气候变化领域集成服务门户(CCPortal): Mapping tree carbon density using sentinel 2A sensor on Google Earth Engine in Darjeeling Himalayas: Implication for tree carbon management and climate change mitigation

CCPortal

DOI	10.1016/j.pce.2024.103569
	Mapping tree carbon density using sentinel 2A sensor on Google Earth Engine in Darjeeling Himalayas: Implication for tree carbon management and climate change mitigation
	Singh, Manendra; Arshad, A.; Bijlwan, Amit; Tamang, Mendup; Shahina, N. N.; Biswas, Ankur; Bhowmick, Arpan; Vineeta; Banik, Ganesh Chandra; Nath, Arun Jyoti; Shukla, Gopal; Chakravarty, Sumit
发表日期	2024
ISSN	1474-7065
EISSN	1873-5193
起始页码	134
卷号	134
英文摘要	The Himalayan region is a most fragile ecosystem globally. Trees make up around 90 % of the global biomass carbon pool and previous studies have shown that tree carbon balance cannot be easily assessed by conventional methods. Considering trees as a backbone of the forest ecosystem, present study assessed the heterogeneity in tree carbon density using field-inventoried data and NDVI-based modelling with Sentinel 2 A imagery on Google Earth Engine. The specific aim of the study was to assess the spatial distribution of tree carbon density in the Darjeeling Himalayas using Sentinel 2 A. The object-based classification of forest area using a random forest algorithm showed a high accuracy (Kappa coefficient value of 0.92, OOB error 0.17). The regression model using NDVI as a predictor of tree carbon demonstrated a good fit (R2 = 0.78) for predicting tree carbon density. Validation results show high accuracy of the regression model in predicting tree carbon density with a low RMSE of 9.39 Mg ha-1 (R2 = 0.80, % RMSE = 11.55 %). The classification of tree carbon density into five classes revealed that a significant proportion of the forest area (57.05 %) falls under moderate carbon density (50-75 Mg ha-1). In Darjeeling Himalayas, majority of forests are under the carbon density between 50 and 75 Mg ha-1. Improvement and conservation efforts must be directed for very low carbon density (01-25 Mg ha-1) areas covering 0.05 %, and high carbon density (75-100 Mg ha-1) covering 36.22 % of the forest area, respectively, to balance the overall carbon storage potential of the region.
英文关键词	Darjeeling himalayas; Carbon management; NDVI-Carbon modelling; Tree carbon density; Random forest
语种	英语
WOS研究方向	Geology ; Meteorology & Atmospheric Sciences ; Water Resources
WOS类目	Geosciences, Multidisciplinary ; Meteorology & Atmospheric Sciences ; Water Resources
WOS记录号	WOS:001178235800001
来源期刊	PHYSICS AND CHEMISTRY OF THE EARTH
文献类型	期刊论文
条目标识符	http://gcip.llas.ac.cn/handle/2XKMVOVA/302200
作者单位	Govind Ballabh Pant University of Agriculture Technology; Indian Council of Agricultural Research (ICAR); ICAR - Indian Agricultural Research Institute; Assam University; Indian Council of Forestry Research & Education (ICFRE); Rain Forest Research Institute (RFRI); North Eastern Hill University
推荐引用方式 GB/T 7714	Singh, Manendra,Arshad, A.,Bijlwan, Amit,et al. Mapping tree carbon density using sentinel 2A sensor on Google Earth Engine in Darjeeling Himalayas: Implication for tree carbon management and climate change mitigation[J],2024,134.
APA	Singh, Manendra.,Arshad, A..,Bijlwan, Amit.,Tamang, Mendup.,Shahina, N. N..,...&Chakravarty, Sumit.(2024).Mapping tree carbon density using sentinel 2A sensor on Google Earth Engine in Darjeeling Himalayas: Implication for tree carbon management and climate change mitigation.PHYSICS AND CHEMISTRY OF THE EARTH,134.
MLA	Singh, Manendra,et al."Mapping tree carbon density using sentinel 2A sensor on Google Earth Engine in Darjeeling Himalayas: Implication for tree carbon management and climate change mitigation".PHYSICS AND CHEMISTRY OF THE EARTH 134(2024).