Breakthrough in climate protection – Zoo House News
Rwanda is the first country to present a national inventory based on mapping the carbon stocks of each tree. Researchers from the University of Copenhagen, in collaboration with Rwandan authorities and researchers, have developed a method to accomplish this task.
“There are major uncertainties surrounding current forest assessments at the international level. By mapping the carbon inventory of each individual tree, accuracy is greatly improved. Also, the way different countries compile their inventories is not consistent due to different contexts, goals, and available datasets. We hope that this method will become a standard and thereby allow for better comparisons between countries,” says PhD researcher Maurice Mugabowindekwe, Department of Geosciences and Natural Resources Management (IGN), University of Copenhagen. He is the first author of the scientific paper that the presented a new method and the article was accepted for publication by Nature Climate Change, one of the most well-known journals in this field.
Maurice Mugabowindekwe, who is Rwandan himself, is helpful in the work, but the choice of Rwanda to develop the method was scientifically sound, he points out:
“The country has rich landscape variation, including savannas, forests, subhumid and moist forests, shrublands, agroecosystemic mosaics, and urban tree ecosystems representative of most tropical countries. We wanted to test the method for all of these landscape types. Rwanda is also a signatory to several international agreements on forest protection and climate protection. For example, as part of the Bonn Challenge, Rwanda has committed to restoring approximately 80% of its land by 2030. This is a reliable method for monitoring tree carbon.”
First method for mapping single trees
Preserving natural forests and planting new trees are considered important ways to limit climate change. However, large uncertainties about the carbon content of the trees make it difficult to assess the effectiveness of concrete initiatives. The researchers at the University of Copenhagen have overcome this problem.
The new method benefits from databases that indicate the relationship between the extent of the crown and the total carbon content of an individual tree.
“Mapping individual trees and calculating their carbon stocks has traditionally been done in forestry, albeit on a much smaller scale. The bottom line is to expand these approaches from a very local to a national level,” says researcher Ankit Kariryaa, who is 50 :50 at the IGN and Institute for Informatics (DIKU). Scientists from these two departments at the University of Copenhagen developed the method under the leadership of the IGN in collaboration with other international scientists.
The new method will help Rwanda to check the fulfillment of obligations under programs such as the global climate protection program REDD+ for forestry or the African Forest Landscape Restoration Initiative, AFR 100.
Many trees are found outside of forests
Manually mapping an entire country’s trees would be a huge effort and prohibitively expensive. Thus, the new method represents a breakthrough, as no other method would realistically be able to provide the same information at the individual tree level.
“It’s important to take a holistic approach and also include trees that are outside of forests,” says Ankit Kariryaa, noting that 72% of the trees mapped were in cropland and savannas and 17% in plantations.
At the same time, the relatively small proportion of trees occurring in natural forests – 11% of the total number of trees – accounts for about 51% of Rwanda’s national carbon stock. This is possible mainly because natural forests have a very high carbon content per tree volume, thanks to the very low level of human intervention secured by national legislation.
‘This suggests that the conservation, regeneration and sustainable management of natural forests is more effective than planting in mitigating climate change,’ comments Maurice Mugabowindekwe.
Rainforest seems to be “a giant green blanket”.
It is crucial that the computer can distinguish between the individual trees. This is because the relationship between crown expansion and the total carbon content of a tree varies greatly depending on the size of the tree. A very large tree will have a much higher carbon content than a group of trees with the same combined canopy extent. So if the group were mistaken for a tree, the carbon content would be significantly overestimated. A deep neural network is used to identify each tree.
“Especially for the rainforest, it is a big challenge to determine how many different trees are present in an image. At first glance, the forest only appears as a huge green blanket can also be used to identify the individual trees in the overgrowth of dense forests,” explains Christian Igel, Professor of Machine Learning at DIKU.
Training the computer using verified examples is the core of machine learning. In the Rwandan study, the computer was trained on a set of approximately 97,500 manually delineated tree canopies representing the full range of biogeographical conditions across the country.
The study used publicly available aerial and satellite imagery of Rwanda with a resolution of 0.25 x 0.25 m. These images were collected in June-August 2008 and 2009 and were provided by the Rwanda Land Management and Use Authority and the University of Rwanda for provided. More than 350 million trees have been mapped.
Applications beyond Rwanda
Nine researchers from the University of Copenhagen visited Rwanda in July 2022 with the twin aims of fieldwork and presenting the results of the first nationwide mapping to the Rwandan authorities and other stakeholders in the country’s forest sector.
“The presentation was well received,” reports Maurice Mugabowindekwe. He was promptly tasked by the Rwandan authorities with an updated mapping based on recent aerial photographs from 2019. This work will now continue.
In addition to Rwanda, the method has already been tested for a handful of countries. These include Tanzania, Burundi, Uganda and Kenya.
“We hope that nationwide high-resolution satellite imagery can also be purchased for these and other countries to enable the same approach to be applied. Also, we encourage the inclusion of funding for regular high-resolution imagery along with localized field inventory databases in development packages to enable similar work around the world,” says Maurice Mugabowindekwe, adding:
“The method has shown good results when applied directly to a new country or region. If the model is further trained on a local set of samples, the accuracy becomes even better. Size and carbon inventory represents the first step in monitoring the impact of landscape restoration efforts as well as conservation. If you cannot create an accurate and reliable inventory, you risk missing a framework to track the impact of landscape restoration. This could make the conservation and sustainable management of forests and other tree-dominated landscapes impossible. Therefore, this is a science that is likely to have implications.