Electronic Scientific Paper Archive

Effects of grazing on plant community structure and aboveground net primary production of semiarid boreal steppe of northern Mongolia

Ariuntsetseg Lkhagva1,2, Bazartseren Boldgiv2, Clyde E. Goulden3, Oyunchuluun Yadamsuren4 and William K. Lauenroth1
1 Department of Botany, University of Wyoming, Laramie, Wyoming, USA
2 Department of Ecology, National University of Mongolia, Ulaanbaatar, Mongolia
3 Asia Center, Academy of Natural Sciences of Drexel University, Philadelphia, Pennsylvania, USA
4 School of Agricultural, Forest and Environmental Sciences, Clemson University, Clemson, South Carolina, USA


We studied the effects of grazing on plant community structure and total plant biomass across the landscape while taking account of nutrient gradient, from wet and nutrient-rich sites (north-facing steppe) to dry and nutrient-poor sites (south-facing steppe) in semiarid steppe of northern Mongolia. Livestock grazing increased species richness of wet and nutrient-rich sites, while no significant change was observed in dry and nutrient-poor sites. The species richness increase in the wet and nutrient-poor sites was explained by local colonization of grazing-tolerant species. Species that adapted in the wet and nutrient-rich sites were driven to local extinction as a consequence of competitive exclusion. At a large spatial scale, livestock grazing can have a potential negative effect on a regional species pool, as it excludes species adapted in wet and nutrient-rich sites. Although grazing did not affect species richness in the dry and nutrientpoor south-facing steppe, plant communities under grazing shifted to dominance by short and prostrate forb species. A spatial difference of the total plant biomass across the landscape was higher in non-grazed landscape but this difference lessened in grazed landscape. The greatest percentage reductions of the total plant biomass due to grazing were in wet and nutrient-rich sites. In conclusion, at a community level, plant communities responded differently to livestock grazing, but at an ecosystem level, the total plant biomass decreased under grazing across the landscape of the semiarid boreal steppe. The livestock grazing in the wet and nutrient-rich sites resulted in the disappearance of moss cover, which is the main insulator of permafrost. The loss of moss cover could potentially accelerate a thawing of permafrost and warming of this region. Also, we found different results on dominance of Artemisia frigida Willd. from those
reported in the steppe of Inner Mongolia. At a regional scale, this species might not be considered as an indicator species of livestock grazing.

Keyword: Landscape,livestock grazing,nutrient gradient,plant biomass,plant community structure


Grazing is an inherent part of many arid and semiarid grassland ecosystems. A consistent pattern of grazing effect on vegetation structure and net primary production (NPP) has not been documented in arid and semiarid grassland ecosystems (McNaughton 1985; Milchunas and Lauenroth 1989; Biondini et al. 1998; Fernandez-Gimenez and Allen-Diaz 1999; Oba et al. 2001; Osem et al. 2002; Cingolani et al. 2003; Derner and Hart 2007; Han et al. 2008). Milchunas and Lauenroth (1993) found that ecosystem- environmental variables (e.g. evolutionary history of grazing and environmental moisture) hold particular importance for the effect of grazing on vegetation structure and NPP, rather than grazing variables such as number of grazing animals and grazing intensity; the smallest effect of grazing on plant community structure and NPP has been observed in dry ecosystems with long evolutionary history of grazing because plant avoidance and tolerance characteristics for water and grazing stress are evolutionarily convergent in these ecosystems. Similarly, syntheses and models have emphasized the importance of environmental gradients of soil fertility and precipitation on how grazing affects the plant diversity (Olff and Ritchie 1998). An increase in species richness due to grazing has been documented in wet and nutrientrich ecosystems (Belsky 1992; Collins et al. 1998), while negative effect of grazing on species richness is more common in dry and nutrient-poor ecosystems (Hobbs and Huenneke 1992). Vegetation structure varies across the landscape through differential resource availability (Schimel et al. 1985; Burke et al. 1999; Hook and Burke 2000; Cingolani et al. 2003; Casper et al. 2012). Responses of different plant communities to grazing differ across the landscape (Sala et al. 1986; Milchunas et al. 1989; Milchunas and Lauenroth 1989). Grazing can amplify the difference between plant communities (Anchorena and Cingolani 2002), or converge them (Fuhlendorf and Smeins 1999; Adler et al. 2001). In accordance with the generalized model of the effects of grazing on grassland communities by Milchunas et al. (1988), grazing affects the plant community composition of wet and nutrient-rich ecosystems more intensely than dry and nutrient-poor ecosystems. Effects of grazing on plant community structure and NPP of ecosystems have been recognized by ecologists but few attempts addressing the combined effects of both grazing and nutrient gradient at the landscape level have been done (Proulx and Mazumder 1998; Austrheim and Eriksson 2001; Olofsson et al. 2004). We studied the effects of grazing on plant community structure and the total