Results on this page show the effect of different gridding techniques on the resulting dual Doppler kinematic and reflectivity structure for the 990223 event (21 UTC). In order to capture the convection sampled by the aircraft at the time of the TRMM satellite overpass, radar data were placed onto a cartesian grid extending out to 150 km from the S-pol radar. The beam-crossing angle criteria for dual Doppler synthesis was reduced from 22 to 20 deg in order to retrieve the 3-D wind field as far as possible north of the S-pol radar site.

In order to perform the cartesian gridding, the NCAR software program REORDER was utilized using two different methods to interpolate the radar data in polar coordinates to a cartesian system (i.e., the radius of influence). In the first method, a fixed radius of influence was selected with x=y=1.5 km and z=1.0 km. In this method, the program uses a selected weighting scheme (Cressman) to map data points within the radius of influence zone, independent of range. In the second method, a variable radius of influence was selected with the azimuth and elevation criteria equal to 1.65 deg (beamwidth of the TOGA radar). In this method, the radius of influence distance varies as a function of range (like a cone) so that the search for data points extends farther out at long range compared to short range.

As shown in the CAPPI's below,
the effect of the variable radius of influence is to smooth the
reflectivity field and fill in gaps compared to the fixed radius
of influence (*the smoothing and gap filling effect becomes
more pronounced with height*).

To better compare the different radius of influence techniques on the vertical structure of convection, the plots below show a comparison of volume averaged vertical air motion and reflectivity. In these plots, the red profile represents the results using the variable radius of influence while the blue profile represents the fixed radius of influence.

- The variable radius of influence tends to stretch the cartesian grid in the vertical and smooth/fill gaps in the reflectivity field more than the fixed radius of influence. This effect becomes more pronounced as a function of height;
- The grid stretching observed in the variable radius of influence method can produce large and un-expected differences in the vertical motion profile results;
- Comparisons with EDOP data as well as with PPI's of the ground-based radar data suggest that the fixed radius of influence is a better method for mapping the data onto a cartesian grid, despite the loss of data at upper levels (click here to see upper level CAPPI's using the fixed radius of influence);
- Analysis of all dual Doppler data for the 990223 event is based on cartesian gridding using a fixed radius of influence).