Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Land Surveying and Geo-Informatics | en_US |
| dc.contributor.advisor | Wong, Man Sing (LSGI) | en_US |
| dc.creator | Adeniran, Ibrahim Ademola | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/13515 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Development of an integrated remote sensing and machine learning model for urban heat island analysis | en_US |
| dcterms.abstract | Climate change has become an undeniable reality, with its effects continuing to manifest in the present day. As global temperatures rise and weather patterns shift, the need for this study arises from the significant challenges in obtaining sufficient near-surface air temperature (Ta) data for urban thermal environmental studies, which are crucial for addressing both global and local concerns. Accurate Ta data is essential for understanding and mitigating urban heat islands at the canopy layer (UHICL), this study thus aim to addresses the barriers hindering the effective utilization of remotely sensed data for Ta both at local and global city level. | en_US |
| dcterms.abstract | Initially, a global-scale city-based Ta prediction model was developed for UHICL analysis. A representative sample of 30 cities was selected using stratified random sampling to ensure diverse climate, geographic, and socioeconomic representation. The model utilized NCDC average daily Ta data, supplemented with crowdsourced Netatmo CWS data, addressing the scarcity of urban weather stations. Land Surface Temperature (LST) data from the MODIS satellite, combined with auxiliary data, we reused to build the Ta prediction model. An ensemble model, derived from optimal prediction models for each city, demonstrated high accuracy and outperformed contemporary models. The predicted Ta was employed to estimate UHICL for the sampled cities. | en_US |
| dcterms.abstract | For Ta study within city at local scale, Hong Kong was chosen as area of study due to it high density and heterogenous landscape. Then to generate Ta at fine spatial and temporal resolution suitable for locale scale UHICL analysis, LST data were sourced from the thermal infrared (TIR) band of both sun-synchronous (Landsat-8 and Sentinel 3-SLTR) and geostationary satellite which served as primary data for this study. Cross comparison was then carried out to assess the effect of retrieval algorithm and time of the day on the harmonized use of LST from the respective data. Following the cross-comparison study to address the limitations in existing LST fusion model such as missing data in complex weather regions and increased bias due to land cover changes this study proposed the Integrated Spatiotemporal Fusion AlgoriThm (ISFAT). ISFAT is a three-stage algorithm for seamless fine scale LST prediction at hourly temporal resolution. ISFAT significantly improved the spatial and temporal resolution of LST data, making it suitable for local-scale thermal studies. | en_US |
| dcterms.abstract | Finally, the Federated Learning Neural Network (FLANN) framework was adapted for predicting Ta at a local city scale, addressing the uneven distribution of weather stations in Hong Kong. Weather stations were classified based on Local Climate Zones (LCZs), and spatially continuous Ta was predicted using the FLANN framework. This decentralized approach prevented overfitting or underfitting due to uneven station distribution. The FLANN model demonstrated the highest accuracy in Ta prediction outperforming sate of the art multiple layer perceptron in the prediction of fine scale Tair from satellite data recording an r2 of 0.95 against 0.92 recorded by the MLP and was used for UHICL and Urban Heat Vulnerability Index (UHVI) assessments, facilitating effective hotspot identification and mitigation strategy development. | en_US |
| dcterms.abstract | This study provides valuable tools and frameworks for both global and local-scale urban thermal analysis. The advancements contribute to more sustainable and resilient urban environments, better informing urban planning, climate mitigation strategies, and public health policies. | en_US |
| dcterms.extent | xxv, 218 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2024 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Urban heat island -- Remote sensing | en_US |
| dcterms.LCSH | Urban climatology -- Remote sensing | en_US |
| dcterms.LCSH | Urban geography -- Data processing | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13515