AVSET: A Clever Way to Achieve Faster Volume Scan Updates
Based on a recent field survey, over 62% of respondents rate faster Volume Coverage
Pattern (VCP) updates (more frequent low elevation updates) as the most important VCP
improvement the Radar Operations Center (ROC) could provide. Constructing faster
VCPs sounds easy and straight forward; however, there are only two ways to achieve
faster VCPs: either spin the antenna faster, or sample fewer elevation angles. With
a little thought it becomes apparent that both of these options are problematic.
If one opts to spin the antenna faster, data quality (e.g., clutter
filter performance) and hardware maintenance issues soon overwhelm the discussion. For
example, current VCPs like 12 and 121 are already approaching those rotational
limits. On the other hand, a decision to scan fewer elevations, predefining
multiple VCPs with different combinations of elevations, and the operational
implementation of these new VCPs becomes a daunting task. Addition-ally, this
option would significantly increase the number of VCPs required to fulfill the
myriad of meteorological situations expected across the country.
Given the apparent incompatibility between the need for faster product (VCP)
updates and the reality of VCP design, implementation and management, the ROC
decided to look inside the VCP to the individual volume scan. By treating each
volume scan independently, it was realized that we could dynamically control
the number of scanning angles based on the sampled meteorological return. The
result of this epiphany is the Automated Volume Scan Evaluation and Termination (AVSET) function.
The basic premise of the AVSET function is to terminate the current
volume scan after the radar has scanned all the elevations with important
return. In other words, once the data collection elevation overshoots the
available radar return, the volume scan is terminated, as there is no benefit
realized by continuing the execution of the current volume scan, and a new
volume scan is begun. The net effect of AVSET is to shorten the elapsed
time between data collection on low elevation angles during periods when
no significant data are avail-able on the higher elevation tilts.
When enabled by the operator, the AVSET function evaluates the return
on each elevation above 5° and calculates the areal coverage of return
18dBZ and greater and 30dBZ and greater. If the areal coverage of >18dBZ
is less that 80 km2 (total over the entire radar coverage area) AND the
areal coverage of >30dBZ is less than 30 km2 (total over the entire radar
coverage area) AND the areal coverage of 18dBZ and greater has not increased
by 12 km2 or more since the last volume scan then AVSET terminates the volume
scan after completion of the next higher elevation. This volume scan termination
scheme causes the system to enter its normal transition (RDA antenna retrace,
RPG concludes algorithm processing and product generation, etc.), to prepare
for the start of a new volume scan.
Figures 1 and 2 illustrate the benefit of AVSET processing using
Level II data from a severe thunderstorm event collected by the Norman,
OK, Weather Forecast Office (WFO) WSR-88D (KTLX) radar. Figure 1 is the
Composite Reflectivity product from KTLX at 02:30Z. Figure 2 provides
the 10°, 12°, and 14° Base Reflectivity products from 02:36. Had AVSET
been active during this event, this volume scan would have been terminated
after completion of the 12° slice. The resulting volume scan duration would
have been approximately 250 seconds which is about 45 seconds faster than
the standard VCP 11 volume scan duration.
The amount of time savings achieved by AVSET depends on the active VCP
and the areal coverage of return. Given the best possible situation, AVSET
will terminate the volume scan after completion of the second elevation
above 5°. Table 1 provides the minimum scanning angles, elevation scan times,
and shortest VCP durations for four AVSET-controlled VCPs. For reference,
Table 2 provides the average update times (without AVSET) for the VCPs listed in Table 1.
To provide comparison data and images, several Level II cases have been analyzed. A
representative case is the March 2, 2008, KTLX data set. During this event, severe
convective weather moved through Central OK. For the 4-hour period from 02/2000Z
through 03/0000Z, the Norman, OK WFO (using the KTLX radar executing VCP 11) issued
three separate tornado warnings and multiple severe thunderstorm warnings. One
confirmed tornado occurred at 02/2246Z, approximately 75nm northwest of KTLX.
NOTE: AVSET will start evaluation on the first elevation above 5°. With
the current design, AVSET will always process one elevation cut above the
elevation where the AVSET reflectivity thresholds are met.
The Level II data from this event was reprocessed on a ROC test bed RPG
that executes special code that emulates the AVSET function. For the 4-hour
time period, the AVSET-controlled volume scan times averaged 212 seconds. Focusing
on the 3-hour period prior to, and including, the tornado, the AVSET-controlled VCP
11 scan times averaged 198 seconds (see Figure 3). Had it been available on KTLX, AVSET
would have enabled KTLX to produce 66 volume scans during the 4-hour period; conversely,
only 49 volume scans executing VCP 11 were possible.
As denoted in Figure 3, AVSET terminated the 22:40 volume scan after the
completion of 7.5° (206 seconds). The left image in Figure 4 is the KTLX base
reflectivity product from the "last elevation cut" of this volume scan. The right
image is the KTLX base reflectivity product from 8.7°, the next higher elevation
cut, which would not be sampled with AVSET active.
As one can see, only a small amount of weak return was present above the elevations
sampled by the AVSET-controlled VCP. In many of the test cases, the only return above
the AVSET-sample elevations was attributable to side lobes.
Visual inspection of the products from KTLX (standard VCP 11) and KTLX (AVSET-controlled VCP 11)
showed only operationally insignificant differences (Figures 5 and 6). The observed differences
did not impact the identification and interpretation of the unfolding weather events.
Note: For AVSET, the Velocity Azimuth Display (VAD) Wind Profile (VWP) task
was modified to dynamically determine the elevations and slant ranges needed to achieve
the required heights based on the previous volume scans avail-able elevations. It is
interesting to note that in some instances, the AVSET-induced VWP elevation/slant range
combinations result in additional wind data.
AVSET represents a paradigm shift in operational volume scanning for the WSR-88D. In
the past, WSR-88D VCPs have always automatically and continuously scanned predefined
elevation angles. This scheme resulted in each VCP having a particular periodic update
cycle that never changed, regardless of the sampled meteorological conditions. The only
way to change the update period was to invoke another VCP and accept its elevation scans
and periodic update rate.
When enabled, the AVSET function evaluates the return on each elevation above 5° and
terminates the current volume scan after the radar has scanned all elevations with
important return. The net result of AVSET is a shortened elapse time between data
collection on low elevation angles (and generating volume-based products) during periods
when no significant data are available on the higher elevation tilts.
AVSET has not been approved for operational fielding, yet. For now, AVSET
is currently in OSIP (Stage 2) and is still undergoing testing at the ROC.
Joe Chrisman
ROC Engineering Branch