PSI’s overall ambition is to support management of tropical forest ecosystems under uncertainty. We decompose ecosystem management under uncertainty into three principal sub-issues:
- the risk of land use change (its ‘transitional potential’) at various geographic and temporal scales
- the potential loss of ecosystem services associated with land use change (which can be measured but with significant uncertainty), and
- the management of the resulting risk exposure through information and decision support
1. How much risk is there of land use change? The objective of this component is to develop a dynamically evolving ‘risk score’ that corresponds to the transitional potential of a given plot of land (its likelihood of changing from one land use to another) by considering socioeconomic, biophysical and institutional drivers of land use change. Like a consumer credit rating, this ‘risk score’ does not need to represent an exact probability, but should use multi-factorial data mining analysis to develop ordinal estimates of a plot’s propensity to change land use.
While the initial geographical focus will be on the Brazilian Amazon, the risk scoring approach should be applicable to other geographies in the greater Amazon Basin and ultimately to tropical forests around the world in Africa and SE Asia.
Several existing tools and approaches developed by INPE, NASA and other institutions can be brought to bear as inputs into this analysis. These include:
- Agent-based models drawing on socioeconomic data (including use of cellular automata)
- Geospatial data mining of satellite data and other data sources at high temporal resolutions
- Other machine learning approaches
2. Given land use change how much GHG emissions will take place? The overall objective of this component is to develop a forest GHG estimation approach that quantifies net GHG emissions flux from various types of land use change, at resolutions that can be coupled with the risk models developed in previous component. The estimation approach should reduce estimation errors to no more than +/- 30% in a cost-effective manner. In addition, the estimation approach should support future scenario analysis, i.e. apply ecosystem models like CASA-NASA to help quantify GHG risk associated with future land use change in scenario-driven what-if analysis.
3. What can be done to re-shape risk exposure? The overall objective of this component is to develop an R&D platform that facilitates collaborative decision-making by enabling immersive visualization of context sensitive data & information layers as well as optimal deployment of monitoring stations.