GIS BASED ENVIRONMENTAL IMPACT ASSESSMENT FOR SITING INDUSTRIES

Abstract

Industrialization plays a crucial role in the economic and social development of a country. The production of industrial goods and services involves the exploitation of natural resources which has resulted in various environmental and pollution problems. Harmful emissions and wastes generated from the industries have impacts like air pollution, water pollution, the contamination of soils, acid rain etc. at the local level and climate change, ozone layer depletion and the loss of biodiversity at global level. Therefore, industry and its impact on the environment has become one of the issues of the debate on sustainable development. The development should be in such a way that the desired socio-economic development and safe guarding of environment and maintaining good quality living conditions are obtained simultaneously. Environmental Impact Assessment is a tool to assess the environmental impacts of industries during project planning and implementation. To identify environmentally suitable site for industries, EIA is the most common process for minimizing the harmful effects of pollution on the environment and human health. Current practice in EIA does not allow for a thorough assessment of alternative sites because ElAs are usually implemented when commitments have been made in favour of a particular location. As a result, many ElAs often concentrate on the environmental impacts associated with a few sites and necessary mitigating measures rather than finding the most environmentally suitable site. If the objective of EIA is to be fulfilled, then it would be necessary to select sites of minimum environmental impact rather than the application of mitigation measures to sites of greater environmental consequences. It is therefore becomes important to concentrate effects on regional analysis and evaluation in order that areas of least impact can be defined in a more scientific and objective manner. If planning machinery are required to optimize land use so that environmental impacts are minimized, then EIA would also be needed for the most important planning decision making. EIA must be applied at all level of decision making on order to reduce environmental degradation. The purpose of an EIA is to safe guard the environment; it has often been regarded as a tool of advocacy by environmental interest groups. The conclusions of EIA must be presented in such a manner that they can be taken into account along with other relevant economic and social factors. Most of the EIA methodology involves the scaling- rating (or ranking) of the alternatives on each of the use of numerical scores, letter assignments or linear proportioning, or ranking from best to worst in terms of potential impacts on each factor. The parameters are either beneficial or adverse and of different magnitude; sometimes it is difficult to assign their relative importance precisely and accurately and involves substantial subjectivity in expert judgment. It is, therefore, imperative to use some statistical method to minimise the subjectivity and achieve parameters relative importance with least bias. Moreover, Ad-hoc method fails to organize the information for analysis and presentation into a meaningful way. Checklists are in updated version of the ad hoc approach in which specific areas of potential impacts are listed. The evaluator needs to tick against each environmental parameter for adverse, beneficial, or no effects due to proposed project activities. Checklist defines the parameters to be evaluated, but it is usually very large, very subjective, and provides little guidance that can aid in the decision making process. Mathematical Matrix method is criticized for too much of mathematical operations. Further, it is also difficult to carry out the multiplication of matrices manually. Munn (1975) mentioned that only advance stages of EIA include spatial analysis as it is considered complex and data hungry. But with the advent of user friendly Geographical Information Systems, the complexity and cost of spatial analysis has been reduced. The problems of database storage, integration retrieval and management can be solved with the help of GIS. GIS can produce geographical matrices for spatial decision analysis which can be utilized in Environmental Impact Assessment. The major reason for integration of GIS and EIA is spatial dimension of the environmental information. It is the overlay method of EIA, which has been encouraged by the GIS technology from its early hand drawn technique to the computerized overlays and then its integration with spatial Multi Criteria Decision Making and Artificial Intelligence Approaches for various applications. In GIS, the overlay method is often called as suitability analysis from the planning point of view. As per industrial siting literature, researchers have mainly concentrated on vector GIS for developing Decision Support System for industrial site selection using AHP or Ranking Methods (Jun, 2000; Thomas, 2002; Eldrandaly, 2003; Dudukovic, 2005). Some studies on industrial siting are based of specific parameters like Mendes (2001) on transportation and Lamelas (2006) on