Abstract The concentration of chlorophyll‐ has been estimated from sea color, primarily based on the reflectance ratio between the blue and green spectral bands in multiband images obtained by satellite optical sensors. However, this method is only effective in open, clear water and cannot be reliably applied to shallow coastal waters. In this study, we propose the potential use of double‐peaked features at wavelengths of nm and nm in hyperspectral reflectance data as indicators of the chlorophyll‐ concentration in shallow coastal waters. We conducted data mining to identify spectra exhibiting the double‐peaked features in reflectance data obtained by the Hyperspectral Imager Suite (HISUI) over coastal regions of Okinawa Island, Japan. Our analysis revealed that the detection points with the double‐peaked features are clustered in specific shallow coastal areas, such as reefs and coves, where biological activity is expected to be high, especially around Maeda and Yakata. For these detection points, the spectral relationship between peak strengths at nm and nm can be categorized into distinct trends, which appear to depend on local variations in the chlorophyll‐ concentration. In addition, in situ measurements using FieldSpec confirmed the presence of double‐peaked features at Yakata and Maeda, whereas such features were absent at non‐detection points. These findings suggest that the double‐peaked features in hyperspectral data can be used to reliably detect elevated chlorophyll‐ concentrations in shallow coastal waters. This method is expected to be useful for monitoring coastal ecosystem activity over time in shallow coastal waters.
Yamamoto et al. (Sun,) studied this question.