Updated: Sep 14, 2021
Unfortunately, the answer is not as cut and dried as you might have hoped.
The regulations state in 14 CFR § 91.169 (b)(2)(i) that an alternate is not required to be filed if,
For at least 1 hour before and for 1 hour after the estimated time of arrival, the ceiling will be at least 2,000 feet above the airport elevation and the visibility will be at least 3 statute miles.
Pilots affectionally know this as the 1-2-3 rule. Essentially this says that an airport must be designated in an IFR flight plan to provide a suitable destination if a landing at the intended destination becomes inadvisable. However, waters are muddied because the regulations do not stipulate what weather reports or forecasts to use to make this determination. On a short duration flight filed right before takeoff, surface observations (METARs) should be used to determine the weather "one hour before" arrival. However on a longer flight planned a couple hours in advance, it is the forecast that comes into play most of the time. Even though it's not stated in the regulations, in reading the Instrument Flying Handbook (FAA-H-8083-15B), there's no question that if a Terminal Aerodrome Forecast (TAF) is issued for the destination airport, it must be used to evaluate the 1-2-3 rule. But what if a TAF isn't issued for the proposed destination airport?
How about using the nearest airport where a TAF is issued? Perhaps. As long as the destination airport is within 5 statute miles of that nearby airport, that may be acceptable. That's because a TAF is only valid 5 statute miles from the center of the airport's runway complex according to the Aviation Weather Services advisory circular, AC 00-45H Change 2. There are not that many airports that fall into this category, however. If the destination is greater than 5 statute miles from the other airport, then how does a pilot evaluate the 1-2-3 rule.
To answer that, a little history lesson is in order.
The FAA is pretty clear in the Instrument Flying Handbook (last revised in 2012) that the latest METARs, TAFs and aviation area forecast (FA) along with SIGMETs and AIRMETs, are the primary tools to be used in making this evaluation. However, in May 2015, the FAA formally asked the NWS to cease production of the textual aviation area forecasts (FAs). The FA was the product that pilots officially used to determine if an alternate airport was necessary when a TAF was not issued for the destination airport. The intent of the FAA in retiring the FA and transitioning to the digital Graphical Forecasts for Aviation (GFA) is to allow the Aviation Weather Center (AWC) meteorologists to focus their efforts on maximizing operational benefit to aviation users, resulting in improved weather information to pilots. The GFA, which is fully automated, was designed to ensure the availability of equivalent information, in addition to adding graphical displays of the predominant weather, sky cover, ceiling height, surface visibility and wind speed and direction. On October 10, 2017, the NWS ceased production of the textual FA for the conterminous U.S. So it would seem that the Instrument Flying Handbook needs to be revised to include the GFA. But what about using Model Output Statistics (MOS) instead?
MOS (pronounced moss), isn’t new and has been around since the early 1970s. In recent years it is now is being used by some general aviation pilots on a regular basis for preflight weather planning. Model Output Statistics, as the name implies, is derived from the output of numerical weather prediction models such as the North American Mesoscale (NAM) and Global Forecast System (GFS). These models are run every six hours on the supercomputers managed by the National Centers for Environmental Prediction (NCEP). Both the NAM and GFS offer a MOS product.
These forecast models do not automatically produce a point forecast for a specific town or airport, however. Therefore, one of the jobs that MOS fulfills is to use the “raw” model forecast mentioned above and attempt to improve on it using a statistical approach to produce an objective and more useful site-specific forecast (e.g., for an airport). MOS will remove known model biases and also quantify uncertainty through probabilistic forecasts (e.g., probability of precipitation). Most importantly, it will predict some "sensible" weather elements important to aviation such as ceiling height, cloud coverage and surface visibility that many forecast models do not predict directly (although that is changing as new models are developed). In its most basic structure, MOS is simply a computer program that is executed after the underlying model has completed and incorporates historical weather information called geoclimatic data that gives it an advantage over just the basic model output. In other words, it can respond to local terrain since the geoclimatic data used by MOS "knows" if the airport sits next to a river or in a valley or up on a hilltop. The GFA, on the other hand, uses the High Resolution Rapid Refresh (HRRR) model and does not provide this kind of site-specific guidance, but does provide forecasts for ceiling height, cloud coverage and surface visibility on a map.
As mentioned above, regulatory requirements force a pilot flying under instrument flight rules (IFR) to determine if an alternate airport needs to be included on the instrument flight plan. For airports with a Terminal Aerodrome Forecast (TAF), this is easy. But there are approximately 700 airports in the United States where the NWS issues a TAF. That leaves thousands of airports without the benefit of a TAF. So if you are flying to an airport that is not blessed with such a forecast, can MOS be used in place of a TAF to serve this regulatory function?
