The US Bureau of Land Management (BLM) manages nearly 1 million km² of public lands that support recreation, livestock production, and wildlife habitat. Monitoring the condition of vegetation on these lands is crucial for sound management but has historically been difficult to do at scale. Here we used newly developed remote-sensing tools to conduct an unprecedented assessment of trends in vegetation cover and production for all BLM rangelands from 1991 to 2020. We found widespread increases in cover and production of annual grasses and forbs, declines in herbaceous perennial cover, and expansion of trees. Cover and production of annual plants now exceed that of perennials on > 21 million ha of BLM rangeland, marking a fundamental shift in the ecology of these lands. This trend was most dramatic in the Western Cold Desert of Nevada and parts of surrounding states where aboveground production of annuals has more than tripled. Trends in annuals were negatively correlated with trends in bare ground but not with trends in perennials, suggesting that annuals are filling in bare ground rather than displacing perennials. Tree cover increased in half of ecoregions affecting some 44 million ha and underscoring the threat of woodland expansion for western rangelands. A multiscale variance partitioning analysis found that trends often varied the most at the finest spatial scale. This result reinforces the need to combine plot-level field data with moderate-resolution remote sensing to accurately quantify vegetation changes in heterogeneous rangelands. The long-term changes in vegetation on public rangelands argue for a more hands-on approach to management, emphasizing preventative treatment and restoration to preserve rangeland habitat and functioning. Our work shows the power of new remote-sensing tools for monitoring public rangelands and developing effective strategies for adaptive management and conservation.

Managers tasked with balancing livestock production and wildlife habitat in mesic meadows face a unique set of challenges. These challenges are compounded in the arid western United States because mesic areas comprise only a small portion of the landscape yet provide vital forage and water resources to livestock and wildlife and are essential to underlying ecosystem integrity and function. Our objectives were to compare the effects of short-duration grazing treatments that varied by season and intensity on the quantity and nutrient quality of pasture forage and average daily gains (ADG) of yearling cattle. We established pastures (n = 15 total pastures) in mesic meadows at the University of Idaho Rinker Rock Creek Ranch in south-central Idaho and stocked them with heifers during 2019 and 2020. Heifers grazed six pastures in June (early-season; 16 d) and six pastures in August (late-season; 16 d) at moderate (30–40%) and high (70-80%) relative utilization levels (n = 3 pastures per treatment). Three pastures were not grazed by cattle representing controls. Forage was collected and analyzed for nutrient quality at pregrazing, postgrazing, and after a period of regrowth in late September. Contrasted against all other treatments, early-season grazing at a high-intensity produced forages with the highest crude protein (P < 0.001) and lowest neutral-detergent fiber (P ≤ 0.04) after regrowth. ADGs of yearling heifers were greater (P ≤ 0.01) during the first year of the trial than the second year. Differences in ADG were not associated with grazing season (P = 0.08) or intensity (P = 0.12), despite numeric differences. Forage quantity and quality and heifer ADG responses varied between years, likely due to spring and annual precipitation differences. This study demonstrates the effectiveness of early-season grazing at a high intensity for improving forage quality and increasing ADG of livestock grazing mesic meadows.

The availability and accessibility of adequate and quality fodder is a huge challenge for pastoral and agropastoralist communities in Tanzania. Fodder scarcity, primarily driven by rangeland degradation, population growth, and climatic changes, limits livestock productivity and deepens rural poverty. This study assessed factors influencing agropastoralists' access to fodder and their willingness to conserve fodder sources in Magu District, Tanzania. A total of 210 agropastoralist households were surveyed in three wards covering six villages. The results show that access to fodder during the wet season is positively influenced by a household's income, land size, number of livestock, and physical capital. Dry season fodder accessibility is negatively influenced by households' land size, physical capital, and the number of livestock. Results also show that more than 60% of the agropastoral households accessed fodder sources formally (legally) in open areas, while 35% accessed fodder informally (illegally) in traditionally reserved highland areas, people's farms, and restricted urban areas. Conservation of rangelands and fodder sources is positively influenced by a head of households' level of education and income, and only 40% of households participated in conservation of rangelands and fodder sources. Improved land tenure, education, and income diversification can improve households' access to fodder and their willingness to conserve fodder sources.

