Analysis of Spatiotemporal Variability in Multi-annual Ocean Colour Data of the Southern North Sea
Blaas, Meinte1; De Boer, Gerben J.2
1Deltares, NETHERLANDS; 2Deltares, Delft University of Technology, NETHERLANDS
Ocean colour data possesses the well-known benefit of synoptic spatial coverage under favourable atmospheric conditions. The regular overpass of space-borne ocean colour sensors and the currently available spatial resolution provide additional value for this type of earth observation for studying optical water quality (transparency), transport of suspended solids, origin and fate of algal blooms and dispersion of water masses in large lakes and marine waters.
The current study presents an analysis of multiple years of data of SPM (suspended particulate matter) for the southern North Sea as retrieved from MERIS reflectance measurements with the HYDROPT algorithm (Van der Woerd & Pasterkamp 2008, Eleveld et al. 2008). The data are evaluated in terms of dominant spatio-temporal modes of variation determined by means of the Data-INterpolating Empirical Orthogonal Functions analysis (DINEOF, e.g., Alvera-Azcárate et al. 2007). Compared to regular EOF or principal component analysis, DINEOF has the advantage that it can deal with data sets that contain irregular gaps e.g. due to cloud cover, whitecapping, and sun glint.
From field observations and process-based model studies it is generally concluded that observed patterns of variation of SPM in the southern North Sea are governed by multiple (partly interacting) processes such as resuspension by waves and tidal currents, surface mixing and stratification in the water column, vertical settling and horizontal advection (e.g. Jago et al. 1993, Gayer et al. 2006, Fettweis et al. 2007, Pietrzak et al. 2011).
In the current study we apply the DINEOF technique to an extended state vector that not only includes retrieved SPM but also covarying data such as wave-height, local water depth and other hydrodynamic factors. In this way the contributions of the various processes to the spatiotemporal variability are assessed and regions and periods of interaction of these processes can be determined.
The deconstruction of SPM variability by these empirical methods is valuable for numerical transport model development, but also has potential for the design of validation protocols for ocean colour retrieval.