Embedded Files
Coordinates: Latitude 37,78 Longitude 21,75 / Altitude: 600m.
The forest of Foloi is a particularly popular forest in the prefecture of Ilia in Western Greece. A favorite excursion destination as it has a large area and is flat. It is an oak forest surrounded by arable land and a pine forest.
The forest is large so we decided to study the heart of the forest, a compact triangular area south of the village of Koumanis.
Screenshot from Google Maps showing the area we will be studying
In the following photo (screenshot from EO Browser) you see the triangular area of the forest, 3 km2 in area, which will be our study area.
Screenshot from EO Browser with the area that will be the field of study
So from within the platform we started to study photos on various dates to come to conclusions (all the photos we used can be seen here). So from within the platform we started to study photos on various dates to come to conclusions (all the photos we used can be seen here). First we wanted to confirm our hypothesis that autumn was delayed. We found a first proof of this using the Compare option as you can see in the following photo:
The left part shows the forest at the end of November 2022 while the right part shows the end of November 2023. It is obvious that on the right the green dominates to a greater extent
We reached similar conclusions by studying the timelapse videos produced by EO Browser. The following video is a composite of two timelapses we created on the platform. Essentially it shows in juxtaposition images of the exact same date, on the left from autumn-winter 2020-21 and on the right the exact same dates of autumn-winter 2023-24. In order to maintain the correspondence in the dates, we also left the images with extensive cloud cover, even though they have no information to offer. The difference and delay in the arrival of winter is very easy to distinguish, especially in November. Perhaps the most characteristic frame is that of November 24th.
Let's mention here that we studied the photos by applying various filters. In addition to true color (where we also took the screenshots) we also made observations in false color and in agriculture and of course in NDVI (normalized difference vegetation index).
For example we quote the same photo, the same date from Foloi forest using different filters:
Study through Python
Next we wanted to draw some quantitative conclusions, so it was Python time! First we took the photos that make up our field of study and "painted" them black outside the triangular area we mentioned earlier (we also saved them in the shared folder). After they were processed, we simply ran the python program for each one individually and recorded in a spreadsheet the percentage of green and brown that the code detected. The spreadsheet is below.
Python.xlsx
E.g. in the example you see, green was calculated at 72.01% and brown at 10.21%.
We then drew quantitative conclusions from the spreadsheet we created:
19/11/2023 to 09/11/2022: percentage of green: 95.83 to 94.82. That is, we can detect a time delay of 10 days
09-14/11/2020 to 19-24/11/2023: in this five-day period we had the very large increase of coffee in the foliage of the trees. About a 400% increase for 2020 and a 650% increase for 2023. Here again we therefore see a 10-day time lag in when leaf yellowing dramatically increases
14/11/2022 to 19/11/2023: the rate of brown color in 2022 is 33% more than the corresponding rate 5 days later in 2023. Therefore there is a time delay significantly longer than 5 days
To mention the difficulty in drawing conclusions as many of the critical days in November (which mainly have the yellowing of the leaves) have high cloud cover.
General conclusion that the time delay is at least 10 days.
Page updated
Google Sites
Report abuse