Strategies and mechanisms are needed to enhance encroaching shrub species use by ruminants. Seriphium plumosum crude proteins, neutral detergent fibers, total phenolics, and condensed tannins were determined from combined fine leaves and stems collected during the dry and wet seasons, from burned and unburned areas. S. plumosum crude protein was higher during the wet season on postburned areas. Neutral detergent fiber was higher during the dry season and similar in both postburned and unburned areas. The mean total phenolics and condensed tannins were higher during the dry and wet seasons, respectively. These results have shown that S. plumosum forage quality in the wet season is sufficient to meet ruminants nutrient requirements, but the opposite is true for the dry season; hence, the need to include it in livestock feed with other protein and energy supplements to improve its forage quality beyond ruminant maintenance requirement.

The plateau pika (Ochotona curzoniae) is a keystone species in the alpine rangeland ecosystem on the Qinghai–Tibet Plateau (QTP). However, it has been considered a pest because these mammals dig holes and make bare patches, which reduces grassland productivity. Accordingly, there is a trade-off required between biodiversity conservation and pest management. Habitat assessment is one of the critical approaches in small mammal management and protection. Although a few papers have reported the distribution of plateau pika and predicted its potential distribution under climate change at local scales in some areas, little is known about the synergistic effect of climate change and human disturbance on the wider distribution of this mammal on the QTP. In this study, we recorded 219 points of plateau pika presence on the QTP via Global Positioning System and used the Maxent model to predict potential suitable habitats. The results indicated that, under the current climate, the potential distribution of plateau pika is mainly located in southern Gansu Province, eastern Qinghai Province, and northwestern Sichuan Province, accounting for 4.44% of the total area of the QTP. The most important factors in defining habitat suitability and limiting the distribution of plateau pika include normalized difference vegetation index (29.1%), human footprint (21.6%), seasonal precipitation (11.2%), and elevation (7.8%). In the future, in the 2030s and 2050s, the total area of suitable habitat on the QTP is projected to decrease, while the habitable region in northwest Sichuan and eastern Qinghai will likely increase.

Sustainability of rangeland is adversely affected by many factors in Turkey, as well as in developing countries. The rangeland in Turkey's Kayseri province does not produce the amount of roughage needed by the current number of animals. To improve this situation, the state has encouraged rangeland improvement and management activities to increase the productivity and sustainability of rangelands. The aim of this study is to determine farmers' behaviors and characteristics that affect their use of improved rangelands, especially whether their use maintained the improved level of forage production. Farmers engaged in animal husbandry in 11 villages where rangeland improvement activities were completed in Kayseri constituted the population of the study. Simple random sampling was used to determine the farmers included in the sample with 244 farmers interviewed. Data were collected by administering a questionnaire filled during the face-to-face interviews with each individual respondent. Nonlinear canonical correlation analysis was used in the analysis of the collected data. Results identified farmer characteristics and behaviors affecting sustainable rangeland use, as indicated by maintaining forage productivity greater or equal to 3 200 kg/ha on improved pastures. It has been determined that the contribution of the farmers with a large number of animals, a large land size, producing forage crops, and participating in the rangeland improvement activities is higher in the sustainability and adoption of rangeland improvement and management activities. The results of our study will provide a new and different perspective on how to increase the efficiency and planning of rangeland improvement activities and in the formation of agricultural policies on this subject.

