WMOD00459

• Title: Snow Accumulation Algorithm for the WSR-88D Radar, Version 1
• Date: 6/1/1996
• Weather Modification - Doc Type: Report

Besides the two measurement sites at about 25 kIn from the radar, four other sites were
chosen at greater distances as shown in table 3. As with the two nearest sites, the observing
location at the 49-km range was located in an established neighborhood with solid fenced
backyards in a region that should not have had ground clutter. In addition to operation of
the shielded Belfort gages at these three sites, hourly manual measurements were made of
Sand SD' on 1- by I-ft snowboards laid on the ground or snow cover near the gages. The
same Clear Vu Model 1100 gages used in Albany were used to core snow on the boards after
four depth measurements were made and averaged. Manual measurements were usually
made from early morning until normal bedtime whenever snow was falling. Comparisons of
all hourly snowboard and gage observations from gages No.1 through 3 indicated that the
3 operators were conscientious. Resulting correlation coefficients between S pairs exceeded
0.96 for each data set.

Denver WFO personnel made special hourly observations of Sand SD whenever snowfall
occurred during the 1995-96 winter/spring. These measurements and the snowboard
observations from the 3 gage sites nearest the Denver WSR-88D have yet to be analyzed.

Three additional sites were chosen for operation of Belfort gages but not hourly snowboard
observations. However, snowboards were sampled in the same manner just discussed each
morning after snowfall had occurred. This procedure provided comparisons with gage
measurements and identified days with snowfall. The same practice was used at the five
Cleveland gages to be discussed.

All Belfort gages used in this study had accurate clocks that rotated once per 24 h, which
meant the pen trace overwrote the same horizontal line until precipitation occurred. The
need to identify days with precipitation is obvious. This identification was not difficult
because gage operators filled out a worksheet each morning after snow or rain, noting current
weather conditions including wind speed at gage orifice level. A small Taylor wind speed
meter with a minimum indication of 2 mi h-1 was used to measure wind speed.

Little data resulted from the Black Forest gage listed in table 3 until after late January 1996
because of difficulties finding a reliable operator. The data quality from the other five gages
in the Denver area was generally very good. The three most distant gages were located in
well-protected clearings in conifer forest. The two gages well west of the radar were located
in a large mountain valley, protected from ground clutter by upwind terrain, and in a
mountainous area where the lowest tilt beam intersects the terrain. The latter site, at 3265
m altitude, was located at the Mt. Evans Research Station climatological station.
Measurements from this location will be examined to determine if any useful snowfall
accumulation estimates can be made by radar in a very cluttered region. The 1.50 tilt beam
may be usable for this purpose, but this estimate has yet to be attempted.

Five gage sites were operational by November 1, 1995. The sixth gage was also established
by then, but reliable observations were not obtained until after late January when a new.
operator was trained. Gages (all with Alter-type wind screens) were installed and calibrated,
and antifreeze and other supplies were located at each site. Each observer received training
concerning gage servicing. The two observers nearest the radar received additional training
in making hourly measurements of snow depth and in observing sizes and types of the larger
snowflakes that provide almost all of the meteorological radar returns during snowfall. The
third site with special hourly observations (Lakewood) was maintained by a Reclamation
meteorologist. A snow particle identification guide was prepared for use by hourly observers.

10