Based on 14 years of satellite observations of this place.
Live forecast from the latest update. See the full Bidwell Park page.
How we forecast
Every forest has a rhythm. We learn when each place usually turns, check that against the people and cameras watching on the ground, and tell you how sure we are.
behind a typical forecast
each forecast on its own
places where foliage spotters check us
research cameras at our destinations
Satellites have photographed every American forest, every few days, since the turn of the century. That is a long memory of when each hillside greens up, turns and drops its leaves.
A maple ridge in Vermont and an aspen grove in Colorado keep very different calendars. We work out, place by place, when color usually peaks. Then we nudge that date for this year: a warm stretch before peak delays color, a cool one brings it forward.
Space can’t see what you see standing under the trees. So we grade our dates against state foliage spotters, research forest cameras and volunteer observers, at places we never tuned on.
Nature doesn’t keep appointments, so we never give you just a date. Here is a real one, in its peak window now, as it stood at the latest update.
Peak date: our best single guess at the most colorful day.
Window: the peak date ±5 days, about a weekend and a half. Where state spotters report, peak has landed inside it about two years in three.
Evidence: how much real observation stands behind this exact place.
Based on 14 years of satellite observations of this place.
Live forecast from the latest update. See the full Bidwell Park page.
The map’s color-report view counts days to and from each place’s peak date. Pick a date and every forest shows where it should be in its season.
| Stage | When | What you’ll see |
|---|---|---|
| Not started | 24+ days before peak | Under 5% of leaves turned |
| Early color | 12–23 days before peak | 5–24% of leaves turned |
| Moderate color | 5–11 days before peak | 25–49% of leaves turned |
| High color | 1–4 days before peak | 50–74% of leaves turned |
| Peak color | Peak day to 5 days after | 75% or more turned |
| Fading | 6–18 days after peak | Leaves dropping |
| Past peak | More than 18 days after peak | Mostly bare or brown |
Peak timing: the other map view shows when each forest is expected to peak, using early, mid and late month bands rather than its color stage. Each band spans about ten days.
Each destination page carries one of three labels. These shares are counted across the 605 destinations on the site today.
Years of direct observations of this exact place. Our most reliable forecasts.
99% 599 destinations
This spot is hard to see from above, so we lean on the closest forests we can. A real forecast, with less behind it.
1% 5 destinations
We won’t guess. Usually too little of the forest changes color to track, so we point you to nearby places where we do have a date.
0% 1 destination
Grading a forecast on the data used to build it flatters it. So we compare our dates with what people and cameras actually recorded, season after season.
Two of our dates in three landed within 4 days of the peak state spotters reported. 95% landed within two weeks.
63 places across New York, Pennsylvania, Wisconsin and Missouri
landed within two weeks of the peak the camera recorded.
38 cameras at our destinations, 3+ years each
| Where | Checked against | Two in three within | Places |
|---|---|---|---|
| New York | I LOVE NY foliage reports | 4 days | 36 |
| Pennsylvania | PA DCNR foliage reports | 3 days | 15 |
| Wisconsin | Travel Wisconsin foliage reports | 4 days | 8 |
| Missouri | Missouri MDC foliage reports | 8 days | 4 |
| Mid-Atlantic and Southeast | Research forest cameras | 8 days | 14 |
| Northeast | Research forest cameras | 14 days | 17 |
And the window holds. Where state spotters report, peak has landed inside our ±5-day window about two years in three. The West and Plains don’t yet have enough ground reports near our destinations to grade fairly; we’ll add them as reports come in.
Our dates tell you when a place usually peaks, adjusted for how this year is running. That is what you need to book a cabin or plan a drive.
A storm or an early frost can change things fast. A few days out, a park webcam or a ranger station will know things no forecast can. Ours included.
We adjust for temperature, but a drought, an early hard frost or a windstorm can still move or cut short peak.
Mountaintops turn before valleys. For a large place, the date is the middle of its season: go higher for earlier color, lower for later.
Where pines outnumber maples and aspens there is less color to track. Destination pages say how much of the forest actually changes color.
Every forecast refreshes daily, and each page carries one date: the day it was last updated. The inputs behind it arrive with a short delay. Satellite readings reach us a few days after the satellite passes overhead, and clouds can hide a forest for days at a time. Weather records post a day or two behind. Each update uses the newest clear satellite reading and the latest weather available, so a page’s color reading can be a few days older than its update date.
When a clear reading is delayed, we keep showing the most recent one we have and label it as the latest reading rather than as what the trees look like right now. We watch these delays on our side so every update uses the newest data as soon as it arrives.
The map groups dates into color bands and blends each spot with the forest around it, so it reads at a glance. The destination page uses that place’s own measurements. When they differ, trust the destination page.
Everyone defines “peak” a little differently and uses different evidence. We publish how often ours matches what spotters report, so you can judge for yourself.
No. No destination, park or business pays to appear, to rank higher or to change a date.
Yes, with a link back. Every destination page shows when its forecast was last updated, so a citation can point at the exact date we published.
Yes. Dates move with how warm or cool this year has been before peak, and in the western mountains with how warm spring was. Rain, drought and frost are not in the model yet.
High-country dates in the Rockies, Great Basin, Sierra and Cascades now move with how warm or cool spring was.
East of the Rockies, each date now moves with how warm or cool the weeks before peak have been.
Every window is the peak date ±5 days, and we say how often it has held.
Most destinations moved about five days earlier, closer to what observers report.