Loading…

Deforestation in Michoacan, Mexico, From CYCLOPES-LAI Time Series (2000-2006)

Annual land maps (2000-2006) of the Mexican state of Michoacan were built from the leaf area index series of the Carbon cycle and change in land observational products from an ensemble of satellites project, using a reference map and considering general vegetation types: evergreen forest, deciduous...

Full description

Saved in:
Bibliographic Details
Published in:IEEE journal of selected topics in applied earth observations and remote sensing 2016-12, Vol.9 (12), p.5398-5405
Main Authors: Valderrama-Landeros, Luis Humberto, Espana-Boquera, Maria Luisa, Baret, Frederic
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Annual land maps (2000-2006) of the Mexican state of Michoacan were built from the leaf area index series of the Carbon cycle and change in land observational products from an ensemble of satellites project, using a reference map and considering general vegetation types: evergreen forest, deciduous forest, crops, and grassland shrublands. This was done by calculating 11 phenological variables for each year, and analyzing the principal components. For each pixel, we calculated the distances between the values of the first five principal components and the values of first five principal components of all the pixels in the reference map. Each pixel was assigned the most common type of land cover between the 20 smallest distances. The annual maps were compared year by year to estimate deforestation, namely the pixels corresponding to evergreen forests that changed to another type of cover and stayed that way. We estimate that 35 641 ha of evergreen forest were deforested, of which 19 120 ha became deciduous forest and 16 428 ha became crops. In addition, 458 599 ha suffered nonpermanent changes, mainly at the edge of forests, corresponding to mixed pixels. The results were validated using high-resolution images, with 71% of success in changed pixels, and 81% in unchanged pixels.
ISSN:1939-1404
2151-1535
DOI:10.1109/JSTARS.2016.2597742