Spatiotemporal Variation Of Drought Across The Loess Plateau And Its Effects On Vegetation Phenology

Drought is an important meteorological disaster with frequent occurrence and strong destructive power.In the context of drastic climate change,the drought problem has become more prominent.At the same time,as an important means to study the relationship between climate and vegetation,vegetation phenology has undergone significant changes.In recent years,the warming and drying trend of the Loess Plateau is intensified at the seasonal scale,the drought characteristics of the region and the relationship between vegetation phenology and climate factors may become more complicated.Accurately understanding the characteristics of drought changes and the interaction between vegetation phenology and climate in arid and semi-arid areas is of great significance for studying how local ecosystems respond to climate change.Based on the high-resolution temperature and precipitation data from 1901 to 2100,the standardized precipitation evapotranspiration index(SPEI)was obtained on a 12-month time scale to analyze the spatiotemporal characteristics of drought in the past and in different scenarios of future.Using the Normalized Difference Vegetation Index(NDVI)time series data from MODIS's MOD13A2 product,we extracted the vegetation phenology information from 2001 to 2018 by dynamic threshold method and double logic method to study the spatio-temporal variation characteristics of vegetation phenology on the Loess Plateau,and further analyzed the influence of time-lag and time-accumulation effect of climate factors on vegetation phenology.The main conclusions are as follows:(1)From 1901 to 2017,the interannual variation trend of SPEI index did not reach the significant level,and there was no significant mutation year.Under the 95% confidence level,the average annual SPEI under the three RCP scenarios from 2018 to 2100 showed a significant decreasing trend,with a decreasing rate of 0.120/10,0.096/10 and 0.206/10 years,respectively.There was no severe drought in history,moderate drought occurred in1965,1997,and extreme drought occurred in 1941,and there were a total of 41 years of light drought.The total number of years of drought in the future period will increase,and the number of years of moderate drought will increase obviously,which indicates that the Loess Plateau is expected to be warmer and drier in the future.(2)Drought frequency of different grades on the Loess Plateau has obvious spatial variation characteristics.The frequency of light drought and severe drought was relatively high in the central part of the Loess Plateau during 1981-2010.From 2018 to 2100,the Loess Plateau was mainly affected by mild drought.Under the scenario of RCP8.5,compared with other regions,the frequency of extreme drought in northwest region will be higher in the next two periods,and climate change will increase the spatial distribution difference of extreme drought frequency.Compared with the historical period,80.81%-95.11% of the regions in the last 30 years of this century experienced the grade transition from no drought to light drought.The severity of drought increased in the northwest part of the Loess Plateau,and 7.21%-19.00% of the region will experience drought grade transition from no drought to moderate drought.(3)The beginning time of growing season(SOS)was gradually delayed from southeast to northwest of the Loess Plateau,and the end time of growing season(EOS)was gradually delayed from the periphery to the interior of the Loess Plateau.The spatial distribution of length of growing season(LOS)was not significantly different except in the south part of the Loess Plateau.From the perspective of spatial variation trend,SOS of grassland in the central part of the Loess Plateau showed a significant advance trend.The SOS of croplands in southwest showed a significant delayed trend.The EOS of some croplands in the eastern part of the Loess Plateau had a significant delayed trend.(4)Considering the time-lag and time-accumulation effect of temperature and SPEI,more SOS and EOS changes could be explained,with an average explanatory degree of 81.11% and 75.47%,respectively.Compared with the situation without considering the time effect,the explanation degree improved from 4.02% to 42.94%,and from 16.61%-33.32% for variation of SOS and EOS in different vegetation types.Considering the time-lag and time-accumulation effects of temperature and SPEI,The lag and cumulative months of the SOS in response to temperature ranged from 0.48 to 1.86 and from 1.41 to 4.01.The lag and cumulative months in response to SPEI ranged from 0.91-1.74 and 2.97-5.01.The lag and cumulative monthly of EOS in response to temperature ranged from 1.15 to 1.50 and from 2.08 to 2.49.The lag and cumulative months in response to SPEI ranged from 1.38 to 1.89 and from 1.95 to 4.92.The lag time of SOS of closed shrublands response to time-lag and time-accumulation effects of temperature and SPEI was relatively short,and the lag months of SOS response to temperature and SPEI were 0.48 and 0.91,respectively.The EOS accumulation time of closed shrublands mainly distributed in the southeast region was 2.08 months.The EOS of the deciduous broad-leaved forest was affected by cumulative drought over a longer time scale(4.92 months),and the lag time of response to drought was the longest(1.89 months).(5)The dominant climatic factor of vegetation SOS and EOS in southeast part of the Loess Plateau is temperature.The areas dominated by temperature accounted for 51.16% and18.25%,respectively.There is a positive correlation between SOS of vegetation and temperature,and a negative correlation between SOS and temperature accounts for 35.68%and 35.06% of the area respectively.In addition to the high altitude areas in the west,SOS of most vegetation distributed in the central part of the Loess Plateau is positively correlated with temperature.There is a negative correlation between SOS and temperature in the western and southeastern part of the Loess Plateau.In the,there is a positive correlation between EOS and temperature in the northeast part and middle eastern part of the Loess Plateau,and a negative correlation between EOS and temperature in the middle and eastern part of the Loess Plateau.The areas dominated by SPEI accounted for 3.21% and 2.03%,respectively.In the regional distribution is relatively scattered.The effects of temperature and SEPI on SOS and EOS of grassland were spatially different.