Activity above Earth
Space-weather data describes whether aurora activity is possible in principle. The Kp index is one signal among several, not a local visibility forecast.
Method and system architecture
From input data to a driving decision.
Aurora North processes space weather, weather forecasts and astronomical data for each location and arrival time. Rule-based evaluation combines viewing conditions with data quality, the observation window and reachability.
Description updated · 2 October 2026Input data
Four separate assessment layers.
Each layer describes a different part of the observation conditions. Data is assessed for the relevant location and time; the Kp index alone does not determine local viewing chance.
A clear view from the ground
Cloud, precipitation and visibility are assessed for the relevant time. Where reliable satellite observations are available, they can support the current cloud picture — they do not turn a future condition into a measurement.
Darkness and place
The positions of the Sun and Moon, local brightness, an open northern horizon and the suitability of a viewing spot affect what may actually be visible from the ground.
Time and reachability
A destination only helps if the good window fits the drive. The app therefore compares the expected conditions at arrival and gives a latest useful departure time.
Rules and models
Decision logic and cloud nowcasting.
The driving decision uses defined rules. Reliability, viewing score, departure status and the available observation window determine the result. Insufficient evidence produces a “Wait for data” status. A language model is not involved in this decision.
The experimental cloud nowcast estimates short-term cloud development from consecutive satellite images. Motion estimation provides the baseline; a temporal U-Net with ConvGRU learns corrections and uncertainty. The model is evaluated separately and does not currently affect the driving decision.
- Rule-based decision
- Cloud nowcast disabled by default
- Separate evaluation with no effect on the driving decision
- Missing observations remain unknown
Processing
From data timestamps to an action.
- 01
Estimate potential
Space weather and astronomical darkness describe the underlying observation potential.
- 02
Assess visibility
Cloud, precipitation, visibility and local conditions limit the potential from the ground.
- 03
Compare arrival times
The starting location and reachable destinations are assessed for their respective arrival times and usable observation windows.
- 04
Derive the decision
Reliability, viewing score and departure status are combined into an action with supporting reasons.
Output values
Viewing score and data quality.
The score describes expected conditions. Reliability describes the supporting evidence. The two values are reported separately.
Viewing chance · 0–100
The 0–100 value compares expected viewing conditions within the app. It is not a calibrated probability of seeing the aurora.
Reliability
Data age, completeness and agreement between sources determine reliability. Old, missing or conflicting input data makes the assessment more cautious.
Data quality
Missing values remain unknown.
A gap in the data is kept separate from a favourable measurement or forecast. Missing cloud data is not interpreted as clear sky.
- Older data lowers reliability.
- Conflicting weather models make the guidance more cautious.
- Without a sound basis, the app may ask you to wait for data.
- Without a suitable reachable place, it does not invent a driving destination.
Resolution and forecast horizon
Technical limits of the data.
- Local fog and small cloud fields may fall below a weather model’s spatial resolution.
- Space weather can change significantly within a forecast interval.
- Satellite coverage and usable observations depend on region, time and product.
- Retrieval time, observation time and forecast time are different timestamps. The displayed data age is therefore part of the assessment.
Public data sources
Where the data comes from.
Weather and environmental data is retrieved server-side. Available sources vary by region and data type. Maps, routing and traffic information complement the location- and time-specific assessment.
NOAA Space Weather Prediction CenterSpace weather and aurora activity↗Open-MeteoWeather forecasts↗SMHI Open DataWeather forecasts for Swedish locations↗MET NorwayWeather forecasts, especially for the Nordic region↗EUMETSATSatellite cloud observations where usable↗MapboxMaps and routes started by the user↗TrafikverketTraffic and road information in Sweden↗