FRAT Template: A Flight Risk Assessment With Scoring
01What the FAA publishes, and what it leaves to you
If you need the concept first (what a FRAT is, how weights, thresholds and escalation fit together, and where the tool stops), read what a flight risk assessment tool is. This article is the working template that follows it.
Two FAA documents are worth having open while you build yours. The first is InFO 07015, “Flight Risk Assessment Tool”, an Information for Operators notice dated 3 July 2007. It attaches a sample tool with 38 numbered factors in three sections (pilot qualifications and experience, operating environment, equipment), each with a point value. The second is the FAA Safety Team's “Flight Risk Assessment Tools” fact sheet, which describes a general aviation FRAT built on the PAVE checklist (Pilot, Aircraft, Environment, External pressures) with green, yellow and red results.
Read what they leave open. The InFO says an operator can use the tool as published or modify it, can add items unique to its operation, and “will need to create numerical thresholds that trigger additional levels of scrutiny prior to a go/no-go decision”. It also says risk assessment tools are only part of a safety management system and that the tool cannot guarantee a safe flight. The fact sheet is just as direct: a FRAT is not meant to make the go or no-go decision for you, and no FRAT can anticipate every hazard. So there is no standard score. There is a sample, and the calibration is your job.
02A sample FRAT template
The template below uses the four PAVE families and yes or no questions. A “yes” adds the weight; a “no” adds zero. Several questions echo the kind of factor the FAA sample lists, but the selection and every weight here are our sample, not a standard, and not the FAA's numbers. With all 16 answers at “yes” the sample maximum is 63.
| Factor family | Question (yes adds the weight) | Sample weight |
|---|---|---|
| Pilot | Captain has fewer than 200 hours on type | 5 |
| Pilot | A pilot has flown fewer than 100 hours in the last 90 days | 3 |
| Pilot | Single-pilot flight | 5 |
| Pilot | Rest before the duty day is shorter than 10 hours | 5 |
| Pilot | Planned duty day is longer than 12 hours | 4 |
| Aircraft | A deferred item (minimum equipment list) affects this flight | 2 |
| Aircraft | Landing distance needed is more than 80 percent of the runway available | 5 |
| Environment | Night operation | 5 |
| Environment | Destination ceiling or visibility is forecast close to approach minimums | 4 |
| Environment | Crosswind is above 15 knots | 4 |
| Environment | Runway is wet or contaminated | 3 |
| Environment | Thunderstorms or moderate to severe icing on the route | 5 |
| Environment | Best available approach has no vertical guidance | 3 |
| Environment | Mountainous airport, or an airport the crew has not used before | 4 |
| External pressures | Trip called with less than 4 hours of notice | 3 |
| External pressures | A passenger or schedule commitment makes a delay hard to accept | 3 |
Three rules make the form harder to game. Every question gets an answer, because a blank is not a zero. The score is taken before the crew is committed to the trip, not at the hold-short line. And the form names the flight: date, tail number, departure, destination and crew, so a score can be found again later.
Keep it short. Sixteen questions is near the upper limit of what a crew will answer honestly before every leg. If you add a factor of your own (an offshore deck, a medical crew on board, a student solo), look for one to retire.
03Sample score bands, the action each requires, and who approves
A score with no consequence is a number on a form. Each band needs an action and a signature, written into your procedures before the first flight is scored. The ranges below are samples against the sample maximum of 63. Calibrate them to your own data.
| Band | Sample range | Action required | Who signs |
|---|---|---|---|
| Low | 0-14 | Fly. The crew talks through the two highest-scoring answers and the score is kept on the record. | Pilot in command |
| Medium | 15-24 | Mitigate before departure: change the crew, the time, the fuel or the destination, write the mitigation down, and score the flight again. | Chief pilot, or a second qualified person the operator has named |
| High | 25 and above | No-go as planned. The flight only goes if mitigations bring the score down, and the decision is made outside the cockpit. | Director of operations or the accountable executive |
The worked example in InFO 07015 shows the middle band doing its job. A night flight with a low-time captain, a first officer short on recent hours, a wet runway and a crosswind scores 20 on the FAA sample. That is above the company threshold, so the chief pilot steps in. He swaps in a more current first officer and moves the destination to an airport with no crosswind expected, and the score drops to 13. Nobody signed a waiver. The flight that left was a different, lower-risk flight.
