基于QAA算法的巢湖悬浮物浓度反演研究

QAA(quasi-analytical algorithm)算法基于辐射传输原理,估算水体后向散射系数,是目前较为成熟的半分析模型方法.2009年6月对巢湖水体进行野外实测,在获得遥感反射率、吸收衰减系数、悬浮物浓度(cTSM)等数据的基础上,构建了巢湖水体cTSM的近红外双波段反演模型,并将QAA算法计算得到的后向散射概率作为输入,以期提高cTSM的反演精度.结果表明,①对巢湖水体而言,807 nm和834 nm是构建近红外双波段模型的最佳波段,807 nm处的单位散射系数为0.411 m2.g-1,834 nm处的单位散射系数为0.395 m2.g-1;②利用QAA算法计算得到的巢湖水体的后向散射概率为0.029,将该值输入反演模型后,cTSM的反演精度与输入其他经验值相比得到明显提高,反演值与实测值的均方根误差RMSE和平均相对误差绝对值MAPE分别达到12.143 mg.L-1和24.378%;③在cTSM较高的站点(近似高于30 mg.L-1),利用近红外双波段模型反演的悬浮物浓度结果更为稳定和可靠.

Monitoring the Total Suspended Matter of Lake Chaohu Based on Quasi-Analytical Algorithm

There has been some effort by using quasi-analytical algorithm(QAA), which is based on radiative transfer theory to calculate backscattering coefficients of waters. A field investigation in Lake Chaohu was carried out in June, 2009, and a dataset including many parameters such as remote sensing reflectance, absorption and attenuation coefficients, concentration of total suspended matter(cTSM), had been obtained. Thereafter, a model for estimating cTSM was built, and in order to improve the accuracy, we employed the QAA algorithm to simulate backscattering ratio of Lake Chaohu. The results show that: 1 807 nm and 834 nm are determined as optimum bands for the Two Near-Infrared Bands Method in Lake Chaohu,and the average specific scattering coefficient of suspended matters at the wavelength of 807 nm and 834 nm is 0.411 m²·g^-1 and 0.395 m²·g^-1respectively; 2The backscattering ratio of Lake Chaohu is 0.029, and better accuracy is achieved by using this value rather than other empirical values.The root mean square error(RMSE) and mean absolute percentage error(MAPE) of the estimated cTSM with calculated backscattering ratio are 12.143 mg·L^-1 and 24.378%; 3It is found that stations with high cTSM (nearly greater than 30 mg·L^-1) can have more stable and reliable estimated results.