武汉市植被物候变化规律及影响因素分析

植被物候作为植被生长响应气候变化的指示器，对于研究气候变化以及城市化进程具有重要意义.城市内部植被通常分布破碎，因此公里级的低分辨率遥感影像难以实现植被的精细识别与分析，而十米级空间分辨率的遥感数据在时间分辨率方面又难以满足物候分析的要求.为此，该文基于遥感数据时空融合技术，缓解高时间分辨率与高空间分辨率之间的矛盾，进行城市植被物候变化规律的分析.基于非局部滤波融合方法，生成武汉市空间分辨率30 m、时间分辨率8 d的地表反射率及EVI(增强型植被指数)序列，进一步采用移动加权谐波分析方法对EVI序列进行重建，并通过动态阈值方法提取2006年～2014年武汉市植被物候信息.实验结果表明：1) 武汉市植被由中心向郊区呈现生长期开始时间(SOS)逐渐推迟、结束时间(EOS)逐渐提前、生长期长度(LOS)逐渐延长的空间分布规律，且整体呈现出SOS提前、EOS推迟、LOS延长的时间变化趋势；2) 植被物候和平均气温相关性并不显著，但EOS和LOS受气温年平均日较差影响显著，气温年平均日较差每增加1℃，EOS推迟约12 d，LOS延长约16 d，降水主要影响SOS和LOS，平均降水量每升高100 mm，SOS提前约5 d，LOS延长约9 d.

Vegetation phenology mapping of Wuhan and analysis of the affecting factors

As the indicator of responses of vegetation growth to climate changes， phenology is of much significance in studying climate changes and urbanization. As the distributions of vegetation inside urban areas are usually scattered， low-resolution Remote Sensing images of the kilometer level hardly recognize vegetation and conduct analyses well， while the Remote Sensing data of 10m spatial resolution are difficult to meet the requirements of phenological analysis in terms of time resolution. To solve the above problem， a spatial and temporal fusion method of Remote Sensing data is applied in this paper to alleviate the contradiction between time resolution and spatial resolution， and phenological mapping together with analysis of affecting factors are carried out. Based on the non-local means filter fusion method， the surface reflectance and EVI (Enhanced Vegetation Index) time series with spatial resolution of 30 m and time resolution of 8 d in Wuhan are generated. The EVI time series are reconstructed by Moving Weighted Harmonic Analysis， and vegetation phenology of Wuhan from 2006 to 2014 are extracted by dynamic threshold method. Through analysis， it is found that 1) from 2006 to 2014， the vegetation phenology of Wuhan from urban to suburb shows the spatial distribution pattern of the gradual delay of the start of season (SOS)， the gradual advance of the end of season (EOS)， and the extension of the growing period (LOS)， with the overall trend of SOS advancement， EOS delay， and LOS extension observed on temporal scales； 2) the correlation between vegetation phenology and average temperature is not significant， but EOS and LOS are significantly affected by the annual average daily temperature amplitude. For an average daily amplitude of 1°C， EOS is delayed by about 12 d， and LOS is extended by about 16 d. Precipitation mainly affects SOS and LOS. For every 100 mm increase in precipitation， SOS is about 5 d ahead of schedule， and LOS is extended for about 9 d.