Flight Radiation logoFLIGHTRADIATION

Methodology

Dose figures on a FlightRadiation report come from the FAA's CARI-7 model. This page sets out how the number is produced, what we assume when you do not tell us, how far off it can be, and what it does not cover. Our figures against the FAA's own published route doses are on the validation page.

1. The quantity we report

Reviewed 4 September 2026 · Model: FAA CARI-7 v4.1 · Reference standard: ICRP-103 (2007)

We report effective dose, written E, in microsieverts. Effective dose is not a measurement. It is a calculated whole-body equivalent that accounts for how biologically damaging each radiation type is, and for the differing sensitivity of organs, so that exposures of different kinds can be compared on one scale. We use the ICRP Publication 103 (2007) tissue weighting factors, which superseded ICRP Publication 60 (1990). That weighting is applied inside CARI-7 itself, not by us.

2. The model

CARI-7 is the cosmic radiation transport code published by the FAA's Civil Aerospace Medical Institute. It uses Monte Carlo particle transport, with shower tables derived from MCNPX, and models secondary particle production explicitly. It is the code the FAA names in Advisory Circular 120-61B, and it is one of the four programs named in the UK's Air Navigation (Cosmic Radiation) Order 2019.

Effective dose rate at altitude depends on exactly three quantities: the vertical geomagnetic cutoff rigidity where you are, the atmospheric depth you are flying at, and the state of the solar cycle. Rather than run the Fortran per order, we ran the FAA's own CARI-7 v4.1 build across a grid of those three variables, and the site interpolates that grid. The result is arithmetically the same thing, several thousand times faster.

We tested that the shortcut is faithful. Given the same vertical profile, the interpolated engine reproduces a direct CARI-7 run to 0.5% on average and 1.0% at worst across eight routes spanning polar to subtropical. That comparison, and a route-by-route check against the 27 flights the FAA published, are on the validation page.

3. Cutoff rigidity, not latitude

The single biggest driver of your dose is how effectively Earth's magnetic field deflects incoming cosmic rays where you are flying. That is measured as vertical cutoff rigidity, in gigavolts. A low cutoff means little shielding and a high dose rate.

Cutoff rigidity is not the same as latitude, and the difference is large enough to matter. At 40 degrees north the cutoff over the North Atlantic is about 3.6 GV, but at the same latitude over Asia it is about 9.0 GV. At FL370 that is roughly 4.7 against 2.6 microsieverts per hour, so two flights at identical latitude can differ by nearly a factor of two. We take cutoff from the Shea and Smart one-degree grid that ships with CARI-7, sampled along the true great circle rather than at the endpoints, because a great circle bows toward the pole: London to Los Angeles reaches past 60 degrees north even though neither airport is above 52.

4. Solar conditions

Galactic cosmic radiation is modulated by the sun. When the sun is active it deflects more of the incoming flux and your dose falls; at solar minimum it rises. The effect is large at high latitudes, roughly a factor of two between minimum and maximum at the pole, and small near the equator, where the geomagnetic cutoff already excludes the particles the sun modulates.

CARI-7 handles this through the heliocentric potential, a monthly figure published by the FAA and derived from ground-level neutron monitor counts. If you give us the month you flew, we use the published potential for that month. If you do not, or the month falls outside the published record, we use the most recent published month and the report says so on the row concerned. The series we ship runs from January 1958 to July 2026 and is refreshed from the FAA when they publish.

5. What we take from you

  • Origin and destination. IATA codes. We cover the 4,133 airports that have an IATA code and scheduled service, from the OurAirports database, and compute the great circle between them.
  • Round trip and frequency. Whether the route counts both legs, and how many times a year you fly it.
  • Month flown, optional. Sets the solar conditions, as above.
  • Cruise altitude and block time, optional. If you supply them we use them instead of assuming, and the stated uncertainty on your figure narrows from about 30% to about 25%.
  • Passenger profile. Adult, or pregnant. Pregnant changes what the report compares your dose against; see section 8.

6. The vertical profile we assume

Where you do not supply an altitude and a block time, we assume both. Cruise level is taken from the sector length, following the flight levels the FAA used for comparable sectors. Climb and descent are modeled as rates rather than fixed durations, because climbing to FL430 takes longer than climbing to FL200. On any sector with more than two hours of cruise we assume a step climb, because airliners level off lower while heavy and step up as fuel burns.

Those assumptions were fitted against the FAA's published route table. They are assumptions about how airlines operate, not physics, and they are the largest single source of error in your figure. That is why the form offers you the chance to replace them.

7. How far off it can be

The range on your report says how closely this report reproduces the FAA's CARI-7 software. It does not say how closely CARI-7 reproduces an instrument carried in an aircraft, which is a separate question covered on the validation page.

It is ±25% for a flight where you gave us the cruise altitude and flight time, and ±30% where we estimated them from the route length. Each range held for at least 24 of the 27 route doses the FAA published in AC 120-61B, when the model was tested on flights it had not been fitted to. Measured that way the mean absolute deviation is 13.3% with the flight level and time supplied and 17% without.

These are coverage figures, not confidence intervals, and there is no sigma attached to them. The full route-by-route table is on the validation page, along with what the comparison cannot tell you.

8. Pregnancy

When the passenger profile is set to pregnant, the report compares the flights you listed, flown once through, against 1 mSv. That 1 mSv figure comes from ICRP and FAA Advisory Circular 120-61B, and it is an occupational protection limit for a radiation worker who has declared her pregnancy. It is not a clinical threshold at which harm begins.

The clinical reference is far higher. ACOG Committee Opinion 656 reports no increase in fetal anomalies below 50 mGy, roughly fifty times the protection limit. We compare against the stricter of the two because it is the conservative choice. We also compare the listed itinerary rather than an annual total, because a pregnancy is not a calendar year. This report is not medical advice, and cumulative exposure during pregnancy is a conversation to have with an obstetrician or midwife.

9. What we do not model

  • Solar particle events. These are rare, roughly five ground-level events in the current solar cycle, but during a major one a single high-latitude flight can carry a dose comparable to weeks of ordinary flying. We model the ordinary galactic background only. NOAA's Space Weather Prediction Center publishes live alerts.
  • The route you actually flew. We assume the great circle. Filed routes deviate for weather, traffic and oceanic tracks, typically moving route dose by a few percent.
  • Aircraft shielding differences. CARI-7 models a generic airframe.
  • Geomagnetic drift after 2010. Cutoffs are interpolated to your flight date across the six epoch grids CARI-7 distributes, the most recent of which is 2010. Dates after 2010 use the 2010 field, which is what CARI-7 itself does.
  • Your individual risk. Effective dose is a population quantity. What a given dose means for one person depends on age, sex, prior exposure and factors no dose model addresses.

10. Reproducing our numbers

Every report states the model version, the cutoff grid, the months of solar data used, and the uncertainty band, so it can be checked. CARI-7 is free and the FAA distributes it, with source, from its radiobiology page. Note that the FAA's own online calculator has been unavailable for some time; the downloadable software is the working route. If you reproduce a route and get a materially different answer for the same profile, please tell us. Corrections go on the corrections log.