Testing and documenting effects of wind farm (WF) infrastructure on wildlife are crucial considering increasing development throughout Scandinavia, especially for reindeer, which require large areas for grazing and are vulnerable to disturbances. We present results from 2011 to 2019 for semidomesticated reindeer tracked with Global Positioning System (GPS) transmitters, along with herders' knowledge about reindeers' habitat use and changes following WF development within the Raggonjarga reindeer district summer range in Finnmark, Norway. We tracked up to 36 females (ranging from 19 to 36 individuals per year), from their arrival in the study area in April to their departure in the end of October. We evaluated habitat use before, during, and after WF development at the home range and landscape scales. We also evaluated reindeer habitat use qualitatively based on semistructured interviews with local herders. The herders' reported negative effects of the WF on reindeer, both on general habitat use and intrarange movements, resulting in less use of grazing areas surrounding the WF and increased workload for the herders. The GPS results partly support the herders' experiences. We found negative effects of the WF at the landscape scale, except during summer, where the effect was positive. Results at the home range scale showed negative effects of the WF in spring and summer, but not autumn. Different results at different scales make identifying causality challenging, especially as yearly variation was also large. Different results for summer and autumn may relate to changes in herding activities and larger movement patterns, respectively. Similar and contrasting results from the two methods suggest a need for both sources of data in combination to understand and improve land management. Including herders' knowledge to understand results from GPS data is thus crucial. We also suggest future studies focus on mechanisms behind behavioral changes to better understand cause-and-effect relationships and how effects can be mitigated.

Groundwater-dependent ecosystems are biologically diverse and productive ecosystems but constitute a small fraction of total land area in semiarid regions. Efforts to link remotely sensed data from satellite-based platforms to measurements of vegetation structure and function at smaller spatial scales have increasingly received attention, due to the need to manage diverse landscapes at scales relevant to management. In the semiarid western United States, grazing is a dominant land use and meadows can receive a high degree of grazing pressure. In this study we examined satellite-based and near-surface imagery to determine if they were useful in assessing grazed systems with different grazing management. We compared meadows that were chronically grazed by feral horses in conjunction with periodic cattle grazing to a meadow managed and grazed by cattle only. We examined the agreement between near-surface digital cameras (PhenoCams) and satellite-based indices of greenness and production for meadows in the Central Great Basin, United States. We also verified them with field-collected data on percent foliar cover by dominant functional groups. There was strong agreement between the Landsat normalized difference vegetation index and PhenoCam Green Chromatic Coordinate (GCC) (Pearson's r ≥ 0.61). Gross primary production modeled using Landsat satellite imagery and integrated over the growing season had a strong linear relationship with GCC integrated over the growing season (R² = 0.89). Furthermore, despite differences in spatial and temporal resolution, integrated metrics from both platforms were able to discern differences in grazing pressure. Meadows with chronic feral horse grazing plus 3 mo of livestock grazing had reduced integrated gross primary production and GCC in comparison with a meadow that had short-term grazing (for 2 mo) by livestock only. The ability to detect differences in grazed systems from PhenoCams and satellite platforms provides important tools for quantifying the effects of grazing in groundwater-dependent ecosystems.

The Ogallala Aquifer region, located in the Great Plains of the central United States, is the largest freshwater aquifer in North America, supporting one of the most agriculturally productive regions in the world. In this paper, we discuss the history of settlement and water use in this region, from the Homestead Act and the Dust Bowl to modern irrigation systems. While many improvements to irrigation technology and water-efficient crops have helped to prolong the life of the Ogallala, continued use of this finite resource is leading to a tragedy of the commons, wherein difficult land management decisions will have to be made by this century's end. We posit that the art and science of rangeland management stands uniquely poised to tackle this challenge directly through creative integration, where appropriate, of native rangeland restoration, improved pasture management, integrated crop-livestock systems, and regenerative agricultural practices aimed at preserving soil and rangeland health, thereby providing continuity in the ability of the Ogallala region to continue to provide food, fiber, and other ecosystem services both locally and globally. Furthermore, we provide discussion on future research, extension, and educational needs to consider as the exploration for adaptive solutions are developed and evaluated in the coming decades.

