基于时间序列谐波分析的鄱阳湖湿地植被分布与水位变化响应

采用高时间分辨率遥感信息的谐波分析方法,提取反映鄱阳湖湿地植被指数随水位变化的谐波分量,分别以自然年和水文年的不同周期作为湿地植被指数谐波分析单元,利用时间序列信号的最大振幅谐波分量的变化周期表征湿地植被指数在不同分析单元的变化模式,结合常年水位观测数据和湿地植被群落在不同物候期的时间与空间特征,探讨鄱阳湖国家级自然保护区和南矶湿地国家级自然保护区的植被分布面积与水位变化关系.结果表明:(1)鄱阳湖湿地植被分布受水文状况影响的特征明显,相对于南矶自然保护区,鄱阳湖自然保护区湿地植被分布面积对观测水位的变化更为敏感.(2)两个自然保护区范围内的湿地植被分布面积与对应水文年9和10月的观测水位呈现较强的负相关关系,且在0.05水平上显著.一年两季生长的湿地植被分布面积受退水时间影响大于次年的涨水时间,与枯水期的观测水位无明显的相关关系.(3)两个自然保护区在不同高程区间的湿地植被分布面积与观测水位的相关关系和显著性呈现各自特征.在鄱阳湖保护区,12~13 m高程区间的湿地植被分布面积与9月观测水位的相关性最强,且相关关系在0.05水平上显著;13~14 m高程区间的湿地植被分布面积与10月观测水位相关关系更强.在南矶自然保护区,湿地植被分布面积在不同高程区间均与9和10月观测水位显著相关.采用谐波分析方法分析湖泊湿地的植被分布面积与水位关系有助于基于多时间序列遥感信息的湿地水文节律研究.

Response on wetland vegetation distribution to hydrology regularity based on harmonic-time series analysis

This paper examines wetland vegetation distribution based upon perennial water levels and the time and spatial characteristics of wetland vegetation in various phonological phases. Harmonic components are extracted to describe wetland vegetation index and the changing patterns are used by remote sensing data in a high time resolution. This paper defines one analysis unit by a period from 2000 to 2013 as one hydrological year (from September to the next September). Changing patterns of wetland vegetation during different analysis units are expressed by the period of harmonic component which has the maximum amplitude. Results show:(1) Area of wetland vegetation in Lake Poyang Natural Reserve is significantly influenced by hydrological characteristics. Compared to that in Nanji Natural Reserve, wetland vegetation distribution in Lake Poyang Natural Reserve is more sensitive to the changes in water level. (2) Area of the wetland vegetation within natural reserves is negatively Pearson correlated with water level in a hydrological year (at a 0.05 significance level). Impact of recession date is larger than that of next-year flooding date on distribution of wetland vegetation, while water level in dry season is not significantly Pearson correlated with the distribution of two-growing-period wetland vegetation. (3) The Pearson correlations between distribution of wetland vegetation on different elevations and water level in the two natural reserves show distinct correlations. In Lake Poyang Natural Reserve, wetland vegetation distribution from the elevations 12-13 m has the strongest correlation with water level in September, and the wetland vegetation distribution from the elevations 13-14 m has the strongest correlation with water level in October. In Nanji Natural Reserve, the wetland vegetation distributions on different elevations are significantly correlated with water levels in September and October. A harmonic analysis is conducive to further study on wetland hydrology relations based on multi-temporal remote sensing data.