总体最小二乘法在浮游植物色素吸收系数模型估算中的应用

采用常规最小二乘法(LS)和总体最小二乘法(TLS)对浮游植物色素吸收系数与叶绿素a浓度之间的关系进行了研究。利用2003-2012年6个航次243组实测数据,建立了LS和TLS模型,应用模型估算典型波段的色素吸收系数值,并对两种模型进行了验证。采用平均绝对百分误差和均方根误差两种方法进行模型效果评估,结果显示,两种评估方法均能很好地说明,在400～700 nm光谱范围内,TLS拟合结果优于LS拟合结果。总体最小二乘法在综合考虑了参数误差和测量误差的情况下,拟合结果精度更高,更符合实际情况,估算结果的准确性要优于常规最小二乘法。

Application of Total Least Square Method in the Estimation of Pigment Absorption Coefficient Model of Phytoplankton

The relationship between phytoplankton pigment absorption coefficient and the concentration of chlorophyll a was studied by applying the regular least square (LS) method and total least square (TLS) method.  The LS and TLS models were established based on 243 sets of measured data obtained from 6 cruises during the period from 2003 to 2012. The pigment absorption coefficient values of typical bands were estimated by using the models, and the two models were verified. Two methods, mean absolute percentage error and root mean square error, were used to evaluate the model effect. The results showed that both evaluation methods can well explain that TLS fitting results are better than LS fitting results in the spectrum range of 400 ~ 700nm. When parameter error and measurement error are taken into account in a comprehensive way, the overall least square method is more accurate and more consistent with the actual situation, with the accuracy of estimated results being better than those obtained through conventional least square method.