The Current Icing Product, the Supercooled Large Droplet Layer, and the Probability Number That Does Not Mean What Most Pilots Think It Means

The probability number on your icing forecast - whether it reads 40 percent or 80 percent - describes the likelihood that icing conditions exist somewhere within a roughly 13-kilometer grid box, not your personal probability of accumulating ice. Most pilots misread this figure, and that misunderstanding affects every flight-planning decision they make with the Current Icing Product and Forecast Icing Product. Compound the error by overlooking the separate Supercooled Large Droplet layer, and you have a forecast system fully capable of misleading a well-prepared pilot.

What Are the Current Icing Product and Forecast Icing Product?

The Current Icing Product (CIP) is a real-time, three-dimensional gridded analysis of atmospheric icing conditions across the continental United States, southern Canada, and adjacent ocean areas. The Forecast Icing Product (FIP) extends that picture forward in time, providing hourly forecasts out to 18 hours.

Both products come from the Aviation Weather Center, a branch of the National Weather Service, developed in partnership with the National Center for Atmospheric Research (NCAR) in Boulder, Colorado. When you pull up the icing overlay in ForeFlight or Garmin Pilot, the algorithm those researchers built is running the analysis.

How Is the CIP Built?

The CIP doesn’t rely on a single data source. It fuses multiple inputs simultaneously using a technique called fuzzy logic - a method from control systems theory that allows a computer to work with imprecise, overlapping categories. Instead of a binary “icing: yes or no,” fuzzy logic produces a result like “the probability that conditions support icing is 73 percent,” handling atmospheric ambiguity in a mathematically rigorous way.

The inputs include PIREPs from aircraft flying through conditions, METAR observations from ground stations, satellite infrared imagery estimating cloud-top temperatures and liquid water content, numerical weather model output, and radiosondes - the instrument packages on weather balloons launched twice daily by the National Weather Service. A radiosonde transmits temperature, humidity, and pressure as it ascends, giving the algorithm the vertical profile it needs: where the freezing level sits, whether a temperature inversion exists, and whether the atmospheric column can support supercooled liquid water.

The result is a three-dimensional grid showing icing probability and severity across altitude bands. A cross-section might show icing probability peaking between 8,000 and 14,000 feet, with moderate severity estimated between 10,000 and 12,000.

What Does the Probability Number Actually Mean?

This is the critical distinction most pilots miss.

The probability displayed is the probability that icing conditions exist somewhere within that grid box. Each CIP grid box covers a horizontal area of roughly 13 kilometers by 13 kilometers. A 60 percent probability means there is a 60 percent chance that, within that box, conditions support ice formation somewhere inside it.

It does not mean 60 percent of that area is coated in ice. It does not mean you have a 60 percent chance of accumulating ice on your airframe.

Pilots sometimes see 40 percent and conclude it’s manageable - less than half, after all. But if your route runs directly through the core of that region, your actual encounter probability could be substantially higher. Thread between systems, and it could be much lower. The chart builds situational awareness. It is not a go/no-go threshold.

Use it alongside current PIREPs from aircraft that actually flew your altitude today, the temperature at cruise altitude, your aircraft’s certification basis, and your personal minimums.

Why PIREPs Matter - and Where They Fall Short

PIREPs are critical inputs to the CIP, but coverage is uneven. Over congested airspace like the Northeast corridor, reports are plentiful. Over the Rockies or oceanic routes, they can be hours stale or simply absent.

The CIP algorithm accounts for this by weighting PIREPs heavily when fresh and available, but shifting toward numerical models and satellite data when coverage thins. The product over Kansas City at 9 a.m. - with dozens of jets reporting conditions - is more reliable than the same product over western Nevada at 2 p.m. when the last PIREP was four hours old from an aircraft at a different altitude.

Modern commercial airliners equipped with ACARS (Aircraft Communications Addressing and Reporting System) now transmit automated in-flight reports that increasingly feed aviation weather products. The data density from those aircraft far exceeds human-generated reports. Right now, that coverage concentrates on high-traffic routes. General aviation - which flies exactly the routes where PIREP coverage is worst - contributes the most to the system and receives the least density back.

What Is the SLD Layer, and Why Does It Demand Separate Attention?

SLD stands for Supercooled Large Droplet. This is a categorically different hazard from the standard light-moderate-severe icing scale - different physics, different consequences for your aircraft.

Standard certificated icing protection systems, both pneumatic boots and heated leading-edge systems, are designed and tested against a specific droplet size spectrum defined in Appendix C of Part 25 of the Federal Aviation Regulations. That certification envelope covers most stratiform cloud encounters and some cumuliform situations.

Supercooled large droplets are significantly bigger. They originate primarily from freezing drizzle and freezing rain, where larger drops form higher in the atmosphere and fall into the supercooled zone below the freezing level without themselves freezing.

