Data Assimilation & Prediction: Major Projects
Techniques to Assimilate SSM/I Observations of Marine Atmospheric Storms
Sponsor: Office of Naval Research
Abstract:
The interests of commercial shipping and the operations of the U.S. Navy and Coast Guard, among others, are dictated by marine storms and their influence on the ocean surface. Forecasters depend on accurate prediction of these storms by operational numerical models, including the Navy's Coupled Ocean/Atmosphere Mesoscale Prediction system (COAMPS). However, before a forecast can be generated by a numerical weather prediction model, an analysis of the the current state of the atmosphere must be completed which compares well with observations. Our research focuses upon improving the COAMPS' analyses of atmospheric moisture fields.
In practice, a model analysis is generated by modifying, with new data, the fields of a short-term forecast (called the 'first guess'; typically a 6-h forecast) from the previous run of the model. Our goal is to improve the analysis, and ultimately the forecasts, of the COAMPS model by adjusting the form and intensity of marine meteorological features in the first-guess fields towards the representation of the same features in the imagery of the Special Sensing Microwave/Imager (SSM/I) satellite. The unconventional spatial and temporal nature of satellite data has traditionally hindered its incorporation into model data assimilation systems. Our technique makes use of satellite-derived moisture fields from over the otherwise data-sparse oceans.
The SSM/I sensor provides high-resolution measurements of integrated water vapour (IWV). We intend to compare the SSM/I IWV observations of distinct meteorological features with the corresponding representation in the first guess model-derived IWV field and "distort" the model's field (leaving the SSM/I data fixed) to better fit the observed. The distortion is composed of displacement and amplification fields that are required to vary slowly in space. The distortion field minimizes the misfit between the distorted model field and the SSM/I observations. The distorted IWV model field, along with the remaining model fields, will then be subjected to the standard optimal interpolation data assimilation scheme of the COAMPS model.
The principle of the FCA method is to define displacement adjustments to the satellite data which minimize the differences with respect to the radar data, and which satisfy additional smoothness and magnitude constraints. The parameters describing the calibration and alignment are found using a variational approach. The satellite picture is then modified by displacing each pixel by the appropriate amount, before being overlaid with the radar data. An illustration of the method is given in the figure.
Illustration of the FCA method, showing how similar contours from two different data sources can be adjusted to agree with each other.
The effects of the distortion procedure will be quantified using 48-h forecasts of the model. The 48-h forecast initialized with the non-distorted IWV field will be compared with the forecast initialized with the distorted IWV field. The original forecast error can then be decomposed into errors contained in the displacement and amplification adjustments, as well as in the residual (i.e., the error remaining following the distortion procedure).
For details of the distortion representation of forecast errors
see:
Hoffman, R. N., and C. Grassotti, 1996: A technique
for assimilating SSM/I observations of marine atmospheric storms.
J. Appl. Meteor., 35, 1177-1188.
Z. Liu, J.-F. Louis, and C. Grassotti, 1995: Distortion representation
of forecast errors. Mon. Wea. Rev., 123, 2258-2270.
| |
|
Information
Major Projects
Publications