Terra-i data reaching new horizons in the east side of the globe
By Oscar Bautista
Globally more than 1 billion people depend on forests for their livelihoods. Forests play a crucial role in climate regulation, ecosystem services provision and regulation, water supply, carbon storage and many other functions that support biodiversity. Currently the global rate of deforestation is substantial, and there is a growing need for timely, spatially explicit data that flag natural vegetation changes due to human activities.
Since the official presentation of Terra-i at the Rio+20 conference in 2012, the system, described in some media as “ The First Satellite Deforestation Tracker for Latin America” (see our blog post Terra-i: Strong response to tool’s launch during Rio+20), has been continuously monitoring the land surface of Latin America. It has made its findings freely-available to the public on a regular basis, and it has been an input for numerous analyses, research projects and technical reports, some of them conducted for our more than 1,800 registered users.
On December 2014 at the Global Landscapes Forum the World Resources Institute (WRI) announced the incorporation of CIAT’s Terra-i into the Global Forest Watch platform (see terra-i data in GFW), as part of this partnership Terra-i is now releasing the first version (beta) of the pantropical expansion dataset; in order to achieve this goal the Terra-i team have been working intensively during the last year.
The first step in this expansion process was to acquire a huge MODIS image dataset with data starting from January 2000 until September 2014 (nearly 4,200 Gigabits of data). Using automated processes this task took around 2 weeks of continuous MODIS data downloading.
Once the information was saved, we started the pre-processing stage. This consists of extracting the bands we use from the HDF MODIS files and converting those into compatible formats with Terra-i system. The next step was to process the data to detect vegetation change, and produce maps showing this. The last step was to check the results in order to verify our findings in comparison with other high resolution datasets.
The whole process of downloading, pre-processing, processing and verification takes approximately 16 hours per tile, the equivalent of two working days; the total number of tiles processed was 119, covering Africa (45), Asia (49) and Oceania (25) (Figure 1).
Figure 2 shows the results of the whole area being processed, we also include in this figure Latin America for comparison purposes, thus, the user is able to see the consistency of the data; the figure also indicates three zoom areas in different continents for more detailed visualization of detections.
The obtained data is gladly promising, the verification procedure helped us to identify desirable detection patterns of change, and when compared with Global Forest Change (GFC) a high resolution validated dataset, it was found that results follow similar trends; as expected small areas of change (< 6 hectares) are not detected by terra-i.
The figure 3 parallels GFC and new areas of terra-i data, the display shows similarities in data distribution within a tile, in this image is also possible to identify differences in detections caused by differences in the pixel size between datasets.
These results are part of the first phase of the implementation of Terra-i across the tropics. The next phase will consist of a detailed calibration of these data in combination with the use of high resolution imagery for validation and field activities in zones that we are working to define.
The preliminary continental trends for deforestation for the period January 2004 – September 2014 are shown in figure 4. The most affected countries in each continent are presented in the subsequent figures.
Habitat loss trends are different when looked in relation to the total area of each country (see figure 5) the data highlight the highest annual forest loss rates for each continent.
Dynamics of habitat loss differ between countries and continents. The rates of annual forest loss show a different panorama in comparison with the raw data. This is because countries with small areas may be more affected than bigger countries.
Although deforestation has slowed in the last decade, rates remain high in some parts of the world. Furthermore, during the last three decades, most deforestation has taken place in the tropics. By expanding the coverage of Terra-i´s data we aim to improve access to well-timed data that contribute to better decision making.
We recognize the need of our users to have near real time alerts across the tropics. However, as mentioned above the process of obtaining usable and dependable data takes time and dedication. Despite this, the Terra-i team keeps doing their best to accomplish this goal as soon as possible.
The next step in this endeavour is to jointly expand the use of Terra-i by governments in tropical countries, to do so, our team leader Louis will be visiting CIAT’s Asia office in Hanoi, Vietnam, for the entire month of March 2016, to work closely with colleagues there, who have a clearer idea of the vegetation change processes happening in Southeast Asia. This should help guide the successful implementation of Terra-i in the region.
For now a visualization tool of the beta data will be available online at: www.terra-i.org/terra-i/data/data-tropics.html
English version prepared with help of Neil Palmer (CIAT).