plant biomass across the landscape while taking account of nutrient gradient, from wet and nutrient-rich sites (north-facing steppe) to dry and nutrient-poor sites (south-facing steppe) in semiarid boreal steppe of northern Mongolia. We assessed the total plant mass, because harvests of aboveground biomass have been frequently used to estimate annual net primary production of steppe ecosystems with fast turnover (Sala and Austin 2000). The research aims (i) to characterize the combined effects of grazing and nutrient gradient at the landscape level for the general knowledge of grazing and environmental resource relationship; (ii) to answer the question of how livestock grazing affects plant community structure, and how total plant biomass of landscape positions differed by moisture and nutrient in this semiarid boreal steppe of northern Mongolia? About 83% of Mongolia is semiarid grasslands or desert (White et al. 2000). The grasslands have been occupied by nomadic pastoralists for centuries and have evolved with a long history of grazing by large wild mammal herds. These grasslands have experienced high livestock grazing pressure, especially, for the last 80 years. Nowadays, Mongolian grasslands support 50 million head of domestic livestock and large populations of wild herbivores. Relatively little is known about vegetation and dynamics of Mongolian grasslands and few experimental and observational studies have been conducted to determine the effects of grazing on ecosystem productivity or species richness. The main conclusions from the past research suggest that steppe pastures are resilient and recover rapidly when grazing pressure is relaxed or removed (Fernandez- Gimenez and Allen-Diaz 1999, 2001). In accordance with natural zonation, northern Mongolia is a boreal steppe ecosystem on the basis of ecological conditions, geographical boundaries and landscape patterns (Tsogoo 1990; Lhagvajav 1992; Tserendash and Erdenebaatar 1993; Chognii 2001) and its landscape pattern is more complex than other parts due to its higher elevation, deep valleys with some forest and arid steppe and permafrost distribution (Barzagur 2002). This complex landscape pattern within one valley results in significantly different vegetation communities, different livestock grazing pattern and difference of plant communities in their response to grazing.

Material and Methods

Study site
The study site is located in the Lake Hovsgol watershed occupying a north-south tectonic basin at the southern end of the Baikal Rift System in northern Mongolia (51°1′26.8′′N, 100°45′41.2′′E) (Figure 1).

This region, a part of the Mongolian Long Term Ecological Research network, represents the southern boundaries of the Siberian boreal forest and of continuous permafrost; and a biogeographic and ecological transition zone from taiga forest to steppe and from Central to East Asia. Six valleys along 50 km of the northeastern shore of the lake form the study sites. They range from grazed northern valleys just south of Hanh soum (soum is an administrative unit of Mongolia) center, to southern valleys that are part of a protected zone of Hovsgol National Park. The six valleys allow us to define the effects of grazing, heavy in the northern valleys, Turag (TRG) and Shagnuul (SHL), moderate in the middle two valleys, Noyon (NYN) and Sevsuul (SVL), and no grazing in the southern two valleys, Dalbay (DLB) and Borsog (BRG) (Figure 2). The grazing density was 45 livestock km -2 in TRG, 67 livestock km-2 in SHL, 8 livestock km-2 in NYN and 10 livestock km-2 in SVL. The valleys are asymmetrically shaped with a steep south-facing slope, and a less steep forested north-facing slope. The south-facing slope is exposed to sunlight throughout the year and water is the most limiting factor for plants. The north-facing slope receives limited sunlight during the winter and part of the summer. The slope and aspect affect the spatial configuration of the landscape and result in different vegetation types and soil conditions across the landscape (Table S1). Ridge tops and upper parts of north-facing slopes are predominantly covered by Siberian larch (Larix sibirica Ledeb) forest; valley bottoms and south-facing slopes are open grasslands dominated by grasses, sedges, forbs and shrubs. The climate is continental with warm summers and cold winters. Mean annual temperature is _4.5°C. Average annual maximum air temperature is 17.9°C, and average minimum is _27.9°C. The growing season ranges from 125 to 138 days (Namkhaijantsan 2006). Mean annual precipitation is 300 mm with a range from 200 to 450 mm. Seventy percent of the annual precipitation occurs from June to August as rain. The study area often has a thin (5–10 cm) snow cover from the beginning of October until the end of April.
Permafrost is present in most north facing slopes and in lower stream valleys, but is absent in south-facing slopes. The depth of the active layer, the surface soil that thaws each summer is 4.5 m in the grazed north-facing slope and riparian zone, and 1.5 m in the non-grazed north-facing slope (Goulden et al. 2005; Sharkhuu et al. 2007). Soils in non-grazed north-facing lower steppe and riparian zone that have shallow active layers are wet throughout the summer ranging from 40% to 60% soil moisture. In contrast, soils on south-facing slopes are dry with <10% soil moisture (Sharkhuu et al. 2007). The soils are sandy loam in texture.
