昭苏县土壤含水量高光谱建模预测及其与植被生产力的关系  

Hyperspectral modeling prediction of soil water content and its relationship with vegetation productivity in Zhaosu County

作　　者：徐浩威 张飞 周静茹 张梦如 郭立阳 冯子恒 XU Haowei;ZHANG Fei;ZHOU Jingru;ZHANG Mengru;GUO Liyang;FENG Ziheng(College of Geography and Environmental Sciences,Zhejiang Normal University,Jinhua 321004,China;College of Geography and Remote Sensing Sciences,Xinjiang University,Urumqi 830017,China)

机构地区：[1]浙江师范大学地理与环境科学学院,金华321004 [2]新疆大学地理与遥感科学学院,乌鲁木齐830017

出　　处：《生态学报》2024年第12期5230-5245,共16页Acta Ecologica Sinica

基　　金：新疆维吾尔自治区重点研发任务专项计划(2022B02003)。

摘　　要：土壤是生态系统中不可或缺的组成部分,同时也是植物生长与发育的依托。土壤含水量直接影响土壤—植被生态系统中的水分和养分循环,从而决定植物的生长状态。因此,植被的生长状况与土壤含水量之间存在密切的关联。为了深入了解这种关系,本研究选取昭苏县3种优质牧草(冷蒿草、牛筋草和高羊茅)的地面实测高光谱数据作为研究对象,运用倒数、导数、对数等14种光谱变换方法,通过相关性分析找出与土壤含水量高度相关的特征波段,进行植被高光谱的土壤含水量建模预测,目的是要筛选敏感波长和最佳的反演模型。同时,利用CASA模型计算草地实际净初级生产力(ANPP),以分析土壤含水量与植被生产力的关联程度。结果表明:(1)昭苏县3种优质牧草在可见光波段光谱反射特征相似,而近红外波段差异较大,并且除25%—30%土壤含水量区间外,随着土壤水分增加,植被光谱反射率呈现明显下降趋势。(2)经过14种光谱变换后,昭苏县3种优质牧草的光谱反射率与土壤含水量间的相关性得到提高,尤其在R″、(1/R)′、(log R)′、(log 1/R)′、(R^(1/2))″等光谱变换形式下,相关性较高,且特征波段数量较多。(3)在(1/R)′、(1/R)″、(log 1/R)″、(R^(1/2))″等8种高光谱变换形式中,(R^(1/2))″和(log 1/R)″变换下土壤含水量估算模型精度较好。(4)在过去20年间,昭苏县草地ANPP呈明显下降趋势,草地退化主要集中在平原地带,而植被生产力的减少是多因素作用的结果,土壤含水量只是一个微弱的影响因子。总之,本研究为土壤含水量反演及其与植被生产力关系研究中提供新思路,为土地可持续利用和管理策略的指定提供重要参考。Soil is an indispensable component of ecosystems,serving as the foundation for plant growth and development.Soil water content directly affects the water and nutrient cycles within the soil-vegetation ecosystem,thereby determining the growth status of plants,which leads to is a strong correlation between vegetation growth and soil water content.To further investigate this relationship,this study selected ground hyperspectral data of three high-quality forage grass species(Artemisia frigida willd,Goosegrass and Tall fescue)in Zhaosu County as research samples.Fourteen types of spectral transformation methods,including inverse,derivative,and logarithmic,were applied to identify the characteristic spectral bands that are highly correlated with the soil water content by correlation analysis.Subsequently,a predictive model for soil water content was constructed based on vegetation hyperspectral data,with the aim of screening sensitive wavelengths and identifying optimal inversion models.Additionally,the CASA model was employed to calculate the actual net primary productivity(ANPP)of grassland,aiming to analyze the degree of correlation between soil water content and vegetation productivity.The results showed that:(1)The spectral reflectance characteristics of the three high-quality forage grass species in Zhaosu County were similar in the visible light bands but differed significantly in the near-infrared bands.Moreover,the vegetation spectral reflectance showed a clear decreasing trend with the increase of soil water,except for the 25%—30%soil water content range.(2)After undergoing 14 types of spectral transformations,the correlation between spectral reflectance of the three high-quality forage grass species in Zhaosu County and soil water content has been identified.Particularly,under the spectral transformation forms of R″,(1/R)′,(log R)′,(log 1/R)′,(R^(1/2))″,the correlation with soil water content was relatively high,with a large number of characteristic bands.(3)Among the eight spectral transformati

关 键 词：植被高光谱 光谱变换 土壤含水量 CASA模型 实际净初级生产力 

分 类 号：S152.7[农业科学—土壤学]