Sensor Noise Effects of the SWIR Bands on MODIS-Derived Ocean Color Products

The results of this study demonstrate the effects of sensor noise in the shortwave infrared (SWIR) bands on the satellite-derived ocean color products, in particular, normalized water-leaving radiance spectra data. A simple radiance smoothing technique is used for reducing sensor noise equivalent radiance NEΔL(λ) (or increasing the sensor signal-to-noise ratio) values for the SWIR radiances measured by the Moderate Resolution Imaging Spectroradiometer (MODIS) on the Aqua satellite. Specifically, a simple 5 × 5 box running mean approach for MODIS-Aqua-measured SWIR radiances has been proposed. We show that the noise errors in MODIS-Aqua-derived ocean color products using the SWIR-based atmospheric correction algorithm are mainly from large noise errors in the MODIS-Aqua-measured SWIR radiances due to their significantly larger sensor NEΔL(λ) values. Using the proposed SWIR radiance smoothing method, SWIR-derived ocean color products can be improved with considerably reduced product noise. In addition, a practical method in estimation of the MODIS-Aqua SWIR NEΔL(λ) values has been developed. We show that, for future ocean color satellite sensors, it will require NEΔL(λ) values of less than ~5 × 10^-4 and ~9 × 10^-5 mW · cm^-2 · μm^-1 · sr^-1 for SWIR bands at 1240 and 2130 nm, respectively.