應用層級分析法於利害關係人對土地利用變遷的認知

Abstract

全球土地利用在過去幾十年的變遷改變地表及近地表的生物物理特性，進而影響地球的物質與能量循環、生物環境，而產生糧食安全、疾病、生態系服務劣化等風險。因此科學家發展土地利用模式，試圖理解土地利用變遷的複雜過程並能輔助進行土地利用規劃，以達成環境與人類需求的平衡。然而現有的土地利用模式研究大多僅止於產出模擬結果，少有將模式的模擬透過邀請利害關係人或決策者的參與用於後續土地利用決策中，國內也較少研究結合土地利用模式與利害關係人參與並強化兩方的理解與連結。因此，本研究以土地利用類型多元的濁水溪流域為研究區域，利用層級分析法問卷調查利害關係人對土地利用變遷的認知，並分別與土地利用模式CLUE-S的適宜性權重及轉移彈性等參數進行比較，討論利害關係人及模式在土地利用變遷上的認知差異。 本研究利用層級分析法問卷，分別向一般民眾調查其對土地利用變遷驅動力重要性的認知，以及向決策者調查對不同土地利用類型轉移容易程度的看法。前者的驅動力重要性權重與土地利用模式邏輯迴歸的驅動力適宜性係數進行比較，後者的轉移彈性則作為參數納入土地利用模式中模擬2050年的土地利用，同時也與過往所設定轉移彈性的模式透過效率指數進行模擬表現的比較。 民眾問卷土地利用變遷驅動力權重與模式邏輯迴歸適宜性係數比較的結果顯示，對農地而言，民眾認為水源是重要的驅動力因素，但可能相對忽略與道路距離代表的交通因素的重要性；此外，社會因素的多元性及對各種土地利用類型的重要程度也值得持續深入探討。將決策者問卷的轉移彈性結果納入土地利用模式進行模擬，其驗證所得到的效率指數值為0.1022，與過往使用的轉移彈性參數的模擬結果相比表現較好；然而因成對比較矩陣計算產生的轉移彈性值較小，代表各類土地利用皆容易轉移，因此模式更容易受到未來土地利用需求面積、土地利用變遷驅動力等參數影響土地利用的轉移，此時其他模式參數的設定品質及利害關係人參與架構則能成為之後關注的重點。 本研究為國內首篇將利害關係人對土地利用變遷的認知與土地利用模式運作過程中的參數進行比較的研究，期許能增進利害關係人對土地利用變遷動態過程的理解，也提升土地利用模式模擬的表現，作為後續利害關係人參與土地利用規劃相關研究的參考。

Global land use change in the past few decades has altered the biophysical characteristics of the Earth's surface and near-surface, thereby affecting the material and energy cycles and the biological environment of the planet. This, in turn, has resulted in risks such as food security, diseases, and degradation of ecosystem services. In response, scientists have developed land use models to better understand the complex processes of land use change and assist in land use planning to achieve a balance between environmental and human needs. However, current land use modeling studies often focus solely on producing simulation results, with limited incorporation of stakeholder involvement or decision-makers in subsequent land use decision-making processes. There is also a lack of research in combining land use models with stakeholder engagement to enhance mutual understanding and linkage between the two in Taiwan. Therefore, this study focuses on the diverse land use types in the Jhuoshuei River Basin as the study area. It utilizes the Analytical Hierarchy Process (AHP) questionnaire survey to investigate stakeholders' perceptions of land use change. The study compares these perceptions with the suitability coefficients and elasticity parameters of the CLUE-S model, aiming to explore the differences in perceptions of land use change between stakeholder and model. This study utilizes the AHP questionnaire to survey the general public's perception of the importance of drivers of land use change and the views of decision-makers on the ease of transfer for different land use types. The importance weights of the drivers from the public questionnaire are compared with the suitability coefficients of the logistic regression in the land use model. The elasticity from the decision-makers' questionnaire is incorporated as a parameter in the land use model to simulate land use in 2050. These values were also compared with past elasticity settings using figure of merit to evaluate model performance. The comparison between the weights of land use change drivers from the public questionnaire and the suitability coefficients from the logistic regression model shows that, for agricultural land, the public perceives water availability as an important driver but may overlook the importance of transportation factors represented by distance to road. Moreover, the diversity of social factors and their importance for various land use types deserve further exploration. Incorporating the elasticity results from the decision-makers' questionnaire into the land use model for simulation, the figure of merit obtained during validation was 0.1022, demonstrating better performance compared to simulations using past elasticity parameters. However, because the elasticity values generated through pairwise comparison matrix are relatively small, it indicates that all land use types are more easily interchangeable. Therefore, the model is more susceptible to the influence of parameters such as future land use demand areas and land use change drivers on land use transitions. At this point, the quality of other model parameter settings and the framework for stakeholder involvement can become important considerations for future research. This study is the first in Taiwan to compare stakeholders' perception of land use change with parameters in land use modeling. It aims to enhance stakeholders' understanding of the dynamic process of land use change and improve the performance of land use modeling. The findings can serve as a reference for future studies engaging stakeholder in land use planning.