Analysis of High-resolution Sea-surface Measurements from Orbiting Satellites

Surface measurements of ocean parameters such as dynamic height, temperature, salinity, color (Chlorophyll A), and wind are most commonly gridded at around a quarter horizontal degrees (~28 km) interval and daily frequency to form a synoptic map, adequately representing the ocean meso-scale state for many operational and scientific applications in the last quarter century. Modern satellite sensors can resolve surface features at a resolution as fine as 1km, for example, by the operational infra-red radiometers (retrieving the temperature and color) as well as the up-coming SWOT altimeter. However, creating a "daily synoptic" perspective at such a high resolution is challenging mostly because sub-mesoscale features tend to evolve significantly faster with respect to the analysis period (a day) and satellite repeat cycles. We have employed the "multi-resolution analysis" method, a reversible wavelet decomposition technique, to enhance the traditional 2-D variational interpolation method with scale-dependency. The resulting Multi-Resolution Variational Analysis (MRVA) method can decompose the interpolation task into scale-dependent stages, each with its own interpolation length-scale and time-window, accommodating a 0.25 degree length-scale and multi-day (repeat-cycle) time-scale for interpolating the meso-scale components while the sub-meso-scale components can be interpolated independently at a 1-km and sub-day scales. These scale-dependent MRVA length and duration parameters have been determined objectively based on circulation fields from a submesoscale-admitting global-ocean simulation (1/48-degree nominal horizontal grid spacing).