De-aliasing algorithm Case study (illustration de-aliasing)

De-aliasing algorithm

Case study (illustration de-aliasing)

DUAL DOPPLER RETRIEVAL Po Valley-radars and terrain analysis

Data Gridding

• Radar data

• Polar data
• 4 elevation angles(deg) 0.5, 1.4, 2.3, 3.2
• ray resolutions 0.25km x 440 bins, beam width: 0.9 deg

Calculating wind field

• Three Fundamental Equations:

• Radial Velocities (from each radar)
• Mass Continuity

Numerical procedures

• Iterative method:

• horizontal components
• vertical component

• Boundary conditions:

• zero vertical velocity on ground
• zero horizontal velocity gradient on ground (optional: simplify computation w/o loss of accuracy)

Velocity-Azimuth Display (VAD)

The Doppler radar radial wind observation operator

• Interpolation of the model wind vector to the observation location.

• Projection of the wind vector on the slanted direction of the radar beam.

• Broadening of the radar beam: Gaussian averaging kernel.

• Bending of the radar beam: Snell's law.

One-month (January 2002) comparison of Swedish radial winds and HIRLAM model equivalents

• Figure: rms difference as a function of measurement range.

• Rms difference for thinned raw data is significantly higher than for SOs.

• Method used to deter-mine optimal averaging length scales 10 km for 22 km model)

WP 3: Data assimilation

WP 3: Data assimilation

WP 3: Data assimilation

WP 3: Data assimilation

WP 3: Data assimilation

WP 3: Data assimilation

On-line estimation of error covariances

• Covariances of innovation vectors (v=y-Hxb):