Table 1
List of cost functions currently implemented in PyDDA.
| Cost Function | Symbol | Routine |
|---|---|---|
| Total | J(v) | J_function |
| Radar observations | Jo | calculate_radial_vel_cost_function |
| Mass continuity | Jmass | calculate_mass_continuity |
| Vertical vorticity | Jv | calculate_vertical_vorticity_cost |
| Rawinsonde | Jr | calculate_background_cost |
| Smoothness | Js | calculate_smoothness_cost |
| Model | Jmodel | calculate_model_cost |
| Point | Jpoint | calculate_point_cost |
Table 2
List of cost function gradients currently implemented in PyDDA.
| Cost Function | Symbol | Gradient |
|---|---|---|
| Total | ∇J(v) | grad_J |
| Radar observations | ∇Jo | calculate_grad_radial_vel |
| Mass continuity | ∇Jmass | calculate_mass_continuity_gradient |
| Vertical vorticity | ∇Jv | calculate_vertical_vorticity_gradient |
| Rawinsonde | ∇Jr | calculate_background_gradient |
| Smoothness | ∇Js | calculate_smoothness_gradient |
| Model | ∇Jmodel | calculate_model_gradient |
| Point | ∇Jpoint | calculate_point_cost |
Table 3
List of current initialization options in PyDDA.
| Initialization | Routine name |
|---|---|
| Constant wind field | make_constant_wind_field |
| Py-ART HorizontalWindProfile object | make_wind_field_from_profile |
| Weather Research and Forecasting (WRF) | make_background_from_wrf |
| High Resolution Rapid Refresh (HRRR) | make_initialization_from_hrrr |
Figure 1
An example wind barb plot overlaid on radar reflectivity at 3.5 km altitude for winds retrieved in thunderstorms sampled by 2 radars over Darwin. Contours represent the presence of updrafts with given velocities at the 3.5 km height level. The area inside the two circles indicate where the wind retrieval is most reliable.
Figure 2
(a) An example streamline plot overlaid on radar reflectivity at 3 km altitude observed from 2 radars over Darwin, Australia. Contours represent horizontal wind speed. (b) as (a), but zoomed into the region enclosed by the grey box. Contours represent vertical wind speed.
Figure 3
Winds at 0.5 km retrieved by PyDDA using only data from the 2 NEXRAD radars overlaid on a reflectivity mosaic generated from the two NEXRAD radars. Barbs are in m s1.
Figure 4
As Figure 3, but only using the HRRR model as a constraint.
Figure 5
As Figure 3, but using both the radar and HRRR winds as constraints.