Eclipse Weather: How to Read Cloud Forecasts Like a Chaser
Astronomical predictions are precise but still carry stated model limits; weather remains uncertain. A strong plan uses reproducible climatology cautiously, current official forecasts close to the event, and a safe decision deadline.
The short version: For the 22 July 2028 eclipse, use historical cloud statistics only to choose a region months ahead - inland New South Wales generally beats the coast. From 48 hours out, switch entirely to real forecasts and satellite imagery, commit to a base at 24 hours, and make a final move-or-stay call at dawn on eclipse day using the live satellite loop.
Climatology is for planning, forecasts are for deciding
Climatology can describe a named station and record period; it cannot supply a precise clear-sky probability for a city on 22 July 2028. The weather evidence page publishes the original BOM metric labels, values, record spans, source pages, and downloadable transcription. Missing or two-year cloud records remain excluded rather than inferred from rainfall.
Use long-run evidence only when comparing flexible bases. Use current official forecasts, warnings, satellite imagery, safe access, and the actual horizon to choose the observing site.
Evidence: Bureau of Meteorology climate data.
What does July weather actually look like along the path?
Inland Australia in July is dominated by dry, stable winter air: cold nights, frosty mornings that may need an hour to burn off, and frequently cloudless days. This is why the outback and western plains hold the best statistics on the path.
The NSW coast, including Sydney, is a battleground between that dry interior and the Tasman Sea. Cold fronts sweep through every few days trailing cloud bands, and onshore winds can pin low cloud against the coast while it is blue 100 kilometres inland. Coastal viewers should watch the wind direction as much as the cloud forecast: westerly flow off the land is your friend.
The South Island of Aotearoa New Zealand is the trickiest leg: fast-moving fronts, low coastal stratus around Dunedin, and valley cloud that can fill alpine basins like Queenstown's while ridgetops sit in sunshine — patterns documented in NIWA's climate data. NZ plans need more built-in mobility than Australian ones — the sunset eclipse guide covers the terrain side of the same problem.
Evidence: Bureau of Meteorology climate data, MetService forecasts and warnings.
Which forecast tools should you actually use?
- Official forecasts: the Bureau of Meteorology (Australia) and MetService (Aotearoa New Zealand) for warnings, synoptic outlooks, and road-weather alerts.
- Satellite loops: the Himawari satellite covers the whole path, and BOM publishes its imagery as animated loops. On eclipse morning this is your single most valuable tool — not a model, but the actual sky, moving.
- Model viewers: apps and sites that display model cloud fields (Windy and similar) let you compare the European (ECMWF) and American (GFS) models layer by layer. When both agree on clear sky for your site, confidence is high; when they disagree, keep your options open.
- Webcams: regional airport and resort cameras give ground truth for a specific horizon — invaluable for checking whether Queenstown valley cloud has formed.
Check trends across successive model runs rather than reacting to any single run. A forecast that has said clear for three consecutive runs is worth far more than one optimistic frame.
Evidence: Bureau of Meteorology climate data, MetService forecasts and warnings.
How to read a cloud forecast like a chaser
Total cloud percentage is a blunt number. Split it into layers, because they are not equally fatal:
- Low cloud (stratus, fog): the eclipse killer. Opaque, slow to clear, and common on winter mornings inland and along coasts. If low cloud is forecast at eclipse time, move.
- Mid cloud: usually broken; a gamble. Gaps can be generous, but you cannot steer them.
- High cloud (cirrus): the most survivable. The corona shines through thin cirrus, slightly softened. Annoying for photographers, not fatal for viewers.
Two more inland-specific reads: morning fog and frost-haze usually burn off before the partial phase begins just after 12:30 p.m. — check the satellite loop for shrinkage rather than panicking at dawn grey. And afternoon cumulus, which bubbles up after sunny winter mornings, tends to be scattered rather than overcast — but note that eclipse cooling often dissolves small cumulus as totality approaches, one of the shadow's stranger gifts.
Evidence: Bureau of Meteorology climate data, MetService forecasts and warnings.
The 48-hour decision framework
Write this schedule down before eclipse week and follow it mechanically:
- T–48 hours: compare model forecasts for your base and every alternative region you can reach — for NSW that might be coast, mountains, and plains (see the road-trip guide for the corridor). Pick the winning region.
- T–24 hours: commit. Confirm accommodation or park-up plans in the chosen region and select your named site from its city guide, plus one backup in a different microclimate (inland vs coast, basin vs ridge).
- Dawn, eclipse day: open the satellite loop. Actual cloud, actual motion. If your sky is compromised and clearer air sits within about two hours' drive, leave now — before breakfast, before the crowd reaches the same conclusion.
- T–90 minutes: stop optimising. Be parked, settled, and set up an hour before first contact. A good-enough sky you are standing under beats a perfect sky you are still driving toward.
Evidence: Bureau of Meteorology climate data, MetService forecasts and warnings.
What if cloud wins anyway?
Cloud may prevent a view of the corona even when the observing plan is sound. Stay within site rules, keep solar viewers available whenever the bright Sun may appear, and do not begin an unsafe last-minute drive for a speculative gap.
Totality can still produce a rapid reduction in daylight under cloud, but local lighting, animal behaviour, temperature, and horizon colour vary and are not promised. Record the conditions you actually observe, then review the NASA 21st-century eclipse catalogue for future paths.
Evidence: NASA 21st-century eclipse catalogue.
Common Questions
What are the weather odds for the 2028 eclipse?
Do not use the withdrawn city cloud percentages. The site's reproducible comparison reports named BOM station metrics with their record lengths and limitations; it does not claim an eclipse-day probability or a controlled coast-versus-inland ranking.
When will forecasts for eclipse day be reliable?
Broad pattern guidance emerges about five to seven days out, but trustworthy cloud detail arrives in the last 48 hours. Make your region call at 48 hours, your site call at 24 hours, and your final move at dawn using live satellite imagery.
Can you see a total eclipse through cloud?
Thin high cloud may still permit a diminished view, while opaque low cloud or overcast can block the Sun completely. Brightness, temperature, and horizon effects vary by local conditions, so none is promised.
Which is the best weather app for eclipse chasing?
Use a combination: official Bureau of Meteorology or MetService forecasts, a model viewer such as Windy to compare ECMWF and GFS cloud layers, and - most importantly on the morning - live Himawari satellite loops showing actual cloud motion.