 2-01, nl = 7, n 2 = 6, p < 0"05). R e i n d e e r t h a t w e r e l y i n g d o w n t e n d e d t o b o l t at a g r e a t e r d i s t a n c e ( m e d i a n D I F = 160 m) t h a n g r o u p s t h a t w e r e f e e d i n g w h e n p r o v o k e d ( m e d i a n D I F = 8 0 m ; all g r o u p s c o m b i n e d , z = 2.03, n~ = 74, n 2 = 18, p < 0"05). R e i n d e e r t h a t h a d a n u n b r o k e n v i e w o f the a p p r o a c h i n g s n o w m o b i l e t e n d e d t o b o l t at a g r e a t e r d i s t a n c e ( m e d i a n D I F = 160 m) t h a n g r o u p s t h a t h a d a r e s t r i c t e d v i e w ( m e d i a n D I F = 80 m ; all g r o u p s c o m b i n e d , z = - 2-35, nl = 56, n z = 31, p < 0-02). T h e difference w a s s i g n i f i c a n t in a l l - m a l e g r o u p s ( u n r e s t r i c t e d view, m e d i a n D I F = 2 0 0 m ; r e s t r i c t e d view, m e d i a n D I F = 5 0 m ; z = - 2 . 1 5 , n 1 = 8 , n 2 = 4 , p < 0 - 0 5 ). TABLE 1 Data Used in an Evaluation of Seven Measured Responses by Svalbard Reindeer to Direct Provocation by a Snowmobile Data are given for all of groups combined with respect to location, except when significant differences by group type occur. FFC groups, see Methods Variables by measured response
Adventdalen
Sassendalen
Both areas
Minimum reaction distance (m) All groups a Median 600 Minimum 50 Maximum 1 840 N 48
640 80 1 520 35
640 50 1 840 83
Disturbance distance (m) All groups b Median Minimum Maximum N
480 80 1 280 35
410 50 1 280 82
400 50 1 280 47
Distance at initial flight (m) All groups b Median 80 Minimum - 10c Maximum 480 N 62 All-male groups b Median 50 Minimum 15 Maximum 240 N 7
80 10 480 37 240 80 480 6
80 -10
480 99 8O 15 480 13 (~ ontbnted~
TABLE 1--contd. Variables by measured response Distance of flight (m) All groups b Median Minimum Maximum N F F C groups b Median Minimum Maximum N Total running time (s) All groups b Median Minimum Maximum N F F C groups b Median Minimum Maximum N Total locomotion time (s) All groups b Median Minimum Maximum N
Adventdalen
Sassendalen
Both areas
160 10 1 300 57
*
160 50 1 440 34
160 10 1440 91
80 10 400 11
*
200 50 1 040 12
80 10 1 040 23
25 9 180 33
22 0 210 94
33 10 90 13
23 5 90 26
45 10 640 29
38 5 890 84
290 65 765 27
193 16 2130 84
275 65 765 12
190 35 2 130 24
21 0 210 61 13 5 60 13
35 5 890 55
M a x i m u m duration of disturbance (s) All groups b Median 185 Minimum 16 Maximum 2 130 N 57 F F C groups b Median 116 Minimum 35 Maximum 2 130 N 12 a Frequency distribution skewed b Frequency distribution skewed c One group (adult female and a snowmobile had driven between * p _< 0"05; ** p < 0"02.
*
to the right, p < 0"01. to the right, p < 0-001. calf) did not start to run until after the and continued 10m beyond them.
Responses of reindeer to provocation
187
Distance of flight ( D F )
Reindeer in Sassendalen tended to flee further (mean D F = 303 m) than reindeer in Adventdalen (mean D F -- 210 m, Table 1). Median D F was the same (160m) in both areas but the rank variance of the data differed significantly between them (all groups, z = 2"05, nl = 57, n 2 = 34, p < 0-05). Groups o f adult females which included one or more calves (FFC) in Sassendalen fled further (median D F = 200m) than FFC groups in Adventdalen (median D F = 8 0 m : z=2.30, n ~ = l l , n2=12, p<0"05). Reindeer that bolted early tended to run farther: distance of flight was weakly but significantly (p < 0.002) correlated with distance at initial flight (Fig.
