Grassland degradation has emerged as a serious socio-economic and ecological problem, endangering both long-term usage and the regional biogeochemical cycle. Climate change and human activities are the two leading factors leading to grassland degradation. However, it is unclear what the degradation level caused by these two factors is. Using the normalized difference vegetation index (NDVI) and coefficient of variation of NDVI (CV_NDVI), the spatial distribution features of grassland degradation or restoration were analyzed in Qilian County in the northeast of the Qinghai–Tibet Plateau. The dominant climate variables affecting NDVI variation were selected through the combination of random forest model and stepwise regression method to improve the residual trend analysis, and on this basis, twelve possible scenarios were established to evaluate the driving factors of different degraded grasslands. Finally, used the Hurst index to forecast the trend of grassland degradation or restoration. The results showed that approximately 55.0% of the grassland had been degraded between 2000 and 2019, and the area of slight degradation (NDVI_slope > 0; CV_{NDVI (slope)} > 0; NDVI_value > 0.2) accounted for 48.6%. These regions were centered in the northwest of Qilian County. Climate and human activities had a joint impact on grassland restoration or degradation. Human activities played a leading role in grassland restoration, while climate change was primarily a driver of grassland degradation. The regions with slight degradation or re-growing (NDVI_slope > 0; CV_{NDVI (slope)} > 0), moderate degradation (NDVI_slope < 0; CV_{NDVI (slope)} > 0), and severe degradation or desertification (NDVI_slope < 0; CV_{NDVI (slope)} < 0) were dominated by the joint effects of climate and anthropogenic activity accounted for 34.3%, 3.3%, and 1.3%, respectively, of the total grassland area. Grasslands in most areas of Qilian County are forecasted to continue to degrade, including the previously degraded areas, with continuous degradation areas accounting for 54.78%. Accurately identifying the driving factors of different degraded grassland and predicting the dynamic change trend of grassland in the future is the key to understand the mechanism of grassland degradation and prevent grassland degradation. The findings offer a reference for accurately identifying the driving forces in grassland degradation, as well as providing a scientific basis for the policy-making of grassland ecological management.

The Qinghai-Tibetan Plateau (QTP), one of the most important ecological regions in the world, is experiencing a decline in ecological function as a result of severe grassland degradation. Elymus nutans is one of the ecological grass species for restoring degraded grasslands in QTP. The seed yield and seed quality are often limited by soil nutrients in QTP, so it is very important to optimize the application rates of fertilizer for E. nutans seed production. In this study, a “3414” fertilization experiment (nitrogen: 0 ~ 225 kg N ha− 1; phosphorus: 0 ~ 180 kg P2O5 ha− 1; potassium: 0 ~ 135 kg K2O ha− 1) was established in the third year of planting E. nutans seed field based on the soil nutrient supply situation in the planting area to explore the effects of combined application of N, P and K fertilizers on the seed yield and seed quality of E. nutans. The results showed that the balanced application of N, P and K fertilizers significantly improved the seed yield, seed quality and seed vigour of E. nutans. The seed yield of E. nutans was regulated by the seed yield components. Under different fertilization treatments, the number of fertile tillers and thousand kernel weight had a greater contribution to seed yield, followed by the number of spikelets per panicle. Compared to the no fertilization treatment, the combined application of N, P, and K fertilizers increased the seed yield of E. nutans by 23.05%–90.49%. The seed size, seed nutrient content, germination percentage and seedling growth of E. nutans were positively affected by the combined application of N, P and K fertilizers. The seed vigour of E. nutans increased by 38.23%–89.85% compared to the no fertilizer treatment. Fertilizer application promoted seed nutrient storage and increased seed size thereby promoting seed germination and seedling growth. The crude protein, soluble sugar and starch content of E. nutans seed increased by 9.70%–54.05%, 6.70%–25.70% and 4.31%–35.88%, respectively. Nitrogen fertilizer had a greater effect on seed yield and seed quality of E. nutans than phosphorus and potassium fertilizers. The balanced nitrogen, phosphorus and potassium fertilizers at an application rate of 150 kg N ha− 1, 120 kg P2O5 ha− 1, 90 kg K2O ha− 1 in the planting area of QTP was found to be the optimum fertilization amount for increasing seed yield, seed quality and net income. This study provides theoretical evidence and certain practical suggestions for sustainable forage seed production in alpine regions of QTP.

Seed dispersal by livestock in lowland rangelands has received attention from researchers in recent decades while there has been little research into the dispersal of alpine meadow plants’ seed after ingestion by yaks and Tibetan sheep. This study assessed the recovery and germination of seeds by feeding known quantities of seeds from 20 alpine meadow species to yaks and Tibetan sheep. Seeds from 18 out of 20 of the plant species fed to Tibetan sheep survived ingestion and defecation. All of the species fed to the yaks survived. The recovery rate of undamaged seed ranged from 0.4 to 58.4% for yaks and 0.0 to 28.1% for Tibetan sheep. In general, total recovery of ingested seeds from yaks (28.1%) was significantly higher than that from the sheep (9.4%). Passage time of seeds through the digestive tract ranged from 12 to 96 h for yaks and from 12 to 84 h for Tibetan sheep. Seed recovery was negatively related to seed length (yak, r = –0.59; Tibetan sheep, r = –0.52) but was not related to seed width, seed thickness and seed mass. Seeds defecated by yaks and Tibetan sheep were found to have a range of germination rates which were both higher and lower compared with undigested seed depending on the plant species. It is argued that ingestion of seed by grazing yak or Tibetan sheep may have important effects on the population dynamics of alpine plant species of the Qinghai-Tibetan Plateau.

Trampling by grazing animals exerts a comprehensive and serious effect on grassland vegetation and soil. In order to compare the trampling of yaks and Tibetan sheep under different grazing intensities, we examined the hoof pressure and trampling intensity (based on trampling area and hoof-print count) of white yaks (Poephagus grunniens) and Tibetan sheep (Ovis ammon) in an alpine meadow of Tianzhu County, Gansu Province, China, under conditions where either grazing area or livestock number were controlled. The average areas trampled by yaks and Tibetan sheep were 39.2 and 21.6 cm2 respectively. The average hoof pressure of yaks and Tibetan sheep were 6.89 and 3.13 kg cm–2 respectively. The yak-to-sheep ratio of the average area trampled was 1.81 : 1, whereas the yak-to-sheep ratio of average hoof pressure was 2.20 : 1. Average ingestion and walking trampling intensities of yaks were 384.8 × 103 and 247.1 × 103 kg cm–2, respectively, in controlled grazing areas, and 439.1 × 103 and 756.3 × 103 kg cm–2, respectively, in areas of controlled livestock numbers. These values for Tibetan sheep were 15.3 × 103 and 120.3 × 103 kg cm–2 in controlled grazing areas, and 42.6 × 103 and 128.2 × 103 kg cm–2 in areas of controlled livestock numbers. In controlled grazing areas, the ingestion and walking trampling intensities of yaks were 25.2 and 5.4 times higher, respectively, than those of sheep. Under areas of controlled livestock numbers, these values were 10.3 and 5.9 times higher, respectively, than those of sheep. The average trampling intensity of yaks was 7.3 times higher than that of the sheep. Therefore, under conditions of similar grazing intensity, yaks cause more damage than Tibetan sheep in alpine meadows.