The introduction of reindeer to Br 0 ggerhalv 0 ya , Svalbard : grazing preference and effect on vegetation

In 1978 after about 100 years of absence, 15 Svalbard reindeer, Rangifer tarandus platyrhynchus were reintroduced to Br0ggerhalv0ya, a peninsula on the north-western coast of Svalbard. This stock had increased to about 200 animals in 1989. Studies of reindeer grazing behaviour were carried out in 1979-1980 and 1988¬ 1989. Highly prefered lichen species such as Cetraria nivalis had almost completely disappeared by 1989, whereas a less prefered species, Cetraria delisei, was still abundant. Year round dietary intake of grasses, lichens and herbs e.g. Oxyria digyna, had decreased by 1989, whereas that of mosses had increased. Essentially reindeer showed a more opportunistic grazing behaviour with more balanced utilization of all types of plant associations in 1989. The grazing behaviour of the reindeer on Br0ggerhalv0ya is therefore similar to that observed for Svalbard reindeer in general.


Introduction
Reindeer had been extinct on Brøggerhalvøya, a peninsula on the north-western coast of Svalbard, for about 100 years when 15 reindeer were reintroduced in 1978.By 1989 the population had grown to about 200 animals utilizing approximately 60 km 2 , of potential grazing land.
An initial study of grazing preference was carried out in 1979-1980 (Holand et aL, 1981; Persen et aL, 1983) followed by a comparative study using the same techniques in 1988-89 (Scheie and Grøndahl, 1990).The purpose of the present study was to describe possible changes in food and habitat preference of these reindeer after 10 years of increasing grazing pressure.

Methods
Habitat selection was studied by the pellet group technique (Rogers et aL9 1958;Punsvik et aL, 1980 andPersen et aL9 1983).Ten transects with a combined length of ca. 14 km were localized to cover the most important plant associations on Brøggerhalvøya.At every 20 m along the transects plant associations was described for a one m 2 plot using the categories earlier described by Brattbakk and Senstad (1975), Figs.1-3.Plant coverage was estimated according to the Hult-Sernander scale (Rønning, 1976) and the one m 2 plot was then assumed to be representative of plant associations along the last 20 m of the transect.
In 1980 summer and winter fecal droppings were counted in a 10 m wide belt along the transects, but because of increasing amounts of feces, this belt was narrowed to 5 m in 1989.The same transects were used both years.Samples of 10 summer and 10 winter pellet groups (ue.one defecation) were collected and dried to constant weight.These values were used to calculate total amounts of feces per m 2 in different plant associations.
By directly observing reindeer it was possible to estimate time spent grazing in different plant associations.Based on the percent of total grazing time in each association (U), and the area percent of each association on Broggerhalvoya (A) grazing preference (PR) was estimated according to Jacobs (1974)

Distribution of feces
In both 1980 and 1989 the large accumulation of winter and summer feces were found in moss-tundra and on moraine ridges dominated by lichens, but by the later date fecal distribution had changed (Winter X 2 = 1337, p< 0.001; summer X 2 =1510, p<0.001) and deposits had increased substantially in most other associations (Fig. 1).

Grazing preference
Calculations of a grazing preference index (Eq.1) indicates several shifts in grazing preference from 1980 to 1989 (Fig. 2): Dryadetum associations changing from negative preference to neutral, mosstundra dominated by Tomenthypnum nitens from negative to positive, reduced positive preference for Equisetum fields, Deschampsia fields changed from slightly positive to negative, and also reduced preference for marches.x < 8 S 1 00» i: s s -.

Microhistological analyses
The percentage of moss in the diet increased from 1980 to 1989, both in winter and summer.Salix polaris was heavily used in the summer of both years, but the number of fragments of this plant in feces decreased through the 1980 summer while it remained high in 1989.Its winter percentage was lower in 1989 than in 1980, In

Plant coverage
Based on measurements according to the Hult-Sernader scale total plant coverage was reduced from 1980 to 1989 (Table 1).This reduction was apparently largest for vascular plants, and less for mosses and lichens.
x < 1980 Oxyria digyna, was an important food plant, but not nearly so in 1989.Also, the utilization of grasses decreased from 1980 to 1989 as did the utilization of lichens (Fig. 3).The general trend seems to be that the reindeer have changed their grazing habits from selecting highly prefered plant-associations to a more opportunistic utilization.This information was evidenced by the neutral preference in 1989 for the widly distributed Dryadion association for which 1980 reindeer showed a nega-  Rangifer, 13 (1), 1993 tive preference and by the reduced preferences for moss-tundra, Equisetum dominated associations and march vegetation (Fig. 2).
The findings here agree with conclusions drawn from grazing studies performed during the Man And Biosphere (MAB) program in various other geographical areas of Svalbard; namely that Svalbard reindeer prefer forage plants and grazing habitats of the highest quality, i.e.
plant species high in digestive energy and nutrient content and areas with high plant biomass and relatively flat terrain (Scheie and Grøndahl, 1990;Staaland, 1986;Staaland et aL, 1988).When these qualities are not available the reindeer graze plant associations of lower quality (Staaland;1986).A significant finding on Brøggerhalvøya was the increase in utilization of mosses.This could have resulted from over grazing of the more highly prefered lichens, grasses and herbs and in this context, utilization of mosses could be viewed as a sign of impoverished pastures.Observations elsewhere indicate that not only are mosses more resistent to mechanical destruction by trampling than many other forage plants.They are also reported to be an important food resource for high arctic reindeer and caribou (Parker, 1978).The persistent abundance of the lichen Cetraria delisei may have several explanations.Firstly this species has a low digestibility (Staaland et aL, 1988) and may not be prefered by reindeer, also it may be protected since it is tolerant to being frozen under ice (Hasselrot, 1953).

Conclusion
So far, however, the reindeer on Brøggerhalvøya maintain both a high reproductive rate (63 calves: 69 females counted in July-August) and reproductive longetivity as tagged females 15 and 17 years old produced calves.In addition few animals have been found dead, and only a few young males are reported to have emigrated (Scheie and Grøndahl, 1990).Hence it appears that despite observed changes in vegetation, the reindeer grazing conditions on Brøggerhalvøya are still good.
Fig. 2. Comparison of grazing preference (Jacobs, 1974) of reindeer on Broggerhalvoya in 1980 and 1989.Negative values show negative preference and positive values positive preference.For further explanation see Fig. 1.
Any long term comparative study strive for methodological consistancy and in 1980 we did not foresee the importance of quantifying the occurrence of single vascular species like Oxyria digyna or single lichen species which apparently have measurably declined due to increased grazing pressure.It is now apparent that standard methods for describing plant associations may not be sensitive enough to properly assess the effects of grazing on vegetation.An additional factor which may have confounded study findings is that the goose population on Brøggerhalvøya increased during the 1980-89 period.In the settlement of Ny Ålesund a colony of Barnacle Geese, (Branta leucopsis) became established and also the number of Pink-footed Geese, [Anser fabalis brachyrhynchus) has increased.These species may be competitors of the reindeer as analysis of goose droppings showed 53 % Equisetum fragments, 10 % monocots, 22 % mosses and 14 % forbs (Staaland unpublished).