3).
Total locomotion time, total running time (TRT)
Median total locomotion time was 38 s (Table 1). Reindeer in Sassendalen tended to spend more time running (median TRT = 25 s) than reindeer in Adventdalen (median TRT = 21 s, Table 1) but the difference was significant only in F F C g r o u p s ( z = 1.93, nl--- 13, n 2 13, p = 0"05). ----
M a x i m u m duration o f disturbance ( M D D )
Maximum duration of disturbance was significantly shorter in Adventdalen (median M D D = 185 s) c o m p a r e d t o in Sassendalen (median M D D ----2 9 0 s; all g r o u p s , z = 2.00, n 1 = 57, n 2 = 2 7 , p = 0 - 0 4 6 ) and in FFC g r o u p s (z = 2.34, n I = 12, n 2 = 12, p < 0-02; T a b l e 1). M D D was significantly shorter in o p e n terrain (median = 185 s) c o m p a r e d t o i n b r o k e n terrain (median = 2 9 6 s; all 1500 "E 1200 ,_m ,4-
900
~,
soo
c
•C~ ~
m O0
300
•
• •
•
,
0 0
1 O0
•
,O
• 200
'~ 300
• 400
500
Distance at initial flight (m)
Fig. 3. Distance (m) g r o u p s o f reindeer fled after provocation plotted against the distance between them and the snowmobile at the instant that they fled. Spearman's rank correlation coefficient r s = 0-0335, n = 89, p < 0"002. The correlation remains significant even after the u p p e r right-hand two points are excluded. Overlapping data points are not s h o w n .
188
N. J. C. Tyler
groups, z = 2.28, nl = 51, n2 = 24, p = 0"02) but none of the differences by group type were significant.
Energy cost of disturbance Median total locomotion time after provocation was 38 s, including 22 s running (Table 1). Assuming that reindeer ran at 40 km h - 1, walked at 4 km h-1 and did not break through the snow-crust (which they rarely do in April), the gross cost of locomotion for a 50-kg reindeer during a flight of median duration is 46 kJ (from Nilssen et al., 1984). If the total daily energy expenditure (DEE) of a 50-kg animal in April is 11-7 MJ (Tyler, 1987c), one median flight response represents an increment in DEE of 0"4%. Corresponding values for one maximum and one minimum flight response are 4.7% and 0.01% of DEE, respectively.
Frequency of disturbance During observation of six groups of reindeer (n = 32 animals) over 817 h, in which no continuous record for any animal lasted < 24 h, reindeer were overtly disturbed on only four occasions: (1) three snowmobiles drove by 100 m from four animals which had been lying down for 1"5 h; they stood up and walked slowly away for 20 min before starting to feed for 4.5 h; (2) a single snowscooter drove by 200 m from an adult female, which had been lying down for 30 min; she stood up and began to feed for 40 min before lying down again; (3) three skiers walked 50 m past three reindeer lying on the snow; the animals stood up, walked slowly away for 5 min before starting to feed for 1.5 h before lying down again; (4) thirty minutes later these animals leapt up and stood staring at something beyond the observer's view for 5 min before lying down for 3-5 h.
