A Study of Characteristics for Wildfire Potential Sites Using Drought Indices

• Main Authors: Chiao-Jou Hsieh, 謝巧柔
• Other Authors: 林昭遠
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/17729448620778250798

碩士 === 國立中興大學 === 水土保持學系所 === 103 === By the lack of advocacy of “Leave No Trace” concept, human disturbance continuously causes the different levels of stress to the natural environment. Since 1969, the East trail of Mt. Xue inaugurated in the Shei-Pa National Park has attracted many mountaineers due to the integrated recreation facilities, and served as the main path to the Mt. Xue in the snow season, provided the approachability of human activities. However, previous researches show that several wildfire events had occurred along the trail, affected the forest stands and caused the territory expansion of the fire adapted species which would possibly increase the opportunity of wildfire occurrence. Most of the wildfire events were ignited by the inappropriate human activities such as picnic, smoking, camping and garbage burning. How to use environment indices delineating the spatial and temporal distributions of the potential wildfire to provide the precise locations and proper time for the purposes of fire prevention and strategies making is crucial. Normalized Difference Vegetation Index (NDVI), Normalized Difference water Index and Normalized Multi-band Drought Index (NMDI) derived from radiometric and topographic corrected multi-phase satellite imageries were applied in this study. A map of twelve-phase NDVI variation curves plotted using different land-cover classifications after k-mean analysis, and the analysis of monthly variation from the map coupled with the coverages of national resources inventory to acquire the ground truth information for each land-cover classification were the process of NDVI application in the research. The value of NMDIveg <0.4 could be as a threshold to delineate the spatial distribution of wildfire potential site, which can then be employed to compare with the same seasonal scatter maps of NDWI. Results showed that there will be more accuracy for the delineation of wildfire potential extracted from NMDI. Comparing the wildfire potential sites with the burned locations, which derived from Normalized Burn Ratio, the correct classification rate of 89.9% can be expected in this study.