SR1 Summary.Indd

SR1 A Systematic Review

EviEM Summary 2015

Claes Bernes Kari Anne Bråthen Bruce G. Forbes Impacts of reindeer on James D. M. Speed Jon Moen arctic and alpine vegetation Impacts of reindeer on arctic and alpine vegetation Summary of Systematic Review SR1

EviEM, 2015

Please cite this report as C. Bernes et al. (2015): Impacts of reindeer on arctic and alpine vegetation. Summary of Systematic Review SR1. EviEM, Stockholm.

Printed by US-AB, Stockholm, 2015

Cover: Reindeer in Sarek National Park, northern Sweden Photo: Oskar Karlin

Impacts of reindeer on arctic and alpine vegetation Summary of Systematic Review SR1

By Claes Bernes, Kari Anne Bråthen, Bruce C. Forbes, James D.M. Speed and Jon Moen SR1 | Impacts of reindeer on arctic and alpine vegetation

In alpine parts of the Scandinavian mountains, reindeer are the most numerous large herbivores. Photo: Claes Bernes.

Summary The impact of reindeer on vegetation in arctic and trampling, and herbs are also reduced by expos- alpine areas is not fully known. This hampers both ure to reindeer. The occurrence of other kinds of the management of wild and semi-domesticated vegetation, notably mosses, grasses and shrubs, is reindeer populations and the preservation of bio- on average not signifi cantly affected by reindeer diversity. To synthesise the available evidence, we grazing, although individual studies may show have undertaken a systematic review of published signifi cant effects. The species richness of vascular studies that compare vegetation exposed to dif- plants decreases with increasing grazing pressure, ferent levels of reindeer grazing. We found a total at least in cold climates, where the productivity of of 41 studies that fulfi lled all our relevance and vegetation tends to be low. quality criteria. Almost two thirds of these studies had been conducted in Fennoscandia. There is not suffi cient information to explain the considerable variation that we found among Overall, the studies report on large variations in results from different studies and different sites. the response of vegetation to reindeer. Average The outcomes are context-dependent, which sug- effects are usually small, but there are also both gests that research and management have to take signifi cantly positive and signifi cantly negative local conditions into account, but the evidence is effects within all vegetation categories considered. too weak and scattered to inform environmental policy or reindeer management in a detailed way. Nevertheless, when we combined data from dif- It is unlikely that further studies can improve the ferent studies using meta-analysis, some patterns evidence base much, unless they adopt standard- emerged. Lichens are vulnerable to grazing and ised ways of quantifying reindeer impacts.

4 Impacts of reindeer on arctic and alpine vegetation | SR1

Reindeer play key role foraging is made diffi cult by ice or deep snow, and in northern ecosystems herd sizes can therefore vary considerably from one decade to another. In Sweden, the number of Reindeer (Rangifer tarandus L.) have a natural reindeer has fl uctuated between about 150,000 range extending over much of northernmost Eur- and 300,000 over the last 125 years. asia and North America. In considerable parts of this region, they are the only large herbivores. Being the most numerous large herbivores in the Wild reindeer representing a subspecies known far north, reindeer often play a key role in their as caribou are still numerous in Canada, Alaska ecosystems. Unlike most other herbivorous spe- and Greenland. Another subspecies, the Sval- cies, they are able to digest lichens, which makes bard reindeer, occurs in the wild on Spitsbergen them highly adapted to the vegetation of the tun- and other islands in the Svalbard archipelago. In dra. Reindeer lichens (Cladonia and Cladina), a northern Europe and Siberia, however, the ma- prominent part of the ground vegetation in many jority of reindeer populations have been domest- polar areas, are a major component of the winter icated or semi-domesticated for several centuries. diet of most reindeer populations. In summer, Here, they are to a large extent being herded by reindeer prefer green, vascular plants such as indigenous peoples. Some large populations of graminoids (grasses, sedges and rushes), herbs, wild reindeer are still present in Russia, but most and leaves of shrubs and deciduous trees. Never- of the Norwegian and Finnish herds and all of the theless, since lichens are sensitive to trampling Swedish ones are semi-domesticated. under snow-free conditions, they can be affected by reindeer in summer pastures too. Over the seasons, many reindeer herds migrate over large distances between summer and winter Many studies of tundra ecosystems have shown pastures. Reindeer in Sweden normally spend the that intensive grazing and trampling by reindeer snow-free season on alpine tundra or in subalpine can cause lichens to be replaced by vegetation birch forests, whereas they spend the winter in dominated by bryophytes (mosses and liver- lowland coniferous forests. During some winters, worts). If the grazing pressure is increased further,

