Conservation investments, including payments for ecosystem services (PES), have been increasingly devoted to protecting and restoring ecosystems. However, the efficiency and effectiveness of PES programs depend on the program design, biological, economic and social conditions, and dynamic trends in population and households. This dissertation focused on the efficiency and effectiveness of payments for ecosystem services. My study area is China's Wolong Nature Reserve for giant pandas where human interact with the environment under the Grain-to-Green program (GTGP) and the Natural Forest Conservation program (NFCP). Specific objectives for this dissertation are to: (1) evaluate the effects of social and economic factors on the reenrollment in the GTGP after the current contracts expired, (2) target land for enrollment in the GTGP, (3) assess the impacts of social capital and labor migration on the fuelwood consumption of indigenous people, and (4) model the impacts of dynamics in population and households on the effects of the NFCP. This research use interdisciplinary methods and tools for human- environment interactions in a coupled human and natural system (CHANS).

I used stated choice method with main effects design to measure the effects of conservation payment, social norms and program duration on the reenrollment in the GTGP, controlling for household characteristics and land features. In addition to conservation payment amounts and program duration, social norms had significant impacts on program reenrollment. Farming income had a negative effect on program reenrollment, while income from rural-urban labor migrants had a positive effect on program reenrollment. I then explored cost-effective targeting of land for enrollment in the GTGP. Environmental benefits of lands and opportunity costs for land enrollment were estimated using land features and household characteristics. The efficiency of investments in a discriminative payment scheme (payments differ according to opportunity costs) was substantially higher than in a flat payment scheme (same price paid to all participants). In addition, both optimal targeting and suboptimal targeting achieved substantially more environmental benefits than random selection of land.

To assess the impacts of social capital and labor migration on the fuelwood consumption, I used propensity score techniques. Results suggested that social capital in the form of weak social ties to people in urban settings had significant impacts on rural-urban labor migration. Following the chain of capital substitutions, labor migration increased household income, which in turn reduced fuelwood consumption. Simulation results from the systems model suggested substantial panda habitat can be obtained from both cash payment and electricity payment scenarios. However, electricity payment, as a more direct payment approach for reducing human impacts, can improve the efficiency of conservation investments. The effects of conservation investments are non-linear due to increases in human population and number of households. In addition, policy effects can be uncertain due to uncertainties in the behavior of new households that have not been included in the NFCP. The approach and methods in this research may also be applied in many other CHANS.