Climate change brings more frequent and more intense climatic extremes. However, it is unclear exactly how extreme weather conditions will affect the spread of vegetation in the future. This is an urgent issue for research to be able to mitigate the effects of extremities and their impact on vegetation.
A study published in Biology of global change explores large-scale relationships between vegetation and climatic characteristics through machine learning. This demonstrates that the combination of climate and remote sensing of land cover data with a tree-like structure predictable models Called decision trees can effectively highlight the climatic thresholds involved in structuring the distribution of dominant vegetation at different spatial scales.
The results of this study emphasize the importance of extreme climatic conditions in shaping the distribution of several major vegetation types. For example, drought or severe cold are necessary for the dominance of savannas and deciduous forests.
“One of the most important questions to be answered in further research is whether the climate thresholds recognized in this study are static or change depending on climate change in the future,” said researcher Hui Tang of the Department of Geosciences at the University of Oslo. . .
Collaboration between machine learning and vegetation experts
Predicting future vegetation distribution in response to climate change this is a challenging task that requires a detailed understanding of how large-scale vegetation distribution is related to climate. The research team, consisting of computer scientists, vegetation developers and vegetation experts, is studying the rules that come from decision tree models to see if they are informative and if they can provide additional ideas that could be included. in mechanistic models of vegetation.
“It is a difficult task to test whether it is based on data model is informative and reliable. This study emphasizes the importance of interpretable models that allow for such meaningful collaboration with experts in the field, ”says PhD researcher Rita Beigaite of the Department of Computer Science at the University of Helsinki.
“The main climate constraints identified in the study will be useful for improving process-based vegetation models and combining them with Earth system models,” says Hui Tang.
Rita Beigaite et al., Determining Climate Thresholds for Dominant Natural Vegetation Types on a Global Scale Using Machine Learning: Medium Climate vs. Extreme, Biology of global change (2022). DOI: 10.1111 / gcb.16110
University of Helsinki
Citation: machine learning helps determine the climatic thresholds that shape the distribution of natural vegetation (2022, February 25), obtained February 25, 2022 from https://phys.org/news/2022-02-machine-climatic-thresholds-natural -vegetation.html
This document is subject to copyright. Except for any honest transaction for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.