Lightning Strike Probability Calculator

Introduction: why lightning risk is really a collection-area problem

Lightning does not pick targets at random from the whole sky. A downward leader groping toward the ground commits to whatever object first offers a competitive upward connection, and the taller and more isolated that object is, the larger the patch of ground from which it can capture flashes. Lightning protection engineering turns that physical picture into a single number called the equivalent collection area: the area of flat ground that would receive the same number of flashes as the object actually intercepts. Once you have that area, the rest of the arithmetic is straightforward, because flashes to ground are counted per square kilometre per year in almost every published climatology.

This calculator implements the collection-area method set out in IEC 62305-2, the international risk-management standard for lightning protection, and mirrored in Annex L of NFPA 780 in the United States. You supply the ground flash density for your region, the footprint and height of the target, a location factor that accounts for shielding by the surroundings, how much of the time the target is actually there, and the length of the assessment period. The calculator returns the collection area, the expected number of direct strikes over that period, and the probability that at least one of them occurs.

The reason engineers bother with the collection area rather than the footprint is that a structure attracts lightning from far outside its own outline. The standard models that attraction as a border of width equal to three times the height, drawn around the plan of the structure. A one-storey house barely changes the ground it stands on. A 45 metre mast, by contrast, collects flashes from a disc more than 270 metres across, which is why radio towers, wind turbines, church spires and floodlight pylons are struck year after year while the buildings beside them are not.

The final step converts an expected count into a probability. Flashes arrive independently and at a low rate, which makes the Poisson distribution the natural model. If the expected number of strikes in the period is $N$, the chance of a strike-free period is $e^{-N}$ and the chance of at least one strike is the complement of that. For small values the two numbers are almost identical, so an expected count of 0.02 strikes per year and a 2 per cent annual probability mean effectively the same thing. For tall towers the distinction matters a great deal, because an expected count of three strikes per year still leaves a five per cent chance of a completely quiet year.

How to use the lightning strike probability calculator

Start with the ground flash density $N_G$ for the site, in flashes per square kilometre per year. National standards bodies, lightning detection network operators and meteorological services publish these maps; a mid-latitude continental figure of four to eight is typical, while the most active tropical land areas exceed twenty. Enter the plan dimensions of the target next. For a rectangular building use the actual length and width in metres. For a slender mast, a chimney or a wind turbine tower, use the base dimensions; the height term will dominate anyway.

Height is the input that moves the answer most, so measure it to the highest metallic point rather than to the eaves. A lightning rod, an antenna, an aircraft warning light or a lift overrun all count. Then choose the location factor that describes the surroundings: a target ringed by taller trees or buildings collects only a quarter of the flashes it would collect in the open, while the same target on an exposed knoll collects twice as many.

The exposure hours field lets the same model describe people and temporary equipment as well as permanent structures. A building sits in the storm field for all 24 hours of every day, so leave the field at 24. A surveyor, a groundskeeper or a golfer is only exposed for part of the day, and entering the average hours outdoors scales the expected count in direct proportion. Finally, set the assessment period. One year gives the familiar annual probability; a 25 or 50 year period answers the question a designer actually cares about, which is whether the structure will be hit at some point during its service life.

Read the result panel from the top down. The collection area tells you how large a target you have really built. The expected strike count is the quantity that scales linearly with every input, so it is the number to compare between design options. The probability and the one-in-N odds are the numbers to quote to a client or a safety committee, because they answer the plain-language question directly.

Formula for lightning strike probability from flash density and collection area

The equivalent collection area of an isolated rectangular structure, following IEC 62305-2, adds a border of width $3H$ around the footprint. That border consists of four rectangles along the sides and four quarter-circles at the corners, which combine into a single full circle:

Formula: A_D = L ⁢ W + 6 H ⁡(L + W) + 9 π H^2

$$A_D=LW+6H(L+W)+9\pi H^2$$

Here $L$ and $W$ are the plan length and width in metres and $H$ is the height in metres, so $A_D$ comes out in square metres. The expected number of direct strikes per year then follows by multiplying by the ground flash density and the location factor, with a factor of ${10}^{-6}$ to reconcile square metres with the square kilometres used by flash-density maps:

Formula: N_D = N_G ⁢ A_D ⁢ C_D ⁢ 10^−6

$$N_D=N_G{A}_D{C}_D{10}^{-6}$$

This page adds two everyday extensions to that standard expression. The exposure fraction $f=\frac{h}{24}$ scales the count when the target is present for only part of each day, and the assessment period $y$ in years scales it again, so the expected count over the whole period is $N=N_{D}fy$. The probability of at least one strike is the Poisson complement:

Formula: P = 1 − e^−N

$$P=1-e^{-N}$$

The odds shown beside the percentage are simply $1/P$, rounded for readability. The location factor $C_D$ takes the four values tabulated in the standard: $0.25$ when the target is surrounded by taller objects or trees, $0.5$ when the neighbours are the same height or lower, $1$ for an isolated target, and $2$ for an isolated target on a hilltop or knoll.

Worked example: a 45 metre mast on open farmland

Take a guyed communications mast 45 metres tall on a 10 metre by 10 metre fenced compound, standing alone in farmland where the ground flash density is 6 flashes per square kilometre per year. The footprint contributes $10\times 10=100$ square metres. The side border contributes $6\times 45\times (10+10)=5400$ square metres. The corner circle contributes $9\pi \times {45}^2\approx 57256$ square metres, giving a collection area of about 62,756 square metres, or 0.0628 square kilometres.