Reindeer numbers in Fennoscandia

Sweden 300,000

200,000 Norway

This diagram shows total 100,000 populations of semi- domesticated reindeer in Finland Sweden, Norway and Fin- land following the autumn slaughter. After calving in spring, herds are signif- 0 icantly larger. 1880 1900 1920 1940 1960 1980 2000

5 SR1 | Impacts of reindeer on arctic and alpine vegetation

bryophytes may be replaced by grasses and other Shifting opinions on the impacts graminoids. Changes of the latter kind have been of reindeer grazing observed in experiments or where reindeer often crowd together along fences and other obstacles to In Sweden, public opinion on how reindeer their migration, but not in studies of rangelands grazing affects alpine vegetation has shifted with freely roaming semi-domesticated reindeer. during the last few decades. In the 1990s, several

If reindeer ranges are enclosed by fences, grazing and trampling along the fence may become so intense that vege- tation is strongly impacted. This picture was taken at Ifjordfjellet in Finnmark, Norway. Photo: Kari Anne Bråthen.

6 Impacts of reindeer on arctic and alpine vegetation | SR1

well-publicised records of vegetation degrada- Moreover, there is still no scientifi c consensus tion in areas grazed by reindeer helped to form on how important reindeer really are for the a widespread perception that some mountain diversity of alpine or arctic vegetation. While areas were overutilised. There was a concern some studies have found the impacts of reindeer that Swedish reindeer husbandry was not sus- grazing to be strong, others suggest that they tainable. In other parts of Fennoscandia, severe are relatively minor, at least in comparison to overexploitation of reindeer ranges was noted those of other grazers (such as voles, lemmings during these years, not least on lichen heaths or insects). in Finnmark in northernmost Norway and in Finnish Lapland. Need for a synthesis of the available evidence More recently, however, the impact of reindeer grazing on alpine vegetation was subject to Since the documented impacts of reindeer on re-evaluation in Sweden. Analyses of reindeer vegetation differ between studies and between numbers and grazing effects indicated that the regions, it is not easy to predict the ecological fears of overgrazing were based on local damage consequences of various forms of management of around a few enclosures and fences. No evidence reindeer populations, whether wild or domestic- of large-scale overutilisation of reindeer ranges ated. This suggests a need to synthesise and as- in the Swedish mountains could be found. The sess all the available evidence on how vegetation current consensus is that overgrazing of Swedish is affected by reindeer, and to identify knowledge reindeer ranges has been temporary and local, gaps that could be fi lled in by further research. and that it has rarely caused permanent damage. Recent evidence from Finnmark’s winter range- In Sweden, the recent re-evaluation of what lands points to a similar conclusion. reindeer grazing means for arctic and alpine ve- getation makes an examination of the scientifi c A common view today is that, on the whole, reindeer grazing benefi ts the montane vegeta- What is a systematic review? tion and enhances its species richness. There are indications that even heavy grazing during In this review, we used a systematic ap- a limited period can promote the diversity of proach to synthesise available evidence on alpine vegetation in the long run. Fears have ac- the impacts of reindeer herbivory. System- tually been expressed that parts of the mountain atic reviews are entirely based on existing region will become overgrown with shrubs be- studies – in this respect, they do not differ cause of inadequate grazing pressure, although from ordinary literature reviews of scientifi c the number of reindeer in Sweden is not much questions. The difference lies, instead, in the smaller now than it was around 1990. way of working. A systematic review is char- acterised by meticulous planning, methodical The current attitude to reindeer grazing is refl ec- procedures and a transparent, objective and ted in an environmental quality objective for the complete documentation of all assessments mountains adopted by the Swedish Parliament. carried out in the course of the work. This One of the specifi cations of this objective de- approach is designed to increase reliability clared that it is essential to preserve ‘a mountain and repeatability, avoid bias and permit landscape characterised by grazing’. However, meta-analysis (quantitative conclusions no details were given on how this was to be in- based on data from several different studies). terpreted in ecological terms.

