Remote-sensing-based Methodology For Red Tide Detection And Its Applications

Qinhuangdao, which is heartland of Economic Zone Round BoHai,It is one of important developing area for ocean economy. However, thered tide existing for years is quite bad for development of ocean economy.Therefore, making effective and real-time monitoring, forecasting oceandisasters has significant meanings for the ocean economy here.The applicability of previous red-tide detection methods is assessed,building on the field observations of the red-tide extent of the red tideoccurred in coastal Qinghuangdao in June,2010. The result reveals thatall of the previous methods are involved in some kind of shortcomings.Here we analyze the band settings of MERIS(Medium ResolutionSpectral Imager) and MODIS(moderate-resolution imagingspectra-radiometer) sensors, and present a new red-tide detection methodbased on MERIS and MODIS data. The thresholds for distinguishingred-tide zone and non-red-tide zone are determined respectively throughcalculating the scales of the remote-sensing reflectance ratio between blueand green bands. This method is applied to analyze the red tide occurredin coastal Qinghuangdao in June,2010, and the result reveals thethreshold for distinguishing red-tide and non-red-tide is2.According to the conclusions above, further tests were carried out.First, the measured spectral data from Marine Optical Buoys before and after the Dangan Island red tide of Pearl River estuary was tested, resultsconfirmed that the threshold of2is reasonable to differentiate red tidewater from red tide water. Second, because turbid water (such as regionof high concentration of suspended sediment) is common to interferewhen monitoring red tide, it is necessary to verify whether the thresholdis able to distinguish between turbid water and red tide water, resultsshow that the method is feasible. In addition, band ratio of blue and greenwas calculated for MERIS data containing sea ice and the ratio of sea icearea is less than2. It indicated that red tide and sea-ice containing waterbody can also be discriminated with this threshold. Finally, the methodmentioned in this paper was compared with the other remote sensingmethods of red tide distinguishing. Results showed that, the proposednovel method was only the most accurate, but also generated the smallesterror, when discriminating the area of red tide.This method has successively detected the massive red tide occurredin coastal Qinghuangdao in June，2009. Given the complicated causationsof red tide, such environmental factors as the SST (sea surfacetemperature), precipitation, wild velocity etc. for the two red-tide eventsin2009and2010are further analyzed.