The Effectiveness of the Grain for Green Program in Restoring the Loess Plateau

Context

The Loess Plateau, located in northern China, is widely recognized as one of the most environmentally degraded regions in the world due to centuries of deforestation, intensive cultivation, and overgrazing (Xing, 2023). To counter soil erosion and its compounding impact on regional poverty, the Chinese government launched the Grain for Green Program (GFG or GFGP) in 1999. Officially known as the Sloping Land Conversion Program (SLCP), the GFGP is the largest ecological restoration and rural development program in the world (Zhang, 2024). The program aimed to convert sloped, erosion-prone farmland into forest and grassland, offering subsidies and food compensation to farmers who retired their land from cultivation.

While the GFGP made significant gains in afforesting the region, its efficacy in overall ecological restoration is questioned: proponents highlight the increase in tree cover, decrease in soil erosion, and other environmental benefits while critics question its effects on the region’s hydrological cycle, impact on biodiversity and food security, and long-term viability. These conflicting perspectives raise fundamental questions about whether the Grain for Green Program should be expanded, adjusted, or phased out and whether its ecological benefits outweigh its tradeoffs in the context of climate adaptation, rural livelihoods, and national food security.

Literature Review

This essay will review two primary articles with opposing views on the impact and future recommendations for the GFGP. He et al. (2021) argue that the GFGP was a resounding success that should be expanded in their article, Transformation of agriculture on the Loess Plateau of China toward green development, stating that “Grain for Green Program had an overwhelming advantage in protecting the natural ecological environment” (He et al., 2021). Meanwhile, Ge et al. (2020) “caution against further revegetation over the Loess Plateau given the reduction in water available for agriculture and human settlements” in their Impact of revegetation of the Loess Plateau of China on the regional growing season water balance (Ge et al., 2020). Additional articles will be utilized to assess the impact of the GFGP and supplement their arguments.

He et al. (2021) view the GFGP as a “synergistic improvement in food, economy, and environment” and praise it as “successful policymaking” that balances agricultural output to meet China’s food security needs, economic incentivization to promote sustainable rural livelihoods, and ecological restoration, noting an increase in forest area of 3% and reduced soil erosion (He et al., 2021). While the authors list some recommendations, such as increasing efficiency through technology and chemical fertilizer, as well as increased education and capacity building for local populations, they largely do not mention the any of the negative impacts of the GFGP.

On the other hand, Ge et al. (2020) note that due to the biophysiology of forest root systems, which are deeper than cropland root systems, afforestation has increased evapotranspiration, drawing moisture from deep soil, and decreased both surface and underground runoff (Ge et al., 2020). Models suggest weak or no hydrological feedback from increased evapotranspiration to rainfall in the region, meaning rainfall has not increased to compensate for this water loss. This decrease in runoff and soil moisture has led to a lower water availability for agriculture and other human demands. Ge et al. conclude that further afforestation could worsen regional water stress and recommend against the expansion of the GFGP.

Additional research shows that the GFGP led to “an increase in ecological water stress,” with increased vegetation significantly reducing streamflow (Han et al., 2022). The current state of vegetation is already close to the threshold of the water resources’ carrying capacity, especially given that the Loess Plateau is naturally a semi-arid region. These hydrological constraints suggest that further expansion of the GFGP could exacerbate water scarcity and undermine ecological stability.

Critics also point out unmet expectations on biodiversity as another environmental indicator of the program’s shortcomings. The GFGP is mostly monocultures, leading to significantly less biodiversity and species richness. GFGP forests generally have 17–61% fewer bird species and 49–91% fewer bee species compared to native forests (Hua et al., 2016). The monocultural stands also compounded the negative effects on the hydrological cycle, slowly draining the Yellow River system (Davidson, 2025).

Another major criticism of the GFGP is its implications for food security, as the program converts cropland into forests. As one local farmer commented, “What about the next generation? They can’t eat trees” (Davidson, 2025), reflecting growing concerns about the program's long-term effects on agricultural yields. Proponents of the GFGP note that the initial decline in grain yield and food self-sufficiency was temporary and that agricultural yield has recovered since the launch of the program (Shi et al., 2020). However, the recovery was largely driven by the intensification of agricultural activity on the remaining cropland, such as heavy chemical fertilizer usage and high tillage through increased mechanization (Shi et al., 2020). These practices go against fundamental agroecological principles and may cause long-term soil degradation due to overexploitation of limited arable land.

Proponents of the GFGP have advocated for the construction of check dams and terrace farming as strategies to maintain food production without exacerbating erosion (Shi et al., 2020). However, evidence suggests these measures are often ineffective: check dams frequently fail due to rapid sediment accumulation, are costly to maintain, and can disrupt local hydrology and ecosystems (Chen et al., 2007). 

Other food security and agricultural yield concerns stem from how land-use conversion was implemented. Detailed studies of land-use conversion in the Loess Plateau shows many instances of “highly productive, flat cropland” converted into forests while “less productive sloped cropland” remained in cultivation, indicating that the GFGP achieved afforestation at the cost of efficiency and food production (Yan, 2019). This misallocation undermines food security and pressures the remaining farmland into unsustainable practices to bridge the production gap.

