This repository demonstrates how the choice of spatial data resolution can dramatically affect estimates of tree cover and subsequent ecological interpretations.
Using two global land-cover datasets — MODIS (500 m) and ESA WorldCover (10 m) — this notebook walks through how to calculate, visualize, and compare forest extent for the same study area.
This project outputs two different calculations for tree cover.
High-resolution basemap used for visual context.
Comparison of MODIS (500 m) and ESA WorldCover (10 m) forest-cover estimates.
The area shown in the images covers approximately 370 km².
A large portion of the upper right appears densely tree-covered, while the lower left contains noticeably less forest in the satellite image from World Imagery.
This spatial pattern is reflected in both output maps, where green indicates tree cover and black represents all other land-cover types.
Placing the MODIS and ESA WorldCover products side by side makes the difference in resolution and classification readily apparent.
The MODIS dataset estimates 105 km² of tree cover (about 28% of the total area), whereas the ESA dataset estimates 318 km² (approximately 86%).
This contrast highlights how data resolution and classification methods can dramatically affect forest-cover estimates and their ecological interpretation.
The notebook performs three main tasks:
- Visualize the study site — display a high-resolution reference image from Esri World Imagery.
- Calculate forest area using two datasets with different spatial resolutions.
- Compare results to illustrate how data resolution influences ecological conclusions.
This can be reproduced with your own area of interest and sampling sites.
- Clone or download this repository.
- Install dependencies:
pip install -r requirements.txt