中文摘要：

近几十年来，许多研究表明植物始花期随气候变暖普遍提前，但对植物花期长度变化的研究仍较少。利用1978—2014年牡丹江地区40种植物的始花期和末花期观测资料，分析该地区主要植物花期的时间分布、变化特征及与气候变化的关系，并评估两种物候模型对花期模拟的适用性。结果表明：①牡丹江地区40种植物的花期开始日在4月13日-8月27日之间，结束日在4月25日至9月13日之间，且均集中分布在5月份。花期长度的变化范围在6～69天。大部分植物（62．5％）花期长度在10～20天。②在研究时段内，植物花期物候出现了一定程度的变化，但大多数植物的变化趋势并不显著。始花期平均推迟速率为0．06天／10a，只有1种植物变化趋势显著（p〈0．05）；末花期平均提前速率为0．28天／10a，没有植物变化显著；花期长度平均缩短0．35天／10a，只有4种植物显著缩短。③绝大多数植物始花期和末花期的年际变化与季前温度呈显著负相关关系，温度敏感度分别在-6．2天／℃-2．3天／℃和-5．0天／℃-1．2天／℃。花期变化趋势不显著与牡丹江地区春季增温趋势不显著有关。④回归模型能够准确地模拟始花期、末花期的年际变化，平均拟合优度R^2分别为0．65和0．38，对花期长度年际变化的模拟效果稍差（平均R^2为0．17）。相比之下，GDD（Growing Degree Days）模型对花期模拟的效果更好，无论是对始、末花期还是花期长度均提高了拟合优度。该研究可为认识植物花期对气候变化的响应以及花期的模拟预报提供依据。

英文摘要：

Over the last several decades, many studies proved that the first flowering date （FFD） of most plant species has become earlier in response to global warming. However, the existing results about the impact of climate change on the flowering duration （FD） were quite limited. In this study, we investigated the temporal distribution and trends in FFD, end of flowering date （EFD） and FD of 40 plants from 1978 to 2014 in Mudanjiang, China. Correlation and regression analyses were used to examine the relationship between the flowering phenophases and the preseason temperature. Meanwhile, we have evaluated the applicability of two phenological models （including the regression model and the growing degree day model） in simulating the flowering phenophases. The results showed that： （1） During the observation period, the mean FFD and EFD of the 40 plant species ranged from April 13 to August 27 and from April 25 to September 13, respectively. However, most of FFD and EFD were concentrated in May. The FD ranged from 6 to 69 days, with most of them （62.5%） ranging between 10 and 20 days. （2） During the study period, FFD became later at a mean rate of 0.06 days decade-~ with only one species showing a significant trend （P〈0.05）, while EFD occurred earlier at a mean rate of 0.28 days decade-＇ （no one significantly）. The averaged FD has shortened by 0.35 days decade^-1 with only four species showing a significant shortening trend （P〈0.05）. （3） Most species showed a significant correlation between FFD （or EFD） and preseason temperature. The temperature sensitivities of FFD ranged from -6.2 to -2.3 days/℃^-1, while those of EFD varied from -5.0 to - 1.2 days/℃^-1. The unapparent trends of flowering phenophases were probably due to the unobvious variation of spring temperature in the study area. （4） The regression model has successfully simulated the interannual changes in FFD and EFD with the mean goodness of fit （R^2） ranging from 0.38 to 0.65, but failed to sim