A line transect with 100 m length was established across each valley, from the ridge-top across the steppe and riparian zones. These areas were divided into four landscape positions: north-facing lower steppe, riparian zone, south-facing lower steppe and south-facing upper steppe.
The south-facing lower steppe is 30–40 m lower than its associated and steep sloped upper steppe. The riparian zone is on the south part of the floodplain; the north-facing lower steppe is just above the north-facing riparian area. Elevations of the riparian and north-facing lower steppe are similar. Within each landscape position, a plot 50 m 9 50 m was gridded for random selection of sampling points and the number of sample points was based on a power analysis using an initial set of data. Seven 0.25 m2 (0.5 m 9 0.5 m) quadrats within each plot were selected. A total of 168 quadrats were sampled. We performed our sampling in July of 2004. Quadrats were divided into 100, 5 9 5 cm grids (81 intersection points) to estimate canopy cover. If a plant was touching or below an intersection point, it was identified to species and ground cover was categorized as lichen, dung or bare ground. Nomenclature followed Gubanov (1996) and Grubov (2000). Information on life history traits, palatability, nutrient value and response to grazing of individual species was obtained from the literature (Ulziikhutag 1985; Jigjidsuren and Johnson 2003). Harvesting for the total plant biomass was performed at the time of peak standing crop that usually occurs in mid July for the region. All plants within the randomly chosen quadrat were harvested at ground level. Harvested plants were separated into live or dead plant tissue. Live plants were identified to species and also classified into functional groups: grass (species of Poaceae), sedge (species of Cyperaceae), forb (herbaceous dicots), shrub (dwarfs) and moss. Plant biomass samples were dried at 80°C for 20 h and weighed.
Statistical analyses
We plotted a rank-dominance curve, which represents relative dominance of species in the community, from the most important to the least important by relative abundance (Whittaker 1975) based on biomass with log transformations for data normalization for each landscape position in the valleys. Species number, canopy cover of green plants, biomass of functional groups, total plant biomass and litter mass (response variables) were analyzed by analysis of variance (ANOVA). The experimental design included two factors (grazing and landscape) and their interaction, and conformed to a split-plot design. The main factors were treated as fixed effect. The grazing factor had three levels: heavy grazing (TRG, SHL), moderate grazing (NYN, SVL) and no grazing (DLB, BRG). The landscape factor had four levels (south-facing upper and lower, riparian zone, and north-facing lower). According to the split-plot design, grazing was considered as main plot and valleys were considered as subplot (two subplots for each grazing level). The valleys were treated as random effect and were nested within grazing. We ran the ANOVA using a restricted maximum likelihood (REML) as a variance estimation method because the main factors were fixed effects. Canopy cover data were transformed with an angular transformation (Sokal and Rohlf 1995).


Plant community structure
We recorded 162 vascular plant species belonging to 84 genera in 29 families in our total sampling quadrats (168 quadrats 9 0.25 m2). Additionally, a moss species, Selagnella sibirica (Milde) Hieron, was recorded in riparian and north-facing lower steppe. The top three families Poaceae, Asteraceae and Fabaceae accounted for 58 species (36%). For functional groups, the total species included 21 grasses, two sedges, two semi-shrubs and 121 forbs (herbaceous dicots). Significant differences of species richness were not found among grazing levels (P = 0.20), but they were found among landscape gradients (P < 0.0001) as well as an interaction of the two main effects (P < 0.0001) (Table 1). This result indicated that the grazing effect on species richness depends on landscape positions.

North-facing lower steppe
There were two distinct groups in the shape of rankdominance curves. Non-grazed and moderately grazed north-facing lower steppe had lower species richness than heavily grazed north-facing steppe. In the non-grazed and moderately grazed areas, mosses were the main contributor for the total plant biomass constituting 75–80% of that. Species with distribution that is restricted to only permafrost area accounted for 60–85% of the total species richness of non-grazed and moderately grazed valleys. Species richness increased as grazing intensity increased in this landscape position (Figure 3a). Sedge species are important in north-facing lower steppe of the valleys that have no grazing pressure, whereas, xerophytic grasses and forbs were dominant in plant communities of the north facing lower steppe of grazed valleys.
Riparian zone
Species richness was lower in the valleys without grazing and it increased as grazing intensity increased; TRG was highest in species richness, while few sedges, Kobresia sibirica (Turcz. ex Ledeb.) Boeckeler and Kobresia bellardii (Vill.) Fiori, and mosses contributed most of the total plant biomass in the non-grazed riparian zones (26%). Mosses formed a thick layer (5–8 cm) in the riparian and north-facing lower of the non-grazed valleys, while the moss layer had disappeared in grazed valleys (Figure 3b).
