The weather picture: data (WS-214, the storm on Wednesday 11 November 2026)
Kept by incident.meteorologist@web-summit.dinis for the incident team, the dashboard and, as material, Vicky's weather desk. JSON first: read these files, not pages. Every record carries source (the URL read) and read_utc (when). Facts, press and analysis are labelled (kind). Lisbon is on UTC in November (WET): IPMA's local times are UTC.
Every file carries next_read_utc, the time of the next scheduled refresh: trust it until then. Refreshed hourly (at :22) with python3 -I incidents/2026-11-10-storm-charles/data/weather/tools/refresh.py (standard library only). Each run keeps what earlier runs saw and appends what changed to changes.json. A refused host is recorded under errors in that run, and as an ask to Dinis in agents/meteorologist/status.json.
| File | What | Source |
|---|---|---|
warnings.json | IPMA's non-green warnings for the Lisbon district (LSB) now; red and orange warnings elsewhere; history: how Lisbon's warnings changed, seeded with what happened before this desk began (each with its own source and reader), then every change each refresh sees | IPMA warnings open data |
communiques.json | IPMA's communiqués (CHARLES no. 2, BRUNA no. 3), full Portuguese text, kept across reads; no. 1 (no longer in the feed) from the newsroom's record | IPMA communiqués RSS |
stations.json | Hourly observations at eight stations in and next to Lisbon: mean wind, direction, rain, temperature, pressure; the latest, the maximum, kept across reads. No gusts in this feed | IPMA observations |
airport.json | Lisbon airport (LPPT): the current TAF decoded (kt and km/h), every TAF issue seen, every METAR seen with its gust, the maximum gust | NOAA AWC API: TAF, METAR (IPMA issues both) |
forecast.json | IPMA's daily forecast for Lisbon and its hourly model forecast (mean wind, rain, sea); each new model run recorded in hourly_runs | daily, hourly |
gusts-vs-forecast.json | The latest observed wind (airport gust, station means) beside what was forecast for that hour (TAF group, the red warning, the model's peak) | the above |
cap.json | IPMA's warnings for Lisboa as CAP messages via MeteoAlarm, with the time IPMA sent each and the message it replaces (the open data has no issue time and lagged CAP by 2 h+ on 10 Nov) | MeteoAlarm Portugal feed |
forecast.json → second_models | Other models for Lisbon, Wed and Thu morning, one block per source, never merged with IPMA: met_no (MET Norway, NLOD/CC BY 4.0), open_meteo (CC BY 4.0; rate-limited at first try) | met.no, Open-Meteo |
changes.json | First: thursday (Lisbon warnings touching Thu 12 Nov; IPMA model, daily and TAF for Thursday morning); then gusts_above_warned, every airport gust from Wed 05:00 above the gust warned for that hour (the IPMA wind warning in force, else the TAF), each with its METAR; then what each refresh found changed | this tool |
analysis.md, analysis.json | ANALYSIS: when the worst arrives and ends on Wednesday; how far ahead each warning could be known | reasoning from the files |
Notes:
- IPMA's warnings feed carries no issue time. An issue time here comes from a read (two reads bracket it) or from press (labelled
press). - The TAF is a point forecast for the airport, for pilots; quoted beside IPMA's warning, never instead of it. Gusts at the airport and the warning's gusts (120, up to 140 km/h, for the district's most exposed places) are not the same measure.
- The maritime authority's figures (AMN: gusts to 168 km/h at sea) are not IPMA's land warning; they are not mixed in here.
- Public sources only; no contact details.