應用地理資訊系統與資料探勘技術於崩塌地特性及空間關連性分析之研究─以莫拉克風災之阿里山溪集水區為例

Using GIS and Data Mining Techniques for LandslideCharacteristics and Spatial Correlations Analysis-A Case Study in Alisan Creek Watershed Struck by Morakot Typhoon

林文賜(Wen-Tzu Lin) ； 黃碧慧(Pi-Hui Huang)

《水保技術》 7卷1期 (2012/01) Pp. 23-29

https://doi.org/10.30040/JSWCT.201201.0003

摘要

九十八年八月八日莫拉克颱風（又稱為八八水災）造成中南部及東部地區嚴重山崩、地滑、土石流、淹水及村毀等災情，本研究係以遭受莫拉克颱風災害之阿里山溪集水區為範圍，結合地理資訊系統及地形單元概念，劃分崩塌地之地形分區及分析崩塌地之空間特性，並以資料探勘技術之K-Means演算法，依地形分區之崩塌地規模分為高、中及低潛勢區，以及輔以DBSCAN演算法分析地形分區之崩塌地空間關連性。分析結果顯示，研究範圍以10公頃面積門檻值劃分為596個崩塌分區，其中高、中及低潛勢區之比例分為4.87%、14.43%及80.70%，利用DBSCAN模式評估顯示高潛勢區可分為4群，中潛勢區可分為5群，而低潛勢區因數量多且分布廣，其群聚效果較不明顯。而高、中潛勢崩塌分區通常都具有相鄰性或同屬上下游連結或之同一支流之緊密關連性，一旦發生二次災害，這些高、中潛勢之群聚區位對集水區將造成較大衝擊。因此，從集水區治理觀念而言，這些群聚之高、中潛勢崩塌分區應予以優先治理。本研究藉由崩塌地空間關連性分析，可評估應優先治理之崩塌區位，研究成果可供集水區整體治理決策及崩塌地防災之參考。

關鍵字

崩塌地 ； 資料探勘技術 ； K-Means演算法 ； DBSCAN演算法

並列摘要

Morakot typhoon, occurred on August 8, 2009, caused serious damages such as villages and bridges destroyed by landslides, debris flow and flood hazards in south-central and eastern Taiwan. The Alisan creek watershed, struck by Morakot typhoon, was chosen as studied area for landslide characteristics and spatial correlations analysis using GIS and data mining techniques. First, the landslide slope units (LSU) were delineating using threshold-area watershed delineation. Then, the LSU were clustered as high, medium and low potential zones by K-Means algorithm, and the spatial correlations for three potential zones were analyzed by DBSCAN algorithm. The studied area was delineated as 596 SLU using the threshold area of 10 ha, in which the percentages for high, medium and low potential zones are 4.87%, 14.43% and 80.70%, especially. Based on the DBSCAN analysis, high and medium potential zones could be clustered as four and five groups. However, clustering effect for low potential zones was not apparent due to the scattered distribution. The high and medium potential zones show tight correlation due to adjacent to each other and linked in the same sub-watershed. While second disasters were occurred, both potential zones will cause a significant impact on the watershed. The landslides at the hazardous zones should be treated in high priority. The analyzed results can be used as decision making of watershed management for the landslide disaster treatment.

並列關鍵字

Landslide ； Data MiningTechnique ； K-Means algorithm ； Density-Based Spatial Clustering of Applications with Noise Algorithm