On 29-30 January 1996, the Cheyenne area experienced a snowfall event. During this period, official measurements by the National Weather Service Forecast Office (NWSFO) were 0.13" of liquid precipitation and 4.8" of snowfall. The official liquid precipitation was measured using the Automated Surface Observing System (ASOS) and the snowfall was derived using the "New Snowfall to Estimated Meltwater Conversion Table" (Table 1), aided by measurements taken using a standard ruler by NWSFO personnel. This will be referred to as the "Table" henceforth.
However, a backup weighing gauge recorded 0.26" of liquid precipitation and core samples from the event yielded 0.4" to 0.5" of liquid. Therefore, it is likely that at least twice the liquid equivalent recorded by ASOS actually fell.
Snowfall measurements at the time of observations were deemed accurate via a ruler as there was generally less than 10 knots of wind throughout the event. Surface temperatures were quite cold (near 0 degrees Fahrenheit), with warmer temperatures aloft. Using the more representative 0.26" of liquid equivalent with a temperature of 0 degrees F through the Table gives an estimate of 11" of snow (almost 3 times what actually fell). However, using a temperature of 27 degrees F above the inversion (where the snow was likely generated) gives an estimate of approximately 3 to 3.5".
Therefore, this study highlights a typical snow event in the High Plains where the applicability of the New Snowfall Table is questionable. Use of surface temperature alone to determine liquid-to-ice ratios can be extremely misleading, if the temperature at the level where snow is generated is significantly warmer. The study also describes some of the inherent difficulties in measuring snowfall in this region.
The precipitation event began at 0140 GMT on 30 January and ended at 0247 GMT on 31 January. The synoptic weather pattern at the time of the snowfall included the presence of a very cold entrenched Arctic air mass over the eastern slopes of the Rocky Mountains, reaching into the northern and central Plains. Aloft, a warm, moist Pacific air mass had moved across the Rockies into the central High Plains. According to the sounding from Denver, Colorado (DNR) at 12GMT, 30 January (Figure 1), a saturated layer was located in the 780-730 mb layer, with the warmest temperature of 28 degrees F at about 720 mb. In this environment, the snow crystals formed and grew aloft in the warmer air, then fell through the cold air in the lowest levels.
Figure 1. Denver, CO (DNR) 12GMT, January 30, 1996.
During the event, surface wind speeds were generally less than 10 knots. At the onset, southeasterly winds at 8 knots were measured, and north-northwest winds at 6 knots were present at the conclusion of the snowfall. Cheyenne observers deducted that during the snowfall the light wind speeds were probably not significantly affecting how the gauge captured falling snow.
Throughout the event, visibilities reported by ASOS were frequently less than two miles in light snow and fog. At times, ASOS visibilities were reduced to as low as three-quarters of a mile in light snow and fog with lowest cloud ceilings down to 500 feet, and liquid precipitation rates as high as 0.02" per hour. In the brief periods during which snow was not falling, visibilities quickly increased to 7 miles or greater, indicating that the snow, not the fog, was primarily responsible for reducing the visibilities (ASOS measurements typically include fog during significant snowfall).
Surface temperatures at the beginning of the event were 8 degrees F, but steadily dropped to 0 degrees F around 0815 GMT on 30 January. Temperatures continued to fall, reaching a low of -5 degrees F from 1300 to 1400 GMT. This was followed by an increase in temperature to 0 degrees F at 2000 GMT, and +3 degrees F at the conclusion of the event. Temperatures ranged from +3 degrees F to -3 degrees F during the heaviest snowfall (2.8"), and 0.06" liquid was measured by observers for this period (06 GMT through 12 GMT 30 January).
ASOS is the official measuring device of precipitation at Cheyenne, and was used by NWSFO personnel in this event. As mentioned above, ASOS measurements were 0.13" liquid during the event. However, ASOS does not directly measure snowfall or snow depth.
ASOS frequently does not accurately measure liquid precipitation amounts during snowfall, especially in very cold or windy conditions, mainly attributable to the lack of an effective wind shield. Also, the heated tipping bucket used in the system creates a heat-island effect which deflects falling crystals, thus further reducing the amount of snow which the gauge catches. This has been documented previously by DeRung (1990). This system also can cause significant amounts of evaporation to occur before liquid is measured during very cold conditions (Middleton and Spilhaus 1953).
The conventional weighing gauge at Cheyenne is located about ½ mile from the ASOS and is not known to be inaccurate during snowfall in light winds. This measuring device is utilized as a backup in most cases, but is used for recording official precipitation when ASOS is deemed inaccurate by the observer. The device does not use heat. During this event, the weighing gauge recorded 0.26" of liquid, or about twice that estimated by ASOS and observers.
