Urban Growth, Function, and Sustainability: Multi-Scale Assessment and Diagnostic Frameworks Using Remote Sensing and GIS

Correct PDF of the thesis is coming

Abstract

This dissertation addresses a fundamental challenge at the nexus of urban geography, sustainability science, and remote sensing & GIS techniques: how can we move beyond merely predicting where cities expand to rigorously evaluating whether they develop sustainably? Its overarching contribution is to articulate and demonstrate an integrated analytical framework that links spatial simulation, functional diagnosis, distributional scrutiny, and multidimensional sustainability assessment. Rather than treating urban expansion, land-use performance, intra-urban inequality, and ecological–social balance as separate research problems, the dissertation shows how they can be connected within a coherent diagnostic logic. It develops this framework through empirical work on China and India—the world's two most consequential urbanizing nations. 

To achieve this, the research pursues four specific objectives, each corresponding to a logical step in the intellectual progression: (1) simulate alternative urban growth scenarios to reveal the spatial consequences of different policy orientations; (2) develop a diagnostic index to evaluate whether constructed land actually serves its population and economic purposes; (3) unpack the limitations of aggregate indicators by examining intra‑urban heterogeneities; and (4) synthesize economic, ecological, and social equity dimensions into a parsimonious and transferable sustainability assessment framework.

Methodologically, the work integrates cellular automata–Markov chain modeling with landscape metrics (Paper I); applies the Nighttime Light Development Index (NLDI) to distinguish centralized versus decentralized construction land across all Chinese municipalities (Paper II); employs deep learning on street‑view imagery and Lorenz curve analysis to classify and compare six Indian megacities at the intra‑urban scale (Paper III); and constructs a vector‑based Economic‑Ecological‑Equilibrium (3E) framework using exclusively remote‑sensing proxies—nighttime light intensity, Leaf Area Index, and spatial Gini coefficient—applied to over 2,800 Chinese counties across 24 years (Paper IV). These methods are applied at multiple scales and in different national contexts, but they are not presented as separate methodological exercises. Together, they operationalize a single overarching argument: that urban sustainability assessment requires moving from expansion-centric mapping to performance-oriented diagnosis, and from aggregate indices to distributional and multidimensional evaluation.

The results demonstrate a coherent deepening of analytical capability. Scenario simulations show that different policies produce distinct urban morphologies but remain silent on functional performance. The NLDI reveals systematic urban‑rural disparities in China and a “same value but different spectrum” phenomenon, where identical numerical indices conceal radically different development realities. The India study confirms this phenomenon at fine scales and shows that Lorenz curve shapes carry critical diagnostic information beyond scalar values. The 3E framework, validated against the Sustainable Development Goals Index (correlations 0.6–0.8), identifies sustained improvement in Chinese counties, uncovers four distinct development trajectories, and can provide, under certain conditions,  directional diagnoses—pinpointing whether unsustainability stems from economic lag, ecological degradation, or social inequity.

Collectively, the dissertation’s overall contribution is threefold. First, it reframes urban sustainability assessment as a progression from spatial prediction to functional diagnosis to multidimensional evaluation. Second, it demonstrates that aggregate indices, however useful, should be supplemented with distributional and structural analysis in order to avoid masking critical intra-urban and urban–rural heterogeneities. Third, it shows that remote-sensing proxies, when combined intelligently, may offer rapid and policy-relevant diagnostics, subject to contextual validation. Future research could integrate artificial intelligence for automated feature extraction, couple the 3E framework with dynamic simulation models to explore alternative policy pathways, expand ecological indicators with emerging satellite products, and test the framework's applicability across diverse global regions to support sustainable urbanization worldwide.

Link to DiVA