Geographic Weighted Regression Model Combined with Remote Sensing for Estimating Forest Aboveground Carbon Storage of Songyang County

doi:10.13466/j.cnki.lyzygl.2023.04.016

Abstract

Abstract:

Aboveground carbon(AGC)is an important indicator of the basic characteristics of forest ecosystems and a theoretical basis for evaluating the functional structure and productive potential of forests.As one of the nine key forestry counties in Zhejiang Province,Songyang County has a very important ecological status,so the accurate estimation of forest AGC in Songyang County is an important reference value for the stability evaluation of forest ecosystems and forest management in Songyang County.However,Songyang County is dominated by medium and low hilly areas surrounded by mountains,and how to consider the influence of complex terrain on the temporal and spatial variation of AGC is an urgent problem to be solved to achieve accurate estimation of AGC in mountainous forests.Therefore,based on Landsat TM satellite imagery and forest AGC survey data in Songyang County,a geographically weighted regression model(GWR)combined with spatial variation characteristics was constructed to estimate forest AGC,and compared with the results of ordinary least squares(OLS),finally,the optimal model was selected to predict forest AGC and its spatial distribution in Songyang County.The results were asfollows:Texture information from Landsat TM satellite imagerywas important for predicting forest AGC in Songyang County;The GWR model accurately estimated the AGC and spatial distribution of forests in Songyang County,and improved the accuracy by 9% over the OLS model,with an R² of 0.71.The total AGC of forests in Songyang County was 3.901×10⁶ Mg,with an average AGC of 23.70 Mg/hm²,accounting for about 10% of the total AGC of forest vegetation in Lishui City,which had a relatively important position in serving regional ecological functions.The study will provide advanced technical tools for accurate estimation of forest AGC in Songyang County,as well as scientific data for evaluating the function of forest carbon sink in Songyang County.

Key words: forest aboveground carbon storage, remote sensing, geographic weighted regression, model, Songyang County

CLC Number:

S771.8

ZOU Weimin, CHEN Chao, HUANG Lei, SONG Meixuan, LI Xuejian, DU Huaqiang. Geographic Weighted Regression Model Combined with Remote Sensing for Estimating Forest Aboveground Carbon Storage of Songyang County[J]. FOREST RESOURCES WANAGEMENT, 2023, 0(4): 132-140.

