You may be one of the many people wondering why meteorologist and forecasters struggle with snow prediction. Well here are some of the many things we look at when making a snowfall forecast.
Step 1: Determining Precipitation Type
Aside from forecasting the track of the storm, and trust me that can be difficult at times, this is the first thing a forecaster must do to determine if an event will be rain, snow, sleet, freezing rain or even a mix of all. We first need to get an understanding of how a particular storm will evolve and move through the region. While many forecasters may simply look at surface temperatures and wait until they drop to or below freezing (32° F or 0° C), this is wrong and the forecaster may be setting them self up for failure. We need to look at a deeper layer of the atmosphere, mainly the lowest 10,000 feet. The variation of temperatures through this layer of the atmosphere can be rather significant so we like to look at the average temperature of this layer and to do this we look at the thickness of the layer.
Forecasters like to look at the thickness of the atmosphere between two pressure levels. Around here those pressure levels are 1000 millibars(mb) and 500 mb. This layer gives us the best representation of the part of the atmosphere that will generate winter precipitation. If the thickness of this layer is 5400 meters or lower, frozen precipitation looks like a good bet. If the thickness of the 1000 mb to 500 mb layer is greater than 5400 meters, liquid precipitation will be more likely. For Cheyenne, since we are rather elevated, the thickness of this layer can be a little thicker, perhaps up to 5460 meters and we would still see snow. For an example of this, look at Figure 1 which shows the 1000 mb to 500 mb thickness and precipitation. Everything north of that 540 decameter line will likely be frozen precipitation. We also look at Forecast Soundings that give us a look at the atmosphere from top to bottom. These give us good ideas of how friendly the environment will be for snow development. We will look more in to this later in the article.
Obviously, there are cases where we could see sleet or freezing rain depending on the temperature profile between the base of the cloud and the ground, but for the sake of this article lets say the entire temperature profile is below freezing and snow is forecasted.
Alright, we know it's going to snow! Now it's time for the hard part, figuring out how much snow will fall and how much of that will actually stick to the ground.
Step 2: Looking at Surface Temperature
This is when the temperature of the ground and air near the ground comes more in to play. If the ground is at or below 32°, then we will likely see accumulating snow. If ground temperatures are above freezing then snow will likely melt before the ground becomes cold enough for snow to accumulate. Temperatures of the recent past can also play a role in snow accumulation. If temperatures have been really warm especially for a long duration before a storm arrives, then snow will melt a lot before it accumulates. However, if the intensity of the snow is heavy, then snow can accumulate faster than it could melt leaving the area under some significant accumulation. This tends to happen around here in the early spring months.
Step 3: Forecasting the Snow to Liquid Ratio
There is a general rule of thumb out there that says "for every 10 inches of snow you will get 1 inch of rain." This is a very generic assumption and a very horrible one at that. The climatological liquid to snow ratio for Cheyenne is 11:1. That is 11 inches of snow is equal to 1 inch or rain. However, this ratio is useless when forecasting your week to week storm. You see, the ratio varies from storm to storm and depends on the temperatures within the storm. A colder airmass holds less moisture than a warmer airmass. So, a Clipper system that originates in Canada may have a snow to liquid ratio of 25:1 where as a warmer, early spring storm system could have a ratio of 5:1.
So lets say models are predicting 0.50" of liquid precipitation. If the storm is colder and has a ratio of 25:1, then we could see 12.5" of snow accumulation while a warmer storm with a ratio of 5:1 will yield only 2.5" of snow, all with the same amount moisture. If the forecaster were to go by the "Rule of Thumb of 10:1" then he or she would have forecasted 5.0" of snow, much different than what would have fallen, especially if it's a colder storm with a higher snowfall ratio!!!
Unfortunately, the ratio doesn't just change from storm to storm. It often changes during the storm. For instance, temperatures at the start of the storm are warmer and will likely give us a lower snow to liquid ratio than the end of the storm when temperatures are colder and ratios are higher. If temperatures are too cold and the cold air is too deep, then the crystal type is not conducive to higher snow ratios. This is something else we as meteorologists and forecasters need to be mindful of, especially during long duration snow events. Keep in mind, a change in the temperature by even a couple degrees can significantly alter the snowfall amount.
Step 4: Snowflake Size
This is one of the more difficult aspects to forecasting snow and goes hand in hand with forecasting the correct snow to liquid ratio. Large snowflakes know as dendrites can pile up higher as opposed to smaller flakes known as plates, needles and columns. You may have heard the saying, "It's too cold to snow." While that saying is false as it is written, there is some truth behind it. Very cold temperatures lead to snowflakes developing as columns and small plates which will result in lower snow to liquid ratios and take a lot longer to pile up.
To get large dendrites we need the snow growth region to be located where we have good lift, high moisture content and temperatures within that region to between 14° F and -8° F. If this region is deep with ample lift, we could be looking at large snowflakes. You can see a chart that shows snowflake size relative to moisture and temperature in Figure 2.
Now in order for us to even see what the future might hold for us, we need to rely on numerical weather models. These models do the best that they can do at trying to replicate the atmosphere, though they never will be able to do it 100%. Because of this, weather models often put out many different solutions in regards to what will happen which only adds to the uncertainty of a forecast. You can see an example of this in Figure 3 where models show nearly a 10 inch spread in snow accumulation for Cheyenne.
The Importance of Wind Direction and the Local Topography
If you're an avid viewer of CBS NewsChannel 5 or a weather enthusiast, you've probably heard us talk about Upslope and Downslope winds. I was told once that "75% of a forecast is accurately predicting the wind direction and speed, if you can do that the rest will usually fall in to place." Thanks to our close proximity to the mountains, forecasts become a bit more complex.
Westerly winds move downhill and as air sinks, it gets compressed, warms up and becomes drier. Winds that move uphill, usually around here that's a wind with an easterly, northerly, and in few areas, southerly component, do the exact opposite. Air that rises will expand, cool and condense in to water droplets or ice crystals.
If the predicted wind direction is slightly off, say 20-45 degrees, it can blow a forecast completely! You can see this effect in the Visible Satellite image from January 2nd, 2014 in Figure 4. You can see where downsloping limited snow across parts of Northern Colorado and where upslope wind enhanced snow in parts of SE Wyoming. During this particular event, heavier snow bands developed east of the Front Range. This was due to a nearby jet streak that enhanced lift over the area. While the development of these bands were somewhat expected, the exact locations of these bands were very difficult to forecast and typically can't be nailed down until they begin to form.
Now this article wasn't written as an excuse for a blown forecast. When meteorologists such as myself make a snowfall forecast, we expect it to be correct. Hopefully this article helps explain the complications of forecasting snow!
This image is an example of the GFS weather model. This image has an arrow that shows the 540 dm (5400 meter) thickness line.
Diagram relating the snow crystal type to the temperature of formation. Adapted from Snow Crystals, Natural and Artificial by Ukichiro Nakaya (1954).
This image shows how different and inconsistent models can be in regards to snow accumulation.
Raw Visible Satellite Image Courtesy of NCAR.