Predictive Modelling of Built-Up Settlement Expansion in the West Bank using Machine Learning and CA-Markov Methods

Read full text Summarise with VIRAT
Authors: Nolan Kressin
Cited by: 0
References: 72

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This thesis investigates the built-up expansion of Israeli settlements in the contested West Bank using multi-temporal satellite imagery, random forest (RF) modelling, and machine-learning (ML) assisted CA-Markov (Cellular Automata) modelling with TerrSet Land Change Modeler (LCM). The research aims to quantify historical land use and land cover (LULC) changes in the West Bank between 1994, 2014, and 2024 and to predict future urban settlement growth for 2034 and 2054. Utilizing Landsat-5 and Landsat 8-9 imagery at 30-meter resolution, the study employs a supervised RF algorithm to classify LULC and analyze settlement urban expansion. We reveal a significant transformation of the study area, with settlements more than doubling in area over the thirty-year study period, primarily at the expense of agricultural and bare land. To predict future growth, the research utilizes a ML-assisted CA-Markov model within the LCM framework. Three policy-driven scenarios: Business-as-Usual (BAU), Low-growth/High-constraint, and High-growth/Low-constraint, are simulated to predict future urban settlement growth for 2034 and 2054. We identify proximity to existing 1994 settlement footprints and road networks as the most influential drivers of growth. Validation of the predictive model against actual 2024 data illustrated moderate agreement, with a Figure of Merit of 22.05%, highlighting challenges in accurately allocating change in a volatile environment. We find that historical settlement growth was more concentrated near the Jerusalem metropolitan area, while future projections suggest a more dispersed pattern across the study region. By bridging remote sensing methodology with the analysis of politically contested land transitions, this thesis provides a novel, automated framework for monitoring settlement expansion in data-scarce and politically volatile areas. These findings offer valuable insights for regional planning and provide a transparent, reproducible method for documenting spatial transitions in conflict-affected, data-scarce landscapes.

Fields of Study & Subjects

Land Use and Ecosystem Services Remote Sensing and Land Use Remote-Sensing Image Classification

Publication details

Journal / source
Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)
Publication date
26 February 2026
Language
EN
Document type
Thesis