Assessment of suspended particulate matter concentration retrieved by Aqua-MODIS and SeaWiFS in the East China Sea

Abstract

The East China Sea (ECS) is the 11th largest marginal seas around the world. ECS has widely continental shelf, and has relatively high concentration of suspended particulate matter (TSM) affected by the terrestrial material from the large rivers, including the Changjiang River, and also affected by the resuspension in the winter. Recently, several regional algorithms for the TSM retrieval in the ECS have been proposed, such as the algorithms developed by Zhang et al. (2010) and Han et al. (2006). Due to the variation of the optical properties of TSM, it is significant to study whether existing algorithms are adequate and reliable for the inversion of the concentration of TSM in the ECS for all seasons. Yet, up to now, the validation of the satellite retrieved TSM is still lack due to the insufficient of the in-situ data and the standard TSM product in the ECS. In the past three years, we have carried out four seasonal cruises covering the whole ECS, including the spring cruise from May to June 2011, the summer cruise in August 2009, the autumn cruise from November to December 2010, and the winter cruise from December 2009 to January 2010. In this paper, we firstly analyzed the spatial-temporal distribution of the TSM in the ECS. The results showed that there was remarkable seasonal variation with higher concentration in the winter half year and lower concentration in the summer half year. The concentration of TSM was higher inshore than that of offshore. The isolines were parallel to the shoreline as a whole. There was a turbid water tongue with notably seasonal variation spreading to southeast at the 29°N in the middle of the ECS. Finally, based on the remote sensing reflectance retrieved by the Aqua-MODIS and SeaWiFS data, the performance of the four inversion algorithms of TSM were evaluated using the in-situ measured TSM data in the ECS, including the Clark’s model in the SeaDAS, Zhang’s model, Han’s model and Tassan’s model. The results show that all of the TSM retrieved by the four models have large error as compared with the in-situ data, indicating the strong requirement of the improving the TSM retrieval in the ECS.