核电站温排水分布卫星遥感监测及验证

以大亚湾核电站附近海域为研究区，基于HJ-1B卫星热红外遥感数据IRS4，对核电站温排水空间分布特征进行了识别与验证。首先利用时空插值后的NCEP大气廓线数据，结合近地面气象观测信息对IRS4数据进行大气订正;其次查找表数据建立IRS4宽通道辐亮度与辐射温度转换公式，完成研究区海表温度反演。利用温盐深测量仪，与卫星同步采集了84个离散点的水温，经空间插值获得地面调查海水“体温度”（bulk temperature, BT）的分布，并与海表温度（sea surface temperature, SST）反演结果进行了对比。结果表明，调查区平均BT比平均SST高0.47 ℃。在远离热排放口区域，SST高于BT，两者差异在1.0 ℃以内;在靠近热排放口区域，SST低于BT，并且越接近排放口两者之间差距越大。遥感与地面调查两种手段获得的温升分布基本一致，总温升区域面积相近。遥感手段获得的温升等级较少（小于4级），地面调查获得温升等级较多（大于5级）;后者高温升等级（2级以上）区域面积较大，但对于1级温升区面积而言，两种手段差异较小。上午10时左右海表BT与SST差异较小，“肤-体”温差效应可以忽略，在IRS4业务化SST应用中，这一时段的卫星遥感可以作为核电站温排水分布监测的常规手段。

Monitoring the Thermal Plume f rom Coastal Nuclear Power Plant Using Satellite Remote Sensing Data:Modeling and Validation

Thermal plume from coastal nuclear power plant is a small-scale human activity, mornitoring of which requires high-frequency and high-spatial remote sensing data. The infrared scanner (IRS), on board of HJ-1B, has an infrared channel IRS4 with 300 m and 4-days as its spatial and temporal resolution. Remote sensing data aquired using IRS4 is an available source for mornitoring thermal plume. Retrieval pattern for coastal sea surface temperature (SST) was built to monitor the thermal plume from nuclear power plant. The research area is located near Guangdong Daya Bay Nuclear Power Station (GNPS), where synchronized validations were also implemented. The National Centers for Environmental Prediction (NCEP) data was interpolated spatially and temporally. The interpolated data as well as surface weather conditions were subsequently employed into radiative transfer model for the atmospheric correction of IRS4 thermal image. A look-up-table (LUT) was built for the inversion between IRS4 channel radiance and radiometric temperature, and a fitted function was also built from the LUT data for the same purpose. The SST was finally retrieved based on those preprocessing procedures mentioned above. The bulk temperature (BT) of 84 samples distributed near GNPS was shipboard collected synchronically using salinity-temperature-deepness (CTD) instruments. The discrete sample data was surface interpolated and compared with the satellite retrieved SST. Results show that the average BT over the study area is 0.47 ℃ higher than the retrieved skin temperature (ST). For areas far away from outfall, the ST is higher than BT, with differences less than 1.0 ℃. The main driving force for temperature variations in these regions is solar radiation. For areas near outfall, on the contrary, the retrieved ST is lower than BT, and greater differences between the two (meaning >1.0 ℃) happen when it gets closer to the outfall. Unlike the former case, the convective heat transfer resulting from the thermal plume is the primary reason leading to the temperature variations. Temperature rising (TR) distributions obtained from remote sensing data and in-situ measurements are consistent, except that the interpolated BT shows more level details (>5 levels) than that of the ST (up to 4 levels). The areas with higher TR levels (>2) are larger on BT maps, while for lower TR levels (≤2), the two methods perform with no obvious differences. Minimal errors for satellite-derived SST occur regularly around local time 10 a.m. This makes the remote sensing results to be substitutes for in-situ measurements. Therefore, for operational applications of HJ-1B IRS4, remote sensing technique can be a practical approach to monitoring the nuclear plant thermal pollution around this time period.