Complexity and performance of urban land use change models in rapidly growing developing World cities
Student name: Mr Neeraj Garg Baruah
Guide: Dr P K Joshi
Year of completion: 2010
Host Organisation: The School of Forestry and Environmental Studies, Yale University
Supervisor (Host Organisation): Dr Karen C. Seto
Abstract: Urban expansion and spatial patterns of urban land have a large effect on many
socioeconomic and environmental processes. Although there exist numerous urban
growth models, most have significant data input requirements, limiting their utility
in a developing-world context. Yet, it is precisely in the developing world where
there is an urgent need for urban growth models and scenarios since most expected
urban growth in the next two decades will occur in such countries. The main
objective of this paper is to evaluate the performance of different modelling
approaches for simulating spatial patterns of urban expansion in two diverse rapidly
growing developing world cities. We applied two peer-reviewed urban land use
change models – the artificial neural network based Land Transformation Model
(LTM) and the logistic regression based Hybrid Statistical Model (HSM) over
Bengaluru, India and Shenzhen, China. The predictive accuracy of both models
exceeded 50% in both Bengaluru and Shenzhen. Predictive accuracy results based
on the goodness-of-fit measures like the percent correct change (PCM) and Kappa
suggest both the models using the same drivers of land use change were able to fit
observed data relatively well, in the two different urbanizing areas. Both the models
have been able to replicate the urban growth forms in Bengaluru (compact) and
Shenzhen (fragmented) fairly well. It has also been observed that applications that
have larger amounts of observed net change in the reference maps tend to have
larger predictive accuracies. Comparing the performance of the models under a
criteria of magnitude, location and pattern of urban change revealed sharp
differences between the dynamics of the model outputs, especially regarding the
location and pattern of urban change. The differences between the model
performance that resulted from the multi-criteria ‘magnitude-location-pattern’
comparative assessment pointed out that it is not useful to make a statement about
the validity of a model based on only one or two goodness-of-fit measures and in
certain cases can lead to misleading results which may have dire consequences for
issues like global environmental change. Moreover, it is important to recognize that
no single modeling approach can meet with all possible validation criteria, in all
different sorts of environments. Land use modelers should select the modeling
approach and validation criteria that best fit the research question and characteristics
of the study area.
