Search Results

You are looking at 1 - 6 of 6 items for :

  • Author or Editor: Rodger A. Brown x
  • Journal of Atmospheric and Oceanic Technology x
  • Refine by Access: All Content x
Clear All Modify Search
Stephan P. Nelson
and
Rodger A. Brown

Abstract

Actual data are used in one case to investigate the nature and source of vertical velocity errors resulting from analyses of multiple-Doppler radar measurements. Consistent with earlier analytical works, larger errors are found than would be expected from previous theoretical studies. It is shown that the reconstructed maximum updraft speed in strong updrafts (>20 m s−1) is accurate, on the average, to within about 10% (standard deviation of 10%). Storm advection, incomplete sampling of low-altitude divergence caused by the radar horizon, top boundary errors, and uneven terrain are studied and all are dismissed as dominant sources of error in the case considered here. The inability to determine a dominant error source has important consequences for the formulation of vertical velocity adjustment schemes.

Full access
Vincent T. Wood
and
Rodger A. Brown

Abstract

When a thunderstorm mesocyclone changes range relative to a Doppler radar, the deduced core diameter and mean rotational velocity of the Doppler velocity mesocyclone signature oscillate back and forth, even though the radar beam’s physical width changes uniformly with range. The authors investigated the oscillations using a model mesocyclone and a simulated Doppler radar that collected data with an azimuthal sampling interval of 1°. They found that the oscillations are a consequence of changing data point separation with range relative to the Doppler velocity peaks of the mesocyclone signature.

Full access
Vincent T. Wood
and
Rodger A. Brown

Abstract

A variety of single Doppler velocity patterns that simulate those observed in a nondivergent environment is presented. Measurements in optically clear air and/or widespread precipitation are simulated, using horizontally uniform wind fields that vary with height. Vertical profiles of wind speed and direction indicated by the simulated Doppler velocity fields agree well with Doppler radar measurements. Horizontally uniform winds veering with height produce a striking S-shaped pattern, indicative of warm air advection; winds backing with height produce a backward S, indicative of cold air advection. A maximum in the vertical profile of wind speed is indicated by a pair of concentric contours, one upwind and one downwind of the radar. The presence of a frontal discontinuity is indicated by rapid variation of wind direction within the frontal zone. The wind speed profile controls the overall pattern including the spacing between contours, whereas the vertical profile of wind direction controls contour curvature.

Full access
Vincent T. Wood
,
Rodger A. Brown
, and
David C. Dowell

Abstract

Low-altitude radar reflectivity measurements of tornadoes sometimes reveal a donut-shaped signature (low-reflectivity eye surrounded by a high-reflectivity annulus) and at other times reveal a high-reflectivity knob associated with the tornado. The differences appear to be due to such factors as (i) the radar’s sampling resolution, (ii) the presence or absence of lofted debris and a low-reflectivity eye, (iii) whether measurements were made within the lowest few hundred meters where centrifuged hydrometeors and smaller debris particles were recycled back into the tornadic circulation, and (iv) the presence or absence of multiple vortices in the parent tornado.

To explore the influences of some of these various factors on radar reflectivity and Doppler velocity signatures, a high-resolution tornado numerical model was used that incorporated the centrifuging of hydrometeors. A model reflectivity field was computed from the resulting concentration of hydrometeors. Then, the model reflectivity and velocity fields were scanned by a simulated Weather Surveillance Radar-1988 Doppler (WSR-88D) using both the legacy resolution and the new super-resolution sampling. Super-resolution reflectivity and Doppler velocity data are displayed at 0.5° instead of 1.0° azimuthal sampling intervals and reflectivity data are displayed at 0.25-km instead of 1.0-km range intervals.

Since a mean Doppler velocity value is the reflectivity-weighted mean of the radial motion of all the radar scatterers within a radar beam, a nonuniform distribution of scatterers produces a different mean Doppler velocity value than does a uniform distribution of scatterers. Nonuniform reflectivities within the effective resolution volume of the radar beam can bias the indicated size and strength of the tornado’s core region within the radius of the peak tangential velocities. As shown in the simulation results, the Doppler-indicated radius of the peak wind underestimates the true radius and true peak tangential velocity when the effective beamwidth is less than the tornado’s core diameter and there is a weak-reflectivity eye at the center of the tornado. As the beam becomes significantly wider than the tornado’s core diameter with increasing range, the peaks of the Doppler velocity profiles continue to decrease in magnitude but overestimate the tornado’s true radius. With increasing range from the radar, the prominence of the weak-reflectivity eye at the center of the tornado is progressively lessened until it finally disappears. As to be expected, the Doppler velocity signatures and reflectivity eye signatures were more prominent and stronger with super-resolution sampling than those with legacy-resolution sampling.

