The High Snow to Liquid Ratio Storm of December 23, 2000 in Green Bay, WI
An unusually low water content snowfall on Saturday December 23, 2000 produced 5.5 inches of snow at the National Weather Service Office in Green Bay, WI. The Green Bay ASOS recorded only 0.1 inches of liquid, suggesting a 55 to 1 snow to liquid ratio. The strength of the winds can greatly affect gauge catch and this issue is addressed in a later section. Overall, a ratio of 50 to 1 appeared to be a reasonable conclusion. The snow was so fluffy that taking a measurement was like lowering the ruler through air. Ratios of 20 or 30 to 1 are not unusual with arctic air in this part of the country, but 50 or more to 1 is excessive. The following sections present a brief overview of storm features. Next an analysis of storm characteristics favorable for high ratio snow is presented then forecast applications and recommendations are made. The 12 UTC Dec 23 Eta model run data was used for most of the analyses, along with RAOB plots valid 12 UTC Dec 23 and hand analyzed surface maps.
2. Verification of Snow to Liquid Ratio
Core samples were not taken for this storm to augment the precipitation amount, but winds were very light, ranging between 0 and 2 knots on the observations. A chart borrowed from Goodison (1978) by Nolan J. Doesken and Arthur Judson (1997) for “The Snow Booklet”, shows the affects of wind on gauge catch of snowfall (Figure 1). This chart indicates that the vast majority of the snow should have fallen into the rain gauge. Since even the slightest wind causes loss of measured precipitation, one can assume that the actual snow to liquid ratio would be a little less than 55 to 1.
3. Storm Overview
The storm began as a very light grainy snow shortly before 6 AM. Little accumulation occurred through mid morning. By 11 AM, radar echos of 28+ dBZ developed in Central Wisconsin and were moving into Green Bay and other areas of the Fox River Valley of East Central Wisconsin. The snow then fell as large fluffy flakes, many between 1 and 2 inches diameter, composed of large dendrite and a few hexagonal plate ice crystals.
The 500 mb 12 UTC absolute vorticity analysis and the 18 UTC forecast ( Figures 2a and 2b) showed a strong maximum associated with an upper level trough moving southeast toward Northeast and East Central Wisconsin. A surface low pressure center and trough moved from Western Minnesota and Western Iowa at 12 UTC to Eastern Wisconsin and Lake Michigan by 23 UTC (Figures 3a, 3b, and 3c).
The snow occurred in a very cold air mass but still with enough available moisture to support at least a light snowfall. The 12 UTC 1000 to 500 mb thickness values ranged from 516 to 119 decameters (Figure 4). The 281 oK isentropic surface showed mixing ratios of 1 to 2 g/kg available for the storm over and upstream of Green Bay (Figure 5).
4. Characteristics of this Storm Favorable for High Snow to Liquid Ratio
a. Structure of Snowflakes Composed of Dendritic Crystals
Flakes composed of dendrites contain a lot of air in their structure and thus accumulate in a loosely packed fashion. Flakes composed of needles and columns will pack more densely, especially if the flakes break up into individual crystals as they hit the ground.
b. Deep Layer of Temperatures Favorable for Dendritic Crystal Growth
Temperatures in the -16 to -12 oC range favor the dendritic crystal habit. Crystal growth rates also maximize in that temperature range. A deep layer of favorable temperatures existed in this storm. The surface analyses for 12, 18, and 23 UTC December 23 (Figures 3a, 3b, and 3c) featured surface temperatures ranging from just below zero to around 10 oF during the event. Specifically at Green Bay temperatures ranged from 3 to 12 oF (or -16 to -11 oC). RAOB plots for 925, 850, and 700 mb at 12 UTC (Figures 6a, 6b, and 6c) showed temperatures of -13, -15, and -18 oC. The -18 degrees C at 700 mb is a little colder than -12 to -16 oC but the 700 mb 12 UTC RAOB plot (6c) showed warm advection occurring over Green Bay ahead of the mid level trough axis. Also see the 700 mb temperature advection forecast valid at 18 UTC (Figure 7).
c. Vertical Motion Coincident with Favorable Dendritic Crystal Growth Temperatures
Relatively weak vertical motion existed early in the storm around 12 UTC at 850 and 700 mb (Figures 8a and 8b). Vertical motion was forecast to increase considerably at both levels by 18 UTC (Figures 9a and 9b) due isentropic lift from the warm advection (Figure 7) and the approach of the upper level vorticity maximum ( Figures 2a and 2b). Also note that weak warm advection was indicated at 925 mb at 12 UTC (Figure 6a) from Minneapolis, MN to Green Bay, WI ahead of the low level trough implying possible weak vertical motion below 850 mb. Note that the suface trough slowed down as it moved across Eastern Wisconsin (Figures 3a, 3b, and 3c), indicating a prolonged lower tropospheric influence on the production of snow. The storm thus appeared to feature vertical motion superimposed over a deep layer of temperatures favorable for dendritic crystal growth. The dendrites would then have a lot of time to grow to a large size.
d. Light Winds Resulting in Limited Compaction
Strong winds produce a sifting effect by causing snowflakes, and the ice crystals that compose them, to break into smaller pieces and pack more tightly. The almost calm winds of 0 to 2 knots reported by the Green Bay ASOS allowed snow to accumulate as loosely as possible.
Vertical motion over a deep layer of temperatures favorable dendritic crystal growth, plus limited compaction due to very light winds, resulted in an unusually low water content snowfall for a place like Green Bay, WI. Arctic air provided a vertical temperature structure favorable for dendritic crystal growth from the surface through 700 mb. Dendrites tend to form very fluffy loosely connected snowflakes resulting in a similarly low density snowpack on the ground, especially if winds are light.
The moral of the story? Check the observed and forecast atmospheric profile for favorable combinations of temperatures, vertical motion, and saturation. The original forecast for the December 23, 2000 event contained snow amounts only half of what was observed. One of the forecasters victimized by this event had a chance for redemption. Prior to issuing a forecast for a different storm to affect North Central Wisconsin, the forecaster took a quick look at a model time-height section. The analysis showed a brief period in which dendritic crystal growth temperatures would correspond to a layer of maximum vertical velocity. The expected snow totals were boosted above the amounts initially considered and the forecast was on target. Obviously some confidence must exist concerning the accuracy of the models chosen or the ability of the forecaster to make adjustments to those models.
Thanks go to Scott Cultice, Tim Kieckbusch, Tom Helman, Tasos Kallas, Allen Lagree, Dan Clark, and Gene Brusky of the National Weather Service Office at Green Bay, WI. They all provided input for the information presented in this case study.
Baxter, M. A., C. E. Graves, and J. T. Moore, Oct. 2005: A Climatology of Snow-to-liquid Ratio for the Contiguous United States, Wea. Forecasting, 20, 729-744.
Doesken, N. J., A. Judson, 1997: The Snow Booklet: A Guide to the Sciences, Climatology, and Measurement of Snow in the United States. 2d ed. Colorado State University Department of Atmospheric Science, 86 pp.
Goodison, B. E., 1978: Accuracy of Canadian snow gauge measurements, J. Appl. Meteorol., 27, 1542-1548.
Many of the graphics are compliments of the National Centers for Environmental Prediction (NCEP).
“The Ice Crystal Process” section of “Mesoscale Aspects of Winter Weather Forecasting: Topics in Winter Wx Forecasting” by COMET’s Meteorological Education and Training website.
List of Figures