城市群建成区变化对台站温度记录影响的多源遥感研究

One of the unsolved issues in climate change analysis is the contribution of urbanization signal to fluctuation and trends of air temperatures recorded at meteorological stations, especially related to previous methods using rural stations as reference. Here we assume that air temperature records are largely contaminated by changing fraction of urban built-up, instead of fixed urban areas. With support of long-term satellite data and air temperature time series, this study proposes a dynamic way to quantify the contribution of urban signals using urbanization information of every station acquired from remote sensing data and in-situ measurements, instead of defining reference rural stations. This study examined the impacts of urban expansion on the trends of air temperature around meteorological stations in major urban clusters, including Beijing-Tianjin-Hebei, Yangzi River Delta, Pearl River Delta, and Chengdu-Chongqing over last three decades. Two different spatial scales, 8km and1km, will be considered to examine the impact of urbanization at local scales and relative contribution of urban heat island effects to the temporal trends of temperature records. Possible errors produced during this analysis will also be considered and a statistical test will be developed to examine the credibility of key results from the study. Key science question is how much urban signal may contribute to local air temperature records and how well those contaminated records can be used to represent regional temperature trends.

学术界在认识台站温度记录中城市化的影响这一问题上仍有较大争议，特别是以乡村参考站代表背景增暖信号，去除城市化增暖信号的方法。我们假设台站气温的年际变化趋势在很大程度上被城市建成区变化率，而不是被静态的面积比率所影响。进而利用多时段卫星遥感数据序列分析，提出了一种动态的量化建成区变化对台站温度趋势影响的方法。项目拟利用30年多时段的卫星遥感数据和国家气象信息中心最新订正发布的全国台站60年逐日温度记录，提出了一种动态的量化自然背景增暖的方法，即利用区域内所有站点从遥感影像得到的城市化速率，结合台站记录的气温变化趋势，综合统计出背景增暖。以京津冀、长三角、珠三角、成渝等4个主要城市群为实验研究区，通过遥感土地利用分析、野外强化观测、和台站温度时间序列分析来研究气象站点周边过去三十年的城市化速率对温度记录和趋势的贡献，试图回答处于城市化进程中的台站温度数据在多大程度上可以代表区域变化特征。

台站温度记录中的城市化信号对于气候变化研究影响重大并仍存在争议，尤其是在经历快速城市化的区域。以乡村参考站代表背景增暖信号，去除城市化增暖信号的方法未充分考虑城市动态扩张过程，本项目假设台站气温的年际变化趋势在很大程度上被城市建成区变化率，而不是静态的面积比率所影响。项目借助多时段卫星遥感数据获得气象台站周边城市土地利用变化信息，提出了一种动态量化自然背景增暖的方法，建立了台站周边城市土地利用比率变化与温度变幅之间的定量关系模型，定量评估了全国和城市群尺度台站温度记录中城市化影响和城市化影响贡献率。中国区域90%的气象台站周边发生了城市土地利用扩张，评估得到的全国台站周边平均城市扩张速率为0.33%/yr。中国由北向南分布的三个特大城市群，京津冀城市群、长江三角洲城市群和珠江三角洲城市群台站平均城市扩张速率高于全国平均水平，分别达到0.44%/yr，0.53%/yr和0.36%/yr。在全国尺度，年平均温度城市化影响为0.034±0.005"°C/10a" ，城市化影响贡献率为7.54%~10.82%；在城市群尺度，长江三角洲年平均温度记录中城市化影响贡献率最高（16.52%~21.7%），珠江三角洲次之（8.20%~20.70%），最后是京津冀城市群（8.16%~14.53%）。