9:00 AM - 9:15 AM
[ACG49-01] A new method for identifying the Kuroshio/Kuroshio Extension and its application to high-resolution ocean reanalysis data around Japan (FORA-JPN60)
★Invited Papers
Keywords:Front identification, Kuroshio/Kuroshio Extension, Ocean reanalysis data, Dynamic sea surface height , Python script
Ocean models with a horizontal resolution of 1/10 degree or better can realistically reproduce the Kuroshio Current and the Kuroshio Continuation Current. Statistical processing of the Kuroshio current axis information is very important to analyze the results of such calculations.
Since the Kuroshio and the Kuroshio Extension are quite distinct fronts, it is often possible to draw them even if we use an appropriate value of sea surface height (SSH). For example, Nakano et al. (2018) identify the Kuroshio Extension using a time-varying but spatially constant SSH value. However, while this is appropriate for the open-oceanic follow-on currents, a constant SSH value is not appropriate for the western coastal boundary because the SSH field decreases from south to north and has large seasonal variations. For example, the figure shows an example of Kuroshio Current/Kuroshio Extension identification on Jul 1962. The red thick dashed line shows the case where a constant value of sea surface height is used. In this case, the value of sea surface height is too high for the Kuroshio Extension, which is identified to the south of the original Extension, but too low for the Ryukyu Islands, which reaches the East China Sea. This is an example where a single value of sea surface height cannot be used for identification.
Another method other than using a constant SSH value, such as the method of Ambe et al. (2004), is to use local flow velocities to determine the direction of flow and then sequentially search for the next location. While this method is reasonable in principle, it has the disadvantages of being computationally expensive as the resolution increases, and of being susceptible to local mesoscale eddies. This is a problem that is difficult to avoid for identification methods that use local values.
Based on these considerations, although the Kuroshio axis is obtained from the SSH field, instead of using a single SSH value from Taiwan to the Kuroshio continuation, constant SSH values for the Kuroshio are obtained east of 140E and west of 130E, where the SSH does not significantly, and then a function combining tanh is assumed between these SSH values to connect them. The implementation is based on scikit-image, a python image data processing library.
This method was adapted to FORA-JPN60, the latest 60-year ocean reanalysis data for the seas around Japan with 2 km resolution, and tuned until there were no failures in the 60-year monthly and daily average data. Kuroshio Current/Kuroshio Extension is reasonably identified in cases that could not be represented using constant values of sea surface height, as shown by the black bold lines in the figure. Using the information on these current axes, we classified the Kuroshio/Kuroshio Extension paths. Using this Kuroshio extension path, we also obtained the latitude of the Kuroshio Extension. The results of these reanalysis data will be briefly discussed at the conference, in addition to the details of the methodology.
Since the Kuroshio and the Kuroshio Extension are quite distinct fronts, it is often possible to draw them even if we use an appropriate value of sea surface height (SSH). For example, Nakano et al. (2018) identify the Kuroshio Extension using a time-varying but spatially constant SSH value. However, while this is appropriate for the open-oceanic follow-on currents, a constant SSH value is not appropriate for the western coastal boundary because the SSH field decreases from south to north and has large seasonal variations. For example, the figure shows an example of Kuroshio Current/Kuroshio Extension identification on Jul 1962. The red thick dashed line shows the case where a constant value of sea surface height is used. In this case, the value of sea surface height is too high for the Kuroshio Extension, which is identified to the south of the original Extension, but too low for the Ryukyu Islands, which reaches the East China Sea. This is an example where a single value of sea surface height cannot be used for identification.
Another method other than using a constant SSH value, such as the method of Ambe et al. (2004), is to use local flow velocities to determine the direction of flow and then sequentially search for the next location. While this method is reasonable in principle, it has the disadvantages of being computationally expensive as the resolution increases, and of being susceptible to local mesoscale eddies. This is a problem that is difficult to avoid for identification methods that use local values.
Based on these considerations, although the Kuroshio axis is obtained from the SSH field, instead of using a single SSH value from Taiwan to the Kuroshio continuation, constant SSH values for the Kuroshio are obtained east of 140E and west of 130E, where the SSH does not significantly, and then a function combining tanh is assumed between these SSH values to connect them. The implementation is based on scikit-image, a python image data processing library.
This method was adapted to FORA-JPN60, the latest 60-year ocean reanalysis data for the seas around Japan with 2 km resolution, and tuned until there were no failures in the 60-year monthly and daily average data. Kuroshio Current/Kuroshio Extension is reasonably identified in cases that could not be represented using constant values of sea surface height, as shown by the black bold lines in the figure. Using the information on these current axes, we classified the Kuroshio/Kuroshio Extension paths. Using this Kuroshio extension path, we also obtained the latitude of the Kuroshio Extension. The results of these reanalysis data will be briefly discussed at the conference, in addition to the details of the methodology.