StormlinesStorm tracks, explained
Nerd mode: every forecast, readouts, filters and range & bearing for weather enthusiasts

Sections

View
Nerd mode: every forecast, readouts, filters and range & bearing for weather enthusiasts

Storm · Forecast to formEast Asia

A new low could form east of Beijing on Saturday evening

Beijing — Could pass near Beijing Sat evening; 1 in 10 forecasts bring it within 160 km.

Updated Tues., Sept. 29, 8:47 a.m. GMT+8 · 124 forecast tracks from NOAA, ECMWF and Environment Canada: medium

Beijing — Could pass near Beijing Sat evening; 1 in 10 forecasts bring it within 160 km.

Quick answers

As of the Mon 28 Sep 12Z–18Z model runs · updated Tues., Sept. 29, 8:47 a.m. GMT+8

Where will the Sea of Japan low form?
Forecasts that develop it have the centre forming about 270 kilometres east of Beijing, over north China on Saturday evening. The most likely path takes it about 400 kilometres north of Sapporo by Tuesday evening, Oct. 6.
When will the Sea of Japan low be closest to Beijing?
Most likely Saturday, Oct. 3, around 8 p.m. GMT+8. Across the forecasts the comes between Sat 2 p.m. and Sun 4 a.m. GMT+8. Most forecasts pass the centre 100–1,550 kilometres east of Beijing.
What is the chance it passes within 160 kilometres of Beijing?
14%: 1 in 10 forecasts bring the storm’s centre within 160 kilometres of Beijing. Elsewhere: Vladivostok 20%, Seoul 11%. Rain, wind and coastal flooding can reach well beyond 160 kilometres from the centre.
How much rain could Beijing get while the Sea of Japan low passes?
From Tue 8 a.m. (when the forecast’s rain totals start) to Sun 8 a.m. GMT+8, the end of the window when its centre is closest to Beijing, the middle of NOAA’s 31 GEFS forecasts brings 1.5 mm of rain and melted snow to Beijing; fewer than 1 in 10 bring 13 mm. By Sat 2 p.m., when that window opens, the middle forecast already has 0.8 mm. That is all the rain in that time, not only the Sea of Japan low’s, from one model (GEFS) on a coarse grid; local official forecasts can differ. Beijing: rain and wind this week →
How sure is the forecast?
Path certainty: medium. 124 of 185 show the storm, and they on where it goes.
Do the AI and physics models agree?
Largely, yes. Chance within 160 kilometres of Beijing: AI ensembles (AIFS ENS, AIGEFS) 6%, physics ensembles (GEFS, ECMWF ENS, GEPS) 19%. By model: GEFS 32%, ECMWF ENS 14%, AIFS ENS 6%, AIGEFS 6%, GEPS 14%. By Monday evening, Oct. 5 the AI ensembles’ middle track is about 140 kilometres east of the physics ensembles’. At its deepest, the AI ensembles take it to about 993 and the physics ensembles to about 995 mb.
Will the Sea of Japan low become a bomb cyclone?
7% of forecasts (1 in 10) deepen it fast enough to count as a (a pressure fall of roughly 24 in 24 hours, adjusted for latitude).

New to these maps? Read how to read a spaghetti map, or see the week’s rain and wind around the world on the globe.

What we know

Where it forms. The forecasts that show it have the centre taking shape about 270 kilometres east of Beijing, over north China on Saturday evening, at about 1,014 .

How sure. 124 of 185 show the storm, and they on where it goes.1

How strong. It is still strengthening: forecasts deepen it to about 998 by Tuesday afternoon, Oct. 6. One in 10 forecasts make it a , deepening about 24 millibars or more in a day.

What’s uncertain

Whether it forms. 124 of 185 develop it; the rest do not. Storms that exist only in the forecast can change a lot from run to run.

How close it comes. For Beijing, most forecasts put the between 100 and 1,550 kilometres away, most likely Saturday 8 p.m.

