A study on red tide surveillance system in the waters around the Korean coastal waters using GOCI

Abstract

Harmful algal blooms (HABs) in Korea have been reported every year in the last decade, and satellite remote sensing data has continuously been used to study the detection and monitoring. Existing algorithms for red tide detection have been developed for specific red tide species and occurrence waters. However, the probability of false positives is high with existing algorithms for red tide detection because red tide species and seawater environment are different in each sea areas. In particular, since the waters in Korea have optically complex seawater environment, it is necessary to find suitable algorithm for each sea area and specific red tide species. Here, using measurements from Geostationary Ocean Color Imager (GOCI), we have conducted study to find the optimal algorithm for each sea area and each red tide type. The spectral characteristics of C. polykrikoides and N. scintillans, which occur frequently in Korean waters, were investigated in the West Sea, the South Sea and the East Sea. In case of C. polykrikoides, dominant peaks at all waters occurred in 555nm and 680nm (fluorescence peak). On the other hand, the nLw value in blue region of the East Sea was higher than that of the West Sea and the South Sea due to the absorption of dissolved organics and suspended matter. In case of N. scintillans, dominant peaks at the West Sea occurred in 490nm and 680nm. In consideration of these features, we applied four algorithms for detecting the regions of two species. In the South Sea, the areas of C. polykrikoides was most appropriately detected when the spectral shape and modified RI method was applied. Alternatively, in the East Sea, FRTD method was a suitable method for extracting the region of C. polykrikoides. FRTD method was effective for extracting region of N. scintillans in the West Sea. The results of this study will be useful data for red tide detection and monitoring in Korean waters.