The use of satellite imagery for the assessment of fire risk associated with repetitive armed conflicts in North Lebanon

Abstract

The recent history of Lebanon has known repetitive armed conflicts which had significant impacts in terms of mortality and injuries, displacement, insecurity, economic disruption, and damage to the physical environment. More precisely, repetitive armed conflicts may be directly responsible for severe bio-physical modification (UNDP, 2006) by causing damage to the environment (e.g. littoral pollution from oil spill, impact on natural resources from quarrying, loss of flora, fauna and degradation of ecosystems due to fires, etc.). The environment may also be indirectly affected by conflict as the result of changes in the way of life of inhabitants and their use of natural resources (Mubareka and Ehrlich, 2010). The aim of this work was to assess fire risk associated with repetitive armed conflicts on the coastal zone in North Lebanon. A number of recent armed conflict events (dating between 1982 and 2008) which are deemed as hazards to the communities and the environment on the coastal area of NorthLebanon were considered in this study. The methodology of work involved the use of five multi-temporal Landsat (MSS and TM) satellite imageries acquired between 1975 and 2010. The Object-Based Image Analysis (OBIA) approach (Mitri and Gitas, 2008) was employed in this work. The concept here is that the information necessary to interpret an image is not represented in a single pixel, but in image objects. OBIA, which is based on fuzzy logic, allows the integration of a broad spectrum of different object features such as spectral, shape and texture and contextual values, for image analysis. The satellite images were segmented (a total of 6 segmentation levels) and then classified incorporating contextual and semantic information. This involved the use of image object attributes and the relationship between networked image objects of the different Landsat images. A fire risk map was produced comprising five classes, namely, ―No risk‖, ―Low risk‖, ―Moderate risk‖, ―High risk‖ a.