We explain the workings of existing methods for ambiguity removal and describe in detail how we use the known spatial structure of the synoptic scale wind field, as used in meteorological analysis, to improve ambiguity removal skill in a new scheme called 2D-VAR.
2D-VAR is a variational scheme that solves the ambiguity removal problem by minimizing a cost function. The cost function is formulated in terms of wind increments and penalizes deviations from both a background wind field and the ambiguous scatterometer wind solutions obtained from ERS scatterometer wind retrieval.
A main feature of 2D-VAR is the use of a discrete grid that extends beyond the scatterometer swath to incorporate wind increments, generated outside the swath due to scatterometer observations near or at the edge of the swath. The wind increments contribute to the cost function and help to obtain a realistic and meteorologically consistent solution. Another feature of 2D-VAR is observation grouping which appoints a mean location to a group of retrieved ambiguous scatterometer winds. This has the effect of observation smoothing that appears to have a beneficial effect in areas with large gradients (fronts).
Validation of its performance shows that 2D-VAR has both strengths and weaknesses. An objective comparison with PRESCAT, a state-of-the-art ambiguity removal scheme, favours PRESCAT. Subjective analysis by meteorologists at KNMI favours 2D-VAR for cases where 2D-VAR and PRESCAT give significantly different solutions.
2D-VAR will be compared to similar schemes in the framework of the Ocean and Sea Ice Satellite Application Facility to assess its suitability for implementation in the SAF processing chain. In the near future, at KNMI, 2D-VAR will be used for the ambiguity removal of QuikSCAT winds.
J de Vries, A Stoffelen. 2D Variational Ambiguity Removal