The impact of domestic dog and coyote depredation on livestock has received little attention in Mexico. We used livestock depredation insurance claims from 2017 to 2020 and landscape attributes to identify the magnitude of depredation and model the risk of depredation by domestic dogs and coyotes throughout Mexico using maximum entropy modeling. Combined livestock depredations by coyotes and domestic dogs comprised 50.3% of all livestock depredation claims in Mexico. Depredation by domestic dogs was associated with landscape attributes indicative of rural villages and subsistence livestock husbandry. More intensive or larger-scale agricultural land uses (e.g., farming, large ranching enterprises) were associated with coyote depredation, in particular increasing presence of small- to medium-sized livestock (i.e., sheep, etc.). Both domestic dogs and coyotes posed a risk of depredation across much of Mexico (i.e., 23% and 40% of Mexico with > 0.50 likelihood of depredation, respectively), with risk likelihood maps for each canid being similar (i.e., similarity indices = 0.75–0.84) for most risk levels but showing little similarity with respect to regions of high (> 0.70) and very high (> 0.90) likelihood of depredation (similarity = 0.38 and 0.19, respectively). By identifying areas of high depredation risk, our models provide a planning tool to facilitate allocation of resources such as insurance programs, identify areas that would benefit from additional mitigation programs, and provide guidance for managing landscape attributes to decrease risk and limit vulnerability of livestock to predation.

As a large-scale Payment for Ecosystem Services program targeting grasslands on a global scale, the Chinese Grassland Ecological Payment and Award Policy (GEPAP) was implemented for two rounds until 2020, but its performance continues to be debated. In addition, equilibrium (E) and nonequilibrium (NE) ecosystem theory, which is believed to be an important perspective for researching grassland management, has not received enough attention. Therefore, based on E and NE ecosystem theory, this study summarizes herdsmen's attitudes and responses to the GEPAP. It then reviews and analyzes the ecological and socioeconomic effects of this policy by reviewing and organizing literature published since the policy was first implemented in 2011. It is found that 1) herdsmen neither felt clearly satisfied with nor actively responded to the GEPAP in both the E and NE systems; 2) In E grassland systems, the GEPAP has generally restored grasslands to some extent, except for the Xinjiang region. It also promoted an increase in herdsmen's total household income, changes in income structure, and reformations of livestock husbandry; 3) In NE grassland systems, the GEPAP did not significantly improve either the grassland ecological environment or local socioeconomic development. Therefore, it is advised that the GEPAP could beneficially continue, albeit with some improvements, in E systems; however, in contrast, it should be stopped and substituted with other grassland management modes such as community-based grassland management in NE systems. These findings are important for the improvement of Payment for Ecosystem Services programs both nationally and globally.

Targeted grazing in the Great Basin has been used to reduce cheatgrass fuel loads and enhance wildfire control. In this project, we evaluate the economic impact of targeted grazing on cow-calf ranches across southeast Oregon, northeast Nevada, and southwest Idaho when practices such as fencing, water hauling, and herding are necessary for producers to accomplish desired grazing outcomes. Large and small representative ranch models were developed for major land resource areas 23, 24, and 25 where applicable. Typical targeted grazing costs and practices were obtained from producer and agency focus groups in each state and introduced into ranch economic models. Targeted grazing periods begin 1 mo before typical Bureau of Land Management turnout in the spring and again in the fall after typical public land grazing ends. In each year, targeted grazing would occur when the previous growing season (September to March and April to August) had more than 25% of median precipitation based on PRISM historical data. Hence, targeted grazing could occur in the spring, fall, or both depending on precipitation. In both seasons, targeted grazing continues until the desired animal unit months of forage are removed. One hundred precipitation data sets were randomly generated using Excel to mimic the actual number of drought years in the spring and fall. The model is a 40-yr recursive linear programming model using 100 cattle price sets and the 100 precipitation sets. Results are averaged over 10 000 model runs and compared with scenarios with no targeted grazing and targeted grazing based on the actual precipitation data set. Results show changes in cattle herd size, hay sales, and the economic impacts to the public land ranch operation for two ranch sizes in each of the three major land resource areas by state.

