伏牛山地森林植被物候及其对气候变化的响应

Phenology of forest vegetation and its response to climate change in the Funiu Mountains

研究植被物候是理解植被与气候关系的重要途径。在植被对气候变化响应的敏感地区,开展植被物候研究有助于揭示气候变化对植被的影响机制。基于2000-2015年MODIS EVI时间序列影像数据,利用Savitzky-Golay(S-G)滤波方法和动态阈值法提取伏牛山地2000-2015年森林植被物候参数,结合气温、降水数据,运用Man-Kendall趋势检验、Sen斜率、ANUSPLIN插值和相关性分析等方法,研究伏牛山地森林植被物候对气候要素(气温、降水)变化的响应。结果表明:(1)伏牛山地森林植被生长季始期主要集中在第105~120 d,生长季末期主要集中在第285~315 d,生长季长度主要集中在165~195 d。从海拔梯度看,随海拔升高,生长季始期、末期和长度整体上分别呈显著推迟、提前及缩短趋势。(2)生长季始期和生长季末期整体上呈推迟趋势,推迟的像元分别占森林植被的76.57%和83.81%。生长季长度整体呈延长趋势,延长的像元占比为61.21%。生长季始期变化特征主要是由该地区的春季气温降低所导致的。(3)研究区森林植被生长季始期与3月平均气温呈显著偏相关,且呈负相关的区域最多,即3月平均气温降低,导致生长季始期推迟;生长季末期与9月降水呈显著偏相关区域最多,且两者主要呈正相关,即9月降水增加,使生长季末期推迟。植被生长季长度由整个生长期的气温和降水来共同作用,对大多数的区域而言,8月的平均气温和降水与生长季长度的关系最为密切。

The study on vegetation phenology is of great importance to understand the relationship between vegetation and climate. In areas where the vegetation is sensitive to climate change, a phenological study is helpful to reveal the response mechanism of vegetation to climate change. Based on Moderate Resolution Imaging Spectro radiometer （MODIS） Enhanced Vegetation Index （EVI） time-series images from 2000 to 2015, we utilized Savitzky- Golay （S-G） filter and dynamic threshold method to extract the phenological parameters of forest vegetation in the Funiu Mountains. Combining temperature and precipitation data, we used Mann-Kendall （M-K） trend test, Theil-Sen estimator, ANUSPLIN interpolation, and correlation analysis methods to analyze phenological changes of vegetation in response to climate factors （temperature/precipitation） in the Funiu Mountains. The results showed that： （1） The start and end of growing season ranged mainly from 105 d to 120 d and from 285 d to 315 d, respectively. The length of the growth season ranged mainly from 165 d to 195 d. There is an evident correlation between forest phenology and altitude. With increasing altitude, the start, end and length of the growing season presented a significantly delayed, advanced and shortened trend, respectively. （2） In terms of the interannual variations, both the start and end of the growing season mainly displayed a delayed trend in 76.57% and 83.81% of the total area, and the length of the growing season exhibited a lengthened trend of 61.21%. The start of forest growing season is mainly affected by the decrease of spring temperature in the region. （3） A significant correlation was found between the start of growing season and mean temperature in March. The negative correlation coefficient indicated that a delayed start of the growing season was primarily due to the tempera^tre decrease in March. The end of growing season was mainly influenced by September precipitation, and their correlation was positive, that is, the increased precipitation in September can delay the end of the growing season. In addition, the length of the growing season was influenced by temperature and precipitation during the whole growing season, for most parts of the region, the mean temperature and precipitation in August were obviously correlated with the length of forest growing season.