植物叶片光谱角对土壤铀污染的指示作用

通过在土壤中添加浓度为0（对照组），25，75，125，175，275，375和485 μg·g-1的铀，进行5种植物（苎麻、印度芥菜、酸模、甘蓝型油菜、玉米）的盆栽实验，研究了不同生长期的叶片光谱角对土壤铀污染的定性和定量指示作用，分析了定量指示作用与叶片铀含量的关系，并筛选出来在苗期即能反演土壤铀含量的植物，为通过野外实测植物叶片光谱快速、安全地进行土壤铀含量的本底调查和动态监测提供了有效的途径。更重要的是，以该叶片尺度的研究为基础，后续可以开展在冠层尺度采用遥感影像进行大面积土壤铀污染监测的相关研究。实验结果与主要结论如下：（1）以实验植物在不同生长期的实测叶片反射光谱为基础，计算了土壤受到铀污染后，在五个波段区间（表征叶片色素的350～716 nm、表征红边和近红外平台的717～975 nm、表征水分的976～1 265，1 266～1 770和1 771～2 500 nm）的光谱角。以盆为单位统计5种实验植物在13个生长期的光谱角，绝大多数情况下均大于阈值。叶片光谱角对铀产生了350～2 500 nm的全面响应，可以定性指示土壤是否受到铀污染。（2）涵盖全部5种实验植物的8个生长期得到了以光谱角作为自变量、通过显著性检验的线性回归方程。其中7个线性回归方程的决定系数R2＞0.64，3个（苎麻-苗期、印度芥菜-开花期和油菜-蕾薹期）的决定系数R2＞0.81。综合其他反演效果评价指标，可以认为叶片光谱角也能定量指示土壤铀污染程度，但定量指示作用随植物种类、生长期的不同而不同，其中苎麻、印度芥菜在苗期的叶片光谱角就可以比较好地反演土壤铀含量。（3）当叶片铀含量较高时，其光谱角对土壤铀污染的定量指示作用较强。

Spectral Angles of Plant Leaves as Indicators of Uranium Pollution in Soil

In this paper, five plants (ramie, Indian mustard, Rumex, Brassica napus and maize) were pot cultured with 0 (control group), 25, 75, 125, 175, 275, 375 and 485 μg·g-1 uranium in the soil. The qualitative and quantitative indicating effects of leaf spectral angle on soil uranium pollution in different growth stages were studied, and the relationship between quantitative indicating effect and leaf uranium content was analyzed It provides an effective way to quickly and safely carry out the background investigation and dynamic monitoring of soil uranium content through field measurement of plant leaf spectrum. The results and main conclusions are as follows: (1) Based on the leaf reflectance spectra of experimental plants in different growth stages, the spectral angles of soil polluted by uranium in five bands (350～716 nm for leaf pigment, 717～975 nm for red edge and near infrared platform, 976～1 265, 1 266～1 770 and 1 771～2 500 nm for water) were calculated. In most cases, the spectral angles of the five experimental plants were greater than the thresholds of the control group. The spectral angles of leaves havecomprehensive responses of 350～2 500 nm to theuranium in soil, which can qualitatively indicate whether the soil is polluted by uranium or not. (2) Eight linear regression equationspassing the significance test with spectral angles as independent variables were obtained, covering all five experimental plants. The coefficient of determination R2 of 7 linear regression equations were ＞0.64, and R2 of 3 linear regression equations (ramie-seedling stage, Indian mustard-flowering stage and rape-bud bolting stage) were＞0.81. Combined with other inversion effect evaluation indexes, it can be considered that leaf spectral angles can quantitatively indicate the degree of soil uranium pollution, but the function of the quantitative indicator varies with plant species and growth period. (3) There was a positive correlation between leaf spectral angles and uranium contents in soil. (4) The leaf spectral angles of ramie and Indian mustard at the seedling stage can be used to retrieve soil uranium content, which is an outstanding characteristic for indicating soil pollution status as early as possible through plant spectrum.