基于地表温度降尺度的植被物候与地表热环境的关系分析

Analysis of the relationship between vegetation phenology and the surface thermal environment based on the downscaling of land surface temperature

植被物候是植被受外界环境影响后,出现周期性变化的现象,是气候变化影响生态系统的关键指示器。在干旱气候影响下,干旱区地表下垫面升温迅速且蓄热能力强,持续升高的地表温度(LST)对植被物候的影响不容忽视。本文以天山北坡城市群为研究区,利用MODIS的MOD11A1和MOD13Q1遥感影像探究2002-2021年植被物候和LST的空间分布格局,选择随机森林回归将LST进行降尺度,并采用Mann-Kendall非参数检验和Theil-Sen斜率分析植被物候19a来变化趋势,基于空间杜宾模型(SDM)分析空气温度、高程、蒸散发、地表温度、降水、土壤水分和大气气溶胶对植被物候的影响,最后探究LST与植被物候的局部相关性。结果表明,LST降尺度结果精度较高,白天和夜间LST的R2分别为0.982和0.985,能够更多反映地物细节。LST白天和夜间差异较大,分别呈“冷岛”和“热岛”现象。植被生长季开始期(SOS)提前、生长季结束期(EOS)推迟、生长季长度(LOS)延长。植被物候变化较为显著的地区主要集中在绿洲,非绿洲地区的变化则较不显著。温度、蒸散发和降水对植被物候各指标均具有较大影响。其中,LST对植被物候的影响更具典型性。该研究结果对于理解干旱区绿洲植被物候变化的复杂机制以及维护绿洲资源可持续发展具有重要意义。

Vegetation phenology is the phenomenon of periodic changes in vegetation as it is affected by the external environment,and is a key indicator of the impact of climate change on ecosystems.Under the influence of arid climate,the sub-surface of the arid zone warms up rapidly and has a strong heat storage capacity,and the influence of the continuously rising land surface temperature on the vegetation phenology should not be neglected.In this paper,the spatial distribution pattern of vegetation phenology and land surface temperature from 2002 to 2021 was investigated using MODIS remote sensing images MOD11A1 and MOD13Q1 in the urban agglomeration of the north slope of Tianshan Mountains as the study area.Random forest regression model was selected to downscale land surface temperature,and the Mann-Kendall non-parametric test and Theil-Sen slope were used to analyze the changing trend of vegetation phenology over the past 19 years.The effects of air temperature,elevation,evapotranspiration,land surface temperature,precipitation,soil moisture and atmospheric aerosols on vegetation phenology were analyzed based on the Spatial Durbin Model(SDM).Finally,the local correlation between land surface temperature and vegetation phenology was explored.The results show that the accuracy of the land surface temperature downscaling results is high,and the R2 values of the daytime and nighttime LST are 0.982 and 0.985,respectively,which can reflect more details of the features.The land surface temperature varies greatly between daytime and nighttime,showing"cold island"and"hot island"phenomena,respectively.The start of season time(SOS)is advanced,the end of season time(EOS)is delayed,and the length of season time(LOS)is extended.The more significant changes in vegetation phenology were concentrated in oases,while the changes in non-oasis areas were less significant.Temperature,evapotranspiration,and precipitation all have a great impact on various indicators of vegetation phenology.Among them,the effect of land surface temperature on vegetation phenology is more typical.The results of this paper are of great significance for understanding the complex mechanisms of vegetation phenology changes in oases in arid zones,and for maintaining the sustainable development of oasis resources.