2023-04-24 06:30:59
Thanks to high-resolution satellite images and artificial intelligence, scientists have managed to count nearly 10 billion trees planted in the dry regions of sub-Saharan Africa, as well as quantify the quantity of carbon stored. This storage is lower than most previous estimates.
Trees in the drylands of sub-Saharan Africa not only represent a stock of carbon, but also provide valuable ecosystem services. They participate in the protection of the soil against erosion, fertilize it, provide shade for tree crops, and can also serve as timber and heating for local populations. Since they are widely dispersed, studying them on a large scale is difficult. Scientists from NASA and the University of Copenhagen, with the participation of those from INRAE, CEA and CNRS, have developed a new method for counting and characterizing these trees. Their research work has been published in Nature and provides insight into the role of these trees in ecological protection, climate change mitigation and restoration of these ecosystems.
The researchers used high-resolution (50 cm) satellite images as well as field surveys carried out at the plot scale to carry out their study. And thanks to the use of artificial intelligence algorithms, they managed to identify and characterize each tree one by one. The machine-learning model was first trained on nearly 90,000 trees, each of which was manually identified and delineated. Then he was able to identify every tree whose green crown is taller than 3 m2 on more than 320,000 satellite images and to observe their characteristics: height, surface of the crown, biomass… Even grouped trees could thus be identified individually.
In the end, more than 9.9 billion trees were counted over an area of nearly 10 million km2. Allometric equations were then used to estimate the carbon stocks corresponding to the dry masses of the foliage, wood and roots of each tree. As a result, stocks amount to an average of 63 kg for each tree located in the arid zone, and 98 kg for those planted in the sub-humid zone, which has an intermediate climate between semi-arid and humid.
Observing the evolution of these trees in near real time
In total, carbon stocks have been calculated at 0.84 billion tonnes in this region of sub-Saharan Africa which extends north of the equator. A figure that appears low compared to most of the previous estimates made by models used in the context of climate simulations. The study shows that the latter tend to overestimate the carbon stocks of isolated trees in arid zones. This overestimation could be explained, among other things, by the fact that they consider groups of trees to be equivalent to dense forests.
The next stage of this research work will consist of adding a temporal dimension to this mapping to observe the impact of droughts and restoration policies undertaken on the individual evolution of these trees. ” High spatial resolution is the key to improving tree inventories in drylandswrite the authors of this study. The ever-increasing availability of satellite imagery will make realistic, near real-time, continental-scale assessments of carbon pools and dynamics at the individual tree level. This will be key to developing strong drylands management programs needed to achieve the United Nations Sustainable Development Goals. Our article is a step in this process. »
For Jean-Pierre Wigneron, researcher at INRAE and co-author of the article: “This study is exceptional, because it is a pioneer in a type of approach that will revolutionize the monitoring of trees and forests on the scale of our planet: in the short term, it will become possible to map the trees of the planet from the center of the Amazon to our schoolyards. Many fields of application will become much more efficient and precise: monitoring carbon stocks, biodiversity, monitoring logging, protection against illegal forest degradation, etc., in near real time. »
1682323387
#quantification #carbon #stocks #trees #subSaharan #Africa