Shipborne oceanic high-spectral-resolution lidar for accurate estimation of seawater depth-resolved optical properties

Abstract

Lidar techniques present a distinctive ability to resolve vertical structure of optical properties within the upper water column at both day- and night-time. However, accuracy challenges remain for existing lidar instruments due to the ill-posed nature of elastic backscatter lidar retrievals and multiple scattering. Here we demonstrate the high performance of, to the best of our knowledge, the first shipborne oceanic high-spectral-resolution lidar (HSRL) and illustrate a multiple scattering correction algorithm to rigorously address the above challenges in estimating the depth-resolved diffuse attenuation coefficient K_d and the particulate backscattering coefficient b_bp at 532 nm. HSRL data were collected during day- and night-time within the coastal areas of East China Sea and South China Sea, which are connected by the Taiwan Strait. Results include vertical profiles from open ocean waters to moderate turbid waters and first lidar continuous observation of diel vertical distribution of thin layers at a fixed station. The root-mean-square relative differences between the HSRL and coincident in situ measurements are 5.6% and 9.1% for K_d and b_bp, respectively, corresponding to an improvement of 2.7-13.5 and 4.9-44.1 times, respectively, with respect to elastic backscatter lidar methods. Shipborne oceanic HSRLs with high performance are expected to be of paramount importance for the construction of 3D map of ocean ecosystem.

Fig. 1. Principle of the shipborne oceanic HSRL.

a Concept of the shipborne HSRL profiling seawater optical properties. b HSRL can work at both day- and night-time. c Schematic diagram of the HSRL system with the iodine absorption cell discriminator

Fig. 2. Flow chart of HSRL data processing.

Steps from raw data to retrieval products

Fig. 3. Continuous underway HSRL measurements (~800 km).

Plots of a water depth from ETOPO1, b Chl, c SSC, d a_CDOM from HY-1C (Methods) with the ship track represented with the red lines. Discrete in situ measurements were obtained at stations S₁-S₅ (Methods). e Values along the ship track in Fig. 3a–d. Profiles of f b_bp, g K_d, h R retrieved by HSRL with 3 optical depths in red lines

Fig. 4. A diel continuous measurement for a fixed station.

Plots of a water depth from ETOPO1, b Chl, c SSC, d a_CDOM from HY-1C (Methods) with the ship track in red lines. Profiles of e b_bp, f K_d, g R retrieved by HSRL with 3 optical depths in red lines. Discrete in situ data were collected at different time T₁-T₅ (Methods)

Fig. 5. The consistency check of HSRL retrievals.

a Comparisons of K_d. b Comparisons of b_bp. The orange represents outputs of HSRL retrieval algorithm and blue refers to K_d of HSRL retrieval and MSC algorithm in Fig. 2. Fernald method based on lidar ratios of 100 sr (gray) and 200 sr (black) and perturbation method (purple) are used for the elastic backscatter lidar. All lidar results are validated by in situ measurements (green) in c statistical analysis of K_d and b_bp

Fig. 6. The ultra-narrow discrimination of the iodine cell.

a The picture of the packed iodine absorption cell integrated with the temperature controller. The light enters the device from the left and leaves from the right. b Illustration of the spectral discrimination, where the iodine cell (black line) filters the signal (blue line) by rejecting the particulate and molecular Rayleigh signals and transmitting molecular Brillouin signal (blue shadow)