WMOD00564

• Title: Snow Accumulation Algorithm for the WSR-80D Radar: Supplemental Report
• Date: 11/1/1999
• Weather Modification - Doc Type: Report

lower than 1250 m. The two gages exposed to snow are of the heated tipping bucket type (known to be
inaccurate) and have no Alter shields to reduce wind effects. The following adaptable parameters are
being used: minimum reflectivity = 4 dBZ, maximum reflectivity = 40 dBZ: alpha = 100, beta = 2.0,
range correction factor for R > 35 km: F = 1.0 - 0.00500 * R + 0.0001428 * R2.

Analysis and verification was difficult because of mountainous terrain and other factors. However, this
SAA with the indicated range correction factor significantly improved the WSR-88D precipitation
estimates over the default Ze = 300 RIA relationship. Using six gages, from 33 to 179 km range, for
26 storm day totals, the default relation yielded radar estimates of precipitation with a median at 17 per-
cent of actual. For the same data set, the SAA with the range correction produced a median at 97 percent
of actual precipitation. Work on this Reclamation project continues, but such work was not part of this
study and is mentioned only for general interest.

5. RANl~E CORRECTION

Two ways were investigated to deal with rangl~ correction. The simplest way is to make alpha a function
of range, calibrated by precision gage data from adequately sheltered locations. Another way is to
examine the vertical profile of reflectivity or calculated S at a particular range and use the seasonal
gradient values at the first five tilts to create a range correction relationship for farther ranges. Other
empirical techniques may be applicable.

Figure 3 shows the combined results of calculating an alpha that is range dependent. Six radars (KCLE =
Cleveland, Ohio; KMPX = Minneapolis, Minnesota; KENX = Albany, New York; KFTG = Denver,
Colorado; KGJX = Grand Junction, Colorado; KBBX = Beale Air Force Base, California) are
represented and all reasonable gage
information is included. While beta was
fixed at 2.0 for most cases, the KBBX points
are for the alpha values with beta close to 2.0
It 0.3), as in table 1; the unusually small
alpha values did not encourage greater
precision with a fixed beta.

5.1 A Range-Dependent Alpha

The straight lines, labeled with the radar
station identifier, are generally close fits to
the array of points (colored triangles) for that
particular radar and its snow gage data.
However, the lines certainly do not overlap;
they are site specific.

The KCLE data are dominated by shallow
lake effect storms, which the radar beam
overshoots at far ranges.

The KGJX and KBBX radars, at elevations
differing by about 10,000 feet, look at
strongly orographic snowfalls which generate

I I 11 I I. I I I I r I I I I I I

300

...

<S>
N
II

'" 200
....:>
<D
CD

L
o
<+-

co 100
-C
n.

-
<(

~

..

o

o

40 80
Ro n 9 e - K m

120

Figure 3.-Plots of individual relations for the variation of alpha
with range. The thick curved line is the VPR range correction
used for the 1998-1999 season for the northern plains states.

7