The short answer is yes and no! First, MOS is issued two ways; (1) primarily as a textual bulletin (see below); and (2) in a gridded form. Second, not all public use airports have MOS guidance. There are a little over 2,000 airports in the U.S. with a MOS forecast, of which, all 700 airports with a TAF also include a MOS forecast (MOS was created to help forecasters build a TAF). That leaves ~1300 airports with MOS guidance and no TAF. Third, MOS forecasts are not monitored. Unlike a TAF that can be amended by the forecaster, MOS has no such amendment criteria. Fourth, the GFS MOS (MAV) and NAM MOS (MET) do not provide absolute ceiling and visibility forecasts. Instead, it's a categorical forecast. For example, instead of providing a cloud height forecast of 600 feet AGL, a category of “3” is assigned that represents a range of cloud base heights from 500 feet AGL to 900 feet AGL as shown in the table above. Therefore, a ceiling height MOS forecast of 4, 3, 2 or 1 one hour before or one hour after the expected time of arrival would force an alternate airport to be filed. Note some EFBs that provide access to MOS might attempt to "pick a number" that falls in the middle of this range (e.g. 700 feet AGL). The problem here is that the GFS and NAM MOS forecasts are not like TAFs that use forecast groups valid over a range of time; instead, MOS has a temporal forecast resolution with an interval of 3 hours. Using the GFS MOS forecast below, does a categorical ceiling forecast of "5" valid at 0300Z on September 5th (second column) mean that an alternate is unnecessary if arriving at 0500Z? Does the 0300Z forecast persist until 0559Z? What happens if the NAM MOS is different? Which MOS should be used? As you can see, there's no easy answer.
If you look at Aviation Weather Services, AC 00-45H Change 2, you will notice that it says nothing about the GFS (MAV) or NAM (MET) MOS. However, the advisory circular does point out that the Localized Aviation MOS Program (LAMP) can be used as supplemental guidance. Specifically it says,
The LAMP weather product is entirely automated and may not be as accurate as a forecast generated with human involvement. However, information from the LAMP weather product can be used in combination with Terminal Aerodrome Forecasts (TAF), and other weather reporting and forecasting products and tools, to provide additional information and enhance situational awareness regarding a particular location.
The key here is that LAMP is recognized by the FAA and suggests that it can be used to enhance situational awareness. However, it does not specifically suggest it can be used as primary guidance. But read on. LAMP is a statistical system that uses observations (surface observations, radar, lightning), GFS and NAM MOS output, and model output (HRRR) to provide additional guidance for aviation. Given that the GFS MOS is only updated every six hours, LAMP was originally created to act as an update to the GFS MOS guidance. It also comes in both a bulletin and gridded form and is refreshed hourly and provides an hourly forecast interval a 25 hour period for ceiling height and visibility for nearly 2000 airports. Sounds like the perfect solution for alternate airport evaluation for those airports without a TAF.
In fact, if you visit the AWC's GFA tool, you will see the LAMP forecasts for ceiling and visibility. Shown below is the gridded version of the LAMP ceiling forecasts nicely contoured (visibility and flight category selections are available as well). Notice the time slider is hourly making it a bit easier to determine the weather 1 hour before and 1 hour after the expected time of arrival at the destination. Although if your ETA is 0230Z, what hour do you use? That's not entirely clear, but it would be safe to assume that you that the weather expected at 02Z persists through 0259Z. Nevertheless, if the pilot were ever challenged in administrative court after a violation of these regulations, the FAA would likely see this as a reasonable attempt to do the right thing.
In the end, the GFA was built to replace the legacy FA. Therefore, it's safe to assume that the gridded LAMP forecast provided in the GFA would be a legal way to determine if an alternate airport is necessary. Moreover, using the official LAMP bulletin such as the one below for a specific airport in the LAMP system would also be acceptable. It is not recommended to use the GFS or NAM MOS for this evaluation. Furthermore, do not use the MOS forecast provided by most EFBs unless they specifically provide the original LAMP bulletin.
As shown below, LAMP generally performs better than the GFS MOS product alone especially in the first 6 hours of the forecast.
Currently, EZWxBrief provides access to the Localized Aviation MOS Program (LAMP) bulletin on the EZAirport page as shown below.
Most pilots are weatherwise, but some are otherwise™
Dr. Scott Dennstaedt
Weather Systems Engineer
CFI & former NWS meteorologist