South-facing lower steppe
Different pattern from north-facing steppe and riparian zone in terms of species richness and dominance was observed in south-facing slopes of the valleys. Rank and dominance curve of BRG was distinct from others because of its higher species richness including 25 rare species. The heavily grazed valleys, SHL and TRG, were similar to each other and lowest in species richness. Species richness decreased with grazing intensity (Figure 3c) in this landscape position. Forbs were dominant in the moderately and heavily grazed valleys while a sedge, Carex pediformis C.A.Mey, was the dominant species in the valleys without grazing.
South-facing upper steppe
Rank-dominance curves for south-facing upper steppe of the valleys were identical among the valleys except for TRG, the heavily grazed valley, which had the lowest richness and highest dominance by a single xerophytic, short and semishrub species, Thymus gobicus Czern (thyme) and Artemisia frigida Willd. Non-grazed valley, BRG, had the highest richness with 32 species and the lowest dominance by a single species. A larger number of rare species contributing <1% of the total plant biomass with 18 and 17 were recorded in moderately grazed valleys, NYN and SHL, respectively (Figure 3d). Xerophytic and less palatable species for small livestock (sheep and goat), which are considered as indicators of heavily grazed areas (Hilbig 1995), dominated in the south-facing upper steppe of two moderately grazed valleys, while in non-grazed valleys, BRG and DLB, mesophytes and palatable species were ranked in first place.
Cover of bare ground, dung and canopy of green plants were significantly different for both the main effects and the interaction of the two-factor ANOVA (Table 1). Canopy cover of green plants was high in both south-facing upper and lower steppe of moderately grazed valleys (Figure 4). In the riparian zones and north-facing lower steppe of the valleys, green plant cover was 100% because of moss cover in non-grazed valleys and high dominance of forbs in the riparian zone of moderately and heavily grazed valleys. Bare ground cover was greater in the upper part of the south-facing slopes compared to the lower part. Bare ground increased along the grazing intensity gradient. The opposite pattern occurred in litter cover. The lichen
Xanthoparmelia camtschadalis (Ach.) Hale cover was high in non-grazed south-facing lower steppe but it was low in other grazing intensities.
Plant biomass
Similar to the observed differences of species richness among the grazing levels and landscape gradients, nonsignificant difference of the total plant biomass was found for grazing (P = 0.20), and significant differences for landscape and their interaction (P < 0.0001, for two of them; Table 1) were found. The total plant biomass gradient was observed from north-facing lower steppe (156 +-7.5 g m -2) to south-facing slope (28 +-1.6 gm -2 in lower steppe, 42 +-6.2 g m -2 in upper steppe) in non-grazed landscape but the total plant biomass gradient diminished within a grazed landscape. More pronounced effect of grazing on the total plant biomass was observed in wet landscape positions, northfacing lower steppe and riparian zone, than dry and warm landscape positions, south-facing upper and lower steppe. The total plant biomass decreased along grazing gradient in north-facing lower, riparian zone and south-facing upper steppe but grazing increased the total plant biomass slightly in the south-facing lower steppe. Also, an increase in the total plant biomass was observed in heavily grazed riparian zone than moderately grazed valley (Figure 5a).
Although we did not find significant differences of grass, sedge and moss biomass between the grazing levels, significant differences of the functional groups between landscape positions and two-factor interaction were found (Table 1).
The decrease in the total plant biomass in heavily grazed north-facing lower was explained by the decrease in moss biomass; moss biomass contributed 75% of the total plant biomass in non-grazed, 38% in moderately grazed northfacing lower while moss biomass was not detected in heavily grazed north-facing lower steppe (Figure 6a). Also, the same pattern of decrease in moss biomass along the grazing gradient was observed in the riparian zones. The decrease in contribution of moss biomass to the total plant biomass was substituted by an increase in percent of grass and forb biomass in moderately and heavily grazed northfacing lower and riparian zone (Figure 6a). A significant change in sedge and shrub contribution for the total plant biomass was not observed in the riparian zone (Figure 6b). No substantial differences of grass and forb percent for the total plant biomass were observed between grazing intensities, but switching of sedge percent by increase in shrub percent was observed in heavily grazed south-facing lower steppe (Figure 6c). Moreover, percent of functional groups in the total plant biomass did not significantly change along the grazing gradient in south-facing upper steppe except slight increase of grass in moderate grazing and increase in shrub in heavy grazing (Figure 6d). Significant differences of litter mass were found between grazing intensities (P < 0.0001), landscape positions (P = 0.0012) and in their interaction (P < 0.0001) (Table 1). For non-grazed landscape, south-facing lower steppe had the highest (76 +-6.5 g m-2) litter mass, whereas northfacing lower had the lowest (23 +-3.1 g m-2) litter mass. This difference in litter mass was not observed along heavily grazed landscape (Figure 5b) indicating potential effects of grazing on decreasing in litter mass.


