整合土地利用變遷預測與多指標分析之綠色基盤設施配置分析—以臺北市為例

Abstract

面對氣候變遷與快速都市化的雙重挑戰，都市熱島效應已成為影響城市永續發展的關鍵議題。本研究以臺北市為例，整合土地利用變遷預測模型（Land Change Modeler, LCM）與多指標分析方法（Multi-Indicators Analysis, MIA），建構綠色基盤設施優先配置的決策支援系統。研究採用2000-2024年的土地利用資料，透過多層感知機（MLP）模型預測2030年在基線發展（BAU）、都市化發展（UD）及環境保護（PT）三種情境下的土地利用變遷。同時整合都市熱島效應、空氣品質、社會脆弱性、綠地可及性及景觀連結性五項指標，以等權重及熱島優先兩種權重設定進行空間分析。 研究結果顯示，臺北市可供土地利用轉變的空間極為有限，三種發展情境的差異主要體現在邊緣地區，對整體環境品質的影響相對有限。當都市熱島權重提高至0.4時，高需求區域（第5-6級）顯著擴張，基線情境下第6級網格數從330個增至1,120個，增幅達3.39%。空間分析發現，中山、松山、信義、萬華、大安等傳統高密度建成區在各情境下均維持高需求等級，顯示既有都市結構對環境品質的決定性影響。環境保護情境雖能維持較大面積的低需求區域，但對既有建成區的熱島問題改善有限。 本研究證實在高度都市化環境中，傳統土地利用規劃的效果有限，需要創新的立體化策略和系統性的網絡建構。研究建議：(1)針對高需求區域採用屋頂綠化、立面綠化等立體化措施；(2)保護中等需求區域避免進一步惡化；(3)強化低需求區域與市區的生態廊道連接。本研究成果不僅為臺北市提供科學化的規劃依據，也為其他面臨類似挑戰的亞洲高密度城市提供可複製的方法論架構。

Facing the dual challenges of climate change and rapid urbanization, the urban heat island effect has become a critical issue affecting urban sustainable development. This study takes Taipei City as a case study, integrating Land Change Modeler (LCM) and Multi-Indicators Analysis (MIA) to establish a decision support system for green infrastructure priority allocation. Using land use data from 2000-2024, this research employs Multi-Layer Perceptron (MLP) models to predict land use changes in 2030 under three scenarios: Business as Usual (BAU), Urban Development (UD), and Environmental Protection (PT). The study integrates five indicators: urban heat island effect, air quality, social vulnerability, green space accessibility, and habitat connectivity, conducting spatial analysis with both equal weights and heat island priority weight settings. Results reveal that available land for conversion in Taipei City is extremely limited, with differences among the three scenarios mainly manifested in peripheral areas, having relatively limited impact on overall environmental quality. When the urban heat island weight is increased to 0.4, high-demand areas (levels 5-6) expand significantly, with level 6 grids increasing from 330 to 1,120 under the baseline scenario, representing a 239% increase. Spatial analysis identifies that traditional high-density built-up areas including Zhongshan, Songshan, Xinyi, Wanhua, and Daan districts maintain high demand levels across all scenarios, demonstrating the decisive influence of existing urban structure on environmental quality. While the environmental protection scenario maintains larger low-demand areas, improvements to heat island problems in existing built-up areas remain limited. This study confirms that traditional land use planning has limited effectiveness in highly urbanized environments, necessitating innovative vertical strategies and systematic network construction. The study recommends: (1) implementing vertical measures such as rooftop and facade greening in high-demand areas; (2) protecting medium-demand areas from further deterioration; (3) strengthening ecological corridor connections between low-demand areas and urban centers. The research outcomes not only provide scientific planning basis for Taipei City but also offer a replicable methodological framework for other high-density Asian cities facing similar challenges.