Mapping by remote sensing of areas invaded by Acacia mearnsii De Wild. in north-eastern Algeria

Acacia mearnsii De Wild. was introduced into the El Kala region (far north-east of Algeria) during reforestation programmes in the 1970s, in plantations with 20% eucalyptus. In the last few decades, increasingly invasive patches of this Australian acacia species have been observed in the same parts of the country. Tools such as a Geographic Information System (GIS) with up-to-date satellite data need to be used in order to better understand this invasion and identify potentially high-risk ecosystems. This study therefore used mapping tools to investigate the distribution of invasions by this species. Two approaches were applied, one based on data collected in the field and the other on remote sensing techniques, using the supervised classification method across a Landsat 8 image. The results from the two methods show that the species has invaded four sites with a total area of about 3,200 ha. The areas affected are occupied almost entirely by eucalyptus and cork oak forests (respectively 65.0% and 17.9% of the total). Classification across the satellite image could only show heavily invaded areas, whereas the field survey method made it possible to map the entire distribution of the invasion. However, when superimposed, the two maps were largely identical (0.90 <= R <= 1). To conclude, the use of satellite data coupled with field survey data reduced the constraints arising from Landsat image resolution while also demonstrating the importance of methods based on fieldwork.