We found two different trends of responses of plant species richness to livestock grazing. Plant species richness decreased in both south-facing upper and lower steppe, but increased in north-facing lower steppe and riparian zones of the heavily grazed valleys. Plant communities across the landscape responded differentially to livestock grazing. The non-grazed north-facing steppe and riparian zones are narrow ecological habitats for plants because of high soil moisture content recharged by seasonal melting of permafrost, low soil temperature, and less solar radiation than south-facing steppe. These conditions contribute to the dominance of mosses, sedges and a horsetail (we classified a horsetail, Equisetum pretense Ehrh, into forb), which are mesophytes and characteristic of vegetation without large effects of herbivory (Hilbig 1995; Gunin et al. 1999). Mosses play an important role in insulating soil by retaining moisture and protecting permafrost in boreal regions (van der Wal et al. 2001). At small spatial scale, species richness of plant community increased in the north-facing slope and riparian zone due to livestock grazing, but livestock grazing could negatively affect the plant species pool at larger spatial scale (Chaneton and Facelli 1991; Olff and Ritchie 1998). Although species adapted to the wet and nutrient-rich ecosystems are relatively few and include mostly tall tussock sedges, they disappear locally and can have the potential to become extinct from the regional species pool. Opposite to this local extinction, local colonization of grazing-tolerant species increased due to loss of moss cover and soil disturbances by livestock grazing, creating a nutrient-rich gap and stimulating germination of seeds of those species (Cingolani et al. 2003). Livestock grazing shifted competition of plants for light to competition for soil nutrients. In conclusion, the net effect of grazing on the plant richness of the wet and nutrientrich ecosystems, in our case north-facing lower steppe and riparian zone, can be negative. Species richness decreased in south-facing upper steppe under both moderate and heavy grazing. The south-facing steppe favors not only plants tolerant to low water and nutrient availability but also plants with characteristics allowing them to avoid herbivory. The characteristics to tolerate and to avoid grazing in dry ecosystems evolve convergently (Stebbins 1981; Coughenour 1985; Milchunas et al. 1988). As a consequence of the convergent evolution, plants in this environment do not respond significantly to grazing because of their high degree and rate of regrowth (McNaughton 1983; Coughenour 1985) and tillering horizontally or spreading by rhizomes. Thus, grazing does not change the competition for soil resources in these dry habitats. Moreover, there is a linear function with a small negative slope describing species richness with increasing grazing intensity (Milchunas et al. 1988). Although our results for grazing effect on species richness across the landscape with sharp nutrient gradient were consistent with the model of Milchunas et al. (1988), we lacked research to compare our results found in the north-facing lower steppe and riparian zone, especially livestock grazing effects on structure and total plant biomass of those communities with permafrost. A common effect of intensive herbivore grazing is a shift in plant community composition towards dominance by unpalatable species as a consequence of selective herbivory by livestock grazing. This pattern has been documented in arid and semiarid grasslands (Facelli 1988; Fahnestock and Knapp 1994; Gomboev et al. 1996; Augustine and McNaughton 1998). Especially, at the high density of grazing animals and under sedentary grazing, the availability of preferred forage per animal is reduced and decreases the selectivity of animals (Augustine and McNaughton 1998). We found an increase in abundance of unpalatable species, which do not have characteristics stimulating selective response by the grazing animals (Baumont 1996), in heavily grazed southfacing steppe (Figure 3). The unpalatable species constituted more than 50% of the total plant biomass. Particularly, abundance of semi-shrubs, T. gobicus and A.frigida, which are unpalatable to livestock (Ulziikhutag 1985), increased in the heavily grazed south-facing steppe. This result was not consistent with the result found in steppe of Inner Mongolia by Barger et al. (2004). They found that A. frigida was decreased in grazed communities. Potentilla acaulis Linnaues, a xerophytic perennial forb, tolerant to grazing and partially palatable to livestock (palatable to horses and cattle in early spring, but unpalatable to sheep and goats; Ulziikhutag 1985; Jigjidsuren and Johnson 2003), increased in south-facing lower steppe. In the heavily grazed riparian zone and north facing lower steppe, increase in abundance of grazing-tolerant species was observed. Livestock grazing resulted in dominance of unpalatable species in the south-facing steppe while grazing- tolerant forbs became dominant in riparian and north-facing lower steppe. Canopy cover of north-facing lower steppe and riparian zones was 100% under all grazing intensity. Although species composition of this part of landscape shifted to forb dominance by grazing, we did not find a