Grassland is one of the most widespread vegetation types worldwide and plays a significant role in global carbon cycling. Understanding the sensitivity of grassland to climate change and the effect of climate changes on the grassland ecosystems is a key issue in global carbon cycling. One of the goals of this study was to evaluate the three net primary productivity (NPP)–climate models, i.e. the Miami model, the Schuur model and the classification indices-based model. Results indicated that the classification indices-based model was the most effective model at estimating large-scale grassland NPP. In this research, changes in the spatial pattern of global potential grassland from recent past (1950–2000) to future (2001–2050) A2a scenario were analysed with the integrated orderly classification system of grassland (IOCSG) approach in a Geographic Information System (GIS) environment. NPP was evaluated with the classification indices-based model. Results indicate that under recent past climatic conditions, the main parts of global grassland are the savanna and tundra and alpine grassland and will be converted into the savanna, steppe and semi-desert grassland in A2a scenario. As a whole, areas of grassland will increase by 31.76 million hectares. The classification indices-based model estimated a 12.40% increase of total NPP in grassland from recent past to A2a scenario. It will impose a new issue for future grassland researches to support sustainable development and to provide action relevant knowledge to meet the challenge of climate change.

The overexploitation of Grasslands without any return-back and compensation is the major cause of degradation and deterioration of the grassland ecosystem. The Subsidy and Incentive System for Grassland Conservation (SISGC) in China aimed to restore grassland ecology by the reduction of overgrazing, promoting carrying capacity, and increasing alternative employment of herders in non-husbandry sectors. However, the ecological response to the SISGC still remains unclear on the national scale. Here, we used systematic sampling, and satellite image time series data revealed a widespread proliferation of major ecological indicators for grasslands, contrasting climate and actual net primary productivity (NPP) before (2004–2010) and after (2011–2017) the implementation of SISGC founded the contributions to policy, as simulated by the Carnegie-Ames-Stanford-Approach (CASA) model. On average, by two-phase comparison, the actual grassland NPP increased by 11.72%. The contribution of policy implementation and climate factors increased grassland NPP by up to 61.14% and 38.86%, respectively, but the response of the NPP growth of various grassland types exhibited divergence, mainly divided into policy-led (contribution rate of 52.28–97.02%) and climate-led (contribution rate of 57.56–96.00%). Hence, the SISGC policies should be renewed for different grassland types.

Accurate estimation of the aboveground biomass (AGB) of grassland is a key link in understanding the regional carbon cycle. We used 501 aboveground measurements, 29 environmental variables, and machine learning algorithms to construct and verify a custom model of grassland biomass in the Headwater of the Yellow River (HYR) and selected the random forest model to analyze the temporal and spatial distribution characteristics and dynamic trends of the biomass in the HYR from 2001 to 2020. The research results show that: (1) the random forest model is superior to the other three models (R²_val = 0.56, RMSE_val = 51.3 g/m²); (2) the aboveground biomass in the HYR decreases spatially from southeast to northwest, and the annual average value and total values are 176.8 g/m² and 20.73 Tg, respectively; (3) 69.51% of the area has shown an increasing trend and 30.14% of the area showed a downward trend, mainly concentrated in the southeast of Hongyuan County, the northeast of Aba County, and the north of Qumalai County. The research results can provide accurate spatial data and scientific basis for the protection of grassland resources in the HYR.

Soil erosion in the Three-River Headwaters (TRH) region has continued to intensify in recent decades due to human activities and climate change. To reverse this situation, the Chinese central government has launched the Subsidy and Incentive System for Grassland Conservation (SISGC). As a sign of the effectiveness of SISGC implementation, the dynamic changes of soil erosion can provide timely feedback for decision makers and managers. The Revised Universal Soil Loss Equation (RUSLE) model was used to simulate the spatial distribution of soil erosion before and after SISGC implementation, and Mann–Kendall (MK) test to reveal the effect of policy implementation. The results showed that: (1) the soil erosion in the TRH was mainly mild (83.83% of the total eroded area), and the average soil erosion rate and the total erosion were 13.63 t ha⁻¹ y⁻¹ and 323.58 × 10⁶ t y⁻¹ respectively before SISGC implementation; (2) SISGC implementation has curbed soil erosion. After SISGC implementation, the total soil erosion decreased by 3.80%, which showed obvious differences between grassland types; (3) The influences of SISGC were mainly because it has increased vegetation cover, further decreasing soil erosion. However, soil erosion in Alpine grassland has deteriorated, indicating direct targeted policymaking should be on the agenda. Furthermore, SISGC should be continued and grassland-type-oriented to restore the grassland ecosystem.

The source park of the Yellow River (SPYR), as a vital ecological shelter on the Qinghai-Tibetan Plateau, is suffering different degrees of degradation and desertification, resulting in soil erosion in recent decades. Therefore, studying the mechanism, influencing factors and current situation of soil erosion in the alpine grassland ecosystems of the SPYR are significant for protecting the ecological and productive functions. Based on the ¹³⁷Cs element tracing technique and machine learning algorithms, five strategic variable selection algorithms based on machine learning algorithms are used to identify the minimal optimal set and analyze the main factors that influence soil erosion in the SPYR. The optimal model for estimating soil erosion in the SPYR is obtained by comparisons model outputs between the RUSLE and machine learning algorithms combined with variable selection models. We identify the spatial distribution pattern of soil erosion in the study area by the optimal model. The results indicated that: (1) A comprehensive set of variables is more objective than the RUSLE model. In terms of verification accuracy, the simulated annealing -Cubist model (R = 0.67, RMSD = 1,368 t km^–2⋅a^–1) simulation results represents the best while the RUSLE model (R = 0.49, RMSD = 1,769 t⋅km^–2⋅a^–1) goes on the worst. (2) The soil erosion is more severe in the north than the southeast of the SPYR. The average erosion modulus is 6,460.95 t⋅km^–2⋅a^–1 and roughly 99% of the survey region has an intensive erosion modulus (5,000–8,000 t⋅km^–2⋅a^–1). (3) Total erosion loss is relatively 8.45⋅108 t⋅a^–1 in the SPYR, which is commonly 12.64 times greater than the allowable soil erosion loss. The economic monetization of SOC loss caused by soil erosion in the entire research area was almost $47.90 billion in 2014. These results will help provide scientific evidences not only for farmers and herdsmen but also for environmental science managers and administrators. In addition, a new ecological policy recommendation was proposed to balance grassland protection and animal husbandry economic production based on the value of soil erosion reclassification.