When pieced together, the evidence outlined above suggests that the portrayal of the GFGP in He et al. (2021) presents a partial and overly favorable narrative. He et al. frame success in a narrow definition of agricultural output and commend practices that negatively impact the overall ecological system. For example, He et al. commend the widespread adoption of intensive agricultural practices such as plastic film mulching, which causes significant microplastic pollution and soil degradation (Liu et al., 2022), and chemical fertilizer usage, which saw a 76% increase in some regions between 2001 and 2015 (He et al., 2021). The government also incentivized deep plowing, mechanization, and intensive tillage to enhance productivity, resulting in a 18-27% increase in wheat and maize yields and a 56% rise in cereal crop yields during that period; He et al. frame these developments as clear successes in agricultural modernization without addressing their long-term ecological costs. The reliance on synthetic inputs and aggressive soil disturbance undermines agroecological resilience, degrades soil structure, and contradicts principles of sustainable land management.

The long-term sustainability of the GFGP has also been called into question, especially if government incentives cease. There is already an emergence of repeated transition patterns: “grassland → forestland → grassland” and “cultivated land → forestland → grassland”, indicating that areas are seeing discontinuity in implementation by local officials and disengagement by famers, signaling “game problems” between farmers and the GFGP policy (Xing et al., 2023). In fact, more than 56% of farmers across the Loess Plateau have indicated that they would recultivate the sloping farmland if the compensation and subsidies from the government stop (Chen et al., 2007). A study by Zhang et al. (2024) showed that eighteen years post-conversion, only 37% of the lands converted from cropland to forest land in 2001 had been preserved, with 50% reclaimed as cropland and 3% degrading into grassland (Zhang et al., 2024). 

In addition to farmer livelihoods, there are questions around the viability of the program itself. Because of water shortages, species mis-selection and poor management, plant survival rates are low, at an abysmal 22-25%, despite initial promising growth (Chen et al., 2007). The trees that do survive face heightened vulnerability to pests and diseases in the challenging conditions of the Loess Plateau; this limits their growth and therefore their effectiveness in reducing runoff and controlling soil erosion. The monoculture plantations were also susceptible to exotic species intrusion, making the protection of existing native plants even harder (Chen et al., 2007), reiterating biodiversity concerns.

There are also unanswered questions about the afforestation’s net radiative forcing and global warming potential. The changes the GFGP brought to the Loess Plateau’s vegetative cover altered the surface energy balance. The net impact on climate is complex, with warming from reduced albedo of the region combined with cooling from increased evapotranspiration (Xiao, 2014). More research and regional climate modeling needs to be done to understand the effect of the GFGP on the energy balance of the region.

Undoubtedly, there have been benefits to the GFGP. Due to the GFGP, the vegetative cover of the Loess Plateau has increased by 25% (Davidson, 2025). The GFGP combatted soil erosion, along with the ensuing loss of soil organic carbon (SOC), effectively offsetting the negative effect of erosion between 1980 and 1999 (Wang et al., 2021) and mitigating the total potential soil loss by 41.3% (Yu et al., 2022). This prevention of soil erosion has led to increased SOC across the Loess Plateau. In the fight against climate change, the GFGP also significantly enhanced carbon sequestration capacity in the Loess Plateau, from 2.379 × 109 Mg/ha to 2.384 × 109 Mg/ha (Yu et al., 2022). Through subsidies and income growth, farmers saw an average increase of 5100 RMB/a per household (Xu, 2022).

However, research also indicates that if the program is to be continued, major changes need to be made to enhance overall ecosystem services.

Recommendations

The program has evolved through its history of implementation, indicating a willingness to incorporate learnings. Hence, the following recommendations have been consolidated from research criticizing the GFGP:

The GFGP must incorporate mixed forest tactics to increase biodiversity. By incorporating basic, existing mixed forest tactics, the GFGP can increase biodiversity significantly at essentially no cost (Hua et al., 2016).

The GFGP must extensively tailor its implementation with spatial and land-quality considerations, specifically rainfall zone and land-use type. For example, converting abandoned farmlands to grasslands is more effective than converting them to forests in certain parts of the Loess Plateau, with no difference in soil sequestration (Deng et al., 2014). The Chinese government should encourage the GFGP in non-cultivable, erosion-prone areas and implement a “variation-selection-replication-retention” model of other areas, outlined in Figure 1, as deemed fit (Guo et al., 2014).

The GFGP must be evaluated in future with a more integrated ecosystem service assessment, moving beyond vegetation productivity and net primary productivity alone to understand broader ecological trade-offs such as water stress, carbon gains, biodiversity, and more (Xing et al., 2023).

Methodology Review

The differing perspectives of He et al. (2021) and Ge et al. (2020) on the success and future of the GFGP lie in their respective scope of program evaluation, political bias (as a result of their funding source), and disciplinary perspectives. While He et al. take a government-aligned, agronomic approach focused on production and income indicators, Ge et al. conduct independent, climate-focused simulations that reveal critical ecological trade-offs.

Figure 2 outlines the differences between the methodologies and background of each article, as well as observations of potential implications.