Remote sensing of chlorophyll-a as a measure of red tide in Tokyo Bay using hotspot analysis

Abstract

Tokyo Bay plays significant roles in Japan׳s economic and social development; however, the bay is facing the threat on water quality degradation due to harmful algal productions. Water quality parameters are measured at few stations in Tokyo Bay by respective agencies but is limited in time and space. This paper presents empirical models for continual Chl-a retrieval and red tide detection in Tokyo Bay water using satellite reflectance data derived from Landsat OLI sensor. The models use regression results from 38 samples, which obtained for a period between January and December of 2014. Based on the model fit, band ratios of blue and green were used to retrieve the Chl-a (R2=0.63, p-value<0.05). Seasonal pattern of Chl-a were then studied using the images obtained for the study period. Results shows that Chl-a concentrations during summer months are associated with high phytoplankton activity, and that for winter months are accompanying with low phytoplankton activity in Tokyo Bay. A simple hotspot model based on Getis-Ord Gi* were then proposed to detect the red tide events. Based on the hotspot analysis, z score for the dates of Landsat images were calculated and mapped. The high z scores obtained from Chl-a maps often corresponds with the measured red tide events. Results confirm the potential of spatial autocorrelation techniques for the detection of red tide breakouts from Chl-a retrieved Landsat 8 OLI.