Towards Crowdsourcing Forest Monitoring: lessons from SAD powered by Google Earth Engine

(Throughout this week we'll be publishing a series of blog posts about our activities at Rio+20, the United Nations Conference on Sustainable Development. Our full schedule at the conference is available here, and follow our activities as they happen at #googleatrio20. This guest post is from Google Earth Engine partner Carlos Souza of Imazon.) 

In the last seven years Brazil has reduced deforestation considerably, and forest monitoring has been a key part of this effort. Imazon, a Brazilian NGO working to promote sustainable development in the Amazon, has developed a forest monitoring system called SAD, and has been working with Google to integrate Google Earth Engine into our work. We’re now ready to share our progress and to take the next step in our efforts to crowdsource forest monitoring in Brazil and, potentially, the rest of the world.

First a bit about Imazon and our work. SAD is an acronym, in Portuguese, for Sistema de Alerta de Desmatamento, or Deforestation Alert System. This system detects and measures deforestation and forest degradation in Brazil by analyzing MODIS satellite imagery. The idea is to track deforestation in the same way the government tracks inflation, making it a regularly updated indicator of the overall health of the country.

SAD analysis is quite complex, involving a great deal of satellite data. The end result of the analysis is a ratio of soil, photosynthetic vegetation, and non-photosyntheic vegetation for each pixel in a MODIS image covering the Brazilian Amazon. SAD tracks and reports deforestation and forest degradation on a monthly basis by calculating changes in this ratio for satellite images acquired at different times. We provide this map, along with deforestation alerts, to key authorities working to fight illegal deforestation in the Amazon.

This past weekend at Google’s “From the Ground to the Cloud” event at Rio+20 we proudly announced the next step in the evolution of SAD: SAD-EE, powered by Google Earth Engine. Starting in July, Imazon's monthly deforestation reports -- which includes deforestation happening while we are here at Rio+20 -- will be generated by SAD-EE. 




SAD-EE improves Imazon’s forest monitoring program in several ways:

• We can now access and process the data using Google’s cloud, which dramatically changes how we work. For instance, during the testing phase of using SAD-EE we reduced the amount of time we spend downloading and managing the very large data sets of MODIS images by 50%, and analysis in the cloud is much faster than on our desktop computers. Getting this information to the authorities faster can be translated into several hectares of forests saved each month 
• SAD-EE is integrated with the Internet, mobile phone and computer tablet technologies, making it easier for local organizations to access it. 
•  The system can be used outside Brazil, allowing other tropical forest countries to monitor their forests. Indeed, there is now a project to make this happen through a partnership with Google and the World Resources Institute, via the Global Forest Watch Program. 

SAD-EE's reporting tool

For me, the most exciting aspect of SAD-EE is that the system demonstrates the potential for Google Earth Engine to become a multi-sensor, multi-algorithm, multi-technology, crowdsourcing environmental monitoring platform. As Google Earth Engine evolves, it is allowing scientists and remote sensing users to share their knowledge and tools and enabling large groups of people to track and report changes in our planet.


Posted by Carlos Souza, Senior Researcher, Imazon