桂北岩溶区与非岩溶区植物叶片含水率高光谱反演

Hyperspectral inversion of leaf water content at karst and non-karst areas in North Guangxi

关键词:  岩溶区, 叶片含水率, 叶片绒毛, 光谱指数, 高光谱

Key words:  karst area, leaf water content, leaf pubescence, spectral index, hyperspectrum

摘要:

叶片水分状况是反映植被生理状况的重要指标,构建高普适性的植物叶片含水率高光谱反演模型对准确评价岩溶和非岩溶植被生态功能具有重要意义。该文以我国西南典型区域内岩溶区与非岩溶区共17种植物694个样品为研究对象,同步测量叶片含水率和反射光谱,采用单波段、差值型、比值型、归一化型等四类光谱指数模型,对反射光谱及一阶导数光谱进行全波段搜索分析。结果表明:光谱指数D₂₀₄₈-D₁₇₃₃的建模与验证结果均最好,为岩溶植物叶片含水率的最佳估计光谱指数; 对于非岩溶植物,光谱指数D₂₃₅₆/D₁₈₈₅和(D₂₃₅₆-D₁₈₈₅)/(D₂₃₅₆+D₁₈₈₅)建模与验证结果均相差很小,均可作为叶片含水率最佳估计模型。该研究结果还表明叶片绒毛对岩溶植物叶片含水率反演光谱指数构建影响不大。总体上,新构建的最优光谱指数对岩溶区与非岩溶区植物叶片含水率的拟合效果优于传统指数,具有较好的普适性,可为精准评估桂北地区植被水分状态提供科学依据。

Abstract:

Leaf water content is a key indicator that reflects the physiological conditions of vegetation. It is of great significance to construct a highly universal hyperspectral inversion model of the water content of leaves to accurately evaluate the ecological functions of karst and non-karst vegetation. As a result, developing hyperspectral inversion models of leaf water content for karst and non-karst vegetation, respectively is becoming more necessary. In total, 694 samples from 17 plant species were taken at karst and non-karst areas in North Guangxi. Leaf water content and reflectance spectrum ranging from 350 nm to 2 500 nm of all leaf samples were measured simultaneously. Four model structures including single-band, difference, ratio, and normalized difference were employed in this study. All possible index models with single waves and couple of two waves based on both reflectance spectrum and first derivative spectrum were related to leaf water content for karst and non-karst vegetation, respectively. The results were as follows: The spectral indexes of D₂₀₄₈-D₁₇₃₃ had the best modeling and verification result, and was the best index model for estimating leaf water content of karst plants; For non-karst plants, the spectral index of D₂₃₅₆ / D₁₈₈₅ and(D₂₃₅₆-D₁₈₈₅)/(D₂₃₅₆ + D₁₈₈₅)had similar results, and both of them could be used as the best estimation indexes of leaf water content. This results also showed that, for karst plants, the leaf pubescence had little effect on the construction of inversion model of leaf water content. Generally, the newly constructed optimal spectral index has a better fitting effect on the leaf water content in karst and non-karst areas than the traditional index, and has a good general applicability, which could provide a scientific basis for the accurate assessment of vegetation water state in North Guangxi.