Forest transitions in Chinese villages: Explaining community-level variation under the returning forest to farmland program

•Afforestation program implementation does not explain vegetation change across communities.•Vegetation gains shift across space and time from high-elevation to low-elevation communities.•Multiple behavioral drivers, linked to human-ecological factors, shift over time.•Tree cultivation and labor outflow join in a “policy plantation” forest transition pathway. China’s Returning Farmland to Forest Program (RFFP) aimed to transform rural landscapes and livelihoods by compensating households for planting trees on retired farmland. The program has been attributed a key role in an apparent forest transition. Studies uncover great local variation in its impacts, but the mechanisms behind them have received little attention. We examine such heterogeneity in 12 communities in northwest Yunnan, assessing the hypothesis that the RFFP catalyzed a state-led forest transition by evaluating the contributions of RFFP implementation and other processes to land cover change. Our dataset combines socioeconomic data from household surveys, focus groups, and intensive interviews with remote sensing data for a linked, cross-scale analysis. Results show no significant relationship between RFFP implementation and community-level vegetation cover change. Between 2000 and 2010, high-elevation communities had larger vegetation gains, while from 2010 to 2014, low-elevation communities had larger gains. Regression analyses and interview data show off-farm labor, tree crop planting, and changing energy sources influenced the rate of community-level vegetation change. This pattern, combining tree crop cultivation with labor outflows, may represent a distinct “policy plantation” pathway of forest gain. Meanwhile, new, high-elevation cash crops may be constraining forest expansion. This analysis suggests limited additionality for the RFFP in this region and highlights how heterogeneous, intersecting land use processes bring uneven forest transitions.