Studies on Forest Succession under Natural Disturbances with Landscape Model -An Example of Jiou-Jiou Peaks Nature Reserve

• Main Authors: Jeng-I Tsai, 蔡正一
• Other Authors: Fong-Long Feng
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/7cp9pr

博士 === 國立中興大學 === 森林學系所 === 102 === 921 earthquake how to impact the landscape and vegetation in steeper areas of Jiou-Jiou Peaks, and the restoration and variation of landscape and vegetation over these years are curious. Jiou-Jiou Peaks Nature Reserve was established in 2000 to protect the unique topography. Although human disturbances were prohibited, the landscape and vegetation succession were still influenced by temporal and spatial factors. We integrated landscape model, terrain factors and field survey data to realize the landscape change, temporal and processing of forest restoration, niche of Arundo formosana, forest succession and prediction of future condition. Ortho-aerial photographs, taken in 1998/07, 1999/12 and 2009/12, were applied to digital land-use / land-cover maps (LU / LC map), including forest, A. formosana, bared soil, water body (including river side) shadow and others, based on image tone (color) and texture. These maps could be become ground truth for SPOT analysis. We also overlapped the maps to analysis the landscape change preliminary. Results of landscape change showed 126.06 ha (10.56％) of forest never change after earthquake to 2009. 114.78 ha (9.57 ％) from the forest became bared soil, and then developed into A. formosana until 2009. Moreover, 10.65 ha (0.89 ％) areas of rivers side was changed from forest during 1999 to 2009 might because of typhoons. Multi-temporal SPOT images, taken in 1999/07, 1999/09, 2000/09, 2001/07, 2006/02, 2008/08 and 2012/09, were collected. The supervised classification, image differencing algorithm and overlapped analysis were applied to estimate the temporal of forest restoration (forest age) and proccessing. A map showing the forest estimation age and location was generated to recognize the tendency of forest changes. The result showed that the forest restoration occurred from the bottom to the top. According to landscape indices and metapopulation theory, the restoration condition of this reserve was nonequilibrium metapopulation, which means almost no relations between patchs and patchs of forest. A. formosana can hang downward grow and thrive from cliffs because of developed roots. Almost all the steeper slopes of the reserve were covered by this species, but knowledge of this species’ niche was lack. Therefore, we applied data of LU / LC and topographic factors to analyze the niche of A. formosana with two way-ANOVA and cluster analysis. The result exhibited that surfaces facing south and southeast, and slopes from 74° to 78° were were more suitable for the growth of A. formosana. We also think the species might help the succession. The sampling locations were chosen with forest estimation age map, Neyman and Deming methods. We tried to identify forest succession based on field survey data of sampling locations near Gan River. We found the species compositions varied temporally, and the compostion was gathered by heliophilous species. The forest succession sere could not recognize completely in such short periods. Finally, all results were integrated, including landscape analysis, field survey data and topographic factors, to predict the future condition with CA-Markov model. Althought the spatial distribution of simulation do not match the actual situation perfectly, the simulation area was believed. The forest area might be stable at 2029 around 504.20 ha (42.05 ％) but forest will be difficult to recover as pre-earthquake condition.