This dissertation links the natural and social sciences, using modeling techniques to enhance understanding of the functioning of a complex ecosystem and its relevance to humans. For this purpose, I developed a Regional Unified Metatomodel of the Brazilian Amazon (RUMBA) to simulate the Amazon forest provision of ecosystem goods and services and their contribution to human economy and welfare. The model was also used to simulate the potential effect of an incentive to reduce deforestation in return for a payment for avoided releases of carbon into the atmosphere.

Simulation was done from 1975 to 2100, with calibration performed for the first 25 years, and for four scenarios: a baseline scenario, based on historical trends, and four alternative scenarios based on different assumptions and policy choices. The baseline scenario shows deforestation proceeding at high rates, leading to decreasing provision of forest goods and services and increasing economic growth. The growth of GRP per capita, on the other hand, remains much smaller than that of GRP. Regional welfare decreases significantly over the simulated period. The overall monetary contribution of ecosystem goods and services to the regional economy is estimated as 5 times the GRP in year 2100.

Scenarios of increased investment in development yielded higher economic growth accompanied by lower levels of welfare, while opposite trends were found for scenarios of higher investment in human, knowledge and natural capital. Finally, results also show that in order for a monetary compensation to represent a significant incentive to land owners to reduce deforestation, higher prices for avoided carbon emissions would have to be set than current prices of the emerging carbon market.

Main research findings are that increasing land use change in the Brazilian Amazon incurs significant losses of ecosystem services without this being adequately offset by increasing monetary income or welfare of people. This reseach has also found that in the absence of significant incentives from global beneficiaries for any one ecosystem service, or a combination of incentives addressing several types of ecosystem services, rational land uses at the local level lead to sub-optimal provision of these services from the global perspective.