基于时序解混模型的植被物候反演及对城市热岛的时空响应研究

Influence of urban heat island on vegetation phenology and the ecosystem is the current hot topic of global change research. Urban heat island as a harbinger of global change, can reflect the influence of global changes on the plant growth, ecosystem and human health and provide information to develop policies for its negative impact. In the study of vegetation phenology, most studies focus on phenological changes of specific plant species, but there is lack of ecological system vegetation phenology research on a macroscopic scale. Shanghai is taken as study area and is divided into the central urban, suburbs and outer suburbs according the spatial position. The NDVI per eight days is calculated based on MODIS 09Q1. The timing curve of different vegetation phenology (one year once and one year twice phenology) is inversed by empirical orthogonal function and temporal unmixing model based on MODIS NDVI. The vegetation phenology time is determined by maximum sloped method and is validated by remote sensing data. The land surface cover is derived by Landsat data. Urban island effect in Shanghai is inversed based on Landsat data and Aster data, and the maximum, minimum and mean surface temperature is calculated in every annular zone. The driving effect of land surface cover on heat island effect is analyzed. The influence of urban heat island on vegetation phenology is analyzed using linear and Geographically Weighted Regression to provide a basis for revealing the urban ecosystem evolution.

城市热岛对植被物候及其对生态过程的影响是当前全球变化研究的热点，城市热岛作为全球变化的预兆，可以及早体现全球变化对植物生长规律、生态系统和人体健康的影响，进而提前对其负面影响作出相应。当前在城市植被物候的研究上，多数的研究只关注于特定植物的物候变化，而缺少城市化对植被物候变化的分析。本研究选择上海作为研究区，将其分为中心区域，近郊和远郊，利用MODIS 09Q1数据计算每8天的NDVI，基于经验正交函数和时序解混模型提取三个区域不同物候类型（一年一季和一年两季）的NDVI时序曲线；利用最大斜率法确定NDVI，一年一季NDVI和一年两季NDVI的生长季始、末期，生长季长度，并利用遥感数据验证精度。利用Landsat和Aster数据反演上海地表覆被和温度；分析上海地表覆被变化对热岛效应的驱动作用。基于统计分析地表温度对植被生长季始期、末期和长度的影响，为揭示城市生态系统的演化规律提供依据。

人类活动通过改变土地用途，是地球上环境变化的最根本来源。城市化和由此产生的“城市热岛”为评估气候变暖对植被物候的影响提供了一种手段。城市环境升高的地表温度和空气温度会影响城市边界内和城市边界之间植被生长动态的时间。物候和相关的冠层发育对地球表面和大气之间的季节性能量和质量交换具有强大的控制作用。卫星测量用于评估全球范围内植被的主要物候期。本研究聚焦上海市中心城区、郊区和远郊城市化梯度上植被物候的变化。上海中心城市的NDVI呈现增加趋势，地表温度减少；郊区的NDVI呈现减少趋势，地表温度增加。上海中心城区的物候起始期和峰值期在近20年间呈现提前的趋势，但总体时间物候起始期和峰值仍晚于郊区和远郊。城市化梯度上的物候变化研究与空间尺度紧密相关，因此在升尺度和降尺度上分别定量比较植被物候变化。