Analyzing red maple leaf color in 10,950‬ iNaturalist observations

Hi everyone,

I’ve been wanting to do something with iNat data for a while now and finally had time to do a side project over winter break.

One of the unexpected consequences of attending college in Connecticut has been experiencing autumn twice a year. I watch the trees of New Haven change color during late October and early November, then return home in Georgia for fall break and watch the leaves change color again.

I understood the basic principle behind my double-dipping of fall. But what would it look like actually to observe leaf color change across the whole continent? I decided to tackle this question.

I had three qualifications choosing a tree species. It had to have a drastic fall color change, a large range, and many observations. After considering Sugar Maple, (Acer saccharum), Sassafras (Sassafras albidum), Trembling Aspen (Populus tremuloides), among other trees, I finally decided on Red Maple (Acer rubrum), which had the highest observation count.

Figure 1: 2019 research grade observations of A. rubrum with leaves. The color of each marker corresponds to the dominant color of the leaves in the respective observation. Each marker disappears 10 days (frames) after it first appears. At the end of April, observations get green as trees “green up,” and fall colors start appearing at the start of October and spread south.


Figure 2: 2016, 2017, 2018, and 2019 research grade observations of A. rubrum with leaves. 2015 omitted so that a nice square could be made for this graphic


Figure 3: Number of observations with “fall” colors in each week of the year for 2015-2019. “Fall” color was defined in an HSV color model as colors with a hue (0-360) > 300 or < 60, a saturation (0-1) > 0.09, and a value (0-255) > 150. Peaks are present at week 17, and weeks 41-44. Week 17 is overrepresented because of the number of total observations in that week, but early leaves can be reddish. Interestingly the 2018 fall colors appear to be about two weeks later as a whole compared to 2019. The late 2018 fall was observed by others, and was attributed to higher than average daily low temperatures.


Figure 3: Overview of process

For a more detailed writeup on the data and the process, check it out on my website. Let me know if you have any other ideas for projects like this! Thanks to everyone who submitted red maple observations. Sorry I had to break it up into multiple replies; new users are only allowed one image.


Neat! Thanks.

1 Like

This is so neat! I know an issue may come up of more than one tree species being in a photo (for instance sometimes there’s a photo of a hillslope that also has sugar maples) but hopefully that all ‘swamps out’.

1 Like

Oh my. This is just marvelous. Love the animations on your website and the description of the process.