DEVELOPMENT OF LANDSLIDE HAZARD ASSESSMENT SYSTEM

Abstract

Man has stood up to many challenges posed by nature over the ages and delved into the depths of the problems to analyse and solve them. It is in human nature to improvise and simplify. The basic philosophy of science and technology is to reduce complexities. Herein lay our motivation for developing a system which combines the fundamentals of the scientific techniques with the ease that is the hallmark of any automated technology. A simple automated information system, capable of addressing the information needs of a novice user looking up an issue, is of value and worth researching. And the issue that is at the centre of discussion in this work is automated landslide hazard zonation (LHZ). A scourge of the mountains, landslides cause huge losses in terms of both life and property. In the mountainous region landslides are widespread and the high susceptibility to landslides in this terrain is mainly due to a complex geological setting, varying slopes and relief, heavy rainfall, along with ever-increasing human interference in the eco-system. To minimize the damage done, landslide hazard in a particular region has to be estimated and it would be worthwhile if even the non-experts could do so. Although the field of LHZ is a highly researched area, the general user has hardly had any tool to keep him abreast of the threat perception of a region he is moving in to. These and more are the objectives that have been reached by the proposed landslide hazard assessment system (LHAS) through the way of this present research work. The work in itself was multifarious in nature what with the principles of geology to be interpreted by an intelligent system dwelling on the fundamentals from artificial intelligence. This meant that the development had to be broken into fragments called modules so that at each interface there is a chance of providing the suitable understanding of next phase thereby creating compatibility. Each layer is interacting with the layer iii below it as well as above it but there is no need of having a rigid rule structure for all of them since each is an independent module. The concept of an automated system derives heavily from the notion of having a robust input facility. This is all the more important from the point of view of a hazard addressing system so a generalized input system has been planned and proposed. The working intelligence behind the system — the brain of the system — is the combination of understanding module and the inference engine in the expert module. They process all the algorithms throughout the operation of the system. Coming ahead from all the technicalities, the system has an easy user interface and asks if any input need to be given separately. Further, it has been tested under various cases to gain insight into the assessment capabilities of the system. The knowledge base (KB) with its current capacity and the expert subsystem in its current strength make the system one step ahead of traditional knowledge based expert systems (KBES). It is true that complex problems like landslides require solutions that are achieved through complex modeling methods but an approach comprising combination of existing techniques through an intelligent interface might just answer the queries in a simple way albeit with a tradeoff somewhere. Although it is not a replacement of advanced LHZ techniques in its present form, but is a viable alternative. The proposed work set out to try and address all these concerns, picking up the trails left by experts, the ultimate goal achieved being to help society in better sustaining itself within the great playground of nature called earth. iv