Abstract

Spatial vs. Temporal Resolution—Implications for Remote Sensing of Deforestation
Track: Modeling
Authors: Robin Kraft

Remote sensing of the environment typically places a high premium on spatial resolution at the expense of temporal resolution, whereas the element of time is crucial for stemming the spread of deforestation. Using the decade-long time series now available from the MODIS sensor aboard NASA's Terra platform, we have developed a tropical forest monitoring system around a restricted and semi-trained machine learning algorithm that analyzes MODIS vegetation index and thermal anomaly data. The rich time-series data allow the system to identify recent deforestation each month, a time-scale that can provide needed feedback on enforcement efforts and changes in forest management policies. The system is currently running for Indonesia and will be applied to other countries in coming months.

Robin Kraft
Center for Global Development
1800 Massachusetts Ave. NW
Washington, District of Columbia 20036
United States
Phone: 202-416-4056
E-mail: rkraft@cgdev.org