With an isolated location factor of 1, the expected number of direct strikes is $6\times 62756\times {10}^{-6}\approx 0.377$ strikes per year. The annual probability of at least one strike is therefore $1-e^{-0.377}\approx 0.314$, that is 31.4 per cent, or roughly one year in three. Over a 20 year service life the expected count rises to 7.53 strikes and the probability of at least one strike reaches 99.95 per cent, which is why nobody designs a mast of that height without a down-conductor system.

Now compare a person. A standing adult is about 1.8 metres tall on a footprint of roughly 0.4 metres square, giving a collection area near 100 square metres. Someone spending three hours a day in the open in the same 6 flash climate, with no shielding at all, accumulates an expected count of about 7.5 in 100,000 per year, or roughly one chance in 13,000. Applying the sheltered location factor of 0.25, which is far more realistic for someone who is usually near trees, vehicles or buildings, moves that to about one chance in 53,000.

Comparison table: how the collection area grows with height

The table below fixes the footprint at 10 metres by 10 metres, the ground flash density at 6 flashes per square kilometre per year and the location factor at 1, and varies only the height. It shows how quickly the quadratic term takes over.

Notice that going from 5 metres to 150 metres, a factor of 30 in height, multiplies the collection area by roughly 465 and the expected strike count with it. That is the single most useful intuition the model offers: lightning exposure is governed by the square of the height of the tallest metallic point, and almost nothing else on a typical site competes with it.

Interpreting the result: probability bands for planning

An annual probability below one per cent describes an ordinary low-rise building on a sheltered site. Direct strikes are possible but rare, and the dominant electrical risk is usually a surge conducted in from an incoming service rather than a strike to the roof. Between one and ten per cent you are in the territory of exposed farm buildings, small towers and rural sites, where surge protection on incoming lines and bonding of metallic services start to earn their keep.

Between ten and fifty per cent a direct strike is an expected event within a few years, and a full lightning protection system with air terminations, down conductors and an earth termination network is normally justified. Above fifty per cent the structure will be struck repeatedly, and the design question shifts from whether to protect it to how to route the current safely and how to protect the electronics inside. IEC 62305-2 makes that judgement formally by comparing computed risk against tolerable risk, and this calculator supplies the first ingredient of that comparison rather than replacing it.

Strike Odds: the storm-cell prediction game on this page

The game below the calculator drops you into the forecaster seat. A storm cell drifts across a landscape toward a single structure, and the panel tells you the structure height, its footprint, the location factor for the site, and the flash density the cell will deposit during its pass. Before the cell arrives you have to commit to a probability that the structure takes at least one direct strike, using the same collection-area arithmetic the calculator performs.

When the cell arrives the outcome is drawn from the true probability, so a lucky hit earns you nothing. Points come from how close your stated probability was to the true value, and the running Brier score tracks how well calibrated you are across the whole session. Coach mode reveals the intermediate collection area and expected strike count while you are still deciding, at half points, which makes the first few storms a guided walk through the formula rather than a guessing game.

Limitations and assumptions behind this lightning strike model

The collection-area method is a screening tool, not a physical simulation. It assumes flashes fall on flat, uniform terrain with a uniform density, that the 3H attraction radius applies regardless of the current in the leader, and that the target is electrically simple. Real attachment depends on the peak current of the descending leader, so a low-current flash may be captured only very close to the structure while a high-current flash can be drawn in from further away. Averaged over a season these effects largely cancel, which is why the standard tolerates the simplification, but the answer for any individual storm can be very different.

Nearby structures are handled only through the coarse four-value location factor. A single tall tower 50 metres away can shield a building far more effectively than a factor of 0.5 suggests, and a building on the edge of a cliff can be more exposed than a factor of 2 implies. Complex roof shapes, spires and rooftop plant are not modelled at all; the standard suggests treating them separately or using the rolling-sphere method, and this calculator simply asks for the height of the highest point.

The Poisson assumption treats flashes as independent. In practice a single thunderstorm often delivers several strikes to the same tall object within a minute, so the number of separate storm events per year is smaller than the number of strikes. This does not change the expected count, but it does mean that the probability of at least one strike computed here is a slight overestimate for very tall structures whose strikes are strongly clustered.

The exposure-hours extension assumes storms are uniformly distributed through the day and the year, which they are not. Convective lightning peaks in the afternoon and in the warm season, so someone who works outdoors only in the morning is less exposed than the raw hours suggest, while an afternoon sports league is more exposed. Finally, the output describes direct strikes only. It says nothing about side flashes, step potential, conducted surges or the injuries that occur several tens of metres from the attachment point, all of which account for a large share of real lightning casualties.

Frequently asked questions about lightning strike probability

Sources checked: the equivalent collection area A_D, the 3H attraction border, the location factor values of 0.25, 0.5, 1 and 2 and the expression N_D = N_G × A_D × C_D × 10⁻⁶ are taken from the International Electrotechnical Commission, IEC 62305-2:2010, Protection against lightning – Part 2: Risk management, Annex A. The same collection-area expression appears in the risk assessment annex of the National Fire Protection Association, NFPA 780, Standard for the Installation of Lightning Protection Systems. The casualty odds quoted above (1 in 1,222,000 in a given year and 1 in 15,300 over an 80 year lifetime, from 2009–2018 United States averages) come from NOAA National Weather Service, How Dangerous is Lightning?, and general safety guidance from the NWS Lightning Safety program.