Later on. The forecasts fan out after Sunday. By Tuesday night, Oct. 6, the is about 2,750 kilometres across.2

When it arrives

When the centre passes closest to each place, earliest to latest across the forecasts. Places are ordered by the most likely time.

Window of Most likely timeNight
is the share of forecasts that bring the centre within 160 kilometres. Shaded columns are night; times are in your time zone.

Rain and wind this week in Beijing, Seoul, Vladivostok, Sapporo and Petropavlovsk-Kamchatsky.

DetailsStrength, sources and the fine print

How strong it gets

Central pressure over time. The line is the middle forecast; the band holds 8 in 10 of them.

(lower is stronger)

Times are shown in

Where the forecasts come from

ModelTracksStatus
NOAA
21 of 3118Z
arrived 2 h 16 min ago
Latest run
ECMWF
35 of 5112Z
arrived 4 h 52 min ago
Latest run
ECMWF
33 of 5118Z
arrived 1 h 6 min ago
Latest run
NOAA
20 of 3112Z
arrived 6 h 34 min ago
18Z arriving
Environment Canada
15 of 2112Z
arrived 8 h 22 min ago
Latest run

Also tracking

  1. Hurricane PoloCould pass near San Diego Wed morning; 2 in 10 forecasts bring it within 160 km
  2. Southern Ocean lowVery deep low in the Southern Hemisphere, near 936 mb now and filling; all 185 forecasts show it
  3. Amundsen Sea lowVery deep low bottoming out near 956 mb Tue night, in the Southern Hemisphere; all 185 forecasts show it
  4. Southern Ocean lowVery deep low bottoming out near 953 mb Thu morning, in the Southern Hemisphere; nearly all forecasts show it
  5. Southern Ocean low south of New ZealandVery deep low near 922 mb now, in the Southern Hemisphere; 9 in 10 forecasts show it
  6. See all 27 storms →

Notes

  1. Forecast tracks from the 18Z model run: GEFS (31), ECMWF ENS (51), AIFS ENS (51), AIGEFS (31), GEPS (21). A single run is one high-resolution forecast; ensemble forecasts are the same model run many times with small changes.
  2. The cone encloses where the centre is expected to be about two times in three at each hour, from the spread of the forecasts, widened by 20 percent; not yet calibrated against past storms. It shows uncertainty in the track, not the size of the storm.
  3. Shading shows the share of forecasts that bring the centre within 160 kilometres of a spot at any time in the next 7 days. Rain, wind and surge can reach much farther.
  4. Times are in your time zone. Arrival windows run from the earliest to the latest closest approach across forecasts that bring the centre within about 300 miles (500 km).
  5. NOAA/NCEP.
  6. Data Source: Environment and Climate Change Canada.
  7. This service is based on data and products of the European Centre for Medium-Range Weather Forecasts (ECMWF). Source: www.ecmwf.int. Licence: CC BY 4.0.
  8. Cyclone tracks computed by Stormlines from the model fields.
  9. Reading these maps: , or the illustrated guide to spaghetti models. The week’s rain, wind and temperature everywhere: the Stormlines globe.
  10. How often these forecasts turn out right: the season’s forecast check, AI vs physics.
  11. Stormlines storm ID: L-2026-0262.

Stormlines draws every storm the world’s ensemble forecasts can see: tracks from NOAA (GEFS and the AI ensemble AIGEFS), ECMWF (ENS and the AI ensemble AIFS) and Environment Canada (GEPS), updated as each model run lands. Coastlines and place names from Natural Earth; sea areas from IHO Sea Areas, version 3 (Flanders Marine Institute, marineregions.org, CC BY 4.0; modified: simplified, some islands removed, names edited).

Not an official forecast. For warnings and decisions, follow your national weather service.

Loading 124 tracks…

Stormlines Tracks: NOAA, ECMWF, Environment Canada

Full forecast
Through Wed 2 a.m. GMT+8