研究山地过渡带植被物候格局以及植被物候变化的驱动机制具有重要意义。基于2000—2018年MODIS NDVI遥感数据,利用双Logistic曲线拟合法提取2000—2018年秦岭山地植被物候参数,结合气温和降水数据,运用趋势分析、相关分析、冗余分析等方法,研究秦岭山地不同物候期的时空变化特征和对气象因子的响应。结果表明:(1) 秦岭山地物候始期集中于60~102 d,物候末期主要集中在315~345 d,生长季长度集中在225~255 d,空间上看,具有明显的垂直地带分布特征,随海拔升高,物候始期、末期、生长期长度分别呈延迟、提前和缩短趋势。(2) 从植被物候的年际变化来看,物候始期62.25%的区域呈现提前趋势,53.42%的区域物候末期呈现推迟趋势,其中59.18%的区域植被生长期长度延长,仅有5.12%位于中东部中高海拔区的部分区域显著延长。(3) 植被物候始期大部分区域与2—5月平均气温和降水呈负相关关系,分别占总面积的50.29%和65.24%;物候末期与8—11月平均气温和月降水量主要呈正相关关系,分别占总区域的66.63%和59.77%;(4) 冗余分析(Redundancy analysis,RDA)结果表明秦岭山地植被物候受当季和前期气象因子的共同影响,春季物候受物候发生期的影响大于上年冬季,物候末期前期的气温和降水对秋季物候的影响强度大于秋季,不同坡向物候变化对气象因子的响应程度有所差异。

It is important to investigate the vegetation phenology pattern and the driving mechanism of vegetation phenology change in a mountain transition zone. In this study, we used the double logistic curve fitting method to extract the phenological parameters of forest vegetation in the Qinling Mountains on the basis of the moderate resolution imaging spectro radiometer and normalized difference vegetation index time-series images from 2000 to 2018. Combining temperature and precipitation data, we performed trend analysis, correlation analysis, and redundancy analysis to study the characteristics of time and space changes in different phenological periods and the response to climate elements in the Qinling area. The following results were obtained: (1) The start and end of the growing season ranged from 60 days to 102 days and from 315 days to 345 days, respectively. The length of the growth season ranged from 225 days to 255 days. From a spatial perspective, it has evident vertical zone distribution characteristics. With an increase in altitude, the start, end, and growth periods of phenology were delayed, advanced, and shortened, respectively. (2) The beginning of phenology was advanced, whereas the end of phenology was postponed. The proportions of pixels in advance and delayed phenologies were 62.25% and 53.42%, respectively. The length of the growing season exhibited a lengthened trend of 59.18%. Of the significantly extended area, 5.12% area was mainly located in the middle and high altitude areas of the central and eastern regions. (3) The initial period of vegetation phenology was negatively correlated with the average temperature and precipitation from February to May, accounting for 50.29% and 65.24% of the total area, respectively. The end of phenology was positively correlated with the average temperature and precipitation from August to November, accounting for 66.63% and 59.77% of the total area, respectively. (4) Redundancy analysis results show that the vegetation phenology in the Qinling Mountains was affected by both the current season and the earlier period of meteorological factors. The spring phenology was more affected in the phenology occurrence period than in the winter of previous year. Compared with autumn weather factors, temperature and precipitation in the early period of phenology exhibit a more significant relationship with the end of phenology. The phenological changes of different slopes exhibit different responses to meteorological factors.