結合遺傳演算法與生物氣候指數在優化胡志明市及未來情境下的熱脆弱性之評估;Integrating Genetic Algorithms and Bioclimatic Indices for Optimal Heat Vulnerability Assessment in Ho Chi Minh City and Future Scenarios

NCUIR > Research Center > Master of Science Program in Remote Sensing Science and Technology > Electronic Thesis & Dissertation > Item 987654321/98472

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/98472

Title:	結合遺傳演算法與生物氣候指數在優化胡志明市及未來情境下的熱脆弱性之評估;Integrating Genetic Algorithms and Bioclimatic Indices for Optimal Heat Vulnerability Assessment in Ho Chi Minh City and Future Scenarios
Authors:	范志強;Cuong, Pham Chi
Contributors:	遙測科技碩士學位學程
Keywords:	熱脆弱性指數;通用熱氣候指數;主成分分析;遺傳演算法;PLUS 模型;局部氣候區;Heat Vulnerability Index;Universal Thermal Climate Index;Principal Component Analysis;Genetic Algorithm;PLUS model;Local Climate Zone
Date:	2025-07-11
Issue Date:	2025-10-17 12:49:03 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著胡志明市 (Ho Chi Minh City, HCMC) 的都市化，造成綠地面積的減少和地表的增溫，加劇了居民暴露於都市熱污染的程度，對不同區塊的社會群體形成了不等的影響與風險.。為正確地評估不同區塊的熱脆弱性，本研究提出了一個綜合性的多尺度框架，用於評估、模擬和優化都市區塊的熱脆弱性，其中通用熱氣候指數 (Universal Thermal Climate Index, UTCI) 使用 ERA5 再分析資料和經過驗證的衛星衍生溫度數據, 透過迴歸模型進行估計，達到了高準確度 (R² = 0.96)。並藉由暴露度 (UTCI)、敏感性和適應能力等指標的整合，構建了一個基於主成分分析 (Principal Component Analysis, PCA) 的熱脆弱性指數 (Heat Vulnerability Index, HVI)，進行胡志明全市各空間區塊熱風險的分析。為了評估都市不同土地利用形態的作用，胡志明市劃分為六個局部氣候區 (Local Climate Zones, LCZs)，其中LCZ 3 (緊密低層建築區) 成為熱脆弱性最重要的貢獻者，在隨機森林模型中佔預測重要性的 83% 以上。在2024年，超過 60% 的居民生活在 HVI 高或最高風險區域，且這一比例將隨著都市的發展而顯著地上升。使用區塊生成土地利用模擬 (PLUS) 模型進行的模擬結果顯示，到 2034 年LCZ 3 將增加 17.8%，特別是在快速發展的地區，例如守德區和第七郡。為了減輕這種熱脆弱性，本研究應用了遺傳演算法 (Genetic Algorithm, GA) 來優化發展計畫下的 LCZ 空間配置。GA 模型成功地將整體適應度值降低了約 12%，並將 HVI 熱點的數量從六個減半至三個。從空間分布的預估結果呈現出高風險和最高風險區域平均減少了 5 平方公里，同時全市低 HVI 區域淨增加了 50 平方公里。此一研究結果對於後續胡志明市的發展策略提供了一個明確的參考訊息，可大幅地降低都會區熱脆弱性與暴露風險。 ;As Ho Chi Minh City (HCMC) continues to urbanize, rising land surface temperatures and reduction of green spaces have intensified residents’ exposure to urban heat stress, posing diverse risks across socio-spatial groups. This study introduces a comprehensive, multi-scale framework for assessing, simulating, and optimizing urban heat vulnerability. The Universal Thermal Climate Index (UTCI) was estimated using a regression model based on ERA5 reanalysis and validated satellite-derived temperature data, achieving high accuracy (R² = 0.96). A Principal Component Analysis (PCA)-based Heat Vulnerability Index (HVI) was then constructed by integrating indicators of exposure (UTCI), sensitivity, and adaptive capacity, enabling spatial mapping of risk across the city. To assess the role of urban morphology, HCMC was classified into six Local Climate Zones (LCZs). Among them, LCZ 3 (Compact Low-Rise) emerged as the most significant contributor to heat vulnerability, accounting for more than 83% of predictive importance in the Random Forest model. Currently, over 60% of residents live in areas with high or highest HVI, and this proportion is projected to rise significantly by 2034. Simulations using the Patch-generating Land Use Simulation (PLUS) model project a 17.8% increase in LCZ 3 by 2034, particularly in rapidly growing districts such as Thu Duc and District 7. To mitigate this heat vulnerability, A Genetic Algorithm (GA) was applied to optimize LCZ spatial configurations under development plans. The GA model successfully reduced the overall fitness value by approximately 12% and halved the number of HVI hotspots from six to three. Spatially, this translated into a 5 km² reduction in high and highest-risk zones and a net gain of 50 km² in low-HVI areas across the city. These findings provide a potential strategy for guiding urban growth in ways that reduce heat vulnerability while supporting continued development.
Appears in Collections:	[Master of Science Program in Remote Sensing Science and Technology ] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	0	View/Open

社群 sharing

Loading...