How to Read a Forecast Cone: The Line Is Not the Forecast, and the Cone Is Not the Storm
Every time a typhoon approaches, the same picture goes around: a blob over the ocean, a red line running toward land, and a widening white shape around it.
Almost everyone reads that picture wrong, and the two most common mistakes are the ones that matter most if you are travelling.
The first is thinking the line is where the storm will go. The second is thinking the cone is how big the storm is.
Neither is true, and understanding why will tell you more than any headline.
What the cone actually is
The cone is a statement about forecast error, not about the storm.
Agencies look at how far off their track forecasts have been over the previous several years. At 24 hours the average miss might be a certain distance, at 72 hours a much larger one. They draw a circle of that size around each forecast point and connect the circles.
That is the cone. It is built from past mistakes, not from this storm's wind field.
The standard is roughly two out of three. The centre of the storm stays inside the cone about 66 to 70 percent of the time. Which means it goes outside it something like one time in three.
That is not a rare event. It is a normal outcome that the graphic does not visually communicate at all.
Why the cone gets wider
It widens with time because the forecast gets less certain with time, not because the storm gets bigger.
A cone drawn for tomorrow is narrow because 24-hour track forecasts are quite good now. A cone drawn for five days out is enormous because five-day forecasts still carry real uncertainty.
So a narrow cone near the start is not confidence about landfall. It is just the near term.
The line in the middle is the most misleading part
The centre line is the single most likely track. It is not a promise, and small shifts in it produce very different outcomes on the ground.
A storm centre passing 80km to one side of a city versus 80km to the other side is the difference between a windy afternoon and a serious event. On the graphic, that difference is a few pixels.
Some agencies have started removing or de-emphasising the centre line for exactly this reason. People anchor on it and stop reading anything else.
If you take one thing from this article: look at the cone, not the line, and then stop looking at the cone too.
The cone says nothing about how big the storm is
This is the mistake that gets travellers caught.
The cone shows where the centre might be. It says nothing about how far out from that centre the damaging winds and rain extend.
A compact storm might produce strong winds only 100km from its centre. A large one can spread tropical-storm-force winds 400km or more. Both get drawn with a cone of the same width, because the cone is about forecast error, not size.
So being outside the cone does not mean you are unaffected. It routinely means the opposite.
What to look at instead
Wind radii. Most agencies publish the radius at which winds reach tropical-storm force, usually plotted as a separate shaded area. That shaded area is the one that tells you whether your city gets weather.
The right side matters more. In the northern hemisphere the storm's forward motion adds to its rotation on the right-hand side of the track. That side gets stronger winds, higher surge and heavier rain. Being on the left of the track is meaningfully better than being on the right.
Rainfall forecasts, which are almost entirely separate from wind. Some of the worst typhoon disasters in Asia have come from slow-moving, unimpressive-looking storms that sat over mountains and dropped enormous rain totals. Intensity category tells you very little about flooding.
Timing. Arrival time of tropical-storm-force winds is published by several agencies and is far more useful for travel decisions than landfall time, because that is when airports stop operating and roads close, often many hours before the centre arrives.
The word landfall is doing too much work
Landfall means the centre crossing the coast. That is all.
Serious conditions begin long before it and often continue well after. In a large storm, the difficult period can start 12 hours before landfall.
Planning around a landfall hour is how people end up at an airport that stopped functioning the previous evening.
Spaghetti models are not what they look like
You will also see charts with dozens of coloured lines, each one a computer model run.
Two things about them.
They are not equally credible. Some of those models are consistently better than others, and the good ones are weighted heavily by forecasters. A chart that draws them all with the same thickness hides that.
Their spread is a rough measure of uncertainty, which is useful. When the lines cluster, confidence is higher. When they fan out across a thousand kilometres, nobody knows yet, whatever a headline says.
They are a tool for reading uncertainty, not for picking the line you like best.
Where this site has been wrong
Being honest about this is part of the point of writing here.
On this site we have described forecast spread as though a tightening cluster of models meant a settled outcome. That is half right. Clustering does raise confidence, but models can agree with each other and still be wrong together, particularly when the steering pattern itself is what is uncertain.
The safer phrasing, and the one we will use going forward, is that agreement lowers the range of likely outcomes without eliminating the chance of a shared error.
Where to check this yourself
Rather than reading about a storm second-hand, these are the sources the forecasts come from.
Japan Meteorological Agency: https://www.jma.go.jp/en/typh/
Korea Meteorological Administration: https://www.kma.go.kr/eng/weather/typoon/typhoon_05.jsp
Joint Typhoon Warning Center: https://www.metoc.navy.mil/jtwc/jtwc.html
Hong Kong Observatory: https://www.hko.gov.hk/en/index.html
China National Meteorological Center: http://www.nmc.cn/publish/typhoon/typhoon.html
Pick the one whose area you are actually in, and read that one. Comparing agencies against each other produces confusion rather than clarity, for reasons worth their own article.
