Extending satellite ocean color remote sensing to the near-blue ultraviolet bands

Ultraviolet (UV) radiation has a profound impact on marine life, but historically and even currently, most ocean color satellites cannot provide radiance measurements in the UV, thus UV penetration, global ocean. We develop system (termed as UVISRdl) this study, based deep learning, to estimate remote sensing reflectance (Rrs) at 360, 380, 400 nm (collectively termed near-blue bands, nbUV) from Rrs visible bands that are obtained by satellites. This is tested using both synthetic field-measured data cover wide range large number of values, with resulted coefficient determination close 1.0 bias 0 between UVISRdl estimated known Rrs(nbUV). These results indicate excellent predictability Rrs(nbUV) Rrs(visible) via UVISRdl. The was further applied VIIRS (the Visible Infrared Imaging Radiometer Suite) evaluated matchup field measurements, mean absolute relative difference (MARD) 360 ~14% for oceanic waters ~ 50% coastal waters. equivalent those reported literature satellite Examples distribution Rrs(nbUV), subsequently diffuse attenuation nbUV (Kd(nbUV)), generated after applying data. lays groundwork generate decade-long Kd(nbUV) data, which will be useful important biogeochemical studies.