DISCUSSION Groups of Svalbard reindeer often bunched when provoked. Bunching is an 'anti-predator behaviour' displayed by many species including ungulates (Curio, 1976). Its occurrence is unexpected in Svalbard reindeer because there are no large predators in Svalbard (except for polar bears Ursus maritimus which do not prey on reindeer there). Moreover Svalbard reindeer have been fully protected by Norwegian law since 1925. Residents of the archipelago, however, were granted dispensation to hunt them by the Royal Ministry of the Environment, Oslo, in 1983 and approximately 100 animals have been shot each year since then. Hunting remains forbidden in
Responses of reindeer to provocation
189
Adventdalen. Nevertheless, despite this protection, the reindeer display typical alarm and flight behaviours when disturbed, including bunching and the 'excitation jump' or 'stott' (unpublished data). Altering and bunching in response to provocation by a snowmobile contrasts with the behaviour of barren-ground caribou at the approach of a three-quarter ton truck: these normally fled on sight (Horejsi, 1981). The difference supports Horejsi's (1981) suggestion that it is the rate at which an object approaches, not its size or movement per se, which causes reindeer/caribou to flee: his truck travelled between three and four times faster than the snowmobile in the present study. In further contrast, unlike Horejsi (1981) I recorded no 'no-flight' responses, a reflection of the fact that the snowmobile was driven straight at reindeer until reaching the spot where they had been to start with. This aspect of the experimental design has important consequences for interpretation of results. Svalbard reindeer, like caribou (Horejsi, 1981), evidently distinguish and respond according to the direction at which a vehicle approaches: I was frequently able to drive close ( < 3 0 m ) but obliquely past animals without putting them to flight. Driving straight at them, therefore, must represent a 'worst-case' situation and distance parameters recorded here ought to be regarded as potential maxima, at least with respect to a single snowmobile travelling at moderate speed. So far as assessing potential short-term effects of ordinary snowmobile traffic on the Svalbard reindeer, however, the bias will be conservative. This is corroborated by the four 'real-life' examples of disturbance (see Results), none of which provoked reindeer into running. Disturbance by motor vehicles can have overt effects on animals, such as trauma, and covert effects, such as causing them to avoid otherwise favoured habitat in which regular traffic occurs. This study was concerned only with overt consequences of disturbance. I observed no injuries nor circumstances likely to produce them. Svalbard reindeer are not nomadic; individual animals occupy small, traditional home ranges for weeks or even months at a time in winter (Tyler & Qritsland, 1989). Each animal, therefore, is presumably familiar with its local topography; flight, consequently, is probably rarely a random, stumbling scramble. On the contrary, animals that live in small home ranges are probably capable of instantly selecting the surest route to the safest ground. This may very largely explain how Svalbard reindeer avoid injury when galloping tightly bunched and at full speed over rough terrain, although the small size of groups may also be significant. Miller and Gunn (1979), for example, never observed Peary caribou, which normally live in groups of less than five, injure themselves when fleeing from a helicopter. By contrast, in wildebeest Connochaetes taurinus--a notably gregarious species--several hundreds of animals can be
190
N. J. C. Tyler
drowned or trampled during stampedes (Sinclair, 1979). Some Svalbard reindeer evidently become completely habituated to snowmobiles and other forms of human activity. Four adult males currently live in Longyearbyen for most of the year. One (marked) is known to have been there for nine years. Other males habitually establish summer home ranges around the airstrip where commercial jets land three times each week and where, in summer, there is helicopter activity around the clock. Despite this, and despite regular and increasing snowmobile traffic each winter, the density of reindeer in Adventdalen remains higher than in any other natural subpopulation in Svalbard (Oritsland & Alendal, 1985). Evidence for habituation in reindeer beyond the settlement is less clear. Animals in Sassendalen tended to be more shy than animals in Adventdalen (Table 1). While this reflects different levels of snowmobile traffic in the two areas it need not have been due to different levels of habituation. Herbivore activity is strongly affected by the abundance of food. Foraging time, in particular, increases with decreasing food availability (Jarman & Jarman, 1979; Owen-Smith, 1979; Trudell & White, 1981; Skogland, 1984). Adult females in Adventdalen feed 30 % more per 24-h day in winter than females introduced onto ungrazed pasture at Ny ,~lesund, 100 km further north (Tyler, 1987a, and unpublished data). The density of reindeer in Adventdalen was 50% higher than in Sassendalen when this study was carried out (Tyler, 1987b; N. A. Oritsland, pers. comm.). There is no reason to assume that the per capita supply of food in winter was the same in each area. Consequently, it is not possible to say to what extent differences in responses to a snowmobile between reindeer in the