摘要
将四维LOESS(locally weighted regression)插值法应用至台湾海峡及其邻近海域,在考虑了研究海域海底地形和岸界对分析结果影响的基础上,通过引入空间和时间拟合函数,有效地解决了调查资料在时间和空间上的不均匀性.通过与气候态WOD01数据及卫星遥感资料的对比分析,可以发现,四维LOESS插值法不仅能够较好地再现研究海域的水文要素整体分布特征,而且对于台湾海峡及其邻近海域的中尺度现象(如夏季浅滩附近的局地低温、东山外海上升流等)也有较好的分辨能力.总的来说,四维LOESS插值法在台湾海峡及其邻近海域的应用是可行、可靠的.
The four-dimensional weighted least squares (LOESS) interpolation method is applied to the Taiwan Strait and its adjacent seas, which has not only considered influences of seabed topography and land borders, but also overcome the spatial and temporal discontinuities by introducing both space and time-fitting functions. Com- pared with the WOD01 and remote-sensed data, the four-dimensional LOESS interpolation method can not only re- produce the general hydrology features in the research zone, but also has good capability in resolving various me- soscale phenomena in the Taiwan Strait and its adjacent regions (e. g. the summertime local cool water adjacent to Taiwan Bank and the upwelling off Dongshan Island). In a word, it is feasible and reliable to apply the four-dimensional LOESS interpolation method to the Taiwan Strait and its adjacent seas.
出处
《应用海洋学学报》
CAS
CSCD
2013年第3期331-339,共9页
Journal of Applied Oceanography
基金
国家海洋局第三海洋研究所基本科研专项资金资助项目(海三科2008015
2008014)
国家海洋公益性行业科研专项资助项目(201105002-4)
国家自然科学基金资助项目(40876014)