Staying on Top of the Freezing Level

Winter and airframe icing are like peanut butter and jelly; it’s hard to imagine one without the other.

Winter and airframe icing are like peanut butter and jelly; it’s hard to imagine one without the other. [Credit: iStock]

Short of an erroneous forecast or calibration issue with your immersion thermometer, if you can remain below the lowest freezing level during your entire flight, there’s typically no chance for an encounter with airframe ice. Induction ice is certainly possible, but not airframe ice.

If you plan an altitude where the temperature aloft is zero degrees Celsius or less, airframe icing becomes exceedingly more likely while flying in visible moisture. Therefore, the freezing level is one key variable that you need to determine during your preflight analysis to better quantify your risk of airframe ice.

If you're not already a subscriber, what are you waiting for? Subscribe today to get the issue as soon as it is released in either Print or Digital formats.

Subscribe Now
Subscribe Now

Let’s clarify something right from the beginning. The FAA likes to use the term “freezing level” in all of its documentation. This is kind of a misnomer, given that water in the liquid state doesn’t necessarily freeze just because the static air temperature is below freezing. We must be concerned about the presence of supercooled water, which leads to airframe icing. On the contrary, water in the solid state (i.e., snow) must melt even if the static air temperature is a hair warmer than zero degrees Celsius. Meteorologists prefer to use the more accurate term of “melting level.” But pilots are stuck with “freezing level” for the foreseeable future.

A Rarely Standard Lapse Rate

One approach that some instructors teach is to use the standard lapse rate to calculate or estimate the lowest freezing level. That is, they use the current surface temperature at the airport and then subtract 2 degrees Celsius for every 1,000-foot gain in altitude. For example, if the surface temperature is 10 degrees Celsius and you are departing from an airport at sea level elevation, then the freezing level should be 5,000 feet. That method seems easy enough, but it’s a bad idea to do this. The standard lapse rate should only be applied to performance tables in the pilot operating handbook as a method to determine the departure from standard.

When Mother Nature is at her worst behavior, the atmosphere is rarely standard. In fact, during the late morning and afternoon, the environmental lapse rate is more often than not greater than standard near the surface. This is the layer of air that is directly influenced by the presence of the earth’s surface and, therefore, is what meteorologists refer to as the planetary boundary layer (PBL). The lapse rate in the PBL is often closer to the dry adiabatic rate (DALR) of 3 degrees Celsius for every 1,000-foot gain in altitude. Under these more typical conditions, using the standard lapse rate will cause you to calculate a freezing level that is higher than the actual lapse rate suggests. Therefore, if you expected the freezing level to be at 5,000 feet, you might be surprised during your climb to encounter supercooled liquid water beginning at 3,500 feet instead of 5,000 feet.

[Credit: iStock]

You might say, “That’s crazy—no sane pilot would do this.” Well, even the FAA fell into this line of thinking. In 2005, a student and a flight instructor ended up with a hard landing at Paine Field (KPAE) in Everett, Washington, after accreting airframe ice in a Cessna 172. They departed a nearby airport, Boeing Field (KBFI), to go out and shoot a few practice approaches in actual instrument conditions. This is certainly a noble effort when it is safe to do so. After the first missed approach, the instructor noticed ice accreting on the airframe and directed the student to return to Paine Field to land. During this landing, the aircraft ran off the runway, which resulted in an accident and subsequent FAA and National Transportation Safety Board (NTSB) investigation.

The FAA later determined the instructor, who was pilot in command, busted FAR 91.9 (a), which prohibits pilots from operating an aircraft without complying with its operating limitations. In this case, there’s a placard in the aircraft that states: “Flight into known icing conditions is prohibited.” The instructor was also cited with careless or reckless operation under FAR 91.13 (a). So, a certificate action was taken that included a 90-day suspension. The instructor requested an evidentiary hearing and later appealed the ruling to the NTSB. However, the NTSB agreed with the case the FAA presented.

I certainly don’t take issue with the outcome of this judgment, but one thing struck me as being a bit strange. The FAA argued that the instructor should have been aware of the lowest freezing level and, therefore, the potential for airframe ice at higher altitudes. To my chagrin, they suggested the instructor should have been aware of the surface temperature at the airport of 2 degrees Celsius—based on their briefing prior to departure—and then, they should have used the standard lapse rate to determine that the temperature would be at or below freezing in the clouds aloft. Ugh! Well, this just happened to be convenient for the FAA since the lapse rate near the surface was close to standard on that day and time. Therefore, it worked out, coincidentally, to favor the FAA’s case against this instructor. The lapse rate, as stated previously, varies and may not be the accurate measure to use when figuring the altitudes where icing might be likely.

