| The vegetation cover map is becoming urgent and essential in Cote d’Ivoire. The purpose of this research is developing methods to better monitor the dynamic change of vegetation at country scale using NDVI behavioral analysis on temporal series. At the same time the vegetation cover map at country scale by the use of products from remote sensing has to be produced. In this context annual imagery at250m (MODIS16day composite) and30m (TM) were selected to provide the best temporal and spatial resolution practical for both sensors. In doing so the study mainly illustrated the methodologies. Nevertheless we succeed in producing a vegetation cover map of a pilot area with Landsat TM, scene197/54. The methods and results in this research are therefore applied only to vegetation areas derived from this scene covering the central-northern and northern fraction of Cote d’Ivoire. First of all the MODIS NDVI subset of the country through the NDVI typical color ramp (white, gray to black) displayed only a tiny part of the result in the study. Details on vegetation cover and vegetation use have been extracted with this primary data of Landsat TM with spatial resolution of30m as a reference to assess the changes of vegetation cover, quantify the vegetation cover and use. A Landsat TM change detection map was developed, too. The Landsat TM was acquired on November1986,2006and January2011respectively. The temporal and spatial change was determined once using Visual RGB-Normalized Difference Vegetation Index (NDVI) method. The thematic vegetation use change map was produced with the aid of unsupervised classification and the merge of three layers Landsat NDVI in three channels RGB to a single image. Eight major colors categories were used to represent the changes occurring in the investigation area. The colors depict years of changes (woody cover reduction or regeneration), no change and vegetation stability (low or high). From2006to2011around33.87%of the area experienced vegetation cover change. This high rate is observed with the yellow class. The NDVI composition map results indicate that NDVI can measure vegetative change over time. The digital processing and interpretation of satellite imagery followed by field visit have led to these results: The2011Landsat TM image by visual inspection shows the high use of bush fire in the pilot area at that period (fire pixels were found in the different kinds of crops);The landscape is undergoing changes. Perennial tree plantations (Tectona grandis and Anarcadium indica) spread in the natural landscape increase the areas and are around15years old;The gallery Forest Rivers’arms are heavily farmed;Hybrid classification resulting of Landsat TM2011shows that this area involves eight types of habitats. In the pilot area, it is very difficult to identify them because certain vegetation units are mixed and nested mainly the farmed plots. Finally the maximum likelihood classification technique was used to build a thematic vegetation cover map. The bulk of vegetation cover during the study period is wooded savanna. The perennial crops are gaining the shrubby savanna. The assessment carried out on the produced thematic image indicates accuracy of93.38%and kappa statistic index of0.91were achieved. In conclusion, we notice that the pilot area is affected by a dynamic of deforestation. The creation of class categories red, green, yellow, dark gray supports the assertion of deforestation. The area is characterized by a vegetation mosaic of forest (forest islands, anthropogenic gallery forest, and pre-forest savanna) and savannas (wooded savanna, shrub savanna). The remote sensing was very helpful for considerations of quickly design the vegetation cover and monitor change dynamic. |
PDF Full Text Request