结合高光谱数据反演吉林石头口门水库悬浮物含量和透明度

Estimating total suspended sediments concentrations and transparency with hyper-spectral reflectance in Shitoukoumen Reservoir,Jilin Prvoince

水中悬浮物含量是评价水环境质量的一个重要参数,它可影响水体透明度、混浊度、水色等光学性质,决定太阳光照在水下的分布和浮游植物对光照的利用,并最终影响水体的初级生产力．本文利用长春市石头口门水库的高光谱实测数据和水质采样分析数据,尝试通过一阶微分法建立悬浮物估测模型,再用估测结果反演透明度信息．结果表明:用590 nm处的一阶微分光谱值建立的悬浮物估测模型的决定系数R^2和均方根误差(RMSE)分别为0．76、9．09mg/L,验证模型的R^2和RMSE分别为0．79、7．15mg/L;石头口门水库上游严重的水土流失导致透明度受悬浮物含量影响较大,两者存在明显的指数关系,相关系数r为-0．80,用悬浮物含量建立的透明度估测模型的R^2和RMSE分别为0．79和0．12 m,验证模型的R^2和RMSE分别为0．82、0．11 m．显著水平均为p＜0．01．研究结果表明,该方法用于石头口门水库悬浮物含量和透明度的定量遥感结果较理想．

Concentrations of total suspended sediments （TSS） and transparency （SDD） are important indicators of water quality. In order to determine them of inland water using remote sensing, field hyperspectral data and water samples were colleeted in the Shitoukoumen Reservoir of Changchun on April 27 and June 28, 2006. The serious soil erosion upriver led to close exponential relationship between TSS concentrations and secchi disk transparency （correlation coefficient was -0. 80）, which facilitated the estimation of secchi disk depth （SDD） with TSS concentrations. It was found that the first-derivative model at 590 nm gave the best result to estimate TSS （ determination coefficient R^2 was 0. 76 and root mean square error （RMSE） was 9.09 mg/L）. The estimate result could be used to calculate the SDD （R^2 and RMSE of the model were 0. 79, 0. 12m respeetively at significant level P 〈 0.01 ）. Although the method presented here fits the data for Shitoukoumen Reservoir of Changchun, further analysis of reflectance in relation to water quality should be required from numerous geographic regions before universal equations could be established.