Observational Analysis Of The Kuroshio Inflow At The East Taiwan Channel

The Kuroshio intrudes into the East China Sea (ECS) from northeast of Taiwan，playing an important role in the current circulation，nutrients distribution and materialtransport of the ECS. F0rom the point of view of observational analysis, we studiedthe structures and variabilities of the Kuroshio inflow at the East Taiwan Channel(ETC). Time series used in this study include WOCE moorings measurements, tidalstation records, satellite altimetry, assimilation model outputs, etc. Throughlag-correlation analysis we noticed some new phenomenon based on which a randomjet model was developped. Meanwhile, comparisons show that SODA is unable toreflect the mesoscale variations of the Kuroshio inflow at the ETC.The lag-correlation analysis show that at the western side of the ETC, there areno vertical phase-lag in temperature and velocity. In the middle of the ETC thetemperature of the upper layer lags that of the deeper layer. At the eastern side of theETC，the temperature and velocity of the shallow water lead that of the deep watersignificantly. These results run counter to the temperature phase tilt property inunstable baroclinic waves proposed by Sun (2007), indicating that the Kuroshioinflow at the ETC has no baroclinic instability. In the horizontal direction, thetemperature, velocity and sea surface height at the western side of the ETC all lag thatat the eastern side, reflecting the existence of westward propagating signal and furtherconfirming that the Kruoshio east of Taiwan is influenced by.mesoscale eddiescoming from the interior of the Pacific Ocean. Based on observational results arandom conceptual model of the Kuroshio jet is developed. It statistically shows thatthe shifts of the Kuroshio axis caused by mesoscal eddies and frontal waves exertsignificant influences on the horizontal correlation of the moorings measurements. When a westward propagating anti-cyclonic eddy reaches the eastern coast ofTaiwan, the sea surface height at the Ishigaki increases, the Kuroshio at the ETCflows to the northeast and its axis moves westward, the axisal speed and volumetransport increase. From top to bottom the horizontal velocity vectors tend to rotateanticlockwisely at the western side and rotates clockwisely at the eastern side. On thecontrary, when a cyclonic eddy reaches Taiwan，the sea surface height at the Ishigakidecreases, the Kuroshio axis east of Taiwan meanders anticlockwisely，the axis at theETC moves eastward with the axial speed decreases, the horizontal velocity vectorsfrom top to bottom tend to rotate clockwisely at the western side and rotateanticlockwisely at the eastern side. In the view of temporal mean, the horizontalvelocity vectors at the western side rotate clockwisely in the vertical direction,representing downward flow, and at the eastern side the flow rotates anticlockwisely,representing upward flow. This results is verified by numerical simulation withMITgcm.Long-term sea level records from three tidal stations at Ishigaki, Keelung andSuao are used to estimate the Kuroshio volume transport at the ETC, The results showthat the sea level difference between Ishigaki and Keelung provides the betterestimator with a correlation of0.73and root mean square difference of2.5Sv. Whenwe use the satellite altimeter data to estimate the Kuroshio volume transport at theETC, the best sea level difference is not between the two ends of the PCM-1line butbetween （24°N，123.5°E）and（27.25°N，123.25°E）, the correlation efficient is0.83and the root mean square difference is2Sv. Similar result is also found with SODAdata. The best sea level difference to estimate the Kuroshio inflow volume transport atthe ETC is not at the Ilan ridge, but with one point on the ECS continental shelfnortheast of Taiwan. Further analysis of with MITgcm shows that the key factor ofthis phenomonen is the the unstability of the geostrophic flow and the nonlinear effectcliffy that caused by the ridge topography. Based on these results, we analyzed theKuroshio volume transport through the ETC estimated by satellite altimeter data during the year1992-2010and found that there is a notable interannual variation in it.The results of wavelet analysis suggest the dominat perids are2and5years. Thecorrelation between monthly mean volume transport and PDO index is0.26with theformer lagging by3months.