A short checklist for travellers
Is my location inside the tropical-storm-force wind radius, at any point in the forecast? That is the question, not whether I am in the cone.
Am I on the right side of the track or the left?
When do those winds arrive, not when does the centre arrive?
What is the rainfall forecast, separately from the wind?
Has my airport or airline said anything yet, and have I turned on notifications?
None of those questions are answered by the picture everyone shares.
Photo credits
Photographs in this post come from Wikimedia Commons and are used under their respective Creative Commons licences.
The cone in actual numbers
Everything above is about how to read the cone qualitatively. The National Hurricane Center publishes the exact figures the cone is drawn from, and they are more alarming than the picture suggests. These are the 2026 radii, calculated from the 67th percentile of official forecast errors over the 2021 to 2025 seasons.
| Forecast hour | Atlantic radius (nm) | Atlantic radius (km) | Circle diameter (km) | Area vs the 24-hour circle |
|---|---|---|---|---|
| 12 | 25 | 46 | 93 | 0.4× |
| 24 | 39 | 72 | 144 | 1.0× |
| 36 | 49 | 91 | 181 | 1.6× |
| 48 | 62 | 115 | 230 | 2.5× |
| 60 | 77 | 143 | 285 | 3.9× |
| 72 | 95 | 176 | 352 | 5.9× |
| 96 | 134 | 248 | 496 | 11.8× |
| 120 | 200 | 370 | 741 | 26.3× |
The number that changes how you read the picture
From day one to day five, the radius grows 5.1 times. But a cone is an area, not a line, and area grows with the square of the radius. The five-day circle covers 26.3 times the area of the one-day circle. Measured from the twelve-hour circle it is 64 times the area.
That is why the cone looks like a narrow corridor near the storm and a continent-sized smear at the far end, and why the two ends should not be read with the same confidence at all. The day-five end of a cone is not a slightly less certain forecast. It is a different kind of statement.
In concrete terms, the day-five circle is 741 kilometres across. Put that over Southeast Asia and a single circle spans from Hanoi to well past Da Nang, or covers the whole Luzon landmass with room over the sea on both sides.
The second number, which almost everyone gets backwards
The circles are drawn at the 67th percentile of past errors. That is a deliberate choice and the NHC states it plainly: the size of each circle is set so that two-thirds of historical official forecast errors fall within it.
Read the complement. One time in three, the centre of the storm falls outside the circle.
People treat the edge of the cone as a worst case, as in "at least it cannot be worse than the edge". The edge is not a worst case. It is a line that gets crossed on roughly a third of occasions. And that is per circle; historical data indicate that the entire five-day path of the centre stays inside the whole cone only about 60 to 70 per cent of the time, which is another way of saying that something like one storm in three leaves its own cone at some point.
And the third, which is about what the cone does not contain
The cone describes the probable track of the centre. It says nothing about the size of the storm. A tropical cyclone's damaging winds and its rain shield extend well beyond the centre, so the area that experiences the storm is always larger than the cone, often by a lot.
Stack the three facts and the correct reading of a five-day cone becomes clear:
| What people read | What it actually says |
|---|---|
| The storm will be in here | The centre will probably be in here |
| The edge is the worst case | The centre lands outside about one time in three |
| Day five is a forecast | Day five covers 26 times the area of day one |
| Outside the cone is safe | The storm's effects extend beyond the cone regardless |
Why the Pacific numbers are different
The eastern and central North Pacific radii are consistently tighter than the Atlantic ones. At 120 hours the Pacific circle is 138 nautical miles, about 256 kilometres, which is 69 per cent of the Atlantic figure. Storms there tend to follow steadier tracks, so the same percentile of past errors produces a smaller circle.
That matters if you move between basins and carry your intuition with you. A cone that looks reassuringly narrow in one basin and a cone that looks alarming in another may be expressing exactly the same forecast confidence, because each is scaled to its own history.
The practical rule
| If your decision is | Use the cone at | Because the circle is |
|---|---|---|
| Whether to go to the airport today | 12 to 24 hours | 46 to 72 km radius, tight enough to mean something |
| Whether to move a flight | 48 to 72 hours | 115 to 176 km radius, useful but already 2.5 to 5.9 times the day-one area |
| Whether to cancel a holiday | Do not use day five alone | 370 km radius, 26 times the area, and exceeded one time in three |
The honest summary of a five-day cone is that it tells you a storm exists and roughly which way it is going. It does not tell you where it will be. The decision it should trigger is to start watching daily, not to cancel or to relax.
Sources for this section
- NOAA National Hurricane Center, Definition of the NHC Track Forecast Cone (2026 radii from the 2021–2025 error sample, 67th percentile)
- NOAA National Hurricane Center, Forecast Verification
- NOAA National Hurricane Center, 2026 Products and Services
Conversions to kilometres and the area ratios in the tables above are calculated from the published nautical-mile radii at 1 nm = 1.852 km, and the area ratio is the square of the radius ratio. Both are arithmetic on the NHC figures rather than separate NHC products.
Comments
Post a Comment