Wildfires are a growing threat in a world shaped by climate change. Now, researchers at Aalto University have developed a neural network model that can accurately predict the occurrence of fires in peatlands. They used the new model to assess the impact of different fire risk management strategies and identified a set of interventions that would reduce the incidence of fires by 50-76%.
The study focused on the province of Central Kalimantan in Borneo in Indonesia, which has the highest density peat Fires in Southeast Asia. Drainage to support agriculture or housing expansion has made peatlands increasingly vulnerable to frequent fires. In addition to threatening lives and livelihoods, peat fires release large amounts of carbon dioxide. but, Prevention Strategies Difficulties were encountered due to the lack of clear and specific links between the proposed interventions and fire hazards.
The new model uses measurements made before each fire season in 2002-2019 to predict the distribution of peatland fires. While the results can be widely applied to peatlands elsewhere, a new analysis must be undertaken for other contexts. “Our methodology can be used in other contexts, but this specific model will have to be retrained on the new data,” says Alexander Horton, the postdoctoral researcher who conducted the study.
The researchers used a convolutional neural network to analyze 31 variables, such as land cover type and pre-fire indicators of plants and drought. Once trained, the network predicted the probability of a fire in every spot on the map, resulting in an expected distribution of fires for this year.
In general, neural network predictions were correct 80-95% of the time. However, while the model was usually right in predicting the fire, it also missed many of the fires that actually occurred. The model did not predict about half of the observed fires, which means that it is not suitable as a predictive early warning system. Larger clusters of fires tend to be well anticipated, while isolated fires are often missed by the grid. With more work, the researchers hope to improve the network’s performance so that it can also serve as an early warning system.
The team took advantage of the fact that fire predictions were usually correct to test the impact of different land management strategies. By simulating different interventions, they found that the most effective strategy was to convert shrubs and grasses into swampy forests, reducing the incidence of fires by 50%. If this were combined with blocking all drainage channels except the main ones, the fires would be reduced by 70% in total.
An alternative strategy might be to create more farmer, since good management greatly reduces the possibility of a fire. However, plantations are among the major drivers of forest loss, and Horton notes that “farms are mostly owned by larger companies, often based outside Borneo, which means that profits do not return directly to the local economy beyond providing labor for the workforce. local”.
finally, Fire Prevention strategies must balance risks, benefits and costs, and this research provides the information to do so, explains Professor Matti Cuomo, who led the study team. “We have tried to determine how the different strategies work. It is more about the media policy makers from providing direct solutions.
The results have been published in Earth and Environment Communications.
Alexander c. Horton et al., Targeted land management strategies can halve peatland fire incidents in Central Kalimantan, Indonesia, Earth and Environment Communications (2022). DOI: 10.1038 / s43247-022-00534-2
the quote: A machine learning model that can assess the effectiveness of management strategies for forest fire prevention (2022, September 9), retrieved September 9, 2022 from https://phys.org/news/2022-09-machine-effectiveness-strategies-wildfire.html
This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.