7 SR1 | Impacts of reindeer on arctic and alpine vegetation

support for today’s prevailing opinions on the is- The primary aim of the review was to clarify how sue particularly relevant. The decision to launch grazing, browsing and trampling by reindeer the present review was also motivated by the affect the vegetation of arctic, subarctic, alpine need to interpret and clarify the environmental and subalpine areas (but not coniferous forests). quality objective for the Swedish mountains. The review was designed to include studies from

km

0 250 Lagisduoddar Vestre Seiland Jakobselv Komag- Jeahkkir elva Joahko- njárga Utsjoki Joatka Čearro Raisduoddar Karasjok Jávrioaivit Finnmarksvidda Malla Kautokeino Saana Siččajávri Muotkatakkavaara Jávrresduottar

Vássijávri Vággečohkka R U S I A Geadgečorro Tavvavuoma Boazuvárri Abisko

Ritsem

N O R W A Y Arctic circle

F I N L A N D S W E D E N

Buckland Valley Alaska Toolik Lake

Handölsdalen Daring Lake Snøhetta Rondane C A N A D A Sonfjället Nord-Ottadalen Långfjället + R U S S I A North pole

Fulufjället Yamal Nordfjella Greenland Kangerlussuaq Svalbard Rivière George

Arctic circle These maps show sites or reindeer ranges in Fennoscandia Fennoscandia (top) and the entire circumpolar km region (right) where studies included in this 0 2500 review have been conducted.

8 Impacts of reindeer on arctic and alpine vegetation | SR1

Ground lichens are known to be particularly sensitive to trampling by reindeer. The alpine lichen heath shown here is located in Jotunheimen in southern Norway. Wild reindeer can be found in this area, but the population is relati- vely small. Photo: Mirjam Ekstedt. any such area where reindeer were present, but compared herding districts with different num- to focus on conditions in Fennoscandia. Of the bers of reindeer, but most studies were designed 41 studies that our review was fi nally based as experiments where small areas had been on, 25 had been conducted in Norway, Sweden fenced in to exclude reindeer entirely; compar- or Finland. The others had been carried out in isons were then made of vegetation inside and Svalbard, Greenland, Canada, Alaska or Russia. outside the fences.

The majority of included studies compared areas A few studies were instead based on vegetation where the vegetation had been subject to dif- sampling in a region before and after an increase ferent levels of reindeer grazing. Some of them or decrease of reindeer numbers.

9 SR1 | Impacts of reindeer on arctic and alpine vegetation

No. of No. of Summary of vegetation responses to reindeer herbivory comparisons studies

Herbs/forbs, cover 18 5 Herbs/forbs, biomass 5 2 Herbs/forbs, abundance 14 4

Herbs/forbs, overall effect (cover/biomass/abundance) 40 12

Sedges/rushes, abundance 10 1 Sedges, cover 10 1 Sedges, biomass 5 2 Sedges, abundance 2 1 Grasses, silica-rich, biomass 2 1 Grasses, cover 13 1 Grasses, biomass 5 2 Grasses, abundance 13 3 Graminoids, cover 9 6 Graminoids, biomass 2 2 Graminoids, abundance 2 2

Graminoids, overall effect (cover/biomass/abundance) 75 13

Trees, cover 8 1 Shrubs, evergreen, biomass 3 1 Shrubs, evergreen, abundance 10 1 Shrubs, deciduous, cover 13 1 Shrubs, deciduous, biomass 5 2 Shrubs, deciduous, abundance 13 3 Shrubs (non-forage), cover 2 1 Ericoids, deciduous, abundance 10 1 Dwarf shrubs, evergreen, abundance 2 1 Dwarf shrubs, cover 4 1

Woody plants, overall effect (cover/biomass/abundance) 83 12

Lichens, cover 46 12 Lichens, biomass 5 3 Lichens, abundance 3 2

Lichens, overall effect (cover/biomass/abundance) 54 15

Liverworts, cover 18 2 Liverworts, abundance 2 1 Bryophytes, cover 21 6 Bryophytes, abundance 3 2 Bryophytes, overall effect (cover/abundance) 46 10

Bare ground, cover 34 7

Vascular plants, species richness 31 7

Lichens, species richness 3 2 Bryophytes, species richness 4 2 Non-vascular cryptogams, species richness 15 2

