Tuning algorithms by water type for optical remote sensing: a case study in a tidally stirred shelf sea

Alex Cunningham, Agnes Dudek
University of Strathclyde, Physics Department, Glasgow, G4ONG, United Kingdom
a.cunningham@strath.ac.uk

ABSTRACT

An extensive survey of in situ optical properties and seawater composition was carried out in the Irish Sea in 2001 and 2002 and used to assess the accuracy of SeaWiFS level 2 products. Results showed that the performance of these products, including normalised water leaving radiance nL_w, chlorophyll concentration Chl and the diffuse attenuation coefficient for downward irradiance at 490 nm K_d490, all degraded as suspended sediment concentration increased. A new approach is proposed for improving bio-optical product retrieval, which consists of dividing the Irish Sea into two water types based on relationships between inherent optical properties and applying algorithms with adjusted coefficients to each water type. In this region, the water types can be efficiently differentiated from surface radiometry using a threshold value of 1 mW cm^-2 μm^-1 sr^-1 for nL_w at 555 nm. Results showed that the water type approach produced greatly improved product retrieval when applied to sea surface radiometric data. In addition, mineral suspended sediment (MSS) concentrations were well correlated with the ratio of remote sensing reflectances at 665 nm to 555 nm (r²=0.74), which made it possible to establish an MSS algorithm for the Irish and Celtic Seas. Problems in the SeaDAS atmospheric correction procedure were identified, which limited the application of the dual-algorithm approach over low-reflectance waters. However, it produced realistic values of chlorophyll, mineral suspended sediment and K_d490 when compared against an independent data set in the Bristol Channel, and greatly improved the identification of known fronts and regions of high tidal stirring in the Irish Sea.

Last Update: 2005-03-16