two areas were due to different levels of habituation or differences in nutritional status. Time and energy costs of flight from a snowmobile are surprisingly small. The figures (increments in DEE of 4.7 %, 0.4 % and 0.01% for one maximum, median and minimum flight response, respectively) considerably overestimate the energy costs of 'normal' encounters between snowmobiles and Svalbard reindeer. They are based on responses to highly artificial, direct provocation and, moreover, represent the gross, or instantaneous, energy costs of flight. What matters, so far as the animals' fitness is concerned, is the net cost. Gross and net costs are equal only when animals are unable to compensate for involuntary expenditure of energy by appropriate adjustments in food intake, or are prevented from doing so. Moen et al. (1982) pointed out that changes in animals' behaviour are not a dependable indicator of the effects of disturbance on heart rate, which itself is known to correlate closely with energy expenditure both in Rangifer and other ungulate genera (Nilssen et al., 1984; Fancy & White, 1986). This is a potential limitation of studies of harassment which quantify only sensible responses to disturbance (MacArthur et al., 1979, 1982a, b; Moen et al.,
Responses of reindeer to provocation
191
1982). However, the significance of Moen et al.'s (1982) observation is not clear. Increased heart rate may be a useful indicator of psychological activation but it is not a reliable indicator of oxygen consumption in emotionally excited subjects (Blix et al., 1974; Str~mme et al., 1978). Shideler et al. (1986) suggested that 'the effects of linear facilities and/or off-road vehicles may have important consequences to foraging success of some caribou herds...'. Svalbard reindeer spend approximately 12.9h grazing per 24h during fine weather in late winter: there is no difference between the sexes or between adults and calves in this respect (Tyler, 1987a). The median maximum-duration of disturbance (MDD) following a single provocation by a snowmobile was 193s (Table 1). Thus, assuming all disturbance time represents lost grazing time, then the loss from a single provocation with minimum, median and maximum M D D represents 0.03, 0.4 and 4.6% of daily grazing time, respectively. These values are also likely to overestimate the amount of grazing time that Svalbard reindeer typically lose following an encounter with a snowmobile. The figures are calculated on the basis of the 'maximum duration of disturbance' of each group. They represent, therefore, not the average but the extreme response in each group provoked in this study. Time and energy costs of flight following deliberate provocation are small but, in a marginal environment, the cumulative effect of small costs may have substantial ecological significance. The potential cumulative effect of disturbance can be appreciated by considering the frequency with which reindeer are normally disturbed in late winter in Svalbard. Six groups of reindeer watched continuously for periods > 24 h were disturbed altogether four times, i.e. rather less than one disturbance per group per day at the height of the holiday season. This low frequency reflects the fact that Svalbard reindeer normally inhabit mountain sides and high plateaux in late winter whereas snowmobile routes follow river beds (Tyler, 1987d). Given both the low frequency of disturbance and the low cost of each disturbance, it seems unlikely that snowmobiles may be responsible for substantial increases in energy expenditure or losses of grazing time in Svalbard while the intensity of traffic remains at its present level. This conclusion is contrary to interpretations of results from many previous studies of effects of motor traffic (including aircraft) on reindeer and caribou (see reviews by Shideler (1986) and Shideler et aL (1986)). There are several potential reasons for this. This study considered only immediate, overt responses of Svalbard reindeer to provocation. It ignored entirely all psychological and physiological aspects. But so did all the studies which Shideler (1986) and Shideler et al. (1986) reviewed, so the difference is unlikely to lie here. Alternatively, animals often respond vigorously to motor traffic and it sometimes may be difficult to avoid conclusions being
192
N.J. C. Tyler
influenced by the visual drama of the event. Obviously, galloping is energetically costly. However, the fact that animals gallop when directly provoked by a vehicle tells us little about how normal traffic affects their wellbeing. At the very least we must calculate the gross energy cost of each flight and record how frequently they are put to flight. Perhaps the simplest explanation is that there is no reason to expect Svalbard reindeer to respond to provocation in the same way as other subspecies of reindeer and caribou. Svalbard reindeer are notable for their docile, sedentary behaviour. They normally spend less than 4 % of each 24 h walking and they rarely run (Nyholm, 1976; Kastnes, 1979; Tyler, 1987a): without predators they have no need to. In fact, their activity budget in winter is not substantially different from that of other subspecies (e.g. Skogland, 1984; Boertje, 1985) but caribou and continental wild reindeer live under constant threat of suddenly having to gallop off from biting flies or wolves Canis lupus. Having to escape in a hurry is part of their daily life; for Svalbard reindeer this is not the case.
