At first glance, the apparently barren expanses of the Sahel and Sahara deserts have little green, but detailed satellite images combined with in-depth computer training have revealed a different picture.
In fact, about 1.8 billion trees are part of the West African Sahara and Sahel deserts and the so-called subhumid zone, a previously innumerable generosity that overturns previous assumptions about such habitats, the researchers say.
“We were very surprised that there were quite a few (so many) large trees growing in the Sahara Desert,” lead author Martin Brandt told AFP.
“Of course, there are huge areas without trees, but there are still areas with a high density of trees, and even between the sand dunes here and there grow some trees,”; – added Brandt, associate professor of geography at the University of Copenhagen.
The survey provides researchers and conservationists with data that can help focus efforts to combat deforestation and more accurately measure onshore carbon stocks.
“For conservation, recovery, climate change, etc., such data is very important for establishing a baseline,” said Jesse Meyer, a programmer at NASA’s Goddard Space Flight Center who worked on the study.
“In a year or two or ten studies, it will be possible to repeat … to test whether efforts to revitalize and reduce deforestation are effective or not,” he said in a NASA press release.
Finding and counting trees was not an easy task. In areas with a large number of trees, dense clusters appear relatively clearly in satellite images, even at low resolutions, which are easy to distinguish from bare ground.
But where they are more common, satellite imagery may be too low resolution to highlight individual trees or even small groups.
Higher-resolution images are now available, but even then the problems remain: counting individual trees, especially over large areas, is an almost impossible task.
Brandt and his team came up with the solution by combining very high-resolution satellite imagery with deep learning – essentially teaching a computer program to do the work for them.
But that didn’t mean they could just sit back and wait for results.
Before the deep learning program could begin, it had to be taught, a burdensome process, when Brandt individually counted and marked nearly 90,000 trees himself. It took him a year.
“The level of detail is very high, and the model needs to know what all kinds of different trees look like in different landscapes,” he said.
“I did not accept the misclassification and additional training when I saw the incorrectly classified trees.”
Setting a baseline save
According to him, it was worth the effort to calculate what would take millions of years of work in a matter of hours.
“Other research is based on estimates and extrapolations, here we directly see and count each tree, this is the first estimate from wall to wall.”
Poll published Wednesday in the magazine Nature, covered an area of 1.3 million square kilometers (about 500,000 square miles) and included the analysis of more than 11,000 images.
This technique suggests that “it will soon be possible, with some limitations, to map the location and size of each tree around the world,” wrote Niall P. Hanan and Julius Anchang of the Department of Plant and Environmental Sciences at New Mexico State University in a review of the study.
And accurate information on vegetation in deserts and other arid areas is “fundamental to our understanding of global ecology, biogeography, and the biogeochemical cycles of carbon, water, and other nutrients,” they wrote in a review of the order. Nature.
Better information can help determine how much carbon is stored in these areas, which are not typically included in climate models, Brandt said.
But it is too early to say whether an accurate count of the life of this tree will affect how we understand climate change and its acceleration, he added.
He now hopes to use the technique elsewhere to map more previously hidden trees in 65 million square kilometers (25 million square miles) of arid regions of the world.
© Agence France-Presse