應用UAV空拍技術及HEC-RAS 2D水理模式於河川高灘地植生管理

Application of UAV Aerial Photography Technology and HEC-RAS 2D Hydraulic Model in Floodplain Management

蘇語乾(YU-QIAN SU)

指導教授 ： 施上粟

國立臺灣大學/工學院/土木工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202101304

摘要

人工濕地為都市區域重要的近自然環境，具備水質净化、微氣候調節、生態保育、景觀遊憩等功能。河川高灘人工濕地的水理及棲地條件受到灘地上之植生結構與特徵影響，例如植生密度、種類、樹冠高度等，因此植生是洪氾管理策略中重要的生態因子及水理參數。但許多人工濕地面臨陸化威脅，且密度過高、分佈過廣的植生可能導致河道縮減及水位提高，使洪災風險增加，反而弱化了生態系統防減災的能力（Ecosystem-based Disaster Risk Reduction；Eco-DRR）。本研究針對大漢溪高灘地人工濕地，導入自然解方概念（Nature-Based Solution；NbS），量化分析河川高灘地植生分佈位置、分佈範圍對洪氾影響及生態衝擊程度，以做為人工濕地植生清疏及維護管理的參考。
本研究利用UAV技術擷取研究區域之地形、地貌、地景等資訊，並利用影像判釋技術分析土地覆蓋類型，及建立樹木高度、覆蓋面積、分佈位置等「樹特徵」數據庫，用以建立兩組HEC-RAS 2D模型：傳統曼寧n值模型及植生阻礙物模型。「傳統曼寧n值模型」是以無樹特徵資訊之數值地形模型（DEM）建立水理模式，透過與植生阻礙物模型的模擬結果進行比對，以決定具體的曼寧n值參數；「植生阻礙物模型」則是以樹特徵數據的數值地表模型（DSM）所建立，以植生等阻礙物的特性建立通洪阻礙物（Conveyance Obstructions），用以反應不倒伏高大喬木降低通洪斷面積的阻水效果。研究結果顯示，不同位置的同類型樹木n值並不一致，顯示除了粗糙係數n值外仍有床坡、斷面寬等其他水理條件影響洪水位，其中河道主深槽地形以及斷面阻塞係數BX有著高度相關性。
另外，本研究導入生態河道治理觀念，擬定研究區域內三個區域進行NbS濕地案例的植生管理情境評估。研究發現，「清疏區域 2」為水理高敏感區域，疏伐可以一定程度上降低水流對河道的侵蝕影響，但是會造成該區域水位明顯抬昇，增加洪氾風險。另一方面，「清疏區域 2」為浮洲人工濕地次生林區，為高大喬木分佈面積最廣的區域，此區的次生林提供了重要的生態系統服務之功能，清疏可能造成重大衝擊。經過權衡分析各區植生清疏對生態及防洪的影響程度，「疏 13」方案為最佳方案。

關鍵字

人工濕地 ； 通洪阻礙物 ； 曼寧n值 ； 影像分類及判釋 ； 自然解方

並列摘要

Constructed wetlands are essential near-natural environments in urban areas, which provide wastewater treatment functions, climate regulation, ecological conservation, landscape, and recreation. The flow condition and habitat property of the floodplain are affected by the vegetation structure and characteristics, such as density, coverage, and height. Therefore, vegetation is an essential ecological factor and hydraulic parameter in natural flood management strategy. However, many constructed wetlands encounter the threat of terrestrialization. In addition to excessively high-density vegetation may lead to a reduction of flow conveyance and a increase of water level, which increases the risk of flood disasters weakening the ability of the ecosystem-based disaster risk reduction (Eco-DRR). Unfortunately, there is still a lack of tradeoff research on the causes of vegetation in floodplains. In this study, the concept of a nature-based solution (NbS) was introduced into the Dahan floodplain wetland to quantitatively analyze the impact of tree distribution and coverage area on flood protection and ecological effects as a reference for plant maintenance and management of constructed wetlands.
This study uses an unmanned aerial vehicle to collect floodplain topography, landcover, and tree characteristic data. A modified maximum likelihood classification scheme was developed by incorporating the tree height information to improve the landcover classification. The water-blocking effect of trees, representing the reduced flood conveyance area, was investigated by establishing tree obstructions in hydrodynamic model. Two HEC-RAS 2D models were established, including a conventional model (Adj-n model) and a blocked-obstructions model (BOs model). In the BOs model, Manning's n value of each landcover type was determined by using UAV data, and the land cover types included open water, soil grain, concrete paving, and grass. The Manning's n value of tree patches was set as that of the underlying soil type, as the effects of BOs replaced the roughness caused by trees. The results show that the n values of the same kind of trees in different locations are not consistent. In addition to the value of the roughness coefficient n, there are other factors such as bed slope and section width that affect the flood level. Among them, the topography of the main channel is highly related to the cross-sectional vegetation blockage factor. Besides, this study introduced the concept of NbS and drew up three regions in the study area to evaluate the effects of vegetation management scenarios on flood prevention and ecological conservation. The study found that "Tree-removal Area 2" is a susceptible area of hydraulic. Tree removal can significantly reduce the impact of water flow on river erosion to a certain extent, but it increases the water level in this area and thus raises the risk of flooding. Tree-removal 2 is the secondary forest area of the Fuzhou constructed wetland, the core area with high biodiversity. The secondary forest provides various ecosystem services in this area, and tree removal dredging may cause negative ecological impacts. Based on the analysis of the abovementioned tradeoff from the effects of removing plants in each region on ecological demand and flood control, the "Tree-removal Area 13" reaches the best NbS scenario.

並列關鍵字

constructed wetland ； conveyance obstructions ； Manning's n value ； image classification and interpretation ； NbS

參考文獻

Ambrosia, V., Buechel, S., Wegener, S., Sullivan, D., Enomoto, F., Hinkley, E., Zajkowski, T. (2011). Unmanned airborne systems supporting disaster observations: Near-Real-Time data needs. Int. Soc. Photogramm. Remote Sens, 144, 1-4.

Angela, A., Cohen-Shacham, E., Dalton, J., Edwards, S., Hessenberger, D., Maginnis, S., Maynard, S., McElwee, P., Murti, R., Nelson, C., Ruiz, V., Siikamäki, J., Vasseur, L. (2020). IUCN Global Standard for Nature-based Solutions : first edition.

Arcement, G. J., Schneider, V. R. (1989). Guide for selecting Manning's roughness coefficients for natural channels and flood plains. In: US Government Printing Office Washington, DC.

Baker, R. E. (2012). Combining micro technologies and unmanned systems to support public safety and homeland security. J. Civ. Eng. Archit, 6(10), 1399.

Betsholtz, A., Nordlöf, B. (2017). Potentials and limitations of 1D, 2D and coupled 1D-2D flood modelling in HEC-RAS. TVVR17/5003.