When SLD strikes a wing, it does not accrete only on the leading edge where protection is active. It runs back, accreting aft of the deice boot or heated surface in unprotected regions. The resulting ice shape - called a ridge - can dramatically alter the wing’s aerodynamic characteristics in ways that are not immediately apparent to the pilot. The aircraft can feel entirely normal right up until it doesn’t.

The Accident That Rewrote Icing Certification

On October 31, 1994, an ATR 72 operated by American Eagle crashed near Roselawn, Indiana. 68 people were killed. The aircraft had been holding at 10,000 feet and encountered freezing drizzle. SLD ice accreted aft of the deice boots, aileron control was compromised, and the aircraft rolled out of control.

The NTSB investigation of the Roselawn accident fundamentally changed how the FAA approached icing certification. The result was a new standard called Appendix O, which specifically addresses SLD conditions. Aircraft certified under Appendix O have been tested against the SLD droplet spectrum - not just the classic Appendix C envelope.

If you fly a turbine aircraft or a piston with a known icing certification, you should know which standard it was certified under. That information is in the Airplane Flight Manual.

How to Find the SLD Layer in Your EFB

The CIP and FIP now include a separate SLD probability layer. It is not displayed by default in every electronic flight bag application, but it is there.

Both ForeFlight and Garmin Pilot display the SLD layer as a separate toggle, distinct from the standard icing severity overlay. Find it on the ground before you need it at altitude. When SLD probability is elevated along your route, that is a fundamentally different level of concern than a standard icing probability reading - the droplet spectrum may be outside the design envelope of your certificated protection system.

SLD conditions concentrate in specific regions and seasons: the Great Lakes in late fall and early winter, driven by warm lake water meeting Arctic air; the Pacific Northwest in winter; and the Appalachian corridor. These patterns reflect the specific temperature profile required to produce freezing drizzle and freezing rain - warm air above, cold air below - a structure recognizable in the forecast and one that gives the FIP SLD layer genuine predictive value when a frontal system moves through those areas.

How to Use the Forecast Icing Product for Preflight Planning

The FIP runs CIP inputs forward through the Rapid Refresh numerical weather model, which updates every hour at 13-kilometer horizontal resolution. It produces hourly forecasts out to 18 hours, covering the entire planning window for most operations.

In practice, you can walk through the forecast icing layer hour by hour along your route, watching how probability evolves as a front moves through. You can time a departure or route adjustment to avoid the peak icing window.

The limitation is forecast degradation over time. At hours one and two, the FIP is heavily constrained by recent observations and is reliable. By hour twelve, it is only as good as the model’s ability to predict atmospheric state twelve hours from now. The atmosphere is a chaotic system and model error compounds.

A practical framework:

• Hours 1–6: Planning tool with reasonable confidence
• Hours 6–12: Planning tool with appropriate uncertainty
• Beyond 12 hours: A conceptual sketch - a frontal passage the model places at hour 15 may arrive at hour 12 or hour 18

Always, close to departure, overlay the FIP against current PIREPs and METARs. The forecast is what the model thinks will happen. The observation is what is actually happening. If the two diverge significantly in the hour before you depart, trust the observation.

Where Icing Forecasting Is Headed

Machine learning algorithms trained on historical icing encounter data are beginning to outperform traditional fuzzy logic approaches in certain conditions, particularly in complex terrain and rapidly evolving synoptic situations. NCAR and NOAA have active research programs in this direction. A wholesale replacement of the CIP is not imminent - these products carry certification and validation requirements that take years to satisfy - but the improvement trajectory is clear.

There is also a longer-term opportunity in ADS-B data density. Every general aviation aircraft transmitting ADS-B is a mobile atmospheric sensor. If the moment a pilot activates deice equipment were captured and aggregated across the fleet, the result would be a real-time icing encounter map requiring no one to pick up a microphone. Research programs have explored exactly this. The pilot community flying today is building the dataset that will make icing forecast products of 2030 substantially more accurate than what exists now.

Key Takeaways

• The CIP/FIP probability figure describes whether icing conditions exist somewhere in a 13km × 13km grid box - not your personal accumulation probability.
• The SLD layer is a separate toggle in your EFB and represents a categorically different hazard that standard Appendix C–certified systems were not designed to handle.
• PIREP coverage is sparse over low-traffic routes; weight the CIP product accordingly - it is less reliable over western Nevada at 2 p.m. than over the Northeast at 9 a.m.
• Know your aircraft’s Appendix C vs. Appendix O certification basis - it determines what icing droplet spectrum your protection system was actually tested against.
• Treat FIP forecasts beyond 6 hours as a planning sketch; close to departure, defer to actual PIREPs and METARs over model output.