In this study,we used the Navy's Master Oceanographic Observation Data Set (MOODS),consisting of 116 019 temperature and 9 617 salinity profiles,during 1968~1984 to investigate the temporal and spatial variabilities of South China Sea thermohaline structures and circulation.For temperature,profiles were binned into 204 monthly data sets from 1968 to 1984 (17 years).For salinity,profiles were binned into 12 climatological monthly data sets due to the data paucity.A twoscale optimal interpolation method was used to establish a three-dimensional monthly-varying gridded data set from MOODS,covering the area of 5°~25°N and 105°~125°E and the depth from the surface to 400 m.After the gridded data set had been established,both composite analysis and the Empirical Orthogonal Function (EOF) analysis (for temperature only) were used to identify the major thermohaline fratures including annual mean,monthly anomalies,and interarmual thermal variabilities.The inverted monthly circulation pattern using the P-vector method is also discussed.