Time series of measurement current and intermedaite water
Because the main float of the mooring system was affected by the current, its depth changes greatly, which led to deviations of the entire mooring system. Therefore, all the instruments and equipment designed for use at a predetermined depth were basically in a fluctuating state. The daily average salinity in Figure 2b shows large fluctuations for a maximum floating depth of 300 m.
The results for the current show the velocity structure and variation characteristics of the Kuroshio (Figure 2a). The maximum velocity of the Kuroshio can reach 1.0 m/s, and it can influence vertical depths to 800 m based on 0.2 m/s isotach, which is clear defined as the main poleward flow velocity of the Kuroshio (Chang et al., 2018; Jan et al., 2015). Figure 2b shows the average velocity of each layer with large velocities and velocity variations in the uppermost 400 m, which indicates the main depth of the Kuroshio in the vertical direction. The velocity deviation above 200 m is approximately 0.25 m/s, which is approximately one-half the magnitude of the velocity, indicating the instability of the Kuroshio velocity east of Taiwan. Of course, this instability may be caused by a variety of factors, such as the propagation of mesoscale eddies in the western direction, the movement of the Kuroshio main axis, etc. The most significant variation in the Kuroshio east of Taiwan is expressed as intraseasonal variations with periods of ~85 d, and these variations are mainly modulated by mesoscale eddies propagating westward from the STCC (Ren et al., 2020). The relationship between significant intraseasonal variations in the Kuroshio and IW is the focus of this study. The minimum salinity (\({S}_{min}\)) at the core of the intermediate water, shown in Figure 2c, is found mainly in the potential density range of 26.6–26.8 \({\sigma }_{\theta }\), where \({S}_{min}\) is approximately 31.5 psu and the depth is approximately 600 m. We also found that \({S}_{min}\) at the core showed discontinuous variability; for example, \({S}_{min}\) at the core was approximately 600 m during the period March-April 2016, and \({S}_{min}\) was approximately 550 m by September 2016. There are 7 results for \({S}_{min}\) at the core in the observation period, according to the total measurement time, and an intraseasonal variation period of approximately 70–80 days was estimated. The mean salinity of each layer is shown in Figure 2d, and the black line shows the deviation of the salinity. Although the overall standard deviation was relatively small, the variance was larger at 440 m than in the other layers, indicating a relatively large variation in salinity in the middle layer.
T-S characteristics of Intermediate water
To more clearly analyze the characteristics of IW east of Taiwan, we drew a T-S scatter plot of data obtained from the moored CTDs (Figure 3a). For comparison, historical data from the Argo international project for average temperature and salinity east of Taiwan but away from the Kuroshio area (box A1 in Figure 1.1a) and the South China Sea area (box A2 in Figure 1.1a) represent NPIW and SCSIW, respectively. The \({S}_{min}\) values of NPIW and SCSIW are 34.18 psu and 34.39 psu, respectively.
The main characteristics of IW in the water east of Taiwan are as follows: \({S}_{min}\) varies from 34.15 psu to 34.4 psu, corresponding to a temperature change in the range of 7 to 8°C and a potential density variation of 26.6–26.8 \({\sigma }_{\theta }\) for \({S}_{min}\), respectively. The salinity distribution near the characteristic salinity value of NPIW is denser (near the red curve in Figure 3a), indicating that the overall characteristics of the water mass are closer to those of NPIW during the observations. There were only two moments in time when \({S}_{min}\) exceeded 34.39 psu, as shown in Figure 3a, indicating that observations of IW with typical SCSIW characteristics are relatively infrequent. That is, most of the time, IW in the water east of Taiwan resembles a mixture of NPIW and SCSIW. This water mass was also defined as Kuroshio Intermediate Water (KIW) by earlier studies (Chen, 2005; Chern & Wang, 1998; Nakamura et al., 2013). Mensah et al. (2015) that reported that SCSIW could not flow directly to eastern Taiwan due to blocking by the Green Islands in southeastern Taiwan, but NPIW and SCSIW could mix at the relatively southern location of the Luzon Strait, and the Kuroshio carried this water mass to the east of Taiwan.