Temperature Inversions

At the other extreme is a common situation when there is a formidable surface-based temperature inversion. In this situation, the surface temperature can be a chilly 7 degrees Celsius (45 degrees Fahrenheit) with a freezing level at more than 12,000 feet msl. This is quite common in regions around a warm front. As warm air overruns cold air at the surface, this creates a negative lapse rate (called an inversion) or a scenario where the temperature increases along with altitude before it begins to resume a more normal positive lapse rate in the free atmosphere aloft. When such an inversion exists, using the standard lapse rate may leave you with the impression that the freezing level is quite low when, in fact, it might be a very reasonable day to fly from an icing perspective if you remain below 12,000 feet.

[Credit: iStock]

In this scenario, it is quite common when an aviation accident occurs for the casual observer to quickly conclude that the aircraft encountered icing conditions. If it’s that chilly at the surface, then the freezing level must be just a few thousand feet up, right? That’s what pilots generally thought when a Beechcraft B58 Baron went down after departing the Spirit of St. Louis Airport (KSUS) on the winter evening of January 8, 2022. The flight departed at 7:10 p.m. CST headed westbound toward Denver. They were on an IFR flight plan and were cleared to climb to a cruise altitude of 8,000 feet. Shortly after reaching cruise, the Baron appeared to depart controlled flight with a rapid descent under unknown circumstances. It subsequently impacted the terrain 2.5 miles south of New Melle, Missouri (12 miles west of the Spirit of St. Louis Airport), killing the pilot and another occupant.

The temperature at the surface was 7 degrees Celsius with a dew point temperature of 6 degrees Celsius. There was much speculation and debate within the internet aviation community that airframe icing may have played a role in this fatal accident. This is certainly understandable. The elevation of KSUS is 463 feet msl, and using a standard lapse rate, the altitude of the lowest freezing level should be approximately 4,000 feet as shown here.

  • 1,463 feet --> +5 degrees Celsius
  • 2,463 feet --> +3 degrees Celsius
  • 3,463 feet --> +1 degrees Celsius
  • 4,463 feet --> -1 degree Celsius
This one-hour lowest freezing level forecast clearly shows a freezing level in the range of 11,000 to 13,000 feet msl near the accident site. [Courtesy of Scott Dennstaedt, FLYING Graphics]

Using the standard lapse rate in this way leads to an incorrect freezing level. This is echoed in the one-hour forecast (above), which suggests the lowest freezing level west of the Spirit of St. Louis Airport was between 11,000 feet and 13,000 feet. Using the standard lapse rate instead of the low freezing level forecast cre-ates an error of 7,000 to 9,000 feet in this case.

Shortly after the accident, the NTSB was quick to point out in a press conference that the freezing level was 12,000 feet and icing was unlikely since the aircraft remained below this level. But that didn’t make the internet community all that happy. Some still concluded that the NTSB was premature in its comments and that the causal factor would ultimately be associated with airframe icing, citing the standard lapse rate in their argument.

In fact, one YouTube personality suggested this flight likely encountered freezing drizzle, although no precipitation was reported at KSUS at the time of the accident. The NTSB will release its findings very soon, but the temperature profile on that evening included a healthy surface-based inversion and clearly the standard lapse rate would lead to a much lower freezing level.

The surface analysis chart valid at 00Z depicts a warm front moving north crossing through the flight path of the accident aircraft at the time of departure. [Courtesy of Scott Dennstaedt, FLYING Graphics]

Warm Air Overrun

With a little weather forensics, it was easy to discover that this was the classic case of warm air overrunning cold air—the result of the northerly movement of a warm front through the accident area as shown above. In fact, the temperature at approximately 5,000 feet msl was 9 degrees Celsius as shown below. It is not possible to accrete ice at those static air temperatures. Yes, there was freezing rain reported at the surface about 100 nm to the north-northeast at the Abraham Lincoln Capital Airport (KSPI) in Springfield, Illinois, where it was much colder, and the surface temperature was a chilly 1 degree Celsius.

