利用觀測為主之網格化資料分析台灣熱壓力與人口暴露之時空變化

Abstract

在全球氣候變遷影響下，熱壓力已成為威脅健康與生產力的重要風險。台灣位處高溫高濕的副熱帶氣候區，地形複雜、都市密集與人口高齡化，對熱環境變化尤為敏感。然而，過往研究多受限於觀測資料解析度或聚焦於特定區域，難以在全台尺度下量化熱脆弱性的空間分布，可能低估實際風險。 本研究首次整合中央氣象署的五項氣象網格變數（氣溫、相對濕度、10米風速、地表氣壓與向下短波輻射），透過物理機制模型（Liljegren模型）建構全台逐時、1公里高解析度的濕球黑球溫度（WBGT）網格資料。該資料作為分析2015–2023年台灣暖季（含過渡季節，4月至10月）白天時段（08:00-17:00 LST）熱壓力時空變化特徵及人口暴露風險的核心依據。 研究結果顯示，台灣熱壓力分布具有顯著的地形依賴性，高WBGT值主要集中於西部平原與花東縱谷等區域。趨勢分析指出，2015–2023年間全台平均 WBGT 呈顯著上升趨勢，極端熱危害（WBGT ≥ 32 °C）的發生頻率與影響範圍亦同步擴大。除西南平原等傳統高風險區外，熱壓力增幅最快的區域已逐漸延伸至山麓地帶及花東縱谷。其中，花東縱谷在後期（2020–2023年）的每月極端熱日數，較前期（2015–2018年）增加約6–8天。此趨勢主要受整體氣溫上升、過渡季節（4月與9月）短波輻射增強，以及近地表風速普遍減弱等因子共同驅動。 在人口暴露評估方面，本研究整合1公里解析度之分齡人口資料，發現高暴露風險高度集中於西部都會走廊。其中，65歲以上老年族群落於極端熱區之空間比例為各年齡層最高，顯示脆弱族群之空間分布與高熱危害區域具有高度重疊性。 本研究建立全台首個高解析度 WBGT人口暴露評估框架。研究結果顯示，WBGT指標能有效補足傳統乾球溫度預警閾值（Temp ≥ 36 °C）之限制，更全面地反映濕熱環境下之人體所承受的熱壓力與潛在健康風險。本研究成果可作為未來發展在地化熱預警制度、公共衛生介入措施，以及氣候調適策略的重要科學依據。

Under ongoing global climate change, heat stress has emerged as a major threat to human health and labor productivity. Taiwan, located in a hot and humid subtropical climate zone, is particularly sensitive to thermal stress due to its complex topography, dense urbanization, and rapidly aging population. However, previous studies have often been constrained by limited observational resolution or a focus on specific regions, making it difficult to quantify the spatial distribution of heat exposure and its intersection with demographic structure across Taiwan at the national scale. This limitation may lead to an underestimation of actual heat-related risks. This study examines the spatiotemporal variability of humid heat stress and associated population exposure across Taiwan during the warm-season daytime period (April–October, 08:00-17:00 LST) from 2015 to 2023, using observation-based meteorological data at 1-km spatial resolution. Five key gridded meteorological variables (air temperature, relative humidity, 10-m wind speed, surface pressure, and downward shortwave radiation) from the Central Weather Administration (CWA) were integrated to construct, for the first time, an hourly Wet-Bulb Globe Temperature (WBGT) gridded dataset covering Taiwan’s main island. The results indicate a pronounced topographic dependence of heat stress in Taiwan, with high WBGT values primarily concentrated in the western plains and the Hualien-Taitung Rift Valley. Trend analyses reveal a significant increase in mean WBGT across Taiwan during 2015–2023, accompanied by a concurrent expansion in the frequency and spatial extent of extreme heat hazards (WBGT ≥ 32 °C). While the southwestern plains remain the dominant heat hotspot throughout the study period, the most rapid intensification has extended toward foothill regions and the Hualien-Taitung Rift Valley in recent years. In particular, the Hualien-Taitung Rift Valley experienced an increase of approximately 6–8 extreme heat days per month during the late period (2020–2023) compared with the early period (2015–2018). These changes are jointly driven by global warming, enhanced downward shortwave radiation in April and September, and a widespread weakening of near-surface wind speed. Population exposure was assessed by integrating extreme WBGT hazard frequency with 1-km gridded population data stratified by age groups. Areas of high exposure are primarily concentrated along the western urban corridor. Among all age groups, older adults (≥ 65 years) exhibit the highest spatial proportion within extreme heat exposure hotspots, indicating a strong spatial overlap between climatic heat hazards and population aging. In conclusion, this research presents the first high-resolution WBGT-population exposure assessment framework for Taiwan. The findings demonstrate that WBGT complements existing dry-bulb air temperature thresholds (Temp ≥ 36 °C) by capturing additional physiological heat risks across broader spatial extents and higher seasonal frequencies. These results provide a high-resolution scientific basis for refining heat-health warning systems and supporting targeted climate adaptation strategies in Taiwan.