When evaluating observations for augmentation purposes, the observers utilized the Meltwater Conversion Table (Table 1). It is a part of the DOC, NOAA, NWS Handbook Number 7, Surface Observations (Part IV, Supplementary Observations, Table 2-14). The Table converts water equivalent in hundredths of an inch into new snowfall to the nearest tenth of an inch, using only surface temperature as an input variable. For example, 1.00" of liquid equivalent with surface temperatures of 28 to 34 degrees F is equivalent to 10.0" of new snowfall, or a 10-to-1 ratio of snow to liquid, respectively.
During the event, observers were using the Table to augment snowfall from the ASOS precipitation measurements. At the time, this appeared representative in that the precipitation was seemingly relatively light (0.01" an hour) in conjunction with the cold temperatures (near 0 degrees F). This resulted in hourly amounts of roughly 0.4" of snowfall. During the heaviest snowfall, 0.06" liquid was converted to 2.8" of snowfall at a 45-to-1 ratio using the Table as guidance.
When using the warmer mid-level temperatures, along with 0.26" of liquid the weighing gauge recorded, the Table yields approximately 3 to 3.5 inches of new snowfall, or a 15-to-1 ratio.
Cores of snow on the ground just after the conclusion of this event were sampled, slowly melted, and measured for liquid equivalent. Several cores indicated a range of 0.4 to 0.5" liquid. This is not an unusual case in that all gauge measurements significantly underestimated snowfall. However, the ASOS deficiencies mentioned above led to severe under measurement of liquid (0.13").
The observers during the event were, as previously stated, using the Table in accordance with local policy. At the time this procedure takes into account the surface conditions but not conditions aloft or at larger spatial scales. High Plains snowfall associated with arctic outbreaks can occasionally contain significant amounts of liquid, with relatively low ice-to-water ratios, since mid-level temperatures can be warmer than surface temperatures (Wesley et al. 1990).
A discrepancy of 0.13" precipitation occurred, or 50 percent of the weighing-gauge total and about 30 percent of the core-sampled total. A partial cause of this problem was the adherence to the Meltwater Table. It is likely that this type of error in liquid estimates is a frequent problem in the High Plains, since precipitating Arctic air masses are common occurrences. In a hypothetical situation where 10 inches of snow is measured and falls when the surface temperature is below 0 degrees F, and contains a water equivalent of 0.50", the Table would inaccurately estimate a 50-to-1 ice-to-liquid ratio, or 25" of snow!
Another potential error in utilizing the Table is the assumption that snow always gets fluffier as temperatures get colder. Studies in the Rocky Mountains have shown that the fluffiest, lowest density snows typically fall with light winds and temperatures near 15 degrees F. But at very cold temperatures (near and below 0 degrees F), crystals tend to be smaller so that they pack more closely together as they accumulate, thus producing snow that has a relatively high density (Doesken and Judson 1996).
In this event, a more accurate use of the Table would incorporate a temperature in the 25 to 30 degrees F range with 0.4" of liquid to result in the 4.8" of snowfall received. This is about a 13-to-1 ratio and correlates well to the warmest temperature aloft in this case study.
This snow event illustrates the uncertainty on the reliability of the procedure of utilizing surface temperatures along while employing the Meltwater Table in computing new snowfall from water equivalent. This combined with deficiencies in the capabilities of ASOS in measuring snowfall, led to a serious underestimate of the actual liquid precipitation while maintaining snowfall measurements which agreed with snow depth increases.
Though the Meltwater Table may be more applicable in other geographical regions, for the High Plains temperatures aloft must be factored into the estimation of snowfall, especially during a cold outbreak or overrunning situations. The inherent deficiencies of ASOS precipitation measurements may further complicate this situation.
DOC/NOAA/NWS, 1996: Observing Handbook Number 7, Surface Observations Part IV, Supplementary Observations, Table 2-14, 440pp.
DeRung, D.E., 1990: Snowfall Water Equivalent Comparison of Eight Inch Standard Gauge Versus Heated Tipping Bucket. NWS Central Region Technical Attachment 90-32, NWSFO, Bismarck, North Dakota.
Doesken, N.J., and A. Judson, 1996: The Snow Booklet: A Guide to the Science, Climatology, and Measurement of Snow in the United States. Colorado State University Department of Atmospheric Science, 78pp.
Middleton, W.E., and A. Spilhaus, 1953: Meteorological Instrumentation, 3rd Ed., Univ. of Toronto Press, 286pp..
Wesley, D.A., Weaver, and R. Pielke, 1990: Heavy snowfall associated with an Arctic outbreak along the Colorado Front Range. Nat. Wea. Dig., 15, 2-19.