| The occurrence of drought will greatly restrict the normal operation of industry and agriculture and social development.A full understanding of the causes,changes and influencing factors of drought has a major effect on preventing and alleviating drought.Constructing a reasonable and appropriate hydrological drought index can quantify the characteristics and changing trends of drought,and facilitate people's research and prevention of drought.In recent years,the impact of human activities on drought has continued to increase.Analyzing the impact of land use/cover changes under human activities on drought is of great significance for understanding,restraining,and predicting drought.This paper selects four distribution functions to fit the monthly runoff data of Zhaishang and Lancun Stations from 1958 to 2000,and the monthly runoff data of Shangjingyou and Fenhe Reservoir Stations from 1958 to 2016,and selects the monthly,seasonal,and annual optimal distributions to establish Regional runoff drought index,which analyzes the characteristics of drought at different scales,extracts vegetation and land use information in the study area,and analyzes vegetation from mathematical statistics using Pearson correlation coefficient method,Spearman rank correlation method,cross wavelet method,wavelet coherence method,and GAMLSS model method The correlation between cover and drought index,using the SWAT model to analyze the impact of land use/cover change on hydrological drought in the study area from the perspective of physical causes,the conclusions of the study are as follows:(1)In the optimization of runoff sequence fitting at different scales,the Wakeby distribution is the optimal distribution fitted by runoff sequences at the four stations on the monthly and annual scales.On the seasonal scale,the Wakeby distribution is the optimal distribution of seasonal runoff data from Fenhe Reservoir,Zhaishang and Lancun stations.The Kappa distribution is the optimal distribution of the Shangjingyou station,and the optimal distributions of runoff sequences of different scales are different.(2)Establish the runoff drought index(SDI)based on the optimal distribution of each station,and use the Kriging method to obtain the regional SDI sequence.Divide the regional drought and verify the results based on historical drought records.The results show that the applicability of regional SDI is good,and the drought grade of Galton board experiment is accurate.(3)The monthly,seasonal,and annual regional SDI sequences in the upper reaches of the Fen River all show positive persistence,and droughts will continue to occur in the future.The monthly drought cycle is mainly about 12 months,and there are several months,consecutive seasons,and multi-month droughts locally.Seasonal droughts have 4 seasons and 64 seasons of drought cycles.In the upper reaches of the Fen River,the monthly,seasonal,and annual regional droughts are dominated by non-drought,light-drought,and medium-drought.The three types of droughts at different scales account for more than 85%.Compared with the seasonal scale and the annual scale,the monthly scale regional SDI is more sensitive to the judgment of severe and extreme drought levels.(4)The monthly,seasonal,and annual NDVI series showed an overall increasing trend from 1982 to 2013,and there was a significant increase after 2008,and the autumn vegetation was higher than other seasons.When the significance level is set to 0.05,the vegetation on the monthly,seasonal,and annual scales all have abrupt changes,and the monthly scale is earlier than the seasonal scale,and the seasonal scale is earlier than the annual scale,and the mutation points on the monthly scale are more precise.The monthly scale vegetation has a regular cycle of about 12 months,and the seasonal scale vegetation has a four-season cycle.Vegetation in the growing season showed an increasing trend,with obvious low values and peak