Monitoring and Analysis of Deforestation Process using Satellite Imagery and GIS (Case Study in Myanmar)
4. Annual Expansion of Deforestation
As a second step, we monitored annual expansion of deforestation using AVHRR time series images. In particular, many researches apply the AVHRR temporal products for monitoring of temporal changes associated with vegetation, estimating rangeland productivity and monitoring of forest fire. AVHRR data have two indicators, which are associated with biophysical condition of vegetation. They are Normalized Difference Vegetation Index (NDVI) and surface temperature (Ts). Recent studies demonstrated the advantage of combining NDVI with Ts data for Land Cover Classification (Vasu and shibasaki, 1997) and Land Cover change Analysis (Lamibin and Ehrlish, 1995, 1996; Lambin and Strahler 1994a; Nemani and Running, 1996).
In this process, we compose every January AVHRR NDVI images and the pixel value was stretched 0 to 255 levels. NDVI threshold value for forest and non-forest transition is decided based on the 1995 land cover image overlaid on AVHRR NDVI Image. According to the figure 6 it is clearly seen that spatial expansion of deforestation within 6 years and deforestation was accelerated after 1992.
Figure 6. Yearly Spatial Expansion of Deforestation
In order to define temporal deforestation, pattern, we overlaid the deforested area, which was detected by Landsat TM on the AVHRR NDVI time series images and to compute mean NDVI values in the deforested area.
Figure 7. Comparisons of Annual NDVI value changed in deforested area and wood extraction in cubic ton from study area.
Although deforestation caused by various activities (logging activities, human activities, agricultural expanding), this pattern was quite similar to reverse relation of statistical data of wood extraction in cubic tons from study area. (NOTE: The statistical data does not exactly cover the whole study area).
5. Analysis of Deforestation Process
In order to delineate future deforestation risk area and to clarify relation between deforestation and landscape attributes, we used GIS data (Digital Elevation Model (DEM), transpiration network, and population data). This study is also demonstrated relation of between deforestation process and landscape attributes, which can be analyzed by integration of Remote Sensing and GIS technique.
5.1 Deforestation vs Elevation
This study was made by combination of deforested area image and 1km DEM (Digital Elevation Model).
Figure 8. 3D Visualization of Deforestation on DEM.
Figure 9. Relation of Deforestation and Elevation.
In this process, we assumed every 1km in each land covers were same elevation. First integrate land covers were same elevation. First integrate these two dataset and extract total numbers of deforestation pixels from DEM dataset. The 95% of total deforestation were occurred between 0 to 240m in elevation.
5.2 Deforestation vs Slope Gradient
This analysis was made by combine deforested image and slope gradient dataset which was generated from 1km DEM dataset. The total numbers of deforested pixel wee extracted from slope gradient dataset and plot the graph.
Figure 10. Relation of Deforestation and Slope Gradient
The 95% of total deforestation occurred within 0 to 5 degree in slope gradient.
