Geography as science has provided the greatest theoretical body for studying the configurations produced over the geographical area. The mapping generation depicting soil cover is one of the most important uses of remote perception. Polygons, with oil and livestock activity, were selected from the following landscape units: camps and valleys. The polygons were used to assess multi-temporal changes of soil cover and plant communities over a period of 15 years. Supervised classifications and landscape metric analysis were carried out to examine the possible causes of these changes by comparing: (I) polygons with no oil or livestock activity (witness), (ii) with oil and livestock activity, (iii) with oil activity and no livestock activity, and (iv) with no oil and livestock activity. The results obtained indicated that patch density, total edge, edge density and fragment splitting rate increased across the three landscape units between 2001 and 2016. The contagion rate, by contrast, decreased. An increase in bare soil was observed in the ramps and in the western valleys, with the decline of dominant plant communities. Patch density in undisturbed sites was the lowest, and in places with both riots simultaneously acting at the same time. In undisturbed sites the effective mesh size was maximum, while in sites with some riots it was smaller. When analysing the effect of the riot on landscape features, it was found that its presence generates the highest density of patches and minimum connectivity. The results show that there was a process of fragmentation in the soil cover led by the oil farm and sheep farming, which reduces the patch size and thus the density of patches per unit area.