Even so, there was a massive temperature inversion aloft over Springfield such that the temperature at 3,100 feet msl was 8 degrees Celsius. That’s an increase of 7 degrees Celsius at 2,500 feet above the surface. The temperature didn’t go negative over Springfield until roughly 9,000 feet msl. Whether in St. Louis or Springfield, this would have created a warm-soaked aircraft in the climb.

This 850 mb analysis shows that the temperature was 9 degrees Celsius at approximately 5,000 feet msl near the accident site. Purple is the zero-degree isotherm and isotherms in red are positive temperatures every 3 degrees Celsius. [Courtesy of Scott Dennstaedt, FLYING Graphics]

Where There's Freezing Rain

In a freezing rain scenario, it is common to have two (or more) freezing levels. One way this occurs is in the presence of deep saturated conditions with cold cloud top temperatures. This allows ice crystal growth and creates snow which falls into a melting layer to create rain. These drops then fall into a subfreezing layer near the surface to create supercooled large droplet (SLD) icing called freezing rain. All of this is courtesy of a surface-based temperature inversion with multiple freezing levels aloft. This is the classical freezing rain temperature profile.

But there’s also a more common case where clouds aloft are dominated by liquid when the cloud top temperature is much warmer. In this non-classical case, there may be two or more freezing levels, or the entire temperature profile may be below zero degrees Celsius. In this non-classical case, the saturated layer has a depth usually less than about 10,000 feet. This places the temperature of the cloud top to be warmer than negative-12 degrees Celsius. Warm-topped precipitation events like this—even when the entire temperature profile is below freezing—are dominated by water in the liquid state and often produce drizzle-sized drops with little or no ice crystals that are needed to develop the growth of snowflakes. This kind of non-classical temperature profile produces most of the cases of freezing rain and freezing drizzle.

If you want to avoid making a bad judgment, understand the big weather picture and then use the lowest freezing level forecast, like the one depicted above, which is found on the Aviation Weather Center’s website. This includes an analysis along with hourly forecasts up to 18 hours from the time they are issued.

This forecast is automated but is updated hourly and is generated from the Rapid Refresh (RAP) numerical weather prediction model. The vertical resolution is quite reasonable at 2,000 feet.

The official freezing level forecast is found on the same website at This graphical AIRMET (G-AIRMET) forecast is issued by aviation meteorologists and depicts the freezing level at 4,000-foot intervals. It also indicates where multiple freezing levels may exist, including their height. Its spatial and temporal resolution is not as good as the automated forecast, however.

A Skew-T log (p) diagram can quickly tell you about the altitude of the freezing level over a fixed location by looking for the intersection of the zero-degree Celsius isotherm with the environmental temperature. In this case, a single freezing level is located approximately 7,300 feet msl in this analysis from the Rapid Refresh model. [Courtesy of Scott Dennstaedt, FLYING Graphics]

If you are a weather nerd, you might try learning how to use a Skew-T log (p) diagram like the one above. One of the most interactive websites for these is found at The Op40 input data source used here is the same RAP model that is used to produce the lowest freezing level chart forecast. With such a diagram, you can precisely pinpoint the forecast freezing level over a particular location at a particular time or determine for yourself if multiple freezing levels exist.

The vertical route profile in the EZWxBrief progressive web app ( depicts isotherms every 10 degrees Celsius with the zero-degree isotherm depicted in red. This shows that a portion of a flight from Pittsburgh (KPIT) to Des Moines (KDSM) at 10,000 feet msl will include an altitude higher than the lowest freezing level, thus exposing the aircraft to airframe icing (icing severity shown in blue). [Courtesy of Scott Dennstaedt, FLYING Graphics]

And if you are a bit lazy, you can use a vertical route profile or vertical cross section that depicts the freezing level along your proposed route of flight. Many of the heavyweight apps have such a depiction, including my progressive web app, EZWxBrief. With a profile view, as shown above, it’s painless to see how the freezing level changes across your route of flight so you can quickly compare this to your proposed altitude. In fact, in addition to lines of constant temperature that include the zero-degree isotherm, you can overlay other key elements, such as clouds, icing severity, or even turbulence.

From the December 2022/January 2023 Issue 933 of FLYING

Scott resides in Charlotte, North Carolina, and flies regularly throughout the Mid-Atlantic and Southeast U.S. He is a CFI and former NWS meteorologist. Scott is the author of "The Skew-T log (p) and Me: A Primer for Pilots" and the founder of EZWxBrief.

Subscribe to Our Newsletter

Get the latest FLYING stories delivered directly to your inbox

Subscribe to our newsletter