Figures/Tables 8

Fig.1

Tab.1

Fig.2

Fig.3

Fig.4

Fig.6

Fig.7

References 35

[1]	Mao Fangjie, Zhou Guomo, Du Haqiang, et al. Forest biomass carbon sinks in East Asia,with special reference to the relative contributions of forest expansion and forest growth[J]. Global Change Biology, 2014, 20(6):2019-2030. doi: 10.1111/gcb.12512
[2]	刘腾艳, 毛方杰, 李雪建, 等. 浙江省竹林地上碳储量的时空动态模拟及影响因素[J]. 应用生态学报, 2019, 30(5):1743-1753. doi: 10.13287/j.1001-9332.201905.035
[3]	Zhang Meng, Du Huaqiang, Zhou Guomo, et al. Estimating forest aboveground carbon storage in Hang-Jia-Hu using landsat TM/OLI data and random forest model[J]. Forests, 2019, 10(11):1004. doi: 10.3390/f10111004
[4]	Campbell M J, Dennison P E, Kerr K L, et al. Scaled biomass estimation in woodland ecosystems:Testing the individual and combined capacities of satellite multispectral and Lidar data[J]. Remote Sensing of Environment, 2021, 262:112511. doi: 10.1016/j.rse.2021.112511
[5]	Chang Zhongbing, Hobeichi S, Wang Yingping, et al. New forest aboveground biomass maps of China integrating multiple datasets[J]. Remote Sensing, 2021, 13(15):2892. doi: 10.3390/rs13152892
[6]	陈庆, 郑征, 冯志立, 等. 云南普洱地区思茅松林生物量及碳储量研究[J]. 云南大学学报:自然科学版, 2014, 36(3):439-445.
[7]	刘茜, 杨乐, 柳钦火, 等. 森林地上生物量遥感反演方法综述[J]. 遥感学报, 2015, 19(1):62-74.
[8]	韩雪莲, 张加龙, 刘灵, 等. 基于遥感特征变量的高山松碳储量抽样估算[J]. 西南林业大学学报:自然科学版, 2023, 43(6):1-9.
[9]	Du Huaqiang, Zhou Guomo, Ge Hongli, et al. Satellite-based carbon stock estimation for bamboo forest with a non-linear partial least square regression technique[J]. International Journal of Remote Sensing, 2012, 33(6):1917-1933. doi: 10.1080/01431161.2011.603379
[10]	戚玉娇, 李凤日. 基于KNN方法的大兴安岭地区森林地上碳储量遥感估算[J]. 林业科学, 2015, 51(5):46-55.
[11]	张博, 欧光龙, 孙雪莲, 等. 空间效应及其回归模型在林业中的应用[J]. 西南林业大学学报, 2016, 36(3):144-152.
[12]	邱新彩, 郑冬梅, 王海宾, 等. 结合地统计学与Landsat 8影像的乔木林地上碳储量估算[J]. 中南林业科技大学学报, 2020, 40(11):138-146.
[13]	Wang Jingyi, Du Huaqiang, Li Xuejian, et al. Remote sensing estimation of bamboo forest aboveground biomass based on geographically weighted regression[J]. Remote Sensing, 2021, 13(15):2962. doi: 10.3390/rs13152962
[14]	王海宾, 侯瑞萍, 郑冬梅, 等. 基于地理加权回归模型的亚热带地区乔木林生物量估算[J]. 农业机械学报, 2018, 49(6):184-190.
[15]	叶森土, 金超, 吴初平, 等. 浙江松阳县生态公益林群落分类排序及优势种种间关联分析[J]. 浙江农林大学学报, 2020, 37(4):693-701.
[16]	Bright B C, Hudak A T, Kennedy R E, et al. Examining post-fire vegetation recovery with andsat time series analysis in three Western North American forest types[J]. Fire Ecology, 2019, 15(1):8. doi: 10.1186/s42408-018-0021-9
[17]	周蓉, 赵天忠, 吴发云. 基于Landsat 8遥感影像的地上生物量模型反演研究[J]. 西北林学院学报, 2022, 37(2):186-192.
[18]	郭含茹, 张茂震, 徐丽华, 等. 基于地理加权回归的区域森林碳储量估计[J]. 浙江农林大学学报, 2015, 32(4):497-508.
[19]	刘伟, 潘永柱, 徐肇友, 等. 百山祖国家公园公益林碳储量及分配特征[J]. 生态学杂志, 2021, 40(1):1-10.
[20]	邢晓语, 杨秀春, 徐斌, 等. 基于随机森林算法的草原地上生物量遥感估算方法研究[J]. 地球信息科学学报, 2021, 23(7):1312-1324. doi: 10.12082/dqxxkx.2021.200605
[21]	McFeeters S K. The use of the Normalized Difference Water Index(NDWI)in the delineation of open water features[J]. International Journal of Remote Sensing, 1996, 17(7):1425-1432. doi: 10.1080/01431169608948714
[22]	廖瑶, 李雪, 刘芸, 等. 基于植被指数的高分一号遥感影像火烧迹地提取评价[J]. 自然灾害学报, 2021, 30(5):199-206.
[23]	Haralick R M. Statistical and structural approaches to texture[J]. Proceedings of the IEEE, 1979, 67(5):786-804. doi: 10.1109/PROC.1979.11328
[24]	Dong Luofan, Du Huaqiang, Han Ning, et al. Application of convolutional neural network on Lei Bamboo Above-Ground-Biomass(AGB) estimation using Worldview-2[J]. Remote Sensing, 2020, 12(6):958. doi: 10.3390/rs12060958
[25]	Wang Nan, Sun Ming, Ye Junhong, et al. Spatial downscaling of forest above-ground biomass distribution patterns based on Landsat 8 OLI images and a multiscale geographically weighted regression algorithm[J]. Forests, 2023, 14(3):526. doi: 10.3390/f14030526
[26]	孙钰森, 王维芳, 李国春. 基于地理加权回归克里格模型的帽儿山地区森林碳储量空间分布[J]. 应用生态学报, 2019, 30(5):1642-1650. doi: 10.13287/j.1001-9332.201905.024
[27]	闾妍宇, 李超, 欧光龙, 等. 基于地理加权回归模型的思茅松生物量遥感估测[J]. 林业资源管理, 2017,(1):82-90.
[28]	刘畅, 李凤日, 贾炜玮, 等. 基于局域统计量的黑龙江省多尺度森林碳储量空间分布变化[J]. 应用生态学报, 2014, 25(9):2493-2500.
[29]	卢宾宾, 葛咏, 秦昆, 等. 地理加权回归分析技术综述[J]. 武汉大学学报:信息科学版, 2020, 45(9):1356-1366.
[30]	Zhang Wei, Liu Dan, Zheng Shengjie, et al. Regional precipitation model based on geographically and temporally weighted regression kriging[J]. Remote Sensing, 2020, 12(16):2547. doi: 10.3390/rs12162547
[31]	Chen Huafeng, Qin Zihao, Zhai Deli, et al. Mapping forest aboveground biomass with MODIS and Fengyun-3C VIRR imageries in Yunnan province,Southwest China using Linear regression,K-Nearest neighbor and random forest[J]. Remote Sensing, 2022, 14(21):5456. doi: 10.3390/rs14215456
[32]	Zhang Xiao, Liu Liangyun, Chen Xidong, et al. Fine land-cover mapping in China using Landsat datacube and an operational SPECLib-based approach[J]. Remote Sensing, 2019, 11(9):1056. doi: 10.3390/rs11091056
[33]	Hu Xisheng, Zhang Lanyi, Ye Limin, et al. Locating spatial variation in the association between road network and forest biomass carbon accumulation[J]. Ecological Indicators, 2017, 73:214-223. doi: 10.1016/j.ecolind.2016.09.042
[34]	Cao Lin, Coops N C, Sun Yuan, et al. Estimating canopy structure and biomass in bamboo forests using airborne LiDAR data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2019, 148:114-129. doi: 10.1016/j.isprsjprs.2018.12.006
[35]	Zhang Bo, Li Xuejian, Du Huaqiang, et al. Estimation of urban forest characteristic parameters using UAV-Lidar coupled with canopy volume[J]. Remote Sensing, 2022, 14(24):6375. doi: 10.3390/rs14246375

遥感变量类型	遥感变量因子	参考文献
原始波段	蓝光波段(B)、绿光波段(G)、红光波段(R)、近红外波段(NIR)、短波红外1波段(SWIR 1)、短波红外2波段(SWIR 2)
植被指数	归一化植被指数(Normalized Difference Vegetation Index,NDVI)、增强植被指数(Enhanced Vegetation Index,EVI)、比值植被指数(Ratio Vegetation Index,RVI)、差值植被指数(Difference Vegetable Index,DVI)、归一化水指数(Normalized Difference Water Index,NDWI)、归一化红外指数(Normalized Difference Infrared Index,NDII)	[20?-22]
纹理特征	对比度(Contrast,CON)、相关性(Correlation,COR)、相异性(Dissimilarity,DIS)、熵(Entropy,ENT)、方差(Variance,VAR)、均质性(Homogeneity,HOM)	[23]