change in canopy cover in these landscape positions. Green plant canopy cover tended to increase in moderately grazed south-facing upper and lower steppes because of dominance of P. acaulis, a short and prostrate forb. This species is an indicator of grazing because its cover increases under heavy grazing in semiarid regions of Mongolia and Inner Mongolia (Barger et al. 2004). The short stature and prostrated morphology of this species can explain most of the difference in green plant canopy cover between grazed and nongrazed south-facing lower steppe. The increase of short and prostrated species cover also have been reported in grazed grasslands (Belsky 1992; Barger et al. 2004). A cover loss of the lichen X. camtschadalis, non-nitrogen fixer, was found in grazed south-facing lower steppe. X. camtschadalis is common in grasslands of the Mongolian Plateau. Grazing reduces the biomass of this lichen (Barger et al. 2004; Gao et al. 2004). Moreover, one of the commonly reported grazing effects on grasslands is decrease in litter cover (Knapp and Seastedt 1986; Milchunas et al. 1989; Barger et al. 2004). We observed the same pattern in litter cover and biomass across grazed landscape, although the magnitude of the decrease of litter mass due to livestock grazing was higher in the boreal steppe than the steppe of Inner Mongolia. Litter mass was higher by 87% in the ungrazed south-facing steppe of the boreal steppe (Figure 5), while it was 32% in the steppe of Inner Mongolia. Moreover, litter mass was positively correlated with A. frigida in the steppe of Inner Mongolia (Barger et al. 2004). In terms of mean annual precipitation, the two steppes are similar (320 mm in steppe of Inner Mongolia and 300 mm in the boreal steppe). Dominance of A. frigida in the steppe of Inner Mongolia might be encouraged by a warmer climate (mean annual temperature is 0°C), but in the boreal steppe this species might not compete favorably with C3 grass and forb species, when livestock grazing are excluded. The loss of lichen cover and litter affect the water and nutrient cycling of the ecosystems (Knapp and Seastedt 1986) but detailed studies addressing this particular question are needed. A spatial difference of the total plant biomass was higher across the non-grazed landscape but this difference lessened in the grazed landscape. Livestock grazing resulted in a decrease of the total plant biomass of all landscape positions; the greatest percentage reductions were in riparian zones and north-facing lower steppe (Figure 5a). Positive correlations of the total plant biomass with water availability and nutrient gradients across landscape have been documented in semiarid ecosystems; wet and nutrient-rich sites tend to have more total plant biomass than dry and nutrient-poor areas (Milchunas et al. 1988; Knapp et al. 1993).
In conclusion, at the community level plant communities responded differently to livestock grazing, but at the ecosystem level the total plant biomass decreased to some extent under grazing across the landscape of the semiarid boreal steppe in northern Mongolia. Moreover, livestock grazing caused convergence of the different plant communities across the landscape into grazing-tolerant communities and reduced the spatial heterogeneity of the landscape. Moss cover, which is the main insulator of permafrost, in the wet and nutrient-rich sites, disappeared under livestock grazing. The loss of moss cover could potentially accelerate thawing  of permafrost and warming of this region.


We gratefully thank Dr Peter S. Petraitis (University of Pennsylvania) for his advice and help on statistical analyses, researchers of the Hovsgol GEF project for their cooperation. We also thank students from National University of Mongolia for their field and laboratory assistance. The study was a part of the GEF-financed World Bank project on Permafrost Melt and Biodiversity Loss in Lake Hovsgol National Park and the additional fund Asia Research Center at National University of Mongolia.


  1. Adler P, Raff D, Lauenroth WK (2001) The effect of grazing on the spatial heterogeneity of vegetation. Oecologia 128: 465–479.
  2. Anchorena J, Cingolani A (2002) Identifying habitat types in a disturbed area of the forest-steppe ecotone of Patagonia. Plant Ecol 158: 97–112.
  3. Augustine DJ, McNaughton SJ (1998) Ungulate effects on the functional species composition of plant communities: herbivore selectivity and plant tolerance. J Wildl Manage 62: 1165–1183.
  4. Austrheim G, Eriksson O (2001) Plant species diversity and grazing in the Scandinavian mountains - patterns and processes at different spatial scales. Ecography 24: 683–695
  5. . Barger NN, Ojima DS, Belnap J, ShipingW, YanfenW, Chenet Z (2004) Changes in plant functional groups, litter quality, and soil carbon and nitrogen mineralization with sheep grazing in an Inner Mongolian grassland. Range EcolManage 57: 613–619.
  6. Barzagur D (2002) Territorial organization of Mongolian pastoral livestock husbandry in the transition to a market economy. Focus Geogr 47: 20–25.
  7. Baumont R (1996) Palatability and feeding behavior in ruminants. Ann de Zootechnie 45: 385–400.
  8. Belsky AJ (1992) Effects of grazing, competition, disturbance and fire on species composition and diversity in grassland communities. J Veg Sci 3: 187–200.
