Response Of Vegetation Phenology To Temperature And Precipitation In Xinjiang

Vegetation is a sensitive indicator for global climate change and plays an essential role in regulating climate system through its influence on exchange of energy,water and carbon at the land-atmospheric interface.Vegetation phenology,which refers to the seasonal timing of life cycle events of plants,has been found to be strongly regulated by climate factors.Changes in phenological events have the potential to broadly impact terrestrial ecosystems and human societies,for example,by altering global carbon,water,and nitrogen cycles,interspecific interactions,crop production and distribution of diseases.The development of remote sensing technology has brokenthough the time and space limitations of traditional phenological observation,and provides a new technical means to monitor a wide range of vegetation phenology.There is a typical temperate continental arid climate in Xinjiang,and the ecological environment is very fragile.Due to Xinjiang's special geographical location,the stability of regional ecosystem plays a vital role to the stability of ecosystem in China even in central Asia.Mainly using remote sensing and GIS technology,based on the 8 d MODIS MOD09A1 reflectance product,and combining the Logistic function fitting method and the curvature extremum method,vegetation's start time(SOS)and end time(EOS)of growing season for typical vegetation types in Xinjiang from 2001 to 2015 were extracted.Firstly the spatial and temporal variations of phenology for the typical vegetation types in the 15 years were analyzed.And then combining with temperature and precipitation from 2001 to 2014 in Xinjiang,sensitivity between vegetation phenology and temperature or precipitation was further analyzed to reveal the response and feedback of vegetation to the variations of temperature and precipitation in Xinjiang.The main conclusions in this study could be drawn as follow:(1)Spatial and temporal variations of vegetation NDVIThe distribution of averaged NDVI of the typical vegetation types from 2001 to 2015 in Xinjiang showed significant space variations.NDVI was smaller in Kunlun Mountains in southern Xinjiang,while the NDVI of forests in Tianshan and Altai Mountains was higher.The amplitude of NDVI variations was smaller in Kunlun Mountains in southern Xinjiang,while it was higher in the grassland in western Tianshan Mountains and the cultivation form northren Tianshan Mountains to southern Junggar basin and the partial cultivation in Tacheng region.Temporal variations of NDVI for the typical vegetation types were not significant in Xinjiang from 2001 to 2015,and the variation trends were inconsistent.NDVI of cultivation increased,while that of mixed forest,deciduous needleleaved forest,shrubland and grassland decreased.(2)Monitoring of vegetation phenology based on remote sensingErrors of vegetation phenology based on remote sensing and observation statistics were generally less than 16 d.For vegetation SOS,57%of error absolute values were less than 8 d;93%of error absolute values were less than 16 d;averaged error was 8.2 d,RMSE was 10.2 d,and correlation coefficient(r)value was 0.69(p<0.001).For vegetation EOS,62%of error absolute values were less than 8 d;91%of error absolute values were less than 16 d;average error was 8.1 d,RMSE was 10.4 d,and r was 0.59(p<0.001).(3)Spatial and temporal variations of phenology of typical vegetation typesDistribution of average phenology of typical vegetation types from 2001 to 2015 in Xinjiang showed significant space variations.Vegetation SOS overall displayed a delay trend from south to north,and vegetation EOS overall displayed an advance trend.Vegetation LOS overall displayed a shortened trend from south to north and corresponded to the distribution of vegetation SOS and EOS.The interannual fluctuation of vegetation SOS and EOS was small over 15 years,while the interannual fluctuation of vegetation LOS was greater than that of vegetation SOS and EOS.Temporal variations of SOS for cultivation showed a significant delayd trend(-0.47 d/yr,R2=0.30,p=0.04),while SOS for other vegetation types showed advance trends,in which SOS variation for shrubland was biggest(0.71 d/yr,R2=0.12,p=0.21).Temporal variations of EOS were different among vegetation types.EOS of evergreen needleleaved forest,deciduous broadleleaved forest,mixed forest and cultivation showed advance trends,while that of deciduous needleleaved forest,shurbland and grassland showed delay trends.Among them,the advance trend of cultivation was biggest(-0.39 d/yr,R2=0.34,p=0.02)and the delay trend of grassland was biggest(0.39 d/yr,R2=0.17,p=0.13).LOS for all vegetation types showed delay trends except cultivation,in which the delay trend of shrubland was biggest(0.97 d/yr,R2=0.13,p=0.19)and cultivation LOS showed an advance trend(-0.86 d/yr,R2=0.46,p=0.006).(4)The phenological response to meteorological factorsVegetation SOS was negatively correlated with the averaged temperature during preseason(from the preceding November to April)from 2001 to 2014 in Xinjiang with a percent of 78.5%,and positively correlated with averaged temperature with a percent of 21.4%.SOS of vegetation was positively correlated with cumulative precipitation during preseason with a percent of 75.8%,and negatively correlated with cumulative precipitation with a percent of 24.0%.Vegetation EOS was negatively correlated with averaged temperature during preseason(from May to October)from 2001 to 2014 in Xinjiang with a percent of 45.5%,and positively correlated with averaged temperature with a percent of 54.2%.Vegetation EOS was positively correlated with cumulative precipitation during preseason with a percent of 54.5%,and 45.1%of EOS negatively correlated with cumulative precipitation during preseason.