應用大數據與環境指標探討氣候變遷下台灣西南部海岸林帶變遷

海岸林具有防風、安定飛砂、減輕鹽霧侵蝕及保護農作物功能，是臺灣沿海地區抵抗天然災害之重要綠色防線，但受到生育地天然環境壓力、病蟲害及長期缺乏維護管理，逐漸呈現衰退現象，而林木天然更新不易，再加上人為干擾導致土壤鹽化，嚴重影響整體海岸林帶的環境保護功能。近年來因氣候變遷影響導致缺水與高溫日曬等環境壓力增強，更不利海岸林的生長，對海岸林的經營管理形成高難度挑戰，應透過科學方法，對海岸林應有的配置區位及其壓力來源予以釐清。本研究應用多時序衛星影像結合環境指標、氣候網格資料、IPCC之AR5報告中之RCP情境及隨機森林模型，分析臺灣西南部之雲林縣、嘉義縣及臺南市的海岸林衰退及變遷原因並進行變遷趨勢模擬，探討氣候變遷對海岸林環境各層面的影響，提供擬訂因應策略之依據。研究結果顯示，3縣市之海岸林帶面積均明顯不足，因應氣候變遷衝擊能力有限。以隨機森林模型推估結果，在考量人為與氣候因子模型中，NDBI為最重要影響因子，而溫度與降雨是重要特徵因子。在IPCC之AR5中的RCP情境模擬中，模擬推估2020年至2050年之海岸林衰退及變遷結果，3縣市總體的密生森林面積，在氣候變遷的狀態下是呈現增加趨勢，灌木、草生地或農作物面積則是呈現下降的趨勢。

關鍵字

海岸林；氣候變遷；環境指標；氣候網格資料；隨機森林

並列摘要

Coastal forest in Taiwan has functions of preventing wind, stabilizing sand, reducing salt fog erosion and protecting crops. Coastal forest is also an important ＂green defense line＂ against natural disasters in coastal areas. However, due to natural environmental stress, diseases and insect pests and lack of maintenance and management for while, coastal forest gradually decline, and natural regeneration of trees is not easy. Additionaly, soil salinization caused by anthropogenic interference would affect the environmental protection of coastal forest belt. In recent years, due to climate change, the change in precipitation pattern and the increasing uncertainty in rainfall has led to the increase of environmental stress such as water shortage and high temperature and sun-exposure, which endamage to the growth of coastal forests and poses a difficult challenge to the management of coastal forests in various regions of Taiwan. Given this, it is necessary to clarify the proper allocation location and stress source of coastal forest through scientific methods. In this study, multi-temporal Landsat satellite images were used to identify coastal forest belts in three counties/cities in southwest Taiwan, such as Yunlin County, Chiayi County, and Tainan City. Combined with environmental indicators, gridded data of climate, RCP scenarios in AR5 in the , and Random Forest model, the causes for declining coastal forest were discussed To understand the uncertainty and far-reaching impact of climate change on coastal forest, we can also provide a phased response strategy, measures, and action plan. The results show that the area of coastal forest belts is insufficient in the study area, and the ability to cope with the impact of climate change is relatively limited. Based on the results of Random Forest model, NDBI is the most important controlling factor, and temperature and rainfall are important characteristic factors within the model for considering human and climate factors. Under the RCP scenario in AR5, the results showthat overall dense forest area of three counties/cities would increase, shrubs, grassland or crops would decrease from 2020 to 2050.

並列關鍵字

Coastal forest ； Climate change ； Environmental indicator ； Climate grid data ； Random forest

參考文獻

Alonso-Rego, C.; Arellano-Pérez, S.; Guerra-Hernández, J.; Molina-Valero, J.A.; Martínez-Calvo, A.; PérezCruzado, C.; Castedo-Dorado, F.; González-Ferreiro, E.; Álvarez-González, J.G.; Ruiz-González, A.D. (2021) “Estimating Stand and Fire-Related Surface and Canopy Fuel Variables in Pine Stands Using Low-Density Airborne and Single-Scan Terrestrial Laser Scanning Data,” Remote Sensing, 2021, 13(24), 5170. https://doi.org/10.3390/rs13245170.

Ayrey, E.; Hayes, D.J.; Kilbride, J.B.; Fraver, S.; Kershaw, J.A., Jr.; Cook, B.D.; Weiskittel, A.R. (2021) “Synthesizing Disparate LiDAR and Satellite Datasets through Deep Learning to Generate Wall-to-Wall Regional Inventories for the Complex, Mixed-Species Forests of the Eastern United States,” Remote Sensing,2021, 13, 5113. https://doi.org/10.3390/rs13245113.

Breiman, L. (2001) “Random forests,” Machine learning, 45, 5-32. https://doi.org/10.1023/A:1010933404324.

Chen, D.; Wang, Y.; Shen, Z.; Liao, J.; Chen, J.; Sun, S. (2021) “Long Time-Series Mapping and Change Detection of Coastal Zone Land Use Based on Google Earth Engine and Multi-Source Data Fusion,” Remote Sensing, 2022, 14, 1. https://doi.org/10.3390/rs14010001

Giuliani, G., Chatenoux, B., De Bono, A., Rodila, D., Richard, J. P., Allenbach, K., Dao, H., & Peduzzi, P. (2017) “Building an Earth Observations Data Cube: lessons learned from the Swiss Data Cube (SDC) on generating Analysis Ready Data (ARD), Big Earth Data”, 1:1-2, 100-117, DOI: 10.1080/20964471.2017.1398903

主題瀏覽