Abstract ID: 92

Assessing the response of the MODIS vegetation indices to landscape disturbance in the forested areas of the legal Brazilian Amazon

In the Amazon, the conversion of forest to agriculture and pasture areas is being accompanied by intensive habitat fragmentation(Lovejoy et al., 1986 Laurence and Bierregaard, 1997). Forest fragments are more prone to tree mortality and canopy damage(Laurence et al., 1997). The possibility of remotely assessing the conditions of the forest remnants is certainly instrumental in preserving the long term functioning of the ecosystems as a whole. In fact, some studies do suggest a strong correlation between forest structure and optical vegetation indices (Gamon et al., 1995 Oliveira-Filho and Fonte, 2000 Freitas et al., 2005). In this study, we assessed, by remote sensing means, the impacts of forest fragmentation on the Amazon landscape. Landscape disturbance, obtained for an area of approximately 3.5 million Km2 through simple spatial metrics (i.e. number of fragments, mean fragment area, and border size) and principal component transformation were then compared to the MODIS NDVI (normalized difference vegetation index) and EVI (enhanced vegetation index) seasonal responses. As expected, higher disturbance values prevail in the southern border of the Amazon, near the intensively converted deforestation arc, and close to the major roads. Concerning the respective VI responses, the NDVI seasonal contrast tends to more closely follow the human-induced patterns i.e. forest remnants from areas more intensively converted are usually associated with higher NDVI seasonal values. In fact, the significant correlation between the NDVI seasonal responses and landscape disturbances are clearly corroborated by the analysis of Geographically Weighted Regression (GWR) parameters and predictions. On the other hand, the EVI seasonality seems to behave more independently, and even in an opposite manner i.e. higher variations tends to predominate over less fragmented forest patches. Although further research is needed, our results suggest that the degree of fragmentation of the forest remnants do contribute to the optical remote sensing signal. Thus, it may become possible the upscaling of field-based data on overall canopy condition and fragmentation status for basin wide extrapolations.

Session:  LCLUC and Human Dimensions - Current and future trends of land-use/land-cover change and agricultural intensification.

Presentation Type:  Poster