Probabilistic urban simulation in Ho Chi Minh City, Vietnam

Remote sensing has been recognized as an indispensable tool for urban and landscape planning. Regional or local census and survey data, such as population, income and transportation, are often collected to estimate the urban growth in time and space domain. Considering insufficient detail investigations at locality may not available for developing and undeveloped countries, this study examines the capability of using remote sensing imagery in urban growth prediction, without using other ancillary data. Additionally, to incorporate uncertainties in urban modeling, different from deterministic approaches, this study developed a probabilistic urban simulation, which integrates a statistical model and cellular automata, to predict urban growth for Ho Chi Minh City (HCMC), Vietnam. After quantifying fractions of urban land in HCMC from Landsat imagery for 1990, 2002, and 2010, the model was calibrated by observed urban growth between 1990 and 2002, and validated by 2010. The model was then applied to project the urban growth up to 2030. The simulation represented heterogeneous urban growth in time and space and predicted 10%-15% increase of urban area of HCMC from 2010-2030.