This paper demonstrates an atmospheric correction method to process HJ-1A/B CCD images over Chinese coastal waters with the aid of MODIS-Terra aerosol information.Based on the assumption of zero water-leaving radiance at the two near-infrared (NIR)bands or the shortwave infrared(SWIR)bands,the atmospheric aerosol optical depth(AOD)is firstly retrieved from MODIS-Terra with a simple extension of the NIR-SWIR combined atmospheric correction approach embedded in SeaDAS (SeaWiFS data analysis system).Then the"turbid"and"non-turbid"waters are separated by the turbid water index.Maximum probability of AOD at 551 nm band is derived based on the NIR information over"non-turbid"waters and FLAASH model is selected to do the atmospheric correction of the HJ-1A/B CCD imagery using the AOD values of highest probabilities as input.Similarly,according to the histogram of the AOD distribution at 551 nm band,the studied turbid water area is divided into several blocks and HJ-1A/B CCD imagery is corrected with the corresponding AOD values displayed with the highest frequency in each block.This method has been applied to several HJ-1A/B images over Chinese coastal waters and validated by synchronous in-situ data.The results have shown that this method is effective in the atmospheric correction process of HJ-1A/B CCD images for ocean color remote sensing study and application in the coastal waters.
TIAN LiQiao 1,LU JianZhong 1,CHEN XiaoLing 1,2,3,YU ZhiFeng 1,XIAO JingJing 1,QIU Feng 1 & ZHAO Xi 1 1 State Key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan University,Wuhan 430079,China
Temporal and spatial patterns of inherent optical properties in the Bohai Sea are very complex. In this paper, we used 77 groups of field data of AOPs （apparent optical properties） and IOPs （inherent optical properties） collected in June, August, and September of 2005 in the Bohai Sea, to retrieve the spectral total absorption coefficient a（2） with the quasi-analytical algorithm （QAA）. For QAA implementation, different bands in the region 680-730 nm （in 5 nm intervals） were selected and compared, to determine the optimal band domain of the reference wavelength. On this basis, we proposed a new algorithm （QAA-Com）, a combination of QAA-685 and QAA-715, according to turbidity characterized by a（440）. The percentage difference of model retrievals in the visible domain was between 4.5%-45.1%, in average of 18.8% for a（2）. The QAA model was then applied to Medium Resolution Imaging Spectrometer （MERIS） radiometric products, which were temporally and spatially matched with in-situ optical measurements. Differences between MERIS retrievals and in-situ values were in the range 9.2%-27.8% for a（2） in the visible domain. Major errors in satellite retrieval are attributable to uncertainties of QAA model parameters and in-situ measurements, as well as imperfect atmospheric correction of MERIS data by the European Space Agency （ESA）. During a storm surge in April 2009, time series of MERIS images together with the QAA model were used to analyze spatial and temporal variability of the total absorption coefficient pattern in the Bohai Sea. It is necessary to collect more independent field data to improve this algorithm.
This paper demonstrates an atmospheric correction method to process MODIS/Aqua (Moderate-resolution Imaging Spectroradiometer) ocean color imagery over turbid coastal waters with the aid of concurrent CALIOP (Cloud-Aerosol LIdar with Orthogonal Polarization) aerosol data, assuming that there exists "nonturbid" water in the study area where MODIS aerosol optical properties can be retrieved accurately. Aerosol properties from CALIOP measurements were obtained and related to those from MODIS. This relationship, combined with CALIOP aerosol data, was extended to turbid water to derive MODIS aerosol properties, where atmospheric correction using MODIS data alone often fails. By combining MODIS and CALIOP data, aerosol signals were separated from the total signals at the satellite level, and water-leaving radiances in turbid waters were subsequently derived. This method was tested on several MODIS/Aqua ocean color images over South China turbid waters. Comparison with field data shows that this method was effective in reducing the errors in the retrieved water-leaving radiance values to some extent. In the Zhujiang (Pearl) River Estuary, this method did not overestimate the aerosol effects as severely, and provided far fewer negative water-leaving radiance values than the NASA (National Aeronautics and Space Administration) default methods that used MODIS data alone.