Fire and grazing are key drivers of plant community structure in the Great Plains ecoregion, with critical impacts on the quantity and quality of forage. While the impact of fire on forage quality is well understood in many ecosystems, few studies have examined this question in the northern mixed-grass prairie. We investigated the postburn nutrient content of forbs and graminoids over a 4-yr timeframe, asking 1) how forage quality changed over the 4 yr post fire and 2) whether the effect of fire on quality was impacted by the resumption of grazing. We used near-infrared spectroscopy and generalized partial least squares regression (plsRglm) methods to assess crude protein. We show that the use of generalized models provides improved estimates of percent nitrogen than the standard partial least squares approach. Both crude protein and fiber varied annually after the wildfires, with no effect of grazing on forage quality. In the first and second years, burned sites displayed higher protein than unburned locations. Fire did not affect crude protein between plant functional groups; however, forbs showed higher nutrient values than grasses throughout the 4 yr. Fire affected fiber content with grasses showing higher fiber than forbs. There was a positive relationship between neutral and acid detergent fiber and litter in burned sites. These results indicate no long-term effect of wildfire on protein content, which stabilizes in the third year post burn. The improvement of forb digestibility and retention of protein under dry conditions could modify the grazing capacity of the prairies at large scales. This may also change the foraging behavior of large ungulate grazers through mechanisms such as increasing patch reselection (i.e., intensified grazing cycles) and broadening feeding preferences to forbs when grasses are limited or have high fiber content.

Bighorn sheep (Ovis canadensis) historically inhabited the Black Hills region of South Dakota, but the species was extirpated from the area in the early 1900s. We translocated 26 bighorn sheep from Alberta, Canada to the northern Black Hills. We predicted that translocated bighorn sheep would exhibit similar resource use and selection as populations native to comparable habitats, specifically steep slopes with low overstory cover and high visibility. We used resource selection functions to estimate habitat selection by bighorn sheep and Robel pole and hand-clipped biomass to investigate the correlation between forage estimate methods. Bighorn sheep selected habitat types at varying levels, but were most strongly associated with barren and grassland landscapes while negatively associated with forest landscapes. Bighorn sheep had a positive association with slope and selected for fire-disturbed landscapes and elevation seasonally. Visual obstruction readings and herbaceous biomass at foraging sites (n= 78) were positively associated (r² = 0.62). Herbaceous biomass ranged from 302.07 to 2,487.43 kg·ha^–1. Foraging sites were typically located in areas with little overstory tree canopy cover (mean = 8.41%, standard error [SE] = 1.85), shorter distances to escape terrain (mean = 24.00 m, SE = 3.21), and low amounts of woody debris (mean = 0.25 kg·ha^–1, SE = 0.07). The study area (Deadwood region of South Dakota) provided sufficient landscape attributes and herbaceous biomass to support the newly established bighorn sheep herd. Post-translocation assessments of resource use are crucial for evaluating conservation actions and potential success of future large mammal translocations.