  9. Biondini ME, Patton BD, Nyren PE (1998) Grazing intensity and ecosystem processes in a northern mixed-grass prairie, USA. Ecol Appl 8: 469–479.
  10. Burke IC, Lauenroth WK, Riggle R, Brannen P, Madigan B, Beard S (1999) Spatial variability of soil properties in the shortgrass steppe: the relative importance of topography, grazing, microsite, and plant species in controlling spatial patterns. Ecosystems 2: 422–438.
  11. Casper BB, Goldman R, Lkhagva A et al. (2012) Legumes mitigate ecological consequences of a topographic gradient in a northern Mongolian steppe. Oecologia 169: 1–10.
  12. Chaneton EJ, Facelli JM (1991) Disturbance effects on plant community diversity: spatial scales and dominance hierarchies. Plant Ecol 93: 143–155.
  13. Chognii O (2001) Process of Recovery and Changes of Pasture Used by Nomads. Urlah Erdem, Ulaanbaatar, 1–175. (In Mongolian.) Cingolani AM,
  14. Cabido MR, Renison D, Neffa VS (2003) Combined effects of environment and grazing on vegetation structure in Argentine granite grasslands. J Veg Sci 14: 223–232.
  15. Collins SL, Knapp AK, Birggs JM, Blair JM, Steinauer EM (1998) Modulation of diversity by grazing and mowing in native tallgrass prairie. Science 280: 745–747.
  16. Coughenour MB (1985) Graminoid responses to grazing by large herbivores: adaptations, exaptations, and interacting processes. Ann Mo Bot Gard 72: 852–863.
  17. Derner JD, Hart RH (2007) Grazing-induced modifications to peak standing crop in northern mixed-grass prairie. Range Ecol Manage 60: 270–276.
  18. Facelli JM (1988) Response to grazing after nine years of cattle exclusion in a Flooding Pampa grassland, Argentina. Plant Ecol 78: 21–25.
  19. Fahnestock JT, Knapp AK (1994) Plant responses to selective grazing by bison: interactions between light, herbivory and water stress. Plant Ecol 115: 123–131.
  20. Fernandez-Gimenez ME, Allen-Diaz B (1999) Testing a nonequilibrium model of rangeland vegetation dynamics in Mongolia. J Appl Ecol 36: 871–885.
  21. Fernandez-Gimenez ME, Allen-Diaz B (2001) Vegetation change along gradients from water sources in three grazed Mongolian ecosystems. Plant Ecol 157: 101–118.
  22. Fuhlendorf SD, Smeins FE (1999) Scaling effects of grazing in a semi-arid grassland. J Veg Sci 10: 731–738. Gao YZ, Han XG, Wang SP (2004) The effects of grazing on grassland soils. Acta Ecol Sin 24: 790–797.
  23. Gomboev B, Sekulich I, Pykhalaova T et al. (1996) The present condition and use of pasture in the Barguzin Valley. In: Culture and Environment in Inner Asia (Eds Humphrey C, Sneath D), Wild Horse Press, Cambridge, 124–140.
  24. Goulden CE, Etzelmuller B, Lkhagva A et al. (2005) Permafrost thermal properties and thaw and its relationship to soil and plant cover, Lake Hovsgol, Mongolia, available from URL: http://adsabs.harvard.edu/abs/2005AGUFM.C31A1121G [cited 16 September 2011].
  25. Grubov VI (2000) Key to the Vascular Plants of Mongolia (With an Atlas). Science Publishers, Enfield, 1–817. Gubanov IA (1996) Flora Conspect of Outer Mongolia. Nauka, Moscow, 1–136.
  26. Gunin PD, Vostokova EA, Dorofeyuk NI, Tarasov PE, Blacc CC (1999) Vegetation Dynamics of Mongolia. Kluwer Academic Publishers, Dordrecht, Netherlands, 1–233.
  27. Han G, Hao X, Zhao M, et al. (2008) Effect of grazing intensity on carbon and nitrogen in soil and vegetation in a meadow steppe in Inner Mongolia. Agric Ecosyst Environ 125: 21–32.
  28. Hilbig W (1995) The Vegetation of Mongolia. SPB Academic Publishing, Amsterdam, 1–235. Hobbs RJ, Huenneke LF (1992) Disturbance, diversity, and invasion: implications for conservation. Conserv Biol 6: 324–337.
  29. Hook PB, Burke IC (2000) Biogeochemistry in a shortgrass landscape: control by topography, soil texture, and microclimate. Ecology 81: 2686–2703.
  30. Jigjidsuren S, Johnson D (2003) Forage Plants of Mongolia. Admon Publishing, Ulaanbaatar, 1–563. Knapp AK, Seastedt TR (1986) Detritus accumulation limits productivity of tallgrass prairie. Bioscience 36: 662–668.
  31. Knapp AK, Fahnestock JT, Hamburg SP, Statland LB, Seastedt TR, Schimel DS (1993) Landscape patterns in soil-plant water relations and primary production in tallgrass prairie. Ecology 74: 549–560.
  32. Lhagvajav N (1992) Monthly and seasonal pasture growth in alpine zone pastures. Report of Institute of pasture and forage plants. Ulaanbaatar: Center of Science and Information technology (Mongolia); Report No.: 1999-2. (In Mongolian.)
  33. McNaughton SJ (1983) Compensatory plant growth as a response to herbivory. Oikos 40: 329–336. McNaughton SJ (1985) Ecology of a grazing ecosystem: the Serengeti. Ecol Monogr 55: 259–294.
  34. Milchunas DG, Lauenroth WK (1989) Three dimensional distribution of plant biomass in relation to grazing and topography in the shortgrass steppe. Oikos 55: 82–86.
  35. Milchunas DG, Lauenroth WK (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecol Monogr 63: 327–366.
  36. Milchunas DG, Sala OE, Lauenroth WK (1988) A generalized model of the effects of grazing by large herbivores on grassland community structure. Am Nat 132: 87–106.
  37. Milchunas DG, Lauenroth WK, Chapman PL, Kazempour MK (1989) Effects of grazing, topography, and precipitation on the structure of a semi-arid grassland. Vegetatio 80: 11–23.
  38. Namkhaijantsan G (2006) Climate and climate change of the Hovsgol region. In: The Geology, Biodiversity and Ecology of Lake Hovsgol (Mongolia) (Eds Goulden C, Sitnikova T, Gelhaus J, Boldgiv B), Backhuys Leiden, Leiden, Netherlands, 63–76.
  39. Oba G, Vetaas OR, Stenseth NC (2001) Relationships between biomass and plant species richness in arid-zone grazing lands. J Appl Ecol 38: 836–845.
  40. Olff H, Ritchie ME (1998) Effects of herbivores on grassland plant diversity. Trends Ecol Evol 13: 261–265. Olofsson J, Stark S, Oksanen L (2004) Reindeer influence on ecosystem processes in the tundra.
  41. Oikos 105: 386–396. Osem Y, Perevolotsky A, Kigel J (2002) Grazing effect on diversity of annual plant communities in a semi-arid rangeland: interactions with small-scale spatial and temporal variation in primary productivity. J Ecol 90: 936–946.
  42. Proulx M, Mazumder A (1998) Reversal of grazing impact on plant species richness in nutrient-poor vs. nutrient-rich ecosystems. Ecology 79: 2581–2592.
  43. Sala OE, Austin AT (2000) Methods of estimating aboveground net primary productivity. In: Methods in Ecosystem Science (Eds Sala OE, Jackson RB, Mooney HA, Howarth RW), Springer-Verlag, New York, 31–43.
  44. Sala OE, Oesterheld M, Le_on RJC, Soriano A, Oesterheldt M (1986) Grazing effects upon plant community structure in subhumid grasslands of Argentina. Vegetatio 67: 27–32.
  45. Schimel D, Stillwell M, Woodmansee R (1985) Biogeochemistry of C, N, and P in a soil catena of the shortgrass steppe. Ecology 66: 276–282.
  46. Sharkhuu A, Sharkhuu N, Etzelm€uller B et al. (2007) Permafrost monitoring in the Hovsgol mountain region, Mongolia. J Geophys Res 112: 1–11.
  47. Sokal RR, Rohlf FJ (1995) Assumptions of analysis of variances. In: Biometry: The Principles and Practice of Statistics in Biological Research. WH Freeman, New York, 440–447.
  48. Stebbins GL (1981) Coevolution of grasses and herbivores. Ann Mo Bot Gard 68: 75–86.
  49. Tserendash S, Erdenebaatar B (1993) Performance andmanagement of natural pasture in Mongolia. Nomadic Peoples 33: 9–15. Tsogoo D (1990) Proper utilization methods for mountain forest steppe hay meadows of the MPR (Dissertation). National University of Mongolia, Ulaanbataar, Mongolia.
  50. Ulziikhutag N (1985) Key to Forage Plants of the Pasture and Hayfields of the MPR. Ulsiin Khevleliin Gazar, Ulaanbaatar, 1–445. (In Mongolian.)
  51. Van der Wal R, van Lieshout SM, Loonen MJJ (2001) Herbivore impact on moss depth, soil temperature and arctic plant growth. Polar Biol 24: 29–32.
  52. White RP, Murray S, Rohweder M (2000) Pilot Analysis of Global Ecosystems: Grassland Ecosystems. World Resources Institute, Washington, DC, 1–62. Whittaker R (1975) Community and Ecosystems. McMillan, New York, 1–385.