Forests hiding in plain sight

Posted by

A new global analysis looking at  the distribution of forests and woodlands has ‘found’ 467 million hectares of previously unreported forest in dryland ecosystems – a land area equivalent to 60% of Australia. In this day and age of advanced remote sensing how are such discoveries still possible?

With our current totalC_kHuAUXUAAYrGZ information society, it seems amazing that we can improve our knowledge of the
earths ecosystems any more. Wrong – a new study published in Science has ‘found’ new areas of forest across the world that increases our estimates of forest coverage by nearly 10%, a very significant finding with broad consequences for global carbon budgeting and dryland restoration and management. Check out the Video abstract  here.

The main reason for the change is that previous land type classifications have been based on older and lower resolution satellite imagery without much ground validation. This new reassessment has been made possible by access to higher resolution satellite imagery, through Google Earth Engine – including imagery from more than 210,000 dryland sites – as well as the incorporation of ground validated information from ecological plots established in Australia.

The study has increased the area of dryland forest across all inhabited continents, but concentrated around rainforest, desert and tundra biomes, to the south of the Sahara desert, around the Mediterranean, southern Africa, central India, coastal Australia, western South America, north-east Brazil, northern Colombia and Venezuela, and northern parts of the boreal forests in Canada and Russia. The differences in coverage estimates are most significant in Africa where drylands forest estimates have doubled.

New forest areas neighbouring rainforest, desert and tundra biomes; mainly to the south of the Sahara desert, around the Mediterranean, southern Africa, central India, coastal Australia, western South America, north-east Brazil, northern Colombia and Venezuela, and northern parts of the boreal forests in Canada and Russia (

So what?

This new analysis increases the area of dryland forest by 45%. By revealing that drylands—which make up about 40% of Earth’s land surface—have a greater capacity to support trees and forest than previously perceived, a unique chance is presented to mitigate climate change impacts through large-scale dryland conservation and afforestation actions.

Dryland ecosystems contain some of the most threatened, yet disregarded, ecosystems, while facing pressure from climate change and human activity. Climate change will lead to extended droughts, regional warming, and, combined with a growing human population, an increased risk of land degradation and desertification in dryland biomes. The most recent climate model simulations show that global climate change could cause dryland biomes to expand by 11 to 23% by the end of the 21st century. If this occurs, dryland biomes could cover more than half of the global land surface.

This mapping shows that dryland regions have a greater capacity to support trees than previously perceived and understood. With their low opportunity costs, drylands could therefore provide a unique chance to mitigate climate change through large-scale conservation and afforestation actions. It also shows the potential for improved livelihoods of the people in these areas. Large scale reforestation programs in China (The Great Green Wall) and Africa are currently under construction to plant trees across vast areas to the South of major deserts (Gobi and  Sahara) in an attempt to halt the southern advance of desert. These forest mapping results confirm the excellent potential of such initiatives to establish new forests in dryland regions.

The results also increase our estimates of global forest carbon stocks by 15 to 150 gigatonnes carbon (GtC), or by 2 to 20% (depending on the density of trees). That’s a lot of carbon, especially when you consider that the global emissions of carbon dioxide are approximately 30 gigatonnes.


Not that we can rest easy now and assume that the discovery of this ‘missing sink’ has solved climate change problems. It hasn’t. But these findings will help improve the accuracy of global models of terrestrial carbon sinks and improve our accounting of the global carbon budget and carbon inventories, which are required to be submitted under international climate conventions including the United Nations Framework Convention on Climate Change (UNFCCC) and the Kyoto Protocol. Identifying these significant carbon sinks in dryland areas, some of the poorest regions on earth, could also help support initiatives to return carbon offset benefits to these countries.

Considering the increasing role dryland forests play in preventing desertification, maintaining livelihoods and mitigating the impacts of climate change at regional and global scales, it is imperative that the long term trends in dryland forest cover and quality are continued to be monitored.

The sciency bit

The research, conducted by an international team of scientists and students from the Food and Agriculture Organization of the United Nations (FAO), Australia’s Terrestrial Ecosystem Research Network (TERN), the University of Adelaide and 14 organisations around the world, stems from the FAO’s Global Drylands Assessment phase of the Global Forest Survey.

The work used an innovative new global analysis of the distribution of forests and woodlands across drylands. This class of forest and woodland has previously been difficult to measure globally using satellite imagery or other remote sensing because of the relative low density of trees.

A new photo-interpretation tool developed at FAO, called Collect Earth, overcomes the limitations of automatic and often inaccurate categorisation of forest types of satellite imagery by using a simple validation check for tree number and density.

The team analysed very high spatial and temporal resolution satellite imagery of more than 210,000 dryland monitoring plots to calculate global forest cover and its change over time.

A critical component of the analysis was the use of TERN’s national network of over 500 ecosystem observation plots established in arid and semiarid regions (the AusPlots program) to complete the Oceania component of the study, and was crucial for providing the on-ground verification (or ground truthing) of the global analysis.

The resources provided by TERN were the only global plot-based data source readily available and accurate enough to validate satellite imagery, verify the observed numbers and density of trees, and assess observer accuracy. High quality information on Australia’s ecosystems, including data on vegetation and soils from over 500 dryland sites used in the study are openly available via the TERN Data Discovery Portal.


[Feature image: Earth’s forest just grew by 9% – New analysis finds hidden pockets of forest all over the globe (]

Article is a longer version of one published in The Conversation (12th May 2017)

Article based on following paper:

Jean-François BastinNora BerrahmouniAlan GraingerDanae ManiatisDanilo MolliconeRebecca MooreChiara PatriarcaNicolas PicardBen SparrowElena Maria AbrahamKamel AlouiAyhan AtesogluFabio AttoreÇağlar BassüllüAdia BeyMonica GarzugliaLuis G. García-MonteroNikée GrootGreg GuerinLars LaestadiusAndrew J. LoweBako MamaneGiulio MarchiPaul PattersonMarcelo RezendeStefano RicciIgnacio SalcedoAlfonso Sanchez-Paus DiazFred StolleVenera SurappaevaRene Castro (2017) The extent of forest in dryland biome. Science Vol. 356, Issue 6338, pp. 635-638. 12 May 2017.
DOI: 10.1126/science.aam6527

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s