探討應用空間統計與景觀格局分析於台南縣歷史圖資之時空分析

Exploring the Spatial and Temporal Analysis of Tainan County's Historical Maps by Using Spatial Statistics and Landscape Pattern Analysis

韓威定(Wei-Ting Han) ； 孔憲法(Shiann-Far Kung) ； 陳崇賢(Chung-Hsien Chen)

《航測及遙測學刊》 25卷2期 (2020/06) Pp. 87-114

https://doi.org/10.6574/JPRS.202006_25(2).0003

摘要

本研究透過「臺灣百年歷史地圖」WMTS服務功能，將分別數化能完整覆蓋原台南縣之1904年臺灣堡圖與1989年經建版地形圖第一版，給予歷史圖像資料不同自定義之建成區綴塊作為分析基礎資料，並應用空間統計分析呈現各區域建成區的空間分佈特徵與差異，以及景觀格局分析判識綴塊的數量、面積、空間規模與型態特徵，此外部分類別之歷史統計資料與分析結果具高度相關性，促使在說明研究成果上更具解釋力，並於綜合比較之下得出1904年B與1989年E的自定義建成區相較貼近於實際狀況，雖同年度的不同自定義建成區可發現其空間特徵有著明顯差異，乃不易判定其結果哪個較佳，而是依據研究方向與目的去選擇較符合該框架下的自定義建成區。

關鍵字

台南縣 ； 空間統計分析 ； 景觀格局分析 ； 歷史地圖

並列摘要

With the gradual maturity of geographic information system (GIS) technology, through the combination of GIS and historical data, the study of humanities and society in small-scale spatial regions is more common, but the spatial relationship between various regions in Tainan County is rarely discussed. Therefore, this research digitized the built-up area of Tainan County identified in the 1904 and 1989 historical topographic maps through the WMTS service function of the ＂Taiwan Centurial Historical Map＂ developed by the Academia Sinica. Then customize the different types of built-up areas as the basic data for the method of spatial analysis, landscape pattern analysis, and correlation analysis. This research method used kernel density estimation and hotspot analysis to explore the spatial distribution characteristics and differences of the built-up areas in each region. Besides, the landscape metrics can be used to calculate the number, area, spatial scale and type characteristics of the patch of the built-up areas. Also, the correlation strength and significance between the use of historical statistics and the analysis of analytical results have led to a more explanatory power in the discussion. In conclusion, the built-up areas in the eastern part of Tainan County are generally small and fragmented cold spots, while the western plains are hot spots with large areas. Moreover, the built-up area is highly correlated with the building land of the historical statistical data, indicating that the historical image has high explanatory power for various types of analysis.

並列關鍵字

Tainan County ； Spatial Analysis ； Landscape Pattern Analysis ； Historical Maps