Non-vascular plants, species richness 22 4

Standardised mean difference between high- and low-exposure data −6−5 −4 −3 −2 −1 0 1 2 3 4 5

Decrease with increasing exposure to reindeerNo effect Increase with increasing exposure to reindeer

10 Impacts of reindeer on arctic and alpine vegetation | SR1

Small average effects, but large Several possible reasons variation from case to case for the variation in the results

Overall, the studies that we analysed report Why are our results not as clear and conclusive highly varied responses of vegetation to reindeer. as one might have wished? One reason may be The average effects are usually small, and a that the grazing pressure differs considerably majority of the individual results show no signi- between studies and regions. If this is an im- fi cant response, but there are also both signifi c- portant explanation to the variation in the res- antly positive and signifi cantly negative effects ults, we may expect to fi nd vegetation responses within all vegetation categories considered. and reindeer densities (the number of animals Even within a single study, a certain group of per unit area) to be signifi cantly related. plants may show a positive response to reindeer exposure at some sites, but a negative response Information about reindeer densities is not read- elsewhere. Nevertheless, when we combined data ily available, though. A few authors have estim- from different studies using meta-analysis, some ated local densities based on counts of sighted patterns emerged: animals, but most of the studies included in this review merely compared vegetation at ‘high’ and • We were able to corroborate the fi ndings ‘low’ grazing pressures. What was regarded as within many studies that lichens are vulnerable high in one study may have been considered low to grazing and trampling by reindeer. in another. In the absence of quantitative local • Bryophytes were not found to be vulnerable to data, we calculated average reindeer densities at reindeer exposure. regional levels (e.g. in entire reindeer-herding districts). Such data are inevitably coarse and • We saw no signifi cant overall effect of possibly misleading, however, and we found no reindeer on graminoids (grasses and similar signifi cant relation between vegetation re- growth forms), indicating that these supposedly sponses and estimated reindeer densities. grazing-tolerant plants are not always promoted by reindeer grazing. Nonetheless, the largest changes of lichen cover • We found that herbs (forbs) are reduced by ex- clearly occurred at sites with few reindeer and posure to reindeer, although they usually do not large amounts of lichens. At sites with more occur very frequently in arctic and alpine areas. reindeer, the average lichen cover was very lim- This indicates that herbs are highly selected by ited (perhaps as a result of grazing or trampling reindeer and also vulnerable to grazing. before the studies began), and it usually did not respond much to changes of the grazing pres- • We did not fi nd strong evidence that shrubs sure. This result is an indication that the effects generally are reduced by reindeer. of reindeer herbivory on vegetation are highly • The species richness of vascular plants (which dependent on how the vegetation is composed. include graminoids and herbs as well as shrubs and other woody plants) was found to decrease The global distribution of reindeer covers large with increasing grazing pressure. environmental gradients, ranging from low-pro-

The diagram to the left summarises meta-analyses of how vegetation responds to reindeer herbivory. Positions and widths of the diamond-shaped symbols indicate means and 95% confi dence intervals, respectively. Symbols not spanning the green zero line represent statistically signifi cant effects. Responses are given as standardised mean differences (the difference between mean vegetation properties measured at high and low exposure to reindeer, divided by the pooled standard deviation).

11 SR1 | Impacts of reindeer on arctic and alpine vegetation

ductivity to high-productivity sites, from dry to One reason why average vegetation responses wet soils, and from forests to high alpine tundra. were generally found to be small may be that All different vegetation types that occur here nearly all of the data in our meta-analyses were cannot be expected to respond in the same way sampled in summer ranges. The size of reindeer to grazing and trampling. For instance, it is well herds is usually limited by food resources in the known that dwarf-shrubs (such as crowberry, winter ranges, whereas summer ranges offer heather and dwarf birch) and other woody spe- more forage than the reindeer can use. Their cies have highly varying palatability to reindeer average consumption in summer has been es- and therefore are likely to be differentially affec- timated to amount to 2% or less of the total plant ted by grazing. growth, and such levels are unlikely to cause large changes in the vegetation. Reindeer do not We did investigate how factors such as soil mois- use the landscape randomly, however, and the ture and habitat type infl uenced the impact of grazing pressure may be much higher than aver- reindeer on various groups of vegetation, but no age in preferred sites and vegetation types. consistent patterns emerged. Implication for policy Variations in management systems may be and management another reason for the heterogeneity of our res- ults. Reindeer husbandry operates under many Our review has gathered a large body of data on different forms of legislation, market situations vegetation responses to reindeer grazing, but we and historical legacies, and its ecological con- still have to conclude that the evidence base is sequences may therefore vary strongly between too weak and scattered to inform environmental different herding districts and systems. policy or reindeer management in a detailed way. The available studies are not numerous, extens- We also found indications that impacts of wild ive and representative enough that they can be reindeer on the vegetation cover may differ from used as a basis for specifi c recommendations those of semi-domesticated herds, in some cases regarding reindeer ranges in Sweden, nor in the by being stronger. circumpolar region as a whole. Summarising