The relationship between sampling area of plant community and number of species contained in it has an importance in vegetation ecology. Using a nested quadrat survey, the relationship between plant species and sampling area was discussed under different grazing pressure in the steppe and the determination of minimum sampling area was calculated herein. Ten curve models were selected to fit the species-area relation of plant community in a typical steppe of northern China. The residual standard deviation, correlation index, average value of absolute deviation, and average value of absolute relative deviation were used as the evaluation indexes for model fitting. The results show that the curve equations S = B + CA^Z and S = C(1-e^–BA)^Z have the best simulation effect. According to the simulation results, the minimum sampling areas are determined as 1.16 m² (T0), 1.23 m² (T7), 1.28 m² (T10), and 1.10 m² (T14), presenting the minimum sampling area is feasible when the second derivative is zero. Generally, sheep grazing reduced the species richness of steppe, though some of the plans flourished while some others disappeared or dismissed, but increased spatial heterogeneity, especially on a smaller scale. However, there are also variances in different area scales and grazing intensities. When the sampling area was small (< 2 m²), heavy grazing T14 increased the number of plant species per unit area, yet light grazing T8 reduced the number of plant species per unit area of steppe. When the sampling area was large (> 16 m²), grazing reduced the number of plant species per unit area in the natural steppe. Anyhow, the effect of grazing intensity on plant species fluctuated greatly when the sampling area was medium (2 m² ∼ 16 m²). In general, the similarity of plant communities had an increasing trend with accumulated sampling area, though it varied at different grazing intensities.

Most researchers assume cattle avoid milkweeds due to toxic compounds. However, observations suggest cattle graze common milkweed (Asclepias syriaca L.), an important summer host plant for monarch butterflies (Danaus plexippus L.) east of the Rocky Mountains, as well as showy milkweed (Asclepias speciosa Torr.). This directly relates to efforts to add over 1 billion stems of milkweed for monarch butterfly conservation. We sampled whether cattle regularly graze common/showy milkweed in a central Nebraska restored prairie patch-burn-grazing system with patches burned in the year of the main study (2018) or previous years (2017–2016). We also sampled burned control grasslands without livestock. This sampling therefore compared three different types of burn-graze management. Cow-calf pairs were put onto the grasslands in April, and we sampled every 2 wk through September. We found that wild grazers (e.g., deer) were not grazing common/showy milkweed stems in control grasslands, whereas by August > 73% of stems had been grazed in the patch-burn-grazed grasslands—with no noticeable negative effects on cattle. Also, cattle grazing of common/showy milkweed did not differ significantly in different patches, even though big bluestem grass (Andropogon gerardii Vitman) was grazed significantly more in patches burned in 2018. Finally, milkweed abundance in control grasslands ranged from 34- to 185-fold higher than the average milkweed abundance in patch-burn-grazed grasslands. These results suggest that cattle regularly graze common/showy milkweed and that patch-burn-grazing does not increase milkweed abundance when cattle begin grazing in April. The good news of these results is that cattle safely grazed low amounts of common/showy milkweed, suggesting control efforts for these milkweeds are often unnecessary. Future work should examine how to increase milkweed abundance without affecting cattle weight gain, possibly by deferring grazing in some areas during critical milkweed growth periods.

Seeding of native perennial species commonly fails in the presence of invasive annual weeds. The ability of weeds to deplete soil moisture through early germination and rapid growth gives them a competitive advantage. A possible solution to help native species compete with weeds is to enhance their germination rate and growth. We primed seeds of bluebunch wheatgrass (Pseudoroegneria spicata) and Lewis flax (Linum lewisii) in a matrix of compost, clay, and biostimulants for 4–7 d. We used an extrusion technique to incorporate the seed and priming matrix into pellets for the priming duration that had the quickest germination. We evaluated primed seeds in pellets at two field sites against seeds in pellets that were not primed and seeds left untreated (control). Seed treatments were planted in shallow (1-cm) and deep (15-cm) furrows, in a complete factorial design, with the expectation that the deep furrow treatment would provide an enhanced microsite to improve plant growth. Results indicated that deep furrows were the strongest driver in enhancing seedling establishment, while priming showed a smaller contribution. In the first month after planting, a combination of priming and furrowing increased seedling emergence by 128% and 303%, for bluebunch wheatgrass and Lewis flax, respectively, compared with control seed planted in shallow furrows. The following year, primed bluebunch wheatgrass and Lewis flax seeds in deep furrows increased plant biomass by 158% and 110%, respectively, compared with control seed in shallow furrows. Overall, this study generally indicates that the rapid germination of primed seeds and the use of deep furrows may assist seedlings in establishing, which may allow them to better compete with invasive weeds.

