The effects of human activities on Earth’s vegetation have tended to be negative, mostly because of deforestation1. Restoration efforts are often restricted to small, localized scales. Large ecological-engineering projects aimed at producing regional-scale effects are few, and among these, China’s mega-projects — most notably, the Grain for Green Project (GGP)2 — stand out because of their unparalleled scale (27.8 million hectares of forest re-established as of 2013 across 26 Chinese provinces3). Writing in Nature Sustainability, Tong et al.4 report that the positive effects of these tree-planting projects on vegetation growth can be detected using remote-sensing satellite imagery of a large region of southwestern China (the provinces of Guizhou, Guangxi and Yunnan), in an area associated with highly erodible landscapes called karst. The authors note that these projects, which require considerable investment, will be justified only if the modification of ecosystem properties can be achieved on a large scale.
The government-run GGP, intended to halt soil erosion and desertification2, began in 1999. The project’s goal was to convert land on mountainous terrain prone to erosion (cropland or scrubland) into forested landscapes (Fig. 1). Such forest would be classified as ecological if trees might eventually be logged (subject to permission) as part of a timber quota, and as economical if it contained orchards, or plantations of trees for medical use. Ecological forest accounted for 80% of the planting area, with economical forest making up the remaining 20%4. The GGP was developed partly in response to the consequences of land-use changes during the time of Chairman Mao Zedong, notably the huge areas logged to provide fuel and construction materials during the Great Leap Forward programme, and large-scale conversions of often marginal, sloping land to agricultural use in the 1960s and 1970s to enhance local self-sufficiency — a change that caused severe erosion problems5.
To assess the effects of the tree-planting projects, Tong and colleagues use three independent lines of evidence, and the consistency of the findings convincingly demonstrate the robustness of their results. One approach was the analysis of two complementary properties of vegetation. Satellite-imaging data from 1982 to 2015 allowed the researchers to measure the area of vegetation cover present per square metre of ground (known as the leaf-area index). Other satellite data collected between 1992 and 2012 enabled the authors to assess plant biomass in units of above-ground carbon biomass. Plant biomass can be inferred by converting vegetation optical depth (a property captured by microwave observations that are sensitive to the water content of vegetation) to total carbon using an approach based on the carbon density of above-ground, living, woody vegetation.
Over time, both of these properties revealed a marked transition in the amplitude and/or direction of vegetation trends around the main implementation period of the tree-planting project, between 2000 and 2006. The authors’ calculations indicate that the southwestern region of China that they studied acted as a carbon sink after the GGP implementation, providing a considerable amount of the entire country’s net carbon sequestration. The authors also observed negative vegetation trends in the provinces’ growing urban areas, such as in the cities of Kunming and Nanning. This provides an indirect validation of the team’s satellite-data approach.
The second line of investigation taken by Tong et al. involved the use of dynamic ecosystem modelling to explore what might have happened in the absence of the tree-planting project. The model took into account the effect of the increase in atmospheric carbon dioxide on vegetation during the time frame studied. This modelling exercise highlighted the divergence between the simulated trend of vegetation decrease projected if the tree-planting intervention had not occurred — linked to a long-lasting drought during the previous decade — and the vegetation increase that was observed.
The third approach taken by the researchers was an analysis of the number of hectares on which tree-planting actions were implemented in each of the 295 counties within the 3 provinces studied. These GGP-inventory data showed a correspondence between actions at the county level and positive vegetation trends, as well as stark differences between China’s provinces and the neighbouring countries of Laos, Vietnam and Myanmar, in which the vegetation assessed by satellite imagery decreased over the same period.
Tong and colleagues’ results are encouraging in regard to the large-scale effects of the GGP on vegetation, but should not be taken as a proof of its overall success. As the authors mention, the satellite trends were not validated by measurements taken on the ground. No erosion assessment was undertaken, so one of the main GGP goals was not evaluated directly. Furthermore, the time span of satellite analysis, and of the programme itself, might still preclude the detection of long-term dynamics related to the long lifespan of trees, or might not take into account the role of large but infrequent erosion or disturbance events such as those linked to torrential rains or pest outbreaks.
Most crucial for the overall assessment of the success of the GGP as an ecological restoration project is the fact that satellite data do not distinguish biological composition, such as the presence of different species, and so cannot be used to assess the project’s effects on biodiversity. The GGP focused on the planting of non-native, fast-growing monocultures, which might render the resulting forests more vulnerable to pests3,4. The GGP thus used a narrow view of ecosystem services (the role of vegetation in reducing erosion and desertification rates) that had the additional (and possibly unplanned) benefit of a net carbon-storage outcome. Furthermore, the rationale for GGP actions was based not on previous ecological states or projected overall ecological benefits, but on the potential to reduce the erosion rates on the target land and for the programme to generate income for farmers2.
In the absence of a China-wide assessment of the GGP’s environmental and ecological impacts2, an analysis of data from China based on 258 publications3 identified limited biodiversity benefits of the GGP. This was mainly because the dominant non-native, fast-growing monoculture plantations were linked to a decrease in floral diversity, associated with bee and bird population declines, as observed in Sichuan province. This report3 strongly recommends using native trees when establishing plantations, or at least the establishment of plantations composed of several tree species.
Nevertheless, Tong and colleagues’ work clearly shows a large-scale effect of the GGP on vegetation in southwestern China. This important result needs to be complemented by ground-based studies. Understanding of the GGP’s functional and biodiversity effects is needed to assess its success, and might also identify other interventions that have the potential to enhance or generate wider positive effects of the GGP as an ecological-restoration mega-project. The task set out by Tong and colleagues for how the effects of such massive initiatives can be tested on an adequate scale is valuable and very welcome.
Nature 553, 411-413 (2018)