Bering Sea optical and biological properties from MODIS

The Bering Sea is characterized by unique bio-optical properties, which cause unsatisfactory performance of global ocean color algorithms for retrieval of chlorophyll-a (Chl-a). This study evaluates the normalized water-leaving radiance nLw(λ) and Chl-a in the eastern Bering Sea that are derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) on the satellite Aqua by comparing them to in situ data. MODIS-Aqua ocean color products were derived using the NOAA Multi-Sensor Level-1 to Level-2 (MSL12) ocean color data processing system. The MODIS-derived nLw(λ) showed good agreement with in situ-measured nLw(λ). The mean ratios between them for wavelengths 412, 443, 488, and 551 nm ranged from 1.097 to 1.280, with reasonably accurate blue-green radiance ratios in nLw(λ) that were used as input for deriving Chl-a. However, compared to in situ data, existing global and regional Chl-a algorithms either overestimate or underestimate Chl-a in the eastern Bering Sea. Therefore, we propose a new algorithm for estimating Chl-a using a blended approach that was tested and applied to MODIS-Aqua images. The histogram distributions of MODIS-Aqua-derived and in situ-measured Chl-a data show that Chl-a data derived using the new algorithm agree reasonably well to in situ measurements. Annual, seasonal, and monthly composite nLw(λ) and Chl-a images are produced for the period of 2003 to 2013 in order to interpret the long-term spatial and temporal patterns of nLw(λ) and Chl-a. The nLw(λ) spectra show strong spectral dependence on seasonal variability with distinct spatial patterns. Although strong seasonal and interannual variability has been observed in Chl-a, there is no apparent trend of either increase or decrease in phytoplankton biomass associated with variability in the physical environment for the 11 years of the study period.