UC Berkeley

Satellite imagery is an increasingly valuable data source in agricultural and resource economics. It offers credible, consistent, and globally comparable information on landscape-scale changes to the environment; and the supply of the information is accelerating with innovations in satellite technology. This dissertation presents the critical value of satellite imagery for environmental policy evaluation, focusing specifically on deforestation. Tropical deforestation may account for as much as one tenth of net global greenhouse gas emissions each year. Any viable effort to mitigate the impact of climate change must address deforestation. The effective design of forest policy is dependent on reliable impact analysis. Satellites offer a method of direct observation of forest cover loss, rather than incomplete or biased estimates. The dissertation first presents an algorithm to convert raw satellite imagery into a new and novel data source on tropical deforestation. Next, the data are used to explain the counter-intuitive outcomes from a 2012 conservation policy in Indonesia, relying on the spatial detail afforded by the new data. Finally, a set of empirical results describe the relationship of deforestation to interest rates, commodity prices, and other economic variables, using the temporal detail afforded by the new data. Together, these case studies demonstrate the effective use of a new data source in direct and credible policy analysis for natural resource economics.