联合LiDAR和高光谱数据反演矿山生态修复区植被地上碳储量

Inversion of Aboveground Carbon Stock of Vegetation in Mine Ecological Restoration Area by Combining LiDAR and Hyperspectral Data

准确评估矿区植被地上碳储量,可为矿区生态修复及其碳汇贡献估算提供科学依据。利用无人机采集激光雷达和高光谱数据,分别建立乔、灌、草3种植物群落的地上碳储量估算模型,并在神东矿区大柳塔采煤沉陷生态修复区实现了应用与验证。研究发现:仅利用高光谱数据反演乔、灌、草植物群落地上碳储量的精度R^(2)分别为0.42、0.43、0.41,仅利用激光雷达数据反演乔、灌、草植物群落地上碳储量的精度R^(2)分别为0.64、0.44、0.36,融合LiDAR和高光谱两种数据反演乔、灌、草植物群落地上碳储量的精度R^(2)分别达到0.87、0.73、0.72;LiDAR和高光谱特征中的高度百分位变量和绿色指数分别与地上碳储量的相关性最高,对于提升地上碳储量的反演精度贡献最大,说明融合LiDAR和高光谱数据在反映群落形态结构和光合固碳特征方面的能力,可有效提高矿区植被地上碳储量的反演精度。研究结果表明:融合LiDAR和高光谱数据,可以实现矿区复杂地形和植被配置下的植被地上碳储量估算,为矿区生态修复碳汇贡献评估提供支撑。

Accurate assessment of aboveground carbon stock of vegetation in mining area can provide scientific basis for ecological restoration and carbon sink contribution in mining area. In this study,the aboveground carbon stock estimation models of three plant communities include arbor,shrub and herbs,were established by using unmanned aerial vehicle(UAV) to obtain Light Detection and Ranging(LiDAR) and hyperspectral data. The models were applied and verified in Daliuta coal mining subsidence ecological restoration area of Shendong mining area. It was found that the R^(2)of the aboveground carbon stock of arbor,shrub and herbs were 0. 42,0. 43 and 0. 41 respectively by using hyperspectral data,and the R^(2)of the above-ground carbon stock of arbor,shrub and herbs were 0. 64,0. 44 and 0. 36 by using LiDAR data,and the R^(2)were raised to 0. 87,0. 73and 0. 72 by combining LiDAR and hyperspectral data. Height percentile in LiDAR point cloud and green index in hyperspectral data had the highest correlation with aboveground carbon stock,and made the greatest contribution to improve the accuracy of aboveground carbon stock. It shows that the combination of LiDAR and hyperspectral images can reflect the community morphological structure and photosynthetic carbon sequestration, which can effectively improve the inversion accuracy of aboveground carbon stock of vegetation in mining areas. The study results show that the combination of LiDAR and hyperspectral data can realize the estimation of aboveground carbon stock of vegetation under complex terrain and different vegetation community types in the mining area,and provide support for the assessment of carbon sink of ecological restoration in mining area.