Therefore, we can obtain the typical average characteristics of this water mass, and the average \({\theta }\)-S curves are plotted in Figure 3a (blue curve). \({S}_{min}\) and the corresponding temperature of this water mass were 34.28 psu and 7.5°C, respectively. The core was located at 550 m, which corresponded to a depth between the \({S}_{min}\) core depths of 500 m for SCSIW and 600 m for NPIW.
To explore the proportions of NPIW and SCSIW in KIW east of Taiwan, the mixing ratio of the two water masses was calculated using the concentration mixing formula according to the following method. First, salinity values of 34.19 psu and 34.39 psu were taken to characterize NPIW and SCSIW, respectively, according to the red and green curves in Figure 3a, and the \({S}_{min}\) of each profile measured by the CTDs was taken to be the value of the two mixtures. Then, these three values were introduced into the concentration equation to calculate the mixing ratio. Figure 3b shows that the proportion of NPIW in KIW was relatively high, and 70% of the time, the NPIW mixing ratio exceeded 60%. There were approximately 7 moments in time when the proportion of NPIW reached 100%, which meant that there was basically only NPIW in the area east of Taiwan; these moments were evenly distributed during the observation period. The longest duration lasted for approximately one month in September 2016. Meanwhile, there were 4 moments when the proportion of NPIW was very small, such as March 2016, July 2016, August 2016, and June 2017. The proportion of NPIW in July 2016 and June 2017 was almost zero, indicating that there was basically no NPIW east of Taiwan, and SCSIW was predominant.
The above results show the great variability of IW found east of Taiwan, with some moments where SCSIW was directly observed, and other times when these was almost no evidence of SCSIW. In fact, the results of previous studies basically used the temperature and salt data of cruise sections to study intermediate water masses, but sectional data can only provide characteristics at certain moments, which are variable and cannot represent overall behavior. Therefore, their sectional data are insufficient for analyzing whether SCSIW is present east of Taiwan. Our direct observations explain the ambiguous results about the presence of SCSIW east of Taiwan from the previous study.
Intraseasonal variability of IW
The results of the power spectrum analysis of the salinity in the 500 m layer (Figure 4b) show that IW east of Taiwan had a significant intraseasonal period of 70–80 days. Through a reverse calculation of salinity based on an empirical temperature-salinity formula, Mensah et al. (2015) reported that the intraseasonal period of IW east of Taiwan was ~100 days. Compared with the result of Mensah et al. (2015), the intraseasonal signal of IW obtained from directly measured salinity data in this study may be more realistic reflection of the variation characteristics of the water mass. Of course, it is also possible that the difference in results is due to the timing of the two observations.
To better understand the intraseasonal variability, the meridional velocity anomaly, temperature and salinity anomaly were calculated by subtracting the average value during the observation of each layer from the average daily mooring data displayed in Figure 5b-d. The meridional velocity anomalies are basically consistent and banded in the 0–800 m range, and temperature and salinity anomalies also exhibit synchronization in the 400–800 m range. The alternating band structures of positive and negative shapes are clearly shown in the anomalous temperature and salinity graphs, also indicating an intraseasonal signal of approximately 3 months. The maximum negative and positive salinity anomalies were -0.12 psu and 0.1 psu, and the maximum negative and positive temperature anomalies were -1.5°C and 2°C, respectively. During the observation period, there were 6 negative salinity anomalies in 17 months, March-April, June, September-October, and November-December in 2016 and January-February and April-May in 2017; the positive anomalies occurred during the other observation times. Meanwhile, the temperature and salinity anomalies were consistent and showed synchronous changes. The distribution of the integrated current anomalies showed that at most moments, positive meridional velocity anomalies corresponded to positive anomalies of temperature and salinity, while negative meridional velocity anomalies corresponded to negative anomalies of temperature and salinity. The relatively consistent variation in current, temperature and salinity suggests that all three parameters may be influenced by the same factor.