Methodology
How the BiteForecast index works
BiteForecast is a weather-proxy model for Scottish midge nuisance. It is built to answer a practical planning question: when are conditions likely to feel manageable, annoying, or worth avoiding?
What the model uses
Each forecast combines temperature, wind speed, relative humidity, time of day, UV/sunlight proxy, and season.
The time-of-day curve is explicit: dusk and dawn get crepuscular boosts; bright midday is suppressed; the overnight dead zone is suppressed so still, humid 1am weather does not automatically beat dusk.
Wind is treated as a strong suppressor. Sustained stronger wind can collapse the score even when humidity and temperature look favourable.
What the model does not use
BiteForecast does not use trap counts, live biting reports, on-the-ground observations, or site-specific sensor data.
That means the index can miss small sheltered pockets: campsite edges, woodland hollows, still burns, and lochside corners can feel worse than a broad weather forecast suggests.
The established Scottish midge forecast has historically used trap-count context as well as weather. BiteForecast is deliberately simpler and should be read as a weather-based planning estimate.
Update cadence and forecast horizon
Forecast pages update every 3 hours. The forecast outlook is 5-day.
Weather data comes from Open-Meteo. If live weather data is unavailable, BiteForecast falls back to a seasonal estimate rather than pretending missing data is current.
Current constants awaiting formal sign-off
- Temperature below 7°C returns zero activity in the model.
- Dusk window: one hour before to ninety minutes after sunset.
- Dawn window: one hour either side of sunrise.
- Overnight dead zone: strongly suppressed outside the dawn/dusk windows.
- Wind: 8–12 mph progressively suppresses; sustained 12+ mph strongly suppresses activity.
These values are conservative operational constants derived from Scottish midge literature, Smidge/Scottish forecast context, and site plausibility tests. They should be reviewed after live-season observations.