Our research is editorially independent but we
may earn a commission for purchases made through our links.

It’s the time of year marked by crisp air and sweaters, pumpkin spice, and cinnamon soirées. We’re talking sunkissed September serenades and October Opuses full of cider-sipping sagas and bonfire ballads.

But we fall for fall leaf by leaf, and the fall foliage spectacle improvisationally changes its cue every year. Arrive too early, and you only catch the opening foliage act; too late, and you miss the main peak color event.

To help you achieve your amber aspirations, maple marvels, and rustic reveries, I analyzed a trove of tree data provided by the USA National Phenology Network and the many participants contributing to its “Nature’s Notebook” program.

I crunched the numbers on 60,518 annual tree life cycle observations — or phenophase records, for you botany buffs — covering 305 deciduous tree species. That’s a decade’s worth of data, spanning 2,561 observational sites across the U.S., all to serve you the most accurate, tailored-to-your-needs foliage forecast for 2023! I also designated a few states as “Leaf Leaders” for providing some of the best places to see fall colors in the U.S.

Read on or select a jump link to discover when the leaves in your neck of the woods will put on their grandest show of the year.

Credit: Trips to Discover

(Back to Top)

(Back to Top)

Methodology

Independent: My data analysis research is conducted with independence and impartiality to ensure the integrity of the findings. My research is editorially independent, but Trips to Discover may earn a commission for any purchases made through links on this page.

Raw Data Source: USA National Phenology Network. 2023. Plant and Animal Phenology Data. Data type: Individual Phenometrics. 1/1/2012-12/31/2022 for Region: 71.538800°, -66.934570° (UR); 24.396308°, 172.558333° (LL) USA-NPN, Tucson, Arizona, USA. Data set accessed 9/5/23 at http://doi.org/10.5066/F78S4N1V

Collection: I crafted a tailored dataset using the USA-NPN, featuring 60,518 tree status phenophase observations gathered from a span of a decade. These observations come from 2,561 unique sites across the United States and cover 305 species of deciduous trees. My collection focused on two key phenophase classes: “colored leaves” and “falling leaves.”

Cleaning: I narrowed the range of tree status observations to focus specifically on core U.S. fall foliage from August 1 through November 30 of each year. I refined the original dataset to include only sites within the contiguous U.S. I generated a unique Key ID for each observation by combining the site ID, latitude, longitude, state, common tree name, individual ID, and observation year information. I used this Key ID to merge the “colored leaves” and “falling leaves” phenophase observations into pairs.

Analysis: After an initial examination of over 30,000 paired records, I found that “peak fall foliage” generally falls between the first date leaves start to drop (the initial “falling leaves” date) and the date the last leaf changes color (the final “colored leaves” date). This intersection of phenophases yields the most leaves in vibrant hues still clinging to the trees, also known as peak foliage. I calculated the average date range for peak foliage across the 2,561 observation sites, spanning ten years of data. I then aggregated these site-level averages at the state level. I grouped states into eight regions, determined by geographical proximity and comparable peak foliage dates, with the aid of GPT-4 technology. The “Leaf Leader” states were designated based on my subjective evaluation.

Dates: I then derived four critical metrics for each of the eight regions:

1. Foliage Start: Average of states’ “falling leaves” start dates in a region.
2. Peak Start: The earliest peak foliage midpoint in a regional group.
3. Peak End: The last peak foliage midpoint in a regional group.
4. Foliage End: Average of states’ last “colored leaves” record date in a region.

Replicability: All raw data is publicly available. For questions, email [email protected]. [Google Sheets link] contains the cleaned data and calculations minus proprietary info and raw data. Please credit if sharing.

(Back to Top)