Morteza Najafi; Mojtaba Rafieian; Rama Ghalambor Dezfuli
Abstract
Introduction Nowadays, spatial models and techniques are widely used to analyze challenges at urban and regional levels. These models and techniques can identify the relations between different variables, evaluate their impact on spatial spheres, and thus aid urban planners and managers. Recently, solid ...
Read More
Introduction Nowadays, spatial models and techniques are widely used to analyze challenges at urban and regional levels. These models and techniques can identify the relations between different variables, evaluate their impact on spatial spheres, and thus aid urban planners and managers. Recently, solid waste and the amount of waste generated in urban areas have gained attention as a major global challenge and the World Bank has highlighted the importance of an acceptable global approach to the issue of urban waste in 2016 (World Bank, 2016). Urban waste impacts the city and its urban management system in different ways such as urban environment degradation, economic impacts and the challenges of urban landscape. Different factors impact urban solid waste generation and investigating the relation between these variables can help urban planners and managers formulate general plans and policies to reduce urban waste. But a mere examination of the relationship between factors affecting urban waste generation and the variables proposed by the World Bank cannot provide a good estimate of the future status, since spatial factors always impact the quantity of urban waste generated. Therefore, spatial models and artificial neural networks were proposed and discussed. Geographically Weighted Regression is one of these methods used to investigate the relationship between different factors affecting urban waste generation. Geographically Weighted Regression can investigate the relationship between different variables, examine their impact on the city and predict the relationship between different variable of urban waste generation and their impact on the city in the future. The artificial neural network was also used to assess the nature of data and predict the future status of urban waste. Materials & Methods The study area consists of 22 districts, 123 zones (116 zone due to the availability of supplementary information of 2011-2012 regarding the districts of Tehran), 40323 statistical areas and 895247 land uses of Tehran. Data were classified in three stages. The first phase includes the information collected from Tehran waste management organization regarding urban waste in 1996 to 2016. In the second phase, information was collected from statistical center of Iran regarding demographic segments and social components. Finally, data were collected from Tehran municipality in the third phase providing useful information about urban performance (Land use). Results & Discussion Physical-environmental components and especially land use directly impact urban waste generation. However, results indicate that some land uses such as institutional and publicbuildings gradually stop the increasing process of urban waste generation due to a decrease in their population as compared to residential land use. Population density and income ratio are investigated as the first and second rank variables. These two variables have directly impacted the amount of urban waste generation in most districts of Tehran. From central areas of the 6th district to the southern areas of the 20th district, southeastern areas of the 18th district and eastern areas of the 4th district of Tehran were influenced by population variables. In other words, the amount of urban waste generation is increased with increased population density in these district. However, the amount of urban waste generation in the 22nd and 21st districts do not change with the above mentioned variables. Results indicate that different urban development plans and policies increase population and area dedicated to different land uses and thus, play an important role in urban waste generation. The 22nd and 21st districts are in a desirable status regarding variables such as area, population, and urban waste generation, but predictions indicate that they will reach a similar status and face challenges related to urban waste generation in 10 years. Spatial distribution pattern of urban waste generation in Tehran indicates that the eastern and southern districts produce the highest amount of urban wastes. This pattern is gradually moving from central to western and central districts, and without a plan to control the situation, the pattern will move from east to west and south to north of Tehran in the next 10 years. Based on the results of spatial autocorrelation and a comparison with the results of the least squares method, Geographically Weighted Regression was considered as a suitable method of predicting urban waste variables in Tehran. This indicates that spatial variables affect urban waste generation in Tehran. Moreover, artificial neural network is capable of predicting non-spatial nature of relations among different variables of urban waste generation and thus can predict the amount of urban waste generation in Tehran. Conclusion Results not only identify (physical-environmental, economic and social) variables affecting urban waste generation, but also indicate superiority of Geographically Weighted Regression technique at spatial and non-spatial levels as compared to the least-squares regression and artificial neural network.