ACKNOWLEDGMENTS This study was paid for by Store Norske Spitsbergen Kulkompani A/S and Norsk Hydro A/S. I thank Hhkon Stensrud--the provocateur--for assistance in the field and S. D. Albon, R. T. Bowyer, D. R. Klein, J. P. Kelsall, N. Leader-Williams and F. L. Miller for their comments on the manuscript.
REFERENCES Berg, A. (1987). Response from the snowmobile club. Svalbardposten, Longyearbyen, 11, 11. Blix, A. S., Str~mme, S. B. & Ursin, H. (1974). Additional heart rate--an indicator of psychological activation. Aerospace Med., 45, 1219-22. Boertje, R. D. (1985). An energy model for adult female caribou of the Denali herd, Alaska. J. Range Manage., 38, 468-73. Curio, E. (1976). The Ethology of Predation. Springer-Verlag, Berlin. Fancy, S. G. & White, R. G. (1986). Predicting energy expenditures for activities of caribou from heart rates. Rangifer, Special Issue, 1, 123-30. Horejsi, B. L. (1981). Behavioral response of barren ground caribou to a moving vehicle. Arctic, 24, 180-5. Jarman, P. J. & Jarman, M. V. (1979). The dynamics of ungulate social organization. In Serengeti: Dynamics of an Ecosystem, ed. A. R. E. Sinclair & M. NortonGriffiths. University of Chicago Press, Chicago, pp. 185-220.
Response of reindeer to provocation
193
Kastnes, K. (1979). Svalbardreinens aktivitetsm~nster gjennom ~ret (Activity patterns of Svalbard reindeer throughout the year). Cand. real. thesis, University of Oslo. MacArthur, R. A., Johnston, R. H. & Geist, V. (1979). Factors influencing heart rate in free-ranging bighorn sheep: a physiological approach to the study of wildlife harassment. Can. J. Zool., 57, 2010-21. MacArthur, R. A., Geist, V. & Johnston, R. H. (1982a). Cardiac and behavioral responses of mountain sheep to human disturbance. J. Wildl. Manage., 46, 351-8. MacArthur, R. A., Geist, V. & Johnston, R. H. (1982b). Physiological correlates of social behaviour in bighorn sheep: a field study using electrocardiogram telemetry. J. ZooL Lond., 196, 401-15. Martell, A. M. & Russell, D. E. (eds). (1985). Caribou and human activity. Proc. N. Am. Caribou Workshop, 1st, Whitehorse, Yukon, September 1983. Miller, F. L. & Gunn, A. (1979). Responses of Peary caribou and muskoxen to helicopter harassment. Can. Wild. Serv. Occ. Pap., No. 40, 1-89. Moen, A.N., Whitemore, S. & Buxton, B. (1982). Effects of disturbance by snowmobiles on heart rate of captive white-tailed deer. N. Y. Fish & Game J., 29, 177-83. Nilssen, K. J., Johnsen, H. K., Rognmo, A. & Blix, A. S. (1984). Heart rate and energy expenditure in resting and running Svalbard and Norwegian reindeer. Am. J. Physiol., 246, R963-7. Nyholm, E.S. (1976). Monthly general rhythm pattern variations in the Nordenskiold Land-Sabine Land population of the Svalbard reindeer (Rangifer tarandus platyrhynchus). Norsk Polarinst. ,~rbok, 1974, 129-38. Oritsland, N. A. & Alendal, E. (1985). Bestandens storrelse og livshistorie (Population size and life history). In