As the “roof of the world”, the Tibetan Plateau (TP) is a unique geographical unit on Earth. In recent years, vegetation has gradually become a key factor reflecting the ecosystem since it is sensitive to ecological changes especially in arid and semi-arid areas. Based on the normalized difference vegetation index (NDVI) dataset of TP from 2000 to 2015, this study analyzed the characteristics of vegetation variation and the correlation between vegetation change and climatic factors at different time scales, based on a Mann–Kendall trend analyses, the Hurst exponent, and the Pettitt change-point test. The results showed that the vegetation fractional coverage (VFC) generally increased in the past 16 years, with 60.3% of the TP experiencing an increase, of which significant (p < 0.05) increases accounted for 28.79% and were mainly distributed in the north of the TP. Temperature had the largest response with the VFC on the seasonal scale. During the growing season, the correlation between precipitation and sunshine duration with VFC was high (p < 0.05). The change-points of the VFC were mainly distributed in the north of the TP during 2007–2009. Slope and elevation had an impact on the VFC; the areas with large vegetation change are mainly distributed in slopes <20° and elevation of 3000–5000 m. For elevation above 3000–4000 m, the response of the VFC to precipitation and temperature was the strongest. This study provided important information for ecological environment protection and ecosystem degradation on the Tibetan Plateau.

Potential evapotranspiration (PET) is an important input variable of many ecohydrological models, but commonly used empirical models usually input numerous meteorological factors. In consideration of machine learning for complex nonlinear learning, we evaluated the applicability of three machine learning algorithms in PET estimation in the Yellow River basin (YRB), in addition to determining significant factors affecting the accuracy of machine learning. Furthermore, the importance of meteorological factors at varying altitudes and drought index grades for PET simulation were evaluated. The results show that the accuracy of PET simulation in the YRB depends on the input of various meteorological factors; however, machine learning models including average temperature (T_mean) and sunshine hours (n) as input achieved satisfactory accuracy in the absence of complete meteorological data. Random forest generally performed best among all investigated models, followed by extreme learning machine, whereas empirical models overestimated or underestimated PET. The importance index shows that T_mean is the most influential factor with respect to PET, followed by n, and the influence of T_mean on PET gradually decreased with increased altitude and drier climate, whereas the influence of n shows the opposite trend.

The wind–water erosion crisscross region, where the topography is complicated, is the most severe area of soil erosion on the Loess Plateau. The wind and terrain both have an impact on the soil water erosion process. In order to evaluate the effects of sand cover on runoff and soil loss characteristics, a series of experiments was conducted in two contrasting treatments. One treatment was a bare loess soil slope serving as the control, and the others were sand-covered loess slopes with five different slopes. The results showed that the runoff time, total runoff yield, and total soil loss were different between the sand-covered slope and the loess slope on the slope of 15°. The sediment concentration of the sand-covered slope was significantly higher than that of the loess slope during the entire rainfall process (p < 0.05). The increase in the slope gradient shortened the surface runoff initiation times and enhanced the total runoff volume and soil loss. The total runoff volume and the total soil loss were 39.7 L and 44.3 kg, respectively, on the sand-covered slope of 10°. When the slope gradient increased from 10° to 30°, the total runoff volume and the total soil loss increased by 22.8 L and 42.8 kg, respectively, while the surface runoff initiation times shortened by 300 s. For the sand-covered slopes, the erosion processes appeared to be mainly dominated by sediment transport. The correlation between soil loss rates and slope gradients demonstrated the secondary polynomial function. In addition, the critical slope of sand-covered slopes was from approximately 23° to 28°. The proportion of sand cover and slope responsible for soil erosion was 3:1, which means the wind effect was more important than the terrace factor in terms of soil water erosion in the wind–water erosion crisscross region. The results provide a theoretical basis for soil erosion control in this area.

Over the past few decades, global vegetation cover has obviously changed, particularly in the Loess Plateau, due to vegetation restoration projects in China. This study focuses on the Pisha Sandstone area (PSA) and uses various statistical analysis methods to study the spatiotemporal changes in vegetation coverage (VEC) at different time scales. The effects of topographical and climatic factors on VEC were also quantitatively evaluated using the GeoDetector in the spatial distribution. The results of the study confirm that, on an annual scale, the area with a significant increase in VEC has reached 63.89% (p < 0.05). Change points were diagnosed to have occurred mainly between 2002 and 2012 at different time scales, with the percentage of significant change points in VEC accounting for more than 20% from April to October (p < 0.05). Temporal and spatial changes in precipitation mainly caused VEC changes. In 45.35% of the region, precipitation was significantly and positively correlated with VEC at an annual scale (p < 0.05). Moreover, VEC was most conducive to growth and increase at 1050–1500 m above sea level and 0–21° slope, respectively. In most areas, there was an enhanced interaction relationship between various factors on VEC. Converting farmland to forests in suitable areas, selecting appropriate tree species, and improving soil is conducive to ecological restoration in the PSA in the future.

To analyze the operation risk of check dams under extreme precipitation conditions, taking Ningxia area as an example, this paper carried out a risk assessment of check dams under extreme precipitation conditions in Ningxia through data collection, hydrological statistics, numerical simulation, and other methods. The conclusions are the following: (1) By the end of 2020, about 40% of the silt reservoir capacity of check dams in various water and soil conservation zones in Ningxia has been accumulated. During 1966–2020, the extreme precipitation and frequency of extreme precipitation in Ningxia increased while the intensity of extreme precipitation decreased. The extreme precipitation in Ningxia increased year by year and lasted longer. (2) Under two future scenarios of RCP4.5 (the full name of RCP is Representative Concentration Pathway) and RCP8.5, the extreme precipitation threshold in Ningxia is gradually decreasing from south to north. Extreme precipitation in the future will bring high risk to the operation of check dams in Ningxia. The results of this paper can provide a scientific basis for the operation and management of check dams in Ningxia.