粗面岩质火山碎屑物发育土壤有机质含量的高光谱特征与建模

Hyperspectrum characteristics and Modeling of Organic Matter Content in Soils Developed from Trachytic Tephra

利用ASD Field Spec 4便携式快速扫描分光辐射光谱仪,对采自吉林省长白山地区粗面岩质火山碎屑物发育的土壤进行光谱反射率测定,分析其光谱特征;对土壤原始光谱反射率进行一阶微分、二阶微分、倒数的微分、倒数的对数的一阶微分和倒数的对数的二阶微分等五种数学处理,并应用多元逐步回归分析建立土壤有机质含量的高光谱预测模型。研究表明:土壤有机质含量与原始光谱反射率在565~675 nm波段内呈显著负相关;一阶微分光谱在415 nm、445~605 nm波段内与土壤有机质含量呈极显著负相关,在705~985 nm、1015~1265 nm波段内呈极显著正相关,在865 nm波段处相关系数达到极大值0.87;建立的土壤有机质多元逐步线性回归预测模型中,以一阶微分模型为最优,R2为0.954,可用于粗面岩质火山碎屑物发育土壤有机质含量的快速测定。

Soil samples developed from trachyte pyroclastic collected and their spectral reflectance was measured by 4 matter in Changbai Mountain Region, Jilin Province were ASD FieldSpec spectroradiometers with the characteristics of portable and fast scanning. The original soil spectral reflectance was processed using five mathematical methods of the first order and the second derivatives, the differential of the reciprocal, and the first order and the second order derivatives of the logarithm of the reciprocal. The hyperspectral prediction model of soil organic matter （SOM） content was established by multiple stepwise regression analysis. It was found that： soil organic matter content and the original spectral reflectance showed a significant negative correlation in the 565 - 675 nm band. The first derivative spectra of the spectral reflectance was significantly negatively correlated with SOM content at 415 nm and 445 - 605 nm band and was significantly positively correlated at 705 - 985 nm and 1015 - 1265 nm band. Their correlation coefficient reached the maximum value of 0.87 at 865 nm band. Organic predictive models were established by the stepwise multiple linear regression method in six soil types, and the model of the first order derivative was the most optimal （R^2 of 0.954）. It could be used to rapidly determine SOM content derived from trachyte pyroclastic materials.