This study analyzes the observed subtidal currents, 1/12° global HYCOM model results, and the observed time series to interpret seasonal and interannual patterns in the behavior of the Kuroshio intrusion around the Luzon Strait (LS). The observations include current measurements conducted at mooring station N2 (20°40.441′N, 120°38.324′E) from 7 July 2009 to 31 March 2011, surface geostrophic currents derived from the merged absolute dynamic topography, and the trajectory of an Argo float during the winter of 2010–2011. Results from mooring station N2 confirmed the seasonal changes in the Kuroshio intrusion and the variation of the Kuroshio intrusion during El Niño event from July 2009 to April 2010 and La Niña even from June 2010 to March 2011. The strongest Kuroshio intrusion occurs in the winter, with successively weaker currents in spring, autumn, and summer. Comparison of relative differences (Δ_max (z)) in the maximum absolute value of monthly average zonal velocity components |U_max (z)| showed that the Kuroshio intrusion was stronger during the 2009–2010 winter (El Niño) than the 2010–2011 winter (La Niña). Furthermore, the relative differences (Δ_max (z)) in deeper layers exceed those of the surface layer. Circulation patterns in surface geostrophic currents and the Argo float trajectory confirmed the results of mooring station N2. The Kuroshio intrusion velocity variation modeled using the 1/12° global HYCOM model resembled the observation on both seasonal to interannual scales. Modeled variation in the zonal mean velocity anomaly was also consistent with Niño3, Niño4, and North Equatorial Current (NEC) bifurcation latitude indices, indicating concurrent impacts of the ENSO influence. Monsoon winds strongly affect the seasonal variation while the weak upstream Kuroshio transport induced by El Niño, strongly affects the interannual variation, such as 2009–2010 winter. In 2010–2011 winter, the impact of winter monsoon forcing still exists in the LS. However, the stronger upstream Kuroshio transport during this period did not allow the Kuroshio to penetrate into the LS deeply. This explains why the 2009–2010 winter Kuroshio intrusion (El Niño event) was stronger than that of the 2010–2011 winter (La Niña event).

Within the past 20 years, the international Argo program has acquired more than 2 million temperature and salinity profiles throughout the global ocean. It has become the most efficient means of obtaining ocean observations from upper–intermediate layers. The profile data provided by the Argo Global Data Assembly Centres (GDACs) are submitted by 11 data centres in 9 countries after undergoing quality control. However, because the quality of the submitted data lacks uniformity, users must conduct post-quality-control processing prior to using the data. For this purpose, the China Argo Real-Time Data Centre (CARDC) has developed a system to achieve both rapid global Argo data reception and post-quality-control processing of all temperature and salinity profiles. Already in operation, the system allows daily transfer of global post-quality-controlled Argo data to various operating divisions. The system synchronizes with the GDAC server once each day, extracts all updated Argo profiles, and then automatically reads the data and performs a post-quality-control process comprising 15 quality control tests. The automatic post-quality-control system can detect errors in observation time and satellite fix, as well as identify temperature and salinity spikes, frozen profiles, density inversions and salinity drift/offset. The data quality following the post-quality-control processing has been shown improved effectively. In the future, it is expected that CARDC will update the Argo global ocean observational data set once a quarter.

The international Argo Program was launched at the turn of the millennium. It has since collected over 2 million vertical profiles of temperature and salinity from the upper 2000 m of the global ocean. Gridded interpolation is a technology that gives full play to the advantages of these profiles because they are scattered. This study develops a global gridded Argo dataset, called GDCSM-Argo, by using an improved gradient-dependent correlation scale method. The dataset is theoretically verified, its error-related statistics are recorded, and it is compared with other datasets to establish its reliability. The results show that the maximum mean RMSEs are 0.8 °C for temperature and 0.1 for salinity, and more than 90% of the analysis results are reliable under the statistical probability of 95%. Not only can GDCSM-Argo adequately preserve large-scale signals in the ocean but also retain more mesoscale features than other gridded Argo datasets. Preliminary applications also verify that GDCSM-Argo can systematically describe the spatio-temporal features of multiple elements in the global ocean, and is a useful tool in many areas of research.