Results of a pilot study on seawater color prediction for building an ocean digital twin: representing hue changes caused by wavelength-dependent attenuation

Abstract

This study conducted preliminary data collection, observations, data analysis, and model development to reproduce seawater color from water-quality properties in order to build an ocean digital twin that is as faithful as possible to scientific evidence. We (a) demonstrated that seawater color can be predicted to a certain extent from water-quality variables, and (b) preliminarily confirmed that, even within the same sea area, the attenuation characteristics of light can differ by wavelength by fusing a differentiable physics model that represents the physical processes determining wavelength-specific attenuation (absorption and scattering) with Secchi disk observation results. Photosynthetic organisms such as algae, seagrasses, and seaweeds utilize different frequency bands of light during growth; however, assessment and numerical prediction of their growth environments have largely been limited to evaluating the degree of photon attenuation and water transparency, and comparatively little attention has been paid to the degree of wavelength-dependent attenuation. While aiming to construct a realistic ocean digital twin, this work also seeks to propose methods for more accurate observation and evaluation of biological growth environments.