開放空間配置：環境規劃與設計支持生物多樣性保育

Abstract

由於土地利用變化日益增加及相關基礎設施的發展，棲息地的破碎化及生態連通性的喪失，已對全球生物多樣性的下降造成重大影響。維護和恢復生態連通性的重要性已在近期的保育政策中獲得認可，2022年《昆明-蒙特利爾全球生物多樣性框架》呼籲通過設定中期（2030年）和長期（2050年）目標，保護並恢復所有生態系統的生態連通性。然而，儘管保育政策承認維持生態連通性的重要性，空間規劃和開放空間配置在將連通性需求納入實際設計過程中仍面臨困難。連通性建模仍是評估連通性的主要工具，但其可信度通常受到缺乏可靠數據、假設的主觀性以及缺乏對動物移動及其生態特徵的實證觀察驗證的影響。這些模型有效性的限制進一步限制了模型結果對景觀規劃者、設計師及進行環境影響評估的實務工作者的適用性。 本研究的目的是透過推進規劃方法與設計策略，提升生態連通性在開放空間配置中的納入，以應對前述挑戰。研究重點並非放在提升模型的複雜度，而是強調基於跨數據集驗證分析結果，並結合物種特定的生態特徵與移動行為，進行生態學知識導向的詮釋。分析結果的驗證確保了結果在生物學上的可解釋性，對實務工作者具透明度，且能直接應用於空間規劃與景觀設計，從而提供更有效且具生態相關性的規劃與設計決策依據。 兩棲動物被選為本研究的目標物種，原因在於其受威脅的狀態、對棲息地改變的敏感性，以及其複雜的生命週期，使其成為適合用來指示棲息地多尺度連通性需求的指標物種。本論文過程中開發了三項方法學創新：（1）混合共識建模，用以識別可作為復育規劃目標的兩棲動物族群潛在位置；（2）結合隨機森林與邏輯迴歸的雙重方法，將規劃與設計變數與生態連通性相關聯；（3）結合溪流階層結構的最大熵物種分布模型，提供河岸移動結構的洞見，且無需假設主觀的阻力面或其他專家參數設定。模型結果以公民科學資料及物種特定的生態知識進行驗證，並產生與規劃設計相關的指導方針及未來研究的生態知識基礎。本研究的貢獻在於透過景觀規劃與設計連通性指導方針，改善開放空間的配置。這些指導方針基於有效利用現有有限資源與資料的創新方法，特別針對研究不足的物種。此外，本研究亦支持符合全球生物多樣性框架要求的空間保育規劃之實施。此項工作在快速發展且景觀高度受限且破碎的地區尤為重要，因為規劃與設計必須考量生態過程與生物多樣性需求，以保障永續發展。

Fragmentation of habitat and loss of ecological connectivity due to increasing land-use changes and associated infrastructure development, have contributed significantly to global decline in biodiversity. The importance of maintaining and restoring ecological connectivity has been recognized in recent conservation policy and the Kunming-Montreal Global Biodiversity Framework (GBF), in 2022 called for protecting and restoring ecological connectivity throughout all ecosystems by setting medium- (2030) and long-term (2050) targets. However, despite the recognition of the importance of maintaining ecological connectivity in conservation policy, spatial planning and open space configurations continue to struggle in implementing connectivity needs into practical design processes. Connectivity modeling continues to be the primary tool used to evaluate connectivity; however, its credibility is generally impacted by lack of reliable data, subjectivity in assumptions, and lack of validation against empirical observations of animal’s movements and of their ecological traits. These limitations on model validity further limit the applicability of model results to landscape planners, designers, and practitioners conducting environmental impact assessments. The purpose of this dissertation was to address the aforementioned challenges by advancing methods of planning and design strategies to increase the inclusion of ecological connectivity in open space configurations. Rather than focusing on increased modeling sophistication, emphasis was placed on ecologically informed interpretations based upon validation of analytical outputs against cross-datasets and species-specific ecological traits and movement behaviors. Validation of analytical outputs provides assurance that results are biologically interpretable, transparent to practitioners and directly applicable to spatial planning and landscape design thus providing an opportunity for more effective and ecologically relevant planning and design decision making. Amphibians were chosen as the target species for this study due to their threatened status, sensitivity to habitat modification and their complex life cycle that makes them suitable indicator species of multi-scale connectivity requirements of their habitats. Three methodological advances were developed during the course of this dissertation: (1) Hybrid Consensus Modeling that identified potential locations of amphibian populations that could be targeted for restoration planning purposes, (2) A dual Random Forest and logistic regression approach to relate planning and design variables to ecological connectivity and (3) Maxent species distribution models that incorporated stream order hierarchy to provide insight into riparian movement structures without requiring the assumption of subjective resistance surfaces and other expert’s parameterizations. All modeling results were validated using citizen science data and species-specific ecological knowledge, and generated planning and design relevant guidelines as well as ecological knowledge base for future research. The contributions of this research lie in improving open space configuration through landscape planning and design connectivity guidelines. These guidelines were informed by methodologies that utilize efficiently the existing limited resources and data of understudied species. The work also supports implementation of GBF compliant spatial conservation planning. This is particularly critical in rapidly developing regions with highly constrained and fragmented landscapes where planning and design must consider ecological processes and biodiversity needs to safeguard sustainable development.