温排水监测海面采样与空间插值方法优化研究

本文以辽宁红沿河核电厂为研究对象，提出一种区别于传统放射性（La）的新的钟摆型（Lb）海面数据采集航线方案，同时基于对变分分析插值方法（DIVA）参数的校正，构建了温排水海面监测试验的最优化应用模式。在两种航线方案数据采集基础之上，分别采用反距离权重法（IDW）、克里金法（Kriging）、基于变分分析的插值方法进行空间拟合，分析不同结果的差异性和适用性。通过保留采样点的实测数据验证表明，Lb采集航线中的各方法误差绝对均值和均方差均低于La航线0.02左右，DIVA法误差绝对均值和均方差在Lb中均低于其他方法0.03以上；对基于遥感影像的交叉验证表明，Lb中各方法误差绝对均值在0.2~0.4之间，La在0.5~0.8之间，DIVA法在Lb中误差均值达到最小，为0.27。结果表明，钟摆型海面数据采集航线与基于变分分析插值方法相结合的模式有效的抑制了自然增温对温排水监测的影响，精度较高。

Optimization research on the ocean surface sampling and space interpolation method of thermal discharge monitoring

This paper took Hongyanhe nuclear power plant as the study area to put forward a new acquisition mode Lb that presents pendulum state from near shore to far. Lb was different from traditional acquisition mode La which presents radial state. At the same time, based on the correction of the variational analysis interpolation method (DIVA), the optimal measured model of the sea surface monitoring experiment was constructed. With the two kinds of data acquisition mode, Inverse distance weighted method (IDW), Kriging method and DIVA method were used to fitting space to analyze their applicability and compare the difference of interpolation results. Validated by the retaining sampling points from the measured data, it is showed that the MEAN and MSE of all interpolation methods of Lb mode is less than that of La about 0.03. DIVA method of Lb is lower than both IDW and Kriging methods with 0.02. Based on cross validation of remote sensing detection, it is indicated that the MEAN of all interpolation methods of Lb mode is between 0.2 and 0.4, and La mode is between 0.5 and 0.8. The MEAN of DIVA method of Lb is minimum, which is 0.27. With higher precision, the pendulum type acquisition mode combining with DIVA method can effectively restrain the influence of natural temperature increase during thermal discharging monitoring.