Quantitative tree cover reconstruction during the past 2000 years in the central-northern parts of the Daxing'an Mountains, northeast China

Reconstructing the history of tree cover and exploring the relationship between forest growth and environmental factors are of great significance for predicting future vegetation changes and guiding ecosystem restoration and protection. On the basis of pollen data of two lake cores and three peat profiles in the central-northern parts of the Daxing'an Mountains and using multiple methods, we quantitatively reconstructed the tree cover of the past 2000 years. We found that the reconstruction results of the different methods are quite similar. Among all these methods, random forest had the highest prediction accuracy for modern pollen samples. Therefore, the random forest reconstruction results can better reflect changes in regional tree cover, which was approximately 38%-42% during 2000-200 Cal yr BP. It then obviously increased from a low value of approximately 40% to a modern level of approximately 45%. The precipitation and carbon dioxide (CO2) concentration were the main drivers for the tree cover before and after 200 Cal yr BP, respectively. This indicates that environments with elevated CO2 level in the near future will benefit the growth of vegetation in the central-northern parts of the Daxing'an Mountains. (c) 2022 Elsevier Ltd. All rights reserved.