Quality Asessment And Study Of Sea Surface Salininty Obeserved By Aquarius In The Bay Of Bengal

With the lowest sea surface salinity(SSS) in the Indian Ocean, the variability of the upper salinity structure in the Bay of Bengal(BOB) and its adjacent waters has profound impacts on the ocean circulation and air-sea exchanges, and thus on the climate of the neighbouring continents. However, the cognition of the SSS in these areas was confined by the sparse SSS data. At present the Aquarius/SAC-D mission and SMOS mission offer an unprecedented opportunity for studying SSS variability on many spatial and temporal scales.Aquarius SSS products(ADPS, CAP and CAP_RC) are compared with in situ data from Argo and RAMA in the BOB. NASA’s Tropical Rainfall Measuring Mission 3B42 product and HYCOM salinity are used to assess the discrepancies between salinity retrieved from Aquarius and in situ data. On that basis Aquarius CAP level 3 products are used to discuss the seasonal and interannual variation characteristic of SSS in the BOB. The evaporation, rainfall and wind data were used to analyze the interannual SSS variation in the BOB in the spring 2011 and spring 2012.Assessment analysis shows that level 2 data of three Aquarius products agrees well with in situ data from Argo floats and RAMA in the BOB. The CAP and CAP_RC products outperform ADPS in comparison with in situ data. For ADPS, CAP, and CAP_RC, the overall biases with respect to Argo are-0.14, 0.01, and 0.07; and RMSE(Root-Mean-Square-Error) of 0.60, 0.53, and 0.52, respectively. Compared with RAMA, the biases of ADPS, CAP, and CAP_RC are-0.26,-0.01 and 0.03; and RMSE of 0.47, 0.43 and 0.42. The analysis also reveals the spatial dependence of the discrepencies between Aquarius SSS and in situ data — the system bias and RMSE between three Aquarius SSS products and in situ data(Argo and RAMA) are relatively larger in the east equatorial region and northern part of BOB. Further analysis shows that the discrepency mainly comes from the different ways of Aquarius and traditional in situ measurements. The reasons for this are two-fold:(1) Aquarius radiometer estimats salinity within 2 cm, while the mean depth of the shallowest Argo salinity samples is 3 m or so. In the pluvial regions, such as the northern part of BOB and east equatorial region, the difference in depth of measurements results in Aquarius SSS biased lower than in situ data due to strong rainfall-induced near-surface salinity stratification. Our statistics shows that there is a trend of 0.12 psu/(mm h-1) using linear regression model, and within 6 hours since last precipitation, the differences between Aquarius SSS and in situ measurements due to rainfall are significant.(2) While a in situ measurement provides a local SSS estimate, a satellite retrieval represents a horizontal average over the sensor’s footprint. Our study indicates that variability on short horizontal scales has significant impact on the comparison between Aquarius data and in situ data. Specifically, the impact of small-scale variability for deviation of single point is random, and thus the impact of small-scale variability for system bias could be neglected. Its impact on the comparion between Aquarius data and in situ data are mainly embodied in RMSE: the RMSE increases with the increase of the strength of small-scale variability in SSS. The large discrepancies in the northern part of BOB are mainly caused by the difference in the way of measurements.Aquarius salinity products and Argo data are used to investigate the seasonal and interannual SSS variation in the study area on the basis of evaluation. Aquarius CAP SSS captures main seasonal variation characteristics of SSS in the BOB, including the SSS variation in the southern part of BOB due to the intrusion of salty water from Arabian Sea and the seasonal migration of fresh water in the northern part of BOB. The CAP SSS combined with surface current depicts the southward spreading of the fresh water in the northern part of BOB via East India Coastal Current during November and January. Besides, the SSS reveals anomalous salting(freshening) in the bay during the spring 2011(2012). Our results suggest that the reduced(enhanced) precipitation during the summer 2010(2011) in the northern bay reduced(enhanced) freshening in that region, and the anomalous salting(freshening) water was advected southward by the surface-layer circulation causing anomalous salting(freshening) in the whole bay. Meanwhile, the positive(negative) wind stress curl anomaly also contribute to the anomalous salting(freshening) event during spring 2011(2012) in the bay via impacting the vertical mixing processes.