Two additions are worth considering. First, name one or two critical answers that send the flight to the high band approver whatever the total, for example a landing distance above 80 percent of the runway available. A sum can hide one serious factor among fifteen harmless ones. Second, if you fly single-pilot or you are the chief pilot, name your second person now. The InFO asks small operations to find a strategy that fits, and the FAA Safety Team suggests a designated contact. Our guide to an SMS for small and single-pilot operators covers how to scale the rest.
04How to calibrate weights and thresholds to your operation
Calibration is a desk exercise, and it takes an afternoon. Do it before the template goes live.
- Score flights you have already flown. Take the last two or three months of trips and fill in the template for each from the records: crew, times, weather, airport. Include the flights you cancelled.
- Look at the distribution. If nearly every flight lands in the low band, the tool will never ask for anything. If half of them land in the medium band, crews will learn that medium means nothing. The InFO's test is the one to apply: when every flight is acceptable under any condition, the thresholds are likely wrong.
- Check it against memory. Ask the chief pilot which trips in that period were uncomfortable. Those flights should score in the medium or high band. If one of them scored low, a factor is missing or a weight is too light.
- Adjust one thing at a time. Move a threshold, or change a weight, then score the same set of flights again. The InFO allows an operator to change any item provided it makes a realistic assessment of the hazard being changed, so write down the reason for each change.
- Weight by consequence, not by annoyance. The InFO frames a weight as a product of probability, severity and the value of that event against the rest of the operation. If you already rate hazards on a risk matrix, use the same reasoning so the two tools agree.
Then freeze the template for a few months. A form that changes every week produces scores you cannot compare with each other.
05How to review the scores every month
Each score helps one flight. The pile of scores tells you about the operation, and it is one of the few leading indicators a small operator collects without extra effort. Fifteen minutes a month is enough if you ask the same questions each time.
- Coverage. How many flights were flown, and how many had an assessment completed before departure?
- Band mix. How many flights fell in each band, and is the medium and high share moving compared with previous months?
- Top factors. Which three questions were answered “yes” most often on medium and high flights?
- Mitigations. What was actually changed on those flights, and did the score drop, or was the original plan signed as it stood?
- Decisions. How many flights were delayed, changed or cancelled because of the score? A year with none deserves a question.
- Honesty checks. Are there runs of all-zero forms, forms completed after departure, or one crew that never scores above the low band on trips where others do?
Keep the answers in the same place every month so they form a series. A factor that keeps coming back (one airport, one roster pattern, one season) is no longer a pre-flight question. It is a hazard, and it belongs in the hazard register with an owner and a mitigation. If you chart the band mix, read it the way you read any indicator, with control limits instead of a trend arrow, so one busy week of winter weather does not look like a shift.
Questions
There is a published sample, not a mandated form. FAA InFO 07015, dated 3 July 2007, attaches a sample flight risk assessment tool with numbered factors and point values, and it tells operators to use it as published or modify it for their own operation. The FAA Safety Team also publishes a FRAT for general aviation built on the PAVE checklist: Pilot, Aircraft, Environment and External pressures. Neither document sets thresholds for you. The InFO says each operator creates its own.
There is no standard number. In this template the sample high band starts at 25 out of a possible 63, and that figure is only a starting point. Score your last few months of real flights with the template first, then place the thresholds so routine flights land in the low band and the flights your chief pilot remembers worrying about land above it. FAA InFO 07015 gives the test: if every flight is within the acceptable range under any condition, the thresholds are probably not set correctly.
Someone who is not flying the trip and who has the authority to cancel it. In most operations that is the chief pilot for the middle band and the director of operations or the accountable executive for the top band. The worked example in FAA InFO 07015 has the chief pilot assess and approve flights above the company threshold, and he does it by changing the crew and the destination rather than by signing the original plan.
By naming a second person in advance. FAA InFO 07015 says small operations, where the pilot is also the chief pilot and owner, should consider strategies that fit them. The FAA Safety Team suggests a designated contact, such as a flight instructor, for a score that stays in the yellow after mitigation. Another operator, a mentor or a safety consultant can fill the role. What matters is that the name is agreed before the day you need it.
Look at the scores every month and change the template rarely. A monthly review counts flights by band, lists the factors that drove the medium and high scores, and checks that every flight had an assessment. Change a weight or a threshold only when several months of data, an occurrence or a change in your operation (a new aircraft type, a new base, a new kind of trip) gives you a reason, and record why you changed it.