Svalbardreinen og dens Livsgrunnlag (Ecology of Svalbard Reindeer), ed. N.A. IDritsland. Universitetetsforlaget, Oslo, pp. 52-60. Owen-Smith, N. (1979). Assessing the foraging efficiency of a large herbivore, the kudu. S. Afr. J. Wildl. Res., 9, 102-9. Persen, E. (1986). Snascooteren og naturmilj~et (The snowmobile and the natural environment). Typed report to the Governor of Svalbard. (Available from Sysselmannen p~ Svalbard, N-9170 Longyearbyen, Norway). Shank, C.C. (1979). Human-related behavioral disturbance to northern large mammals: a bibliography and review. Report prepared for Foothills Pipe Lines (South Yukon) Ltd, Calgary. Shideler, R. T. (1986). Impacts of human developments and land use on caribou: a literature review, II. Impacts of oil and gas development on the central Arctic herd. Alaska Dep. Fish & Game, Div. Habitat, Tech. Rep., 86-3. Shideler, R. T., Robus, H. H., Winters, J. J. & Kuwada, M. (1986). Impacts of human developments and land use on caribou: a literature review, I. A worldwide perspective. Alaska Dep. Fish & Game, Div. Habitat, Tech. Rep., 86-2. Siegel, S. (1956). Nonparametric Statistics for the Behavioural Sciences. McGrawHill Kogakshua, Tokyo. Sinclair, A. R. E. (1979). The eruption of the ruminants. In Serengeti: Dynamics of an Ecosystem, ed. A. R. E. Sinclair & M. Norton-Griffiths. University of Chicago Press, Chicago, pp. 82-103.
194
N.J. C. Tyler
Skogland, T. (1984). Wild reindeer foraging niche organization. Holarctic Ecol., 7, 345-79. Sokal, R. R. & Rohlf, F. J. (1969). Biometry: The Principles and Practice of Statistics in Biological Research. W. H. Freeman, San Francisco. Str~mme, S. B., Wikeby, P. C., Blix, A. S. & Ursin, H. (1978). Additional heart rate. In Psychobiology of Stress: A Study of Coping Men, ed. H. Ursin, E. Baade & S. Levine, Academic Press, New York, pp. 83-9. Trudell, J. & White, R. G. (1981). The effect of forage structure and availability on food intake, biting rate, bite size and daily eating time of reindeer. J. appl. Ecol., 18, 63-81. Tyler, N. J. C. (1987a). Natural limitation of the abundance of the high arctic Svalbard reindeer. PhD thesis, Cambridge University. Tyler, N. J. C. (1987b). Environmental studies related to petroleum activities on Svalbard. Report from the reindeer project 1987. Report prepared for Store Norske Spitsbergen Kulkompani A/S & Norsk Hydro A/S, Longyearbyen, Norway. Tyler, N. J. C. (1987c). Body composition and energy balance of pregnant and nonpregnant Svalbard reindeer during winter. Syrup. zool. Soc. Lond., 57, 203-29. Tyler, N. J. C. (1987d). Environmental studies related to petroleum activities on Svalbard. Status and distribution of reindeer in Reindalen. Report from the reindeer project 1986. Report prepared for Store Norske Spitsbergen Kulkompani A/S & Norsk Hydro A/S, Longyearbyen, Norway. Tyler, N. J. C. & Oritsland, N. A. (1989). Why don't Svalbard reindeer migrate? Holarctic Ecol., 12, 369-76.