Generating time-series grid data of sea surface temperature from isotherms in the Northwestern Pacific Ocean using coupled interpolation

Sea surface temperature (SST) isoline charts that were manually mapped using in situ SST data and satellite-derived SST data are valuable because they incorporate oceanographers' knowledge and experience. This type of SST data is useful for studying sea conditions of an area, for analyzing environmental factors that could affect fishing grounds, as a parameter for atmospheric or oceanic models, or as a diagnostic tool for comparison with the SSTs produced by ocean models. However, isoline maps must be digitized and interpolated into grid data in order to be used in these applications. Herein, we propose a coupled interpolation (CI), which couples improved multi-section interpolation and single-point change surface interpolation containing orientation, for generating grid data from SST isolines. We interpolated 1049 SST isoline maps (temperature interval 1°), which cover an area of the northwestern Pacific Ocean (125°E-180°E, 26°N-50°N) and were published by the Japan Fisheries Information Service Center (JAFIC) during 1990-2000, to grid datasets with 15′ grid resolution. We assessed the quality of grid datasets by checking noise points, RMSE analysis, checking offset errors, retrieving percentage of Kuroshio axes and visually comparing inverse isotherms with original isotherms. The quality analysis and comparison with four other interpolators showed the CI interpolator to be a good technique for generating SST grid data from isotherms. We also computed the SST anomaly (SSTA) using the SST grid datasets. The amplitude values of integral SSTA in the area of 31-46°N, 170-180°E were low, whereas they were high in the SW-NE rectangular area of 35-46°N, 142-160°E.