RESULTS

For the interpretation of the data, the following points must be taken into account:

1. The difference between the seasons when the images were recorded makes it difficult to compare the vegetation, due to the different physiognomic conditions and different phenological phases in which the vegetation is found. The first images was recorded October 24, 1992, one month after the rainy season, when the vegetation has the highest amount of foliage ; the second was taken on April 13, 2000, during the dry season of the year, when part of the vegetation tends to lose its leaves and many annual species have died.

2.The results correspond to an unsupervised classification, meaning that it is a mere clustering according to spectral differences and similarities. This implies that some pixels may be in wrong classes.

3. There was no groundtruth available, thus the results are not verified and there is a possibility of confusion and error.

Below is a comparison of the data of the two images:

APPROXIMATE RESULTS

(Data are presented in hectares)

1. The "bare soil" class refers to areas not covered by vegetation when the image was recorded. This class includes seasonal agriculture and pasture.

2. The agriculture or pasture class is reduced due to an increase in the naked soil class.

3. Burned areas belong to zones that might have been used for agriculture and livestock, or to forests.

A zoom to the area showing 8 of the 10 classes is shown below, as per the information provided by the INEGI. Figures 4 and 5 are satellite images, and figures 6 and 7 are the result of the classification.

FIGURE 4

FIGURE 5

FIGURE 6

FIGURE 7

Home page | Michoacán Case | Guerrero Case | Valle de Bravo Case | Naturale desasters casoe Hurricane Pauline

For doubts, questions or comments
Geographical Information Systems Sub-director