MONITORING URBAN DYNAMICS USING HIGH-SPATIAL AND HIGH-TEMPORAL RESOLUTION SATELLITE IMAGES

Abstract

Worldwide economic development and population growth have led to unprecedented urban area dynamics. These changes have an impact on natural ecosystems and influence the development of human society. It manifests a long-term impact on urban sustainability at various scales in terms of economic consequences, environmental degradation, and societal impacts. With the availability of satellite data at higher spatial (3-10 m) and temporal (1-3 days) resolutions, new opportunities arise to monitor and characterize urban changes. This dissertation advances the utilization of high-spatial and high-temporal resolution satellite images to identify and monitor urban area changes, including gradual urban area changes and the recovery process after urban disturbance (a discrete event that disrupts the urban system). However, several knowledge gaps remain: insufficient studies monitoring gradual intra-urban changes, limited geolocation scope and predominately focus on large cities in mapping gradual urban dynamics, only a few studies differentiating multiple urban classes, and a lack of analysis on high temporal frequency especially in the context of urban disturbance. This dissertation addresses these issues targeting three individual case studies. Chapter 2 focuses on evaluating the accuracy of urban dynamics detection across diverse geographic locations globally using Sentinel 2 images based on a deep learning model. This chapter also detects and monitors gradual intra-urban changes in the Washington DC-Baltimore region from 2018 to 2019. The study reveals that in just one year almost 1% of the total urban area underwent changes with the majority coming from the construction of commercial buildings, followed by residential buildings. Almost 10% of changes were attributed to the construction of new or the renovation of existing schools. Chapter 3 expands the geolocation scope to the under-researched country of Ukraine, as well as smaller cities that are under-studies. This chapter analyzed and characterized changes in various urban land use subcategories across 30 Ukrainian cities utilizing Sentinel-2 images. Findings indicate that between 2016 and 2021, approximately 3.5% of the total urban area in Ukraine changed. Among all classes that underwent changes in 2016, the most significant transformation decrease was observed in green urban areas (25.63 〖km〗^2), natural vegetation (12.02 〖km〗^2), and agriculture land (3.85 〖km〗^2). Within urban sub-category transitions, the combined area of the urban fabric (16.61 〖km〗^2), construction sites (14.94 〖km〗^2), dumpsites (5.01 〖km〗^2), and industrial zones (4.42 〖km〗^2) experienced notable increases by 2021. Chapter 4 targets the lack of dense time-series image analysis, especially in assessing recovery after urban disturbances. This chapter examines housing recovery after Hurricane Maria in Puerto Rico utilizing time-series PlanetScope imagery to provide insights into the post-disaster recovery. The locations and duration of temporary protective blue roofs (tarps) are identified and linked with socio-economic data for further analysis. Results estimated 14,767±586 buildings with temporary roof installations across five cities in Puerto Rico, or 6.3% of the total number of buildings, with the average duration of installation 351±10 days (almost 1 year). Additionally, socially vulnerable populations, including people with lower income, were more likely to experience the need for temporary roofs for their damaged property and longer waiting times to substitute those with permanent roofs. Moreover, the in-situ survey in San Juan and Ponce provides a deep understanding of life experiences. The dissertation advances the applications of high-spatial and high-temporal resolution satellite images in urban dynamic monitoring, contributing to the development of inclusive, safe, resilient, and sustainable cities.