geological parameters. Industrial siting is a complex process that requires evaluation of environmental, social and economical factors. Further, there IV are very few studies which utilize advance technology like fuzzy, neural network etc. for industrial siting. Uttaranchal is 27' state of the India and is formed on November 9, 2000 and is in its initial phase of development. In year 2002, the government announced new policy to promote industrialization in the state. With the promotion to industry the development is expected to be in such a way that the desired socio-economic progress and safe guarding of environment and maintaining good quality living conditions are obtained simultaneously. One of the possible ways of minimizing the effects of pollution in environment is by identifying environmentally suitable site for industries in the state. Therefore, Haridwar, one of the districts, is selected as the study area for the research. Industrial siting depends on many factors. Therefore, factors considered for this work are categorized into two main group environmental and socio-economic factors. GIS database is created which include various steps like data capturing, registration, digitization, spatial and non-spatial data linking and spatial queries like buffering, recoding, reclassification, masking etc. GPS survey is done to know the location of industries and boundary of newly declared industrial site. The standardization of factors is done using linear scale transformation and fuzzy membership functions. Knowledge based Multi Criteria Analysis tool is developed to derive weights for the factors. The resulted weights are used in Spatial Multi Criteria Modeling of environmental and socio-economic suitability for industrial siting. The suitability maps are reclassified into various suitability classes according to impact scores of different pollution potential industries categories (i.e. Low, Medium and High). The resulted environmental and socio-economic reclassified maps are cross-tabulated to know the environmental and socio economic suitability condition of each pixel. Artificial Neural Network modeling is done for the condition when expert opinion / knowledge are not available. The results of both the approaches are compared. Further, suitable site for installing of new sugar industry is found out and comparison of suitability of resulted from both the approaches on newly declared industrial site are done. Furthermore, change detection on this site is done to know how industrial development is going on. The purpose of development of Knowledge Based Multi-Criteria Tool is to determine weights for Spatial Multi-Criteria modeling. GIS software do not have tool to determining weights. Users have to utilize in other software packages like MS Excel etc. for weight calculation. Most often, availability of Experts becomes limitation for the multi-criteria analysis. Therefore, the developed tool has capability to store derive weights using different weighting methods like rating, ranking and AHP. Not only this, it can store the opinion of experts as knowledgebase for specific problem, which can be retrieved when required. Also, it has facility to consider knowledge from multiple decision makers simultaneously which can be is combined. The results show that the study area is suitable for medium and low polluting potential industries and is not suitable for high pollution potential industries. 36.91% of area is unsuitable for industrial siting, which comes under constrains as per rules and regulations for siting industries. About 70% of suitable area is common in multi-criteria and neural network modeling for siting new sugar industry. This area lies in the Bhagwanpur, Bahadrabad and Khanpur blocks of the district. On comparing results of multi-criteria and neural network modeling show that Haridwar is suitable for low to medium pollution potential industries. Both modeling techniques have similar trend in different suitability classes. It is also observed that neural network modeling gives more area in high suitability class that multi-criteria modeling. The reason may be that, in multi-criteria modeling, weight determination is based on vi knowledge of experts which may suffer from biasness because human have general tendency to avoid extreme conditions to minimize the risk. Moreover, the aggregation method may not be adequate for representing relations between criteria that are essentially non-linear. In this research work it is demonstrated that the neural network-based GIS modeling approach can approximate an expert's decisions without the explicit elicitation of expert knowledge. Therefore, it is concluded that neural network modeling is a powerful approach for spatial decision making when expert knowledge is not available. Moreover, Bottleneck of conventional cartographic and statistical techniques, such as weight determination, selection of summation functions, and inability to handle noisy and missing data, can be resolved through the learning process and hidden layers in artificial neural network modeling. The planners and decision makers who often have to make complex decisions within a short period of time can utilize this methodology. VII