- New paper by Hansen et al., can visualize changes in forest cover at 30m resolution
- Can help country experts or researchers to better assess forest cover and annual forest cover changes
- Could potentially be useful for REDD+, but needs careful analysis to ensure human-induced forest change is accurately attributed
Editor’s note 20/2/14: The Hansen data have now been made publicly available for download. Click here.
In pioneering new research published in Science, a team led by Matthew Hansen from the University of Maryland, has analyzed 650,000 freely available Landsat images to visualize changes in forest cover at a whopping 30-meter resolution.
“This is the first global wall-to-wall annual forest cover map available at such a high resolution that covers more than one decade,” said Arief Wijaya, post-doc researcher at the Center for International Forestry Research (CIFOR).
Why is it useful?
Many developing countries still lack quality data, the funds to acquire and the capacity to analyze satellite data to estimate annual deforestation rates.
Estimating these rates is needed as baseline data to calculate reference emissions levels (RELs) for carbon accounting in climate change mitigation activities such as reducing emissions from deforestation and forest degradation (REDD+).
The data, expected to become freely available in January 2014, can add transparency and consistency in assessing large area forest changes. For those countries and researchers that currently do not have access to such data, this will be a significant upgrade (previously, the highest resolution that global land cover and forest maps provided was 250-300 meters).
“This could help policy makers and government authorities to analyse or reassess their national estimates,” Wijaya said.
Use data with care for specific applications
While it has been publicized as the most accurate way to estimate global forest loss, experts say that the findings should be used carefully for different purposes and applications.
“Forest cover change can mean many things,” said Martin Herold, professor at Wageningen University. “If there are hurricanes, big fires, pests this could be detected as forest cover change, even though these are commonly not related to human-induced deforestation. Also, sustainable forest management and rotational harvesting can be picked up but could represent a welcome deviation from business as usual.”
“The spatial and temporal detail of the new data is impressive and a real advertisement on what can be achieved with satellite data analysis today. Using the data in the context of specific accounting rules, however, requires attention,” he said.
“The data could potentially be useful for REDD+, but further analysis will be needed to ensure we accurately estimate human-induced forest area change,” he said.
One case study shows that Hansen’s forest loss estimate for Indonesia from 2009-2012 is almost triple the Indonesian Ministry of Forestry’s national deforestation estimates (17,000 square kilometers/year versus 6,500 square kilometres/year).
This discrepancy could be due to various factors such as different definitions of forest and deforestation, use of different minimum mapping units or different levels of uncertainties.
Hansen defines forest loss as the disturbance or complete removal of tree cover canopy (below 25% tree canopy cover). This means any conversion of natural forests, be it plantations, selective logging or shifting cultivation practiced by local communities, would be considered forest loss. The Indonesian Ministry of Forestry, however, considers plantations as a type of forest.
“Therefore, it is essential that forest practitioners systematically analyze and verify the understand the reliability of these maps and the differences to existing datasets with what is observed to be happening on the ground,” Wijaya said.
Explore the new data set here.