Phenological Period Of Different Vegetation In Hebei Province And Correlation Analysis With Net Primary Productivity(NPP) Based On MODIS Data From 2006 To 2016

Vegetation phenology is mainly manifested in the stages of vegetation budding,leaf development,flowering,fruiting,and falling leaves.As an important part of terrestrial ecosystem,the seasonal change of vegetation phenology is deeply influenced by climate factors such as air temperature and precipitation.Therefore,changes of vegetation phenology period can not only reflect seasonal change,but also an indicator of climate change.At the same time,changes of vegetation phenology influence global carbon cycle and vegetation productivity.Therefore,the study of vegetation phenology is helpful to understand climate change and vegetation productivity.Using Moderate Resolution Imaging Spectroradiometer(MODIS)data set MCD12Q2 and NPP(Net Primary productivity)data MOD17A3HGF as data sources,spatial distribution and change trend of annual mean phenological period of grassland,woodland and cultivated land in Hebei Province from 2006 to 2016 were analyzed.The phenological period included start of growing season(SOS),end of growing season(EOS)and length of growing season(LOS).The correlation analysis method was used to study the correlation between phenological period of these three land types and NPP,and the significance of change trend and correlation coefficient at?=0.05 level was judged by significance test.At the same time,time series curves of NDVI(Normalized Difference Vegetation Index)and LAI(Leaf Area index)were respectively generated from MODIS NDVI data and GLASS LAI data,and time series curves of natural vegetation and cultivated land were fitted and smoothed by double logistic method and S-G filter method.Dynamic threshold method was used to extract the relevant phenological period,the result was compared with MCD12Q2 phenological data,and the difference between the results of multi-source remote sensing data was analyzed.Research results were as follows.(1)From the perspective of spatial distribution,93.5%of SOS for grassland appeared on 108-153 DOY(day of year),96.2%of EOS of grassland appeared on 273-304 DOY,89.6%of LOS of grassland was concentrated in 128-190 days.96.2%of SOS of woodland appeared on 107-128 DOY,95.3%of EOS of woodland appeared on 282-306 DOY,89.0%of LOS of woodland was concentrated in 162-194 days.SOS of one-cropping cultivated land in Hebei Province was concentrated in 90-165 DOY,SOS of the first growing season for two-cropping cultivated land a year was concentrated in60-120 DOY,and SOS of the second cropping cultivated land in a year was concentrated in 185-191 DOY.EOS of the one-cropping cultivated land was concentrated in 260-280 DOY in 2016,EOS of the first growing season for the two-cropping cultivated land in a year appeared in 150-170 DOY and EOS of the second season for the two-cropping cultivated land in a year was concentrated in 279-290 DOY.LOS of the one-cropping cultivated land in a year was mainly 105-230 days,LOS of the first growing season for the two-cropping cultivated land in a year was mainly 70-170 days in 2016,and LOS of the second growing season for the two-cropping cultivated land in a year was mainly 82-109 days.NPP was mainly concentrated in 200-700g C/m~2.(2)From the perspective of temporal change,the overall SOS of natural vegetation(grassland and woodland)in Hebei Province showed advance trend but not significant(P=0.08),the overall EOS showed insignificant delay trend(P=0.01),the overall LOS showed significant increase trend(P=0.03).The overall SOS of one-cropping cultivated land showed significant advance trend(P=0.003),the overall EOS showed insignificant delay trend(P=0.69),and the overall LOS showed significant increase trend(P=0.0044).The overall related phenological periods of the first and second seasons of two-cropping cultivated land in a year showed insignificant trend.(3)As for inter-annual change trend,28.7%of SOS for grassland showed significant advance trend,16.2%of EOS for grassland showed significant delay trend,and 20.8%of LOS for grassland showed significant increase trend.23.3%of SOS for forest land showed significant advance trend,11.4%of EOS for forest land showed significant delay trend,and 13.0%of LOS for forest land showed significant increase trend.35.9%of SOS for cultivated land showed significant advance trend,2.2%of SOS for cultivated land showed significant delay trend,and 9.6%of EOS for cultivated land showed significant advance trend.8.0%of EOS for cultivated land showed significant delay trend,16.6%of LOS for cultivated land showed significant increase trend,and9.8%of LOS for cultivated land showed significant decrease trend.(4)The area with significant correlation between SOS,EOS,LOS and NPP of grassland respectively accounted for 14.8%,8.5%and 14.4%of the total grassland area in Hebei Province.The area with significant correlation between SOS,EOS,LOS and NPP of forest land respectively accounted for 12.0%,3.7%and 12.5%of the total forest land area in Hebei Province.The area with significant correlation between SOS,EOS,LOS and NPP of cultivated land respectively accounted for 18.0%,15.5%and 21.1%of the total cultivated land area in Hebei Province.The phenological periods of three vegetaion types showed mainly moderate correlation with NPP,with small proportion of high correlation,and no low correlation area.SOS of vegetation was mainly negatively correlated with NPP in most areas of Hebei Province,EOS and LOS were mainly positively correlated with NPP.However,the correlation for grassland in Bashang Plateau area and two-croppping cultivated land was different from other areas in Hebei province.(5)According to the comparison of multi-source remote sensing products data,results showed that SOS of natural vegetation from GLASS LAI and MCD12Q2 data was the same,EOS had an average difference of 10 days.SOS of natural vegetation from NDVI time series curve and MCD12Q2 data at 20 pixel points had an average difference of 12 days,and EOS had an average difference of 5 days.Compared with ground observation data,the results from remote sensing data relatively lagged.20 pixel points were selected for cultivated land,and phenological periods were extracted.SOS from GLASS LAI time series curve and MCD12Q2 data for one-cropping cultivated land had an average difference of 13 days,and EOS had an average difference of 13days.SOS extracted from NDVI time series curve and MCD12Q2 for one-cropping cultivated land in a year had an average difference of 6 days,and EOS had an average difference of 1 day.Since GLASS LAI was not sensitive to phenological characteristics of two-cropping cultivated land,only phenological periods extracted from NDVI time series curve and MCD12Q2 of two-cropping cultivated land in one year were compared.SOS in the first growing season had an average difference of 16 days,and EOS in the second growing season had an average difference of 22 days.