Loading weather…
Next 2 hours
V4.3.1 blends radar extrapolation with HRRR, NBM, other available short-range guidance, the local learned weights, and the general forecast. Radar dominates the first few minutes and its influence fades with lead time. This is an estimate, not an official warning product.
Wind details
Hour by hour
Daily forecast
Why did the forecast change?
Radar timeline
Model comparison
Our Forecast
Combining available models with neutral weights until verified local performance is strong enough to adjust them.
What the models disagree about
Who is actually getting it right?
V2.12 stores forecast snapshots in a persistent SQLite database, verifies them against nearby NWS observations in the U.S. and worldwide METAR observations elsewhere, and automatically reconstructs missing model runs and observations from the previous 72 hours whenever the PC comes back online.
How scoring works
Mean absolute error plus signed bias. V2 can correct a model that repeatedly runs warm or cold.
0–3h, 4–12h, 13–24h, and 25–72h are scored separately so short-range skill does not automatically control longer forecasts.
Wind MAE, rain hit rate, and Brier probability score are used when the official station observation contains enough data.
20 verified samples are required before a group can influence weights. V4.3.1 blends 7-day, 30-day, 90-day, lifetime, current season/daypart/rain regime, radar rain skill, and a radar-model precision nowcast for the next 2 hours.
Who’s getting what right?
This uses ForecastForge’s verified local history. Lower error is better for temperature and wind; higher radar-verified hit rate is better for rain detection; temperature bias closer to zero is better.
Where each model is strongest
Weather that tells you what to do
Verified model history + radar + NWS/NHC + NOAA tides + marine forecasts + practical work and boating scoring.
Lightning tracker
Rainfall tracker
What happens next
Hurricane & tropical center
Mowing & outdoor work planner
Boating & jet ski conditions
What yesterday's forecasts actually got right
Alerts that matter
Tracked-location dashboard
What ForecastForge trusts right now
Ask, compare, remember context, then make the weather useful
Ask normally. Follow-ups keep the context.
Previous weather decisions
Start, heavy rain, stop, dry again
Can the grass actually be mowed?
Radar around the exact location
Show the evidence behind the blend
Where the forecast starts getting shaky
Models behaving badly
What ForecastForge saw as the storm evolved
Which lawns to mow first
Hyperlocal station bridge
Make the blend prove itself
Buoys, official marine guidance, tides, currents and crossings
Nearby buoy dashboard
Official forecast near the selected point
Tides and currents
Route conditions and Gulf Stream flags
Live cameras where available
Nearby webcam discovery and players are provided by Windy.com. A camera marked LIVE uses the provider's live player. Other cameras may show a recent image or timelapse instead.
Forecast preferences
Locations ForecastForge learns continuously
Every tracked location gets its own hourly model history, observations, scoring, bias correction, and 72-hour recovery after the PC has been offline.
Protected server controls
Public visitors are read-only. Entering the 6-digit PIN creates a private 7-day browser session backed by a random server secret; the PIN is not sent with every owner action. Incorrect PIN attempts and expensive API requests are rate-limited. Your PIN remains on the PC in data\owner-pin.txt.
What powers this build
Never use this app as a replacement for official emergency instructions. Severe-weather decisions should follow NWS/NHC/local authorities.