Time for convening has arrived.

We'll start this uh evening series of meetings.

First item on the agenda is the tree canopy assessment presentation.

Ms.

Wallace.

Thank you.

Good evening, Mayor and City Council.

My name's Bridget Lawler.

I'm the GIS manager, and I will be giving you an update on our tree canopy.

So

just a quick history.

In 2008, we started with 30 meter Landsat data.

So basically every pixel in that imagery was about the size of the infield of a baseball field.

So not very detailed information.

But then since 2010 through 2019, we purchased imagery through the National Agriculture Inventory Program.

That was much higher resolution first at one meter, and then they increased it to 0.6 meter resolution.

And then due to some processing delays, we ended up going with a different satellite imagery in 2021.

And with all of these images, basically what we do is we do pixel classification.

And what you see up here is four images of the exact same area.

We're at kind of like the on Johnson Ferry where the U is that's the Abernathy Arts Center.

You can see the tennis courts right there, and then you have River Valley and Johnson Ferry Abernathy coming in on the right side of the images.

And so basically what you see is you have the natural color imagery on the top left, you have the infrared on the right.

The infrared helps us distinguish a little bit more on trees versus grassy areas and shrubs.

And then we use the pixel classification.

We tell the software to give us one thousand 100 different groups, and that's what that uh image is with all the different colors.

And we go through each color area through different sections of the city, and we tell it is it canopy or is it not canopy?

It's a pretty tedious process.

Um, and so basically what you see here is an output of the canopy assessment.

We have the white area, which is not canopy, the green are those pixels we determined were canopy, and then we have the yellow area where we weren't sure.

And basically, with the pixel classification, we would just go through and whittle away at the yellow areas until we have a complete area of canopy or

not.

As we said in 2021, the NAPE data was not available to us, and so we looked at another satellite image imagery, and we also discovered that near map, which provides our aerial images, um has products derivatives that were um AI generated, and one of which was an AI vegetation layer.

And so we used that for comparison in 2021, and then when we presented that study,

uh

council requested that we go back and look at using the AI data for 2017 to compare the results, and so we did that for 23 as well.

And then this most recent study, we use the AI vegetation again.

And one of the reasons that we continue going with the AI data is the asset the accuracy is much better than our pixel classification.

You can see here with the AI data we generally have with greater than 95% accuracy.

The maps that you see up there, the green dots represent where the model matched the what is occurring in real life, and so you can see there's much more green dots on the AI map than there is on the ISO map.

So again, we have about 95% accuracy with the AI data versus our pixel classification.

And so for this tree canopy study this year, um, as I said, we use the near map imagery and the AI vegetation.

Um for the AI vegetation, it is anything that is greater than two meters in that um in that layer.

So basically anything that's six feet and above, we consider to be canopy.

Um, and then we look at the percent canopy by council district zoning and within the stream buffer, and then we also do a comparison from 2023 based on the council district, the zoning, and then we also looked at the Transform 285400 project.

Um in full disclosure, we had a little bit of an anomaly in the panhandle area, and you can see in this slide, it's up in the northeast corner of district one.

There was one tile that was just not very clear.

It's very fuzzy, you can barely see the houses and the cars, whereas in the other areas outside of it, it's much more crisp.

What we ended up doing when we looked at the AI vegetation, the model didn't pick up much of the of the canopy in that area as well.

So we ended up just uh splicing out that area and filling it in with the 2023 canopy.

It's not the best that in terms of accuracy, but it was much better than looking at what the AI model provided for 2025.

And we have precedent, we had to do that in the past with some of the pixel classifications where over in the Amber Ridge neighborhood several years ago, there was a large cloud coverage that presented prevented us from doing the pixel classification accurately as

well.

So this is our estimated area of canopy.

Uh

result of the models, we are looking at about 56.3% canopy for 2025.

We've noticed that we have a slight uptick since 21 and 23, and we are almost back up to 2017 levels.

Here is a comparison of other metro cities in the Atlanta area.

We're one of the higher canopy

uh

areas for the city.