Temperature dependence of reindeer impacts on the species richness of vascular plants

●

1.0 ●

●

● ● ● 0.5

● ● ●

● ● ● 0.0 ● ● ● ● ● ● ● ● ● ● ● The effect of reindeer grazing on the species richness of −0.5 ● ● ● ● vascular plants was found to depend on the mean annual ● ● temperature. This could mean that exposure to reindeer −1.0 ● tends to increase the number of species at productive sites, which usually occur in relatively benign climates, Standardised mean difference between Standardised mean difference

−1.5 ● but decrease it at colder and less productive sites.

high- and low-exposure vascular species richness Each symbol in the diagram represents one site (or −4 −3 −2 −1 0 1 2 Mean annual temperature ( °C) pair of sites). Symbol sizes are proportional to the preci- sion of the measured response.

12 Impacts of reindeer on arctic and alpine vegetation | SR1

results across all studies in our review was also made diffi cult by a lack of standardised design. EviEM For instance, we found 23 different kinds of data The Mistra Council for Evidence-Based on quantities of woody plants. Environmental Management (EviEM) strives to ensure that environmental management in However, an important point that we can make Sweden is informed by the best possible sci- is that vegetation responses to reindeer grazing entifi c evidence. Through systematic reviews are context-dependent. The variation of effects of relevant research, we aim to improve the found in this review is most likely due to the basis for decisions in environmental policy. variation of vegetation types, productivity and Funded by the Swedish Foundation for Stra- grazing history, just to mention a few important tegic Environmental Research (Mistra) and factors. This suggests that no single conclusion hosted by the Royal Swedish Academy of or recommendation on reindeer and their impact Sciences, EviEM is fi nancially and politically on vegetation can be universally applicable. independent.

How this review was conducted

Originally proposed by the Swedish Environ- Sweden were invited at an early stage to suggest mental Protection Agency, this systematic review modifi cations of the scope and focus of the re- has been initiated and fi nanced by the Mistra view. The fi nal design of the review was estab- Council for Evidence-Based Environmental lished in detail in a protocol that was published Management (EviEM). The review was conduc- in March 2013. ted as a project by a specially appointed team of researchers chaired by Jon Moen, professor We started the review by searching for relevant of ecology at Umeå University. The project was literature using four different kinds of sources: managed by Claes Bernes, EviEM. Stakehold- 17 online publication databases, search engines ers with an interest in reindeer husbandry or such as Google and Google Scholar, websites of environmental aspects of reindeer herbivory in agencies, institutes and organisations with an

The team that conducted this review included (from left to right) James D.M. Speed, Jon Moen (chair), Bruce G. Forbes, Kari Anne Bråthen and Claes Bernes (EviEM project mana- ger). Photo: Bo Söderström.

13 SR1 | Impacts of reindeer on arctic and alpine vegetation

interest in reindeer or the arctic environment, data that were diffi cult to interpret. This kind and bibliographies of selected literature reviews. of critical appraisal is an essential part of the The majority of the more than 6,000 articles systematic review process. found during these searches could be excluded as irrelevant based on their titles or abstracts, The 40 fi nally selected articles describe a total of but we studied 415 potentially useful articles in 41 individual studies. All but two of the articles full text and judged 100 of them to be relevant. had been published in 2000 or later. We had When we assessed the quality of these articles, searched for publications in Swedish, Norwe- however, 60 of them were excluded. Common gian, Finnish and Russian as well as English, but reasons for exclusion were inadequate method- the fi nal selection included only one non-English ological descriptions, and vegetation or reindeer article (written in Norwegian).

