基于因子分析方法的中国植被NDVI与气候关系研究

Study on the Relationship between Vegetation and Climate in China Using Factor Analysis

利用1982～2000年NDVI数据和气象台站资料，对我国几种植被型组和气候的相关关系进行了研究。首先利用NDVI结合植被类型图将我国植被划分为9种植被型组；然后利用因子分析方法进行了气候指标的选择并采用相对湿度、平均最高温度和平均风速作为本研究的气候因子；最后对7种植被型组NDVI值同相应季节及其前三个季节的气候指标进行了相关分析。结果表明，利用因子分析方法选择的气候指标可以较好地进行植被气候关系分析；在我国温度条件比水分条件更明显地影响植被的生长，水分条件较其他气候因素对植被生长表现了更明显的滞后效应；而平均风速则对我国荒漠植被生长有较大的影响。

Multiyear （19 a） NDVI and meteorological station data were used to study the correlative relationship between several vegetation type groups and climate in China. In order to investigate the responses of different vegetation types to climate variability, we classified 9 vegetation type groups using NDVI data and the map of vegetation type. At the same time we selected three variables from the nine meteorological variables to represent climatic factors using factor analysis. And these three meteorological variables are relative humidity; mean maximum temperature and mean wind speed respectively. At last, the correlations between seasonal NDVI of 7 vegetation type groups and current and antecedent three seasonal meteorological variables were analyzed. The results show that the three meteorological variables selected by factor analysis could represent the relationship between vegetation and climate in China. In spring, positive correlations between NDVI and the concurrent temperature occur in all the vegetation types. And the positive correlation is most remarkable in the meadow; the correlations of woods vegetation and scrub vegetation are comparative well. The correlations between spring NDVI and the concurrent relative humidity also put up well positive in all the vegetation types except meadow, but which is not better than the correlation between NDVI and temperature. In summer, several positive correlations between NDVI and the relative humidity of previous spring occur at coniferous forest, broadleaf forest, meadow, artificial vegetation and scrub. And positive correlations between NDVI and the relative humidity of previous autumn occur in all the natural vegetation. In autumn, the correlations between NDVI of all the vegetation types and the concurrent temperature put up positive, while the correlations between NDVI of most vegetation types and the concurrent relative humidity put up negative correlation. In winter, positive correlations between NDVI and the concurrent temperature occur in all the vegetation types except desert, and the correlations of broadleaf forest and artificial vegetation put up remarkable positive correlations and the positive correlation of meadow vegetation put up correlations with p-values 〈 0. 001. These relationships between NDVI and climate indicated that the temperature factor is a stronger force to influence vegetation growth than moister factor and the moister factor represents more obvious lag effects than the other climatic factors in China. Additionally, we found that the mean wind speed influenced the vegetation growth of desert.