Retrieval of Water Optical Properties from Eight-spectral Band Worldview-2 Satellite Images

Conventional ocean color satellites sensors are not very suitable for inland and coastal waters applications due to their low spatial resolutions. High spatial resolution satellite sensors are required to monitor the water quality of inland and coastal waters. The conventional high-resolution satellites such as Landsat, SPOT and IKONOS have limited number of spectral bands. In comparison, the Worldview-2 satellite has 8 spectral bands in the visible to near-infrared (NIR) region with 2-m spatial resolution. The additional spectral bands enable better modeling of water reflectance in terms of the water optical properties. The extra bands also facilitate better atmospheric correction of the images. In this paper, we describe our attempt in deriving water intrinsic optical properties (absorption and backscattering coefficients) from WorldView-2 data. Atmospheric correction was first performed to correct for the effects of Rayleigh scattering and gaseous absorption. The last two NIR bands were used for correcting surface glints and aerosol scattering. The 723 nm band (the so-called red-edge band) was found to be noisy and unusable due to high absorption by atmospheric water vapor. We apply the Quasi-Analytical Algorithm (QAA) on the remaining five bands to retrieve the water absorption and backscattering coefficients from reflectance. The standard QAA (Z. P. Lee et al., IOCCG Report 5, Chapter 10) requires a set of wavelengths different from those available in WorldView-2 data. Thus, we derived the values of the coefficients in the QAA applicable to WorldView-2 spectral bands using the synthesized dataset from IOCCG Report 5. The reflectance data in the test dataset were aggregated to the WorldView-2 spectral bands using the sensor spectral response functions. QAA for WorldView-2 was applied to images of coastal and inland waters in Singapore to derive maps of absorption and backscattering coefficients which can be used to derive some water quality parameters such as turbidity, transparency, penetration depth for photosynthetically active radiation and colored dissolved organic matter.