Retrieval of total suspended matter concentration from Gaofen-1 Wide Field Imager (WFI) multispectral imagery with the assistance of Terra MODIS in turbid water – case in Deep Bay. Issue 14 (17th July 2016)

• Record Type: Journal Article
• Authors: Tian, Liqiao; Wai, Onyx W. H.; Chen, Xiaoling; Li, Wenbo; Li, Jian; Li, Wenkai; Zhang, Haidong
• Is Part Of: International journal of remote sensing. Volume 37:Issue 14(2016)
• Page Start: 3400
• Page End: 3413
• Publication Date: 2016-07-17
• Subjects: Remote sensing -- Periodicals; Télédétection -- Périodiques; 621.3678
• DOI: 10.1080/01431161.2016.1199084
• Languages: English
• ISSNs: 0143-1161

Abstract

ABSTRACT: This study proposed a method for developing high spatial resolution Gaofen-1 satellite (GF-1) Wide Field Imager (WFI)-based total suspended matter concentration ( C TSM ) retrieval model with the assistance of Moderate Resolution Imaging Spectroradiometer (MODIS) data, using the Deep Bay in China as a case. Based on long-term calibrated C TSM measurements of optical backscatter (OBS) 3A turbidity and temperature monitoring system of two stationary stations from January 2007 through November 2008, 33 match-ups were selected to build an exponential retrieval model for MODIS atmospherically corrected remote-sensing reflectance ( R rs ) ratio ( R rs, 645 / R rs, 555 ). Validation of the MODIS model showed well agreement with the seven in situ C TSM measurements with a root mean squared error (RMSE) of 5.06 mg l − 1 and a coefficient of determination R 2 of 0.80. Aided with six MODIS retrieved C TSM products, different band combinations (single band ( R rc, 660 ), band subtraction ( R rc, 660 – R rc, 560 ), band ratio ( R rc, 660 / R rc, 560 ), and total suspended matter index at 660 nm band (TSMI660 ) were evaluated for simultaneous GF-1 WFI Rayleigh-corrected reflectance ( R rc ). The results showed that the exponential model based on the Rayleigh-corrected reflectance ratio ( R rc, 660 / R rc, 560 ) could achieve acceptable accuracy, with RMSE of 14.80 mg l − 1 and R 2 of 0.62. The proposed method would be helpful for dynamic monitoring in the Deep Bay, and more important could also provide an alternative approach for studies when in situ measurements are unreachable.