Landsat 8 Data as a Source of High Resolution Sea Surface Temperature Maps in the Baltic Sea
Landsat 8 Data as a Source of High Resolution Sea Surface Temperature Maps in the Baltic Sea
复制标题
DOI：
10.3390/rs13224619
复制
发表时间：
2021-11
期刊：
影响因子：
--
通讯作者：
K. Bradtke
中科院分区：
文献类型：
--
作者：
K. Bradtke
海表温度(SST)是一个关键的水文变量，可以通过卫星进行监测。具有足够高的空间分辨率以研究沿海水域次中尺度过程的热数据的一个来源可能是陆地卫星任务。陆地卫星8号上的热红外传感器在两个波段收集数据，这允许使用众所周知的非线性分裂窗公式来估计SST(NLSST)，该公式使用大气层顶部(TOA)亮度温度。为了校准它的系数，需要大量的匹配点，代表了广泛的大气条件。在这项研究中，使用了超过1200颗卫星数据和波罗的海浮标和平台6年多来的12个现场测量时间序列来选择匹配点，推导NLSST系数并对结果进行评估。为了滤除受云、冰或陆地影响的像素，使用了IdePix算法和质量评估带并对相邻像素进行了额外的测试。测试了各种旗帜的组合。结果表明，以大气较湿为特征的沿海地区NLSST系数可能高估了较低的SST值。为波罗的海推导的公式产生的偏差接近0°C，RMSE在0.49-0.52°C的范围内。
Sea surface temperature (SST) is a key hydrological variable which can be monitored via satellite. One source of thermal data with a spatial resolution high enough to study sub-mesoscale processes in coastal waters may be the Landsat mission. The Thermal Infrared Sensor on board Landsat 8 collects data in two bands, which allows for the use of the well-known nonlinear split-window formula to estimate SST (NLSST) using top-of-the-atmosphere (TOA) brightness temperature. To calibrate its coefficients a significant number of matchup points are required, representing a wide range of atmospheric conditions. In this study over 1200 granules of satellite data and 12 time series of in situ measurements from buoys and platforms operating in the Baltic Sea over a period of more than 6 years were used to select matchup points, derive NLSST coefficients and evaluate the results. To filter out pixels contaminated by clouds, ice or land influences, the IdePix algorithm was used with Quality Assessment Band and additional test of the adjacent pixels. Various combinations of flags were tested. The results show that the NLSST coefficients derived previously for coastal areas, characterised by a more humid atmosphere, might overestimate low SST values. Formulas derived for the Baltic Sea produced biases close to 0 °C and RMSEs in the range of 0.49–0.52 °C.