| Abstract | Full Paper | PDF | Printer Friendly Format |

 

Abstract
Flood is probably the most devastating, wide spread and frequent natural hazard of the world. This problem is more acute in the areas under strong monsoon regime where 80 percent of the total rainfall is received in just three months. It is evident that the problem of river flooding is getting more and more acute due to human intervention in the flood plain at an ever increasing scale. From the past experience it has been gradually realized that it is more rational to try minimizing the risk and damage involved in floods rather than formulating structural measures for containing the river. Thus we are in the era of all embracing planning.

This paper seeks to formulate an efficient methodology to accurately delineate the flood prone areas in Gangetic West Bengal, India. The investigation proposes to exploit historical records of flood events for last one decade which was systematically archived by the irrigation department of West Bengal government. Basic aim of this effort is to identify the area chronically suffering from river flooding and create a flood hazard map based on socioeconomic and infrastructure related data.

The unit of this study is the development blocks. This is the smallest administrative unit in India as far as water resource management is concerned. Special attention has been given to administrative units rather than physical units because socioeconomic variables are recorded for the administrative units and any remedial measure taken on behalf of the concerned ministry is implemented in the administrative unit level. In technical terms this paper is focused on computing a composite index of flood hazard based on hydrological, topographical and socioeconomic parameter and finally map it using GIS. This index not only considers susceptibility of each block to be inundated but also takes into account the factors that are inherently related to flood emergency management.

Number of flood occurrence in each of the blocks for last decade is considered to be the key variable determining the chance of river inundation for a particular block. Other sets of variable accounts for emergency preparedness and human life and property at risk for a flood of given magnitude. Availability of surfaced roads per square Km and existence of relatively higher ground per square Km in each of these blocks can be considered in the first group of variables. The second group, that is the indicator of potential economic loss and humanitarian disaster, includes density of population, and percentage of villages having safe source of safe drinking water in each of the blocks under investigation.

The key question revolves around the definition of a higher ground or safe place where people can be provided shelter during a major flood event. In the absence of high resolution terrain data this task becomes difficult. For flood plain studies of regional scale such as the current project ASTER relative digital elevation model has been found most suitable terrain data source at almost no cost. The spatial resolution of this data is 30m and elevation data stored in each of the pixels is not absolute but the difference of elevation between two pixels is absolute. Therefore this dataset has been used to extract topographic features like ‘peaks’. Each of these ‘peaks’ are surely areas having higher elevation than its surrounding pixels, thus can be treated as a potential shelters from flood.

It is difficult to assign subjective weightage to each of these variables mentioned above for formulating the composite hazard index. Hence, principal component analysis has been considered as the most suitable technique to perform this task. The main advantage of using principal component analysis for computing composite index lies in its objective judgment based on purely mathematical criteria. A choropleth map has been generated to show how different blocks fall in different hazard index values.

It is well known that availability of very high resolution digital elevation models required to estimate flood depth extremely flat flood plains is very limited in the developing part of the world. This project is a quest for devising a methodology that can perform efficiently for flood hazard mapping without the support of very high resolution digital elevation models. The accuracy of this analysis is likely to be inversely proportional to the size of the blocks constituting the area under investigation and is proportional to the time period for which records of inundated area is available. It has been illustrated that this analysis can be made more accurate by going further down in the administrative level from blocks to revenue villages. The basic merit of this methodology lies in its simplicity and low cost.