A remote sensing and scenario-driven modelling approach for assessing land cover-related impacts of urban growth

Research output: ThesisPhD Thesis

117 Downloads (Pure)


Cities worldwide are struggling with environmental challenges and climate change-induced hazards, including flooding, heat waves and droughts. Defining more resilient urban development strategies requires access to spatial information providing a detailed description of the biophysical state of the urban environment. This information is needed to better understand and quantify the benefits of regulating ecosystem services. Urban planning must also envision possible urban futures and assess impacts of planning decisions on the quality and sustainability of the urban ecosystem. This PhD research addresses these topics by exploring the potential of satellite remote sensing and spatiotemporal modelling for assessing land-cover related impacts of urban growth, focusing on a case study for Brussels and Flemish Brabant.
In the first part of this study, two state-of-the-art airborne remote sensing technologies, i.e. imaging spectroscopy and laser altimetry, are fused to produce detailed urban land cover maps. A synergistic workflow is proposed dealing with challenges in urban land-cover mapping related to within-class spectral variability and presence of shadows. The study demonstrates the added value of structural information derived from LiDAR in improving the distinction between spectrally similar urban material classes. This added value is reflected in an overall kappa increase from 0.80 to 0.87 and 0.65 to 0.69 (compared to using only hyperspectral imagery) for sunlit and shaded pixels respectively. To enable temporal monitoring of urban areas at the regional scale we investigate how airborne imaging spectroscopy can be used to calibrate models for assessing the biophysical composition of urban areas, using medium-resolution satellite data. Results show how the proposed approach could facilitate more automated processing of remote sensing big data, and this while yielding mapping accuracies that are similar or even slightly better than those achieved with a traditional training approach.
To address future urbanization, a multi-scale simulation workflow is developed that draws on historic remote sensing data, socio-demographic data and scenario analysis. The defined scenarios reflect alternative pathways for urban development, i.e. continuation of urban sprawl and sustainable densification. To predict environmental change, a novel Cellular Automata framework is proposed that models quantitative change in urban land cover at sub-cell level. As urban expansion is driven by residential and economic activities, spatial microsimulation and other modelling techniques are used to integrate these dynamics in the simulation workflow. Finally, a synthesis analysis is performed that predicts how each scenario outcome affects important water regulating ecosystem indicators, including infiltration, runoff and evapotranspiration. Our simulations predict increases in surface sealing that amount to 2.8% and 1.8% of the study area’s surface between 2013 and 2040, respectively for the urban sprawl and densification scenario. Correspondingly, both scenarios predict detrimental hydrological changes, that will very likely contribute to more and intenser flooding, drought, and heat waves. To envision a more sustainable future, an additional scenario is considered that combines urban densification with Blue-Green Infrastructure, which does seem to yield improved hydrological fluxes by 2040.
The research demonstrates how integration of remote sensing and spatiotemporal simulation can contribute to the assessment of land-cover related impacts of urban growth and, as such, can assist planners and policy makers in analysing and comparing alternative urban development strategies.
Original languageEnglish
QualificationDoctor of Sciences
Awarding Institution
  • Faculty of Sciences and Bioengineering Sciences
  • Canters, Frank, Supervisor
Award date31 Jul 2021
Place of PublicationBrussels
Print ISBNs9789493079939
Publication statusPublished - 2021


Dive into the research topics of 'A remote sensing and scenario-driven modelling approach for assessing land cover-related impacts of urban growth'. Together they form a unique fingerprint.

Cite this