Overview of literature searches and screening of articles

6,638 articles found 6,020 articles excluded by full-text searches during screening on title in 3 article databases

772 articles found by abstract searches 1,395 articles included 198 duplicates removed in 14 article databases

1,197 articles 821 articles excluded screened on abstract during screening on abstract

39 articles found using search engines, specialist websites, reference lists in literature reviews or 415 articles 319 articles excluded stakeholder contacts screened on full text during full-text screening

4 recent articles added Critical appraisal of 60 articles excluded because when the main search 100 articles included data reported were redundant was updated after full-text screening or susceptible to bias

40 articles providing useful data from 41 studies included in the review

Statistical analysis of data from 31 studies

14 Impacts of reindeer on arctic and alpine vegetation | SR1

Numbers of extracted comparisons

Abundance Biomass Cover Species Shannon Simpson richness diversity diversity

Total vegetation 12 (4) 18 (4) 15 (2) 17 (3) Bare ground 34 (6) Vascular plants 5 (2) 14 (2) 21 (5) 19 (3) 15 (2) Shrubs, deciduous 13 (3) 5 (2) Herbs/forbs 14 (4) 8 (4) 20 (5) Graminoids 2 (2) 4 (3) 11 (7) Grasses 13 (3) 5 (2) Sedges 5 (2) Cryptogams, non-vascular 15 (2) 15 (2) 15 (2) Lichens 3 (2) 7 (5) 50 (14) 5 (2) Bryophytes 3 (2) 21 (6) 5 (2) Mosses 26 (6) Liverworts 20 (2)

Shown here are the numbers of comparisons of different vegetational groups that were extracted from studies inclu- ded in this review. Figures in brackets indicate the number of studies that have contributed with data. Data are shown for groups and aspects of vegetation covered by at least 2 studies.

We extracted quantitative results from 35 of Implication for further research the studies. These data include 455 individual comparisons of vegetational groups (such as As concluded above, the evidence on how ve- herbs, graminoids or lichens) that had been getation in arctic and alpine areas is affected by subject to different grazing pressures. They reindeer herbivory is still quite limited. It is un- also include nearly 1,700 similar comparisons likely that further research will be able to greatly of single vegetation species, but we did not improve the evidence base unless a more rigor- analyse these since an individual species was ous research protocol, specifi cally aimed at the typically not covered by more than 1–3 studies. questions raised in this review, becomes gener- The comparisons are mainly based on measures ally accepted and implemented. Such a protocol of vegetation quantities, such as abundance might, for instance, recommend standardised (number of specimens per unit area), biomass ways of reporting data on vegetation, common or cover, but they also include data on species measures of grazing pressure, detailed descrip- richness (number of species) and other measures tions of study design and statistical analysis, and of species diversity. careful documentation of external factors at each study site, such as the history of land use and The meta-analyses that we fi nally conducted reindeer husbandry. were aimed at determining the average effects of reindeer herbivory on various vegetational Free access to full report groups based on comparative data from all use- ful studies. Studies that were based on a large The full report on this systematic review is number of observations and were therefore able published in the Environmental Evidence to present relatively reliable data were given journal (www.environmentalevidencejournal. greater weight in the meta-analyses than less org/content/4/1/4). The report is also available at extensive studies. EviEM's website (www.eviem.se).

15 EviEM Summary 2015

Are reindeer harming the mountain vegetation of Northern Scandinavia? Or is their grazing, on the contrary, a precondition for the diversity of the montane fl ora? Opinions on this issue have been divided over the years, but in the systematic review that is summarised here, Mistra EviEM has attempted to achieve unequivocal answers.

www.eviem.se

EviEM conducts systematic reviews of environmental issues identifi ed as important by public agencies and other stakeholders. These provide an overall assessment of the state of scientifi c knowledge and help to improve the basis for environmental decision-making in Sweden.

Mistra EviEM Royal Swedish Academy of Sciences Box 50005, SE-104 05 Stockholm, Sweden Visit/Deliveries Lilla Frescativägen 4A, SE-114 18 Stockholm Telephone + 46 8-673 95 00 E-mail [email protected]