Abstract:

Amidst the ongoing global warming and the rapid urbanization process in recent years, examining the relationship between extreme temperature events and urbanization in Sichuan Province is crucial for mitigating the risks associated with these events and for informing future urban planning. Using daily temperature records from 36 meteorological stations in Sichuan Province during 1980-2020, along with data on nighttime lights, land use, and spatial grids of economic and population distribution, we calculated 16 extreme temperature indices recommended by the World Meteorological Organization. By applying a combination of K-means clustering and hierarchical clustering, meteorological stations were classified into urban, suburban, and rural categories. The study analyzed the temporal trends of extreme temperature events in relation to urbanization and explored the regional trends across three time periods of 1980-2010, 1980-2015, and 1980-2020, along with the differences across station types. The results showed a rising trend in extreme high-temperature events and a decline in extreme low-temperature events from 1980 to 2020. Urban stations exhibited the most pronounced increases in maximum temperatures, warm days, and diurnal temperature ranges, while rural stations experienced the sharpest decreases in frost days, cold nights, and cold days. Suburban stations saw the greatest increases in summer days and warm nights. The warming trends of extreme temperature events were evident across all three time periods, with rural stations showing a stronger warming tendency compared to urban and suburban stations. Throughout the periods studied, urban and suburban stations recorded more summer days than rural stations, while rural stations observed more frost days. Additionally, the reduction in cold events was less pronounced at urban and suburban stations than at rural stations. These findings underscore both the shared patterns and regional variations in how urbanization influences extreme temperature events, warranting further research into the underlying mechanisms.

Key words: extreme temperature index, cluster analysis, robust local weight regression method, urbanization impact, Sichuan Province

摘要：

在当前全球变暖背景下，近年来城镇化进程加速，研究四川省极端气温事件与城镇化发展的关联，对于降低四川极端温度事件的灾害风险和未来城市规划布局具有重要参考价值。利用四川省36个气象站点1980—2020年气温日值数据、夜间灯光和土地利用遥感数据、经济和人口空间分布网格数据，通过计算世界气象组织推荐的16个极端温度指数，并采用K-means与层次聚类相结合的方法，将气象站点划分为城市、城郊和乡村3类，分析四川省城镇化进程中极端温度事件的时间变化趋势及其与城镇化的关联影响，并探讨1980—2010年、1980—2015年、1980—2020年3个不同时段极端温度事件的区域变化趋势和不同类别气象站点的变化差异。结果表明，四川省1980―2020年极端高温事件呈增加趋势，极端低温事件呈减少趋势；城市站点的最高气温极大值、暖昼日数、月平均日较差增加趋势最明显，乡村站点的霜冻日数、冷夜日数、冷昼日数减少趋势最明显，城郊站点的夏季日数和暖夜日数增加趋势最明显；极端气温事件的增暖趋势在3个时段都较为显著，且乡村站点的增暖趋势相比城市和城郊站点更为明显；3个时段内，城市和城郊站点的夏季日数均多于乡村站点，而乡村站点的霜冻日数多于城市站点，城市和城郊站点的冷事件减少幅度均小于乡村。研究结果揭示了城镇化对极端气温事件影响的共性和区域差异性，对影响机理的深入研究还有待后续进一步分析和讨论。

关键词: 极端气温指数, 聚类分析, 鲁棒局部权重回归法, 城镇化影响, 四川省