Full access
Rodger A. Brown
,
Vincent T. Wood
, and
Dale Sirmans

Abstract

The magnitude of the Doppler velocity signature of a tornado depends on the effective width of the radar beam relative to the size of the tornado. The effective beamwidth is controlled by the antenna pattern beamwidth and the azimuthal sampling interval. Simulations of Weather Surveillance Radar-1988 Doppler (WSR-88D) velocity signatures of tornadoes, presented in this paper, show that signature resolution is greatly improved when the effective beamwidth of the radar is reduced. Improved signature resolution means that stronger signatures can be resolved at greater ranges from the radar.

Using a special recording device on the National Weather Service's Radar Operations Center's KCRI test bed radar, Archive Level I time series data were collected during the Oklahoma–Kansas tornado outbreak of 3 May 1999. Two Archive Level II meteorological datasets, each having a different effective beamwidth, were created from the Archive Level I dataset. Since the rotation rate and time interval between pulses are common for both Archive Level II datasets, the only parameter that could be changed to reduce the effective beamwidth of the KCRI data was the number of pulses, which also changed the azimuthal sampling interval. By cutting the conventional number of pulses in half for one of the Archive Level II datasets, the effective beamwidth was decreased by about a quarter and the azimuthal sampling interval was decreased from 1.0° to 0.5°. The 3 May 1999 data confirm the simulation results that stronger Doppler velocity signatures of tornadoes typically are produced when the azimuthal sampling interval, and thus the effective beamwidth, is decreased.

Full access
Jessica L. Proud
,
Kelvin K. Droegemeier
,
Vincent T. Wood
, and
Rodger A. Brown

Abstract

Increasing tornado and severe storm warning lead time (lead time is defined here as the elapsed time between the issuance of a watch or warning and the time at which the anticipated weather event first impacts the specified region) through the use of radar observations has long been a challenge for researchers and operational forecasters. To improve lead time and the probability of detecting tornadoes while decreasing the false alarm ratio, a greater understanding, obtained in part by more complete observations, is needed about the region of storms within which tornadoes form and persist. Driven in large part by this need, but also by the goal of using numerical models to explicitly predict intense local weather such as thunderstorms, the National Science Foundation established, in fall 2003, the Engineering Research Center for Collaborative Adaptive Sensing of the Atmosphere (CASA). CASA is developing a revolutionary new paradigm of using a network of small, closely spaced, inexpensive, low-power dual-polarization Doppler weather radars to overcome the inability of widely spaced, high-power radars to sample large regions of the lower atmosphere owing to the curvature of earth given that zero or negative beam elevation angles are not allowed. Also, current radar technology operates mostly independently of the weather and end-user needs, thus producing valuable information on storms as a whole but not focused on any specific phenomenon or need. Conversely, CASA utilizes a dynamically adaptive sensing paradigm to identify, and optimally sample, multiple targets based upon their observed characteristics in order to meet a variety of often competing end-user needs.

The goal of this study is to evaluate a variety of adaptive sampling strategies for CASA radars to assess their effectiveness in identifying intense low-altitude vortices. Such identification, for the purposes of this study, is defined as achieving a best fit of simulated observations to an analytic model of a tornado or mesocyclone. Several parameters are varied in this study including the size of the vortex, azimuthal sampling interval, distance of the vortex from the radar, and radar beamwidth.

Results show that, in the case of small vortices, adaptively decreasing the azimuthal sampling interval (i.e., overlapping beams) is beneficial in comparison to conventional azimuthal sampling that is approximately equal to the beamwidth. However, the benefit is limited to factors of 2 in overlapping. When simulating the performance of a CASA radar in comparison to that of a Weather Surveillance Radar-1988 Doppler (WSR-88D) at close range, with both operating in the conventional nonoverlapping mode, the WSR-88D (with a beamwidth about half that of a CASA radar) performs better. However, when overlapping is applied to the CASA radar, for which little additional processing time is required, the results are comparable. In effect, the sampling resolution of a radar can be increased simply by decreasing the azimuthal sampling interval as opposed to installing a larger antenna.

Full access