Satellite Remote Sensing of Cyanophyte Using Observed Spectral Measurements over the Taihu Lake

The chlorophyll a and cyanobacterial density are important variables for the evaluation of water quality and thus important for red tide monitoring. An evaluation of spectral measurements is implemented for the estimation of chlorophyll a (Chl a) and cyanobacterial density in the Taihu Lake. There are 39 sample points over the Taihu Lake during the experiment from 10 to 12 November in 2008. For each sample point, measurements of spectral reflectance and water quality sampling are conducted. Observation shows that cyanobacterial affect water reflectance greatly, leading to an obvious absorption peak in the red while strong absorption in the blue and near infrared bands. Spectral responses for points with little cyanophyte are similar to that of water reflectance. However, for the cyanobacterial points, spectral responses show the similar trend of vegetation to some extent. Besides, comparison between the reflectance obtained at nadir and at 45° departure indicates that the existence of cyanophytes has great effects on the visible and near infrared regions. This is because the increase of heterogeneity in water will increase the energy that can be acquired by the sensor. To investigate the operational application feasibilities for satellite remote sensing of water quality, equivalent reflectance based on FY 3A/MESRI and AQUA/MODIS band settings is derived using the spectral response functions. Comparison analysis indicates that the equivalent reflectance calculated from FY 3A/MESRI band settings is consistent to that of the AQUA/MODIS. Larger variations are observed for the cyanobacterial water indicating different sensitivity of these bands in water quality evaluation. Furthermore, the Ration Index (RI) model is used for the evaluation of water quality and high determination coefficients (RMS of 0.0174 mg·L^-1 and 0.0188 mg·L^-1 for Chla; 247.21×10⁶ L^-1 and 275.64×10⁶ L^-1 for cyanobacterial density) are observed for chlorophyll a and cyanobacterial density. An important meaning lies in the linear regression for all correlations which indicates the sensitivity for high values of water samples. Generally, RI calculated from MODIS bands is more suitable for water quality assessment. A possible explanation is that the much fine spectral resolution of MODIS bands is more sensitive to chlorophyll signals. This result will be helpful for further evaluation of optical characteristics and water quality using FY 3A/MESRI observations.