ISSN 2348-1218 (print) International Journal of Interdisciplinary Research and Innovations ISSN 2348-1226 (online) Vol. 8, Issue 2, pp: (108-115), Month: April - June 2020, Available at: www.researchpublish.com
Spatial-Temporal Modelling for effective prediction of Spatial Urban growth using Artificial Neural Network: Case Study- Dar-essalaam, Tanzania 1
Joseph Hayola, 2Rigobert Francis Buberwa, 3Beatrice Tarimo Ardhi University, Dar-es-salaam, Tanzania
Abstract: The Dar es Salaam City has been growing at a rapid pace which surpasses the capacity of the City’s local government authority for urban planning and development as a result of expansion of informal, unplanned settlements with inadequate infrastructure and services. Moreover, the Local government authority lacks adequate understanding of the trend of the rapid urban spatial growth with time, which results into little understanding of the rate and extent of the spatial growth due to the complex driving factors. This growth of the urban area affects land cover, hydrology, geochemistry, biodiversity and the socio-economic setup of the City. Understanding the occurrence of urban growth in space and time is very challenging. Thus, due to the complex interactions between humans and the environment, an Urban Growth Prediction Model (UGM) was developed using environmental factors to forecast urban growth by employing Geographical Information Systems (GIS), Remote Sensing (RS) and Artificial Neural Networks (ANNs). The predictive ability of the developed Neural Network model was examined through model accuracy assessment techniques particularly Percentage Correct Match Metric (PCM) and urban growth dispersion metric. A comparison was carried between the actual land cover of 2011 and the simulated land cover of 2011, where the PCM estimate was 67.98%. The study findings show that the model predictive capability is sufficient to be used to forecast the future spatial urban growth. Thus, the proposed ANN Model results may be applied to urban planning practice and urban policy development. Keywords: Artificial Neural Network, Informal Settlement, Geographical Information System, Remote Sensing, Urban Growth.
I. INTRODUCTION Urbanization trends in several less developed Nations have been occurring under the dictates of scarcity and evolution of informal settlements [1]. Urbanization would be an outcome of either a “push” from agricultural yield progression or a “pull” from industrial yield progression leading to “production cities,” through a combination of employees in tradable and non-tradable sectors [2]. According to [3], “Urban growth is a result of a combination of manifold factors: geographical location, natural population growth, rural-to-urban migration, infrastructure development, national policies, corporate strategies, and other major political, social and economic forces, including globalization”. The rapid increase in number of urban settlements has caused the invasion of public and private land by low income groups devoid of planning programs, depriving them of secure tenure and basic physical and social infrastructure. This often results into informal settlements, which is described differently based on the organization and lawful structure of the nation it presents. In this specific study, informal settlements are described as residential buildings fabricated on places without official planning consent [4]. They are mostly branded by high population densities, limited or and low quality housing stock and deficiency of, or insufficient infrastructure and public facilities. Informal Settlement is observed in two ways, as a problem as well as an answer to accommodation requirements to the rapidly emerging cities of several less
Page | 108 Research Publish Journals