Algorithm to estimate daily PAR at the ocean surface from GOCI data: description and evaluation

Abstract

Photosynthetically available radiation (PAR) reaching the ocean surface controls phytoplankton growth, primary productivity, and evolution within marine ecosystems. Therefore, accurate daily PAR estimates are important for a broad range of marine biology and biogeochemistry applications. In this study, hourly data from Geostationary Ocean Color Imager (GOCI), the world’s first geostationary ocean color sensor, was employed to estimate daily mean PAR at the ocean surface around the Korean Peninsula using a budget model based on plane-parallel theory. In situ PAR data collected from two ocean research stations (Socheong-cho and Ieodo) were used to evaluate the accuracy of the GOCI PAR estimates. First, the instantaneous in situ measurements were checked for calibration and exposure errors against Second Simulation of the Satellite Signal in the Solar Spectrum (6S) radiative transfer calculations under the clearest sky conditions and adjusted to eliminate biases. After adjustment, the root-means-square error (RMSE) between 6S and in situ PAR data was reduced from 6.08 (4.81%) and 3.82 (3.93%) mol/m2/day to 2.85 (2.26%) and 1.74 (1.21%) mol/m2/day at the Socheong-cho and Ieodo stations, respectively, and the coefficient of determination R2 was 0.99. Then, the GOCI daily mean PAR estimated by the initial algorithm were corrected using the 2015 adjusted in situ daily PAR measurements collected under clear-sky conditions. The daily mean PAR values derived from GOCI data in all conditions were improved after the correction, with RMSE reduced from 4.58 (8.30%) to 2.57 (4.65%) mol/m2/day and R2 = 0.97. The comparison statistics were similar for 2015 and 2016 combined, with RMSE of 2.52 (4.38%) and mean bias error (MBE) of –0.40 (–0.70%), indicating that the correction was also effective in cloudy conditions. On the other hand, daily PAR estimates from Moderate Resolution Imaging Spectroradiometer (MODIS) and Advanced Himawari Imager (AHI) yielded larger RMSE of 6.24 (10.40%) mol/m2/day and MBE of –2.49 (–4.15%) mol/m2/day (MODIS) and RMSE of 3.71 (6.51%) mol/m2/day and MBE of –2.65 (–4.65%) mol/m2/day (AHI) against in situ measurements. The GOCI-based daily PAR model developed in this study is reliable and suitable for investigating the marine environment around the Korean Peninsula.