The Loess Plateau of China is one of the most severe soil and water loss areas in the world. Since 1999, the Grain to Green Program (GTGP) has been implemented in the region. This study aimed to analyze spatial and temporal variations of ecosystem services from 2000 to 2015 to assess the effects of the GTGP, including carbon sequestration, water regulation, soil conservation and sand fixation. During the study period, the area of forest land and grassland significantly expanded, while the area of farmland decreased sharply. Ecosystem services showed an overall improvement with localized deterioration. Carbon sequestration, water regulation and soil conservation increased substantially. Sand fixation showed a decreasing trend mainly because of decreased wind speeds. There were synergies between carbon sequestration and water regulation, and tradeoffs between soil conservation and sand fixation. It was concluded that ecological projects have contributed significantly to the rehabilitation of the fragile ecosystems of this region. To make these projects more successful and sustainable, long-term management procedures are necessary to maintain and consolidate the improvements.

Grassland gross primary productivity (GPP) is an important part of global terrestrial carbon flux, and its accurate simulation and future prediction play an important role in understanding the ecosystem carbon cycle. Machine learning has potential in large-scale GPP prediction, but its application accuracy and impact factors still need further research. This paper takes the Mongolian Plateau as the research area. Six machine learning methods (multilayer perception, random forest, Adaboost, gradient boosting decision tree, XGBoost, LightGBM) were trained using remote sensing data (MODIS GPP) and 14 impact factor data and carried out the prediction of grassland GPP. Then, using flux observation data (positions of flux stations) and remote sensing data (positions of non-flux stations) as reference data, detailed accuracy evaluation and comprehensive trade-offs are carried out on the results, and key factors affecting prediction performance are further explored. The results show that: (1) The prediction results of the six methods are highly consistent with the change tendency of the reference data, demonstrating the applicability of machine learning in GPP prediction. (2) LightGBM has the best overall performance, with small absolute error (mean absolute error less than 1.3), low degree of deviation (root mean square error less than 3.2), strong model reliability (relative percentage difference more than 5.9), and a high degree of fit with reference data (regression determination coefficient more than 0.97), and the prediction results are closest to the reference data (mean bias is only −0.034). (3) Enhanced vegetation index, normalized difference vegetation index, precipitation, land use/land cover, maximum air temperature, potential evapotranspiration, and evapotranspiration are significantly higher than other factors as determining factors, and the total contribution ratio to the prediction accuracy exceeds 95%. They are the main factors influencing GPP prediction. This study can provide a reference for the application of machine learning in GPP prediction and also support the research of large-scale GPP prediction.

Ecological restoration programs are expected to control environmental deterioration and enhance ecosystem functions under a scenario of increasing human disturbance. The largest ecological restoration program ever implemented in China, the first round of the countrywide Grain for Green Program (GGP), finished in 2010. However, it is not known whether the ecological changes that resulted from the GGP met the restoration goal across the whole implementation region. In this study, we monitored and assessed the ecological changes in the whole GGP region in China over the lifetime of the first round of implementation (2000–2010), by establishing a comprehensive assessment indicator system composed of ecosystem pattern, ecosystem quality (EQ), and key ecosystem services (ESs). Remote sensing interpretation, ecological model simulations based on multi-source images, and trend analysis were used to generate land use and land cover (LULC) datasets and estimate ES and ESs indicators. Results showed that while forest increased by 0.77%, artificial land increased more intensely by 22.38%, and cropland and grassland decreased by 1.81% and 0.68%, respectively. The interconversion of cropland and forest played a primary role in ecosystem pattern change. The increase in ecosystem quality measures, including fractional vegetation cover (0.1459% yr⁻¹), leaf area index (0.0121 yr⁻¹), and net primary productivity (2.6958 gC m⁻² yr⁻¹), and the mitigation of ecosystem services deterioration in soil water loss (−0.0841 t ha yr⁻¹) and soil wind loss (−1.0071 t ha yr⁻¹) in the GGP region, indicated the positive ecological change in the GGP region to some extent, while southern GGP subregions improved more than the those in the north on the whole. The GGP implementation other than climate change impacted ecological change, with contributions of 14.23%, 9.94%, 8.23%, 30.45%, and 18.05% in the ecological outputs mentioned above, respectively. However, the water regulation did not improve (−2283 t km⁻² yr⁻¹), revealing trade-offs between ecosystem services and inappropriate afforestation in ecological restoration programs. Future GGP implementation should change the practice of large-scale afforestation, and focus more on the restoration of existing forest and cultivation of young plantings, formulating rational and specific plans and designs for afforestation areas through the establishment of near-natural vegetation communities, instead of single-species plantations, guided by regional climate and geographical characteristics.

Various multivariate methods were used to analyze datasets of river water quality for 11 variables measured at 20 different sites surrounding Lake Taihu from 2006 to 2010 (13,200 observations), to determine temporal and spatial variations in river water quality and to identify potential pollution sources. Hierarchical cluster analysis (CA) grouped the 12 months into two periods (May to November, December to the next April) and the 20 sampling sites into two groups (A and B) based on similarities in river water quality characteristics. Discriminant analysis (DA) was important in data reduction because it used only three variables (water temperature, dissolved oxygen (DO) and five-day biochemical oxygen demand (BOD₅)) to correctly assign about 94% of the cases and five variables (petroleum, volatile phenol, dissolved oxygen, ammonium nitrogen and total phosphorus) to correctly assign >88.6% of the cases. In addition, principal component analysis (PCA) identified four potential pollution sources for Clusters A and B: industrial source (chemical-related, petroleum-related or N-related), domestic source, combination of point and non-point sources and natural source. The Cluster A area received more industrial and domestic pollution-related agricultural runoff, whereas Cluster B was mainly influenced by the combination of point and non-point sources. The results imply that comprehensive analysis by using multiple methods could be more effective for facilitating effective management for the Lake Taihu Watershed in the future.