What’s Going On Here?

A team of international scientists combined detailed satellite imagery with a deep learning algorithm and managed to find millions of trees in the West African Sahara and Sahel. Using this kind of beneficial AI technology will contribute to a better understanding of the importance of trees outside of forested areas and their role in mitigating degradation, climate change and poverty.

What Does This Mean?

Although a large proportion of dryland trees and shrubs grow in isolation, previously used ‘normal’ satellite imagery was unable to identify those individual trees and assessments so far were based on interpolations, estimations and projections. This supported the assumption that the Sahara Desert is a desolate wasteland, covered only by sand dune fields.

An international team led by University of Copenhagen and NASA researchers set out to refute this conception and revealed that the West African Sahara and Sahel area is in fact home to more than 1.8 billion trees.

But how did they manage to count individual trees across this large dryland region?

First of all, they were able to use high resolution satellite images from DigitalGlobe which could make out individual trees and measure their crown size. The University’s computer science department then developed an AI deep learning algorithm to enable the counting of those now visible trees. By feeding the deep learning model thousands of images of various trees, AI could recognise shapes and shadows which indicate the presence of trees. Within a matter of hours, the model was able to identify and map trees over large areas based on the recognition of tree shapes – monitoring trees outside of forests like this would take thousands of humans several years.

Why Should We Care?

It’s no secret, that trees play a key role in biodiversity and are crucial to our long-term survival. However, most public interest relating to trees to date is devoted to forests, leaving individual trees not well-documented. Researchers plan to expand the count of those non-forest trees and eventually create a global database of all trees growing outside forest areas.

New knowledge about trees in dry-land areas like this is important because currently “trees outside of forested areas are usually not included in climate models and we know very little about their carbon stocks – they are an unknown component in the global carbon cycle” according to Assistant Professor Martin Brandt.

Therefore, those growing efforts to map the Earth’s trees and using this kind of beneficial AI technology has great potential when it comes to documenting changes on a global scale and ultimately, in contributing towards global climate goals – if the number of trees can be mapped, so can the amount of carbon they store.

Be Curious!

  • Read the full scientific article here.
  • Check out the most high-profile existing world tree map which is released annually by Global Forest Watch.
  • Sign up for Climate Change AI’s newsletter or browse through their paper with this interactive summary and discover ways machine learning can reduce greenhouse gas emissions and help society adapt to a changing climate.
  • Last but not least (if you haven’t done so already) switch your search engine from Google to Ecosia and start planting trees for free while you search the web – one, two, tree and go.
Show CommentsClose Comments

Leave a comment