Affecting Factors of Vegetation Recovery at the Post-Fired Sites in a Watershed

• Main Authors: Hang-Hsun Chan, 詹杭勳
• Other Authors: 林昭遠
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/95496873132170944203

碩士 === 國立中興大學 === 水土保持學系所 === 103 === 　　Vegetation recovery assessment model for a fire site can be applied to estimate the vegetation development of the post-fire land, and understand the crucial factors which affect the following vegetation recovery, and it is beneficial for the use of post-fire recovery planning. This study tries to compare the factor’s ability to evaluation, analyze their association and feasibility using literature review for the references of factor priority selection in the model. 　　Wildfire event at the Tataka of Yushan National Park in 1993 was selected as case study. With literature review, remote sensing data collection, geographic information systems and image process techniques coupled with hierarchical regression and association analysis, and factors such as geology, soil, aspect, slope, terrain wetness, altitude, pre-fire land-cover, burn severity and seeding distance from unburned area were surveyed from literatures to explore affecting factors of vegetation recovery at the post-fired sites. Finally, trend analysis was applied to determine the factor which could be served in the models for restoration assessment. 　　Soil, altitude and burn severity are the suitable factors that can be used in the models to assess vegetation recovery at the post-fired sites. Factors of Slope and pre-fire land-cover are recommended to be used in the models under the condition of more than two-year’s succession. However; factors of geology, aspect, terrain wetness and seeding distance from unburned area should be excluded in the models for the sake of association and applicability consideration. Dummy regression models built with the variables of soil, altitude, burn severity, slope and post-fire land-cover show that the determination coefficient increases with time to plant succession, and the value can reach to 0.60 derived from the burning sites after six year’s plant succession, and the value of determination coefficient for this 5- variable model only less than 0.05 comparing to the model with 9- variable derived from literature. This reveals that the model with 5 effective factors could be employed to assess the vegetation recovery at the post-fired sites.