Urban development predictions direction of using a combination GIS and Bayesian the probabilistic model (case study: Ardabil)

Extended Abstract
Introduction
Land is a major element in urban development. Therefore control how we use it, and also calculate the actual needs of the land, In order to satisfy different user in time and generalization And the generalization of matching numbers and the quantities obtained in solving the problem of the future, will work Affordable housing and urban development, (Khakpour et al., 47: 1386). The concept of land and urban space, both natural and economic and social development, qualitative change found and Resulting are richer in a very broad scope and objectives of urban land use. Certainly the use of land and space, as a general source of life and public property, shall be made under the planning principles (Ziyari, 13: 1381). The phenomenon of urbanization is a major concern for urban scholars (sun, Wu.Lv. &Wei, 2013:409). The process of migration of rural population to urban areas is too high (Wu, & Zang, 2012:137) is Unprecedented growth and expansion of the urban (Zing, & et al 2013:754). Urban growth is a spatial process of and social transformation In connection with the change of the urban area and is changing the way people live in different scales (Hayashi & Imura, 2009:133). One of the most important environmental issue related to changes in the twenty-first century is in many countries and regions, especially in developing countries (Jokar& et al, 2013:37). According to planners and experts forecast, in 2020 the urban population of 75% of the entire world will take its place in the absence of approximately 2% of the world's landmass (Bhatta, 2011). Urban development and increased migration to urban areas affect the physical and city-wide level is brought into the surrounding areas Suitable and unsuitable land and urban areas urban development unrestrained influx.
Methodology
In this study, Bayesian techniques have been used in the past to see the future development directions and satellite imagery using is Landsat ETM images of 1987 and 2013 Model and predict future directions for the development of acquired and developed a database of the most important factors that influence the research process.In this study, using 14 parameters of natural and anthropogenic (Elevation, vegetation, land units, main roads, secondary roads, dirt, away from the river, the earthquake happened, geology, industrial zone, steep terrain, airports, land use and distance to fault) Urban development has been done to predict. The first step of this process was obtained topographic maps using the layer elevation model (DEM) with a pixel size of 30 m. With this layer using the various functions available in GIS software packages such as gradient layer, which is used in the model were extracted. Developed areas of the city in the period from 1987 to 2013 using Landsat ETM satellite images were selected for modeling Two-thirds of these areas as areas of modeling and control to find the weight classes were used and a third was used for model evaluation. The Landsat images using bands 3 and 4 was prepared vegetation index (NDVI). For the preparation of land use map of the study area ENVI image processing software and methods of supervised classification is used (maximum likelihood algorithm (MLC)) and Landsat color image.
Discussion
To predict the future development direction were used of the weight of the evidence (Carter et al., 1989). This model is useful as a model and tested in different contexts. The weight of evidence is a statistical method based on Bayesian probability theory (Dennison et al., 2002). The model dependence between an event (physical development in the past) and causal factors (predisposing factors of physical development) estimates. If our causal factors (predisposing factors and physical development of the city) to B, _i and physical development classes each parameter S in the past to consider, In the case of Bayesian theory for calculating the conditional probability of physical development (S) in a given class (B_i), the following equation can be used:
P(s│B_i )= (P(B_i│s)×P(s))/(P(B_i))
After the weight of each class of parameters were considered, the application environment for each weight class were exposed Arc map And the synthesis parameters on the final map was found to predict the development of the city. Using natural fracture maps obtained were classified into 5 classes.
Results
Nowadays, with the rapid growth of cities in developing and developed So that the urban development process and how change at the macro level is One of the most important issues facing researchers on urban issues. Increase of population and the development of non-principal cities, especially in developing countries, the results of will follow such. The loss of resources, lack of compatibility with existing infrastructure in urban growth, changing agricultural land suitable for urban use and increased costs such as housing, Transportation. In this regard the Ardabil Province, especially after being in the process of expanding its Despite laws and regulations inhibitors such as arable land and gardens Act, The preservation and expansion of green space in cities, in the process of expanding its disproportionate impact on the environment with Green and Natural Resources, which includes gardens and city farms and plantations have been lost due to irregular growth and urban diffuse And some villages have been incorporated into the urban fabric.Modeling requirements for future development in the city can be a useful tool, planners can assist in forecasting future needs of the city. This study using satellite images for the years 2013 and 1987 and 14 standard Bayesian theory of natural and man in urban development (Elevation, vegetation, land units, main roads, secondary roads, dirt, away from the river, the earthquake happened, geology, industrial zone, steep terrain, airports, land use and distance to fault), Predict the direction of future physical development of the city of Ardabil, conducted and results were presented as a map Phnh‌Bndy in five classes. The zoning map 17/8 percent of the class with the ability to develop very large area, 29.6% in office with great development potential, 20.4 percent of middle-class capabilities, 18.9 percent and 24.3 percent of the class with low development potential have been in the classroom with very little development capabilities.