Weather drives plant community assembly and seasonal plant production in most rangeland ecosystems. Previous models in the California Annual Grassland have indicated that plant production is largely determined by the thermal accumulation, or degree days, between the start and end of the growing season. These models, however, were relatively limited in spatial and temporal scope and did not include seasonal parameters commonly used in more current production-system models. The purpose of this study is to develop models that account for the annual and seasonal effects of weather inputs on plant production. Weather variables for these models included seasonal effects of accumulated degree days, total precipitation, average temperature, season length, past year's precipitation, and past year's plant production. Our results indicated that sorting precipitation years into low, moderate, and high groupings resulted in better production-model fit compared with a full-suite analysis of all precipitation years. In addition, we found that precipitation, temperature, past year's precipitation, and past plant production significantly improved model fit over models that considered only accumulated degree days and season length. Linking these models to seasonal forecasting applications across western rangeland ecosystems could significantly enhance management of grazing production systems in this environment.

A growing body of work has revealed that differences in body size of ungulates follow ecoregion and soil boundaries and that these size differences are nutritionally influenced. Currently, it is unclear if these patterns of body size result from differences in quantity or quality of forage produced. We quantified differences in white-tailed deer (Odocoileus virginianus Zimm.) body mass and antler size captured at four sites in South Texas, United States. We sampled available forage to determine if long-term differences in average body and antler size could be explained by forage quantity or quality, or some combination of the two. Data collected from 2011 to 2019 indicated female body mass was > 3 kg smaller on the eastern edge of the Coastal Sand Plain ecoregion as compared with those from the western transition zone of the Coastal Sand Plain and Tamaulipan Thornscrub ecoregions. Similarly, male body mass and antler size were > 6 kg and > 20 cm smaller, respectively, in coastal habitats compared with more interior sites. We found that forb biomass, browse and forb diversity, and the nutritional landscape, quantified using digestible energy, crude protein, phosphorus, and neutral detergent fiber, differed between sites. However, differences between sites were inconsistent with predictions that would have supported our hypothesis. Overall, we found no support for the hypothesis that forage quantity drives ecogeographic variation in physical traits of white-tailed deer but were unable to fully refute or support the hypothesis that forage quality, specifically plant diversity, drives ecogeographic variation in physical traits of ungulates.

The presence of seeds in livestock dung can alter the composition of aboveground vegetation. However, there have been few studies of the composition of the dung seed bank of Tibetan livestock on different vegetation types on the Qinghai-Tibet Plateau. The aims of this study were to identify the differences of endozoochorous seed dispersal in dung type and vegetation type and determine the relationships between the dung seed bank and aboveground vegetation in meadow and shrubland. We conducted experiments to assess the density and plant diversity of the seed bank of yak and sheep dung in meadow and shrubland. The dung seed density in Tibetan sheep dung from meadow sites was the highest (239.3 ± 13.8 seeds kg^–1), and the number of species in yak dung from meadow sites was the highest. Among seasons, the livestock dung seed density was highest in November and the lowest in April. The species richness of the dung seed bank was higher in meadows than shrubland and higher in yak dung than that in Tibetan sheep dung only for shrubland. The dispersal of dung seeds has a greater impact on vegetation changes in meadows than shrubland, and the dispersal of seed in yak dung has a greater effect on vegetation changes than does dispersal in Tibetan sheep dung in shrubland. There were more monocotyledons in yak dung of shrubland (42%), whereas there were more dicotyledons in Tibetan sheep dung of meadow (45%). This suggests that yak and Tibetan sheep are complementing each other in their seed dispersal potential. Therefore, we recommend grazing more Tibetan sheep in the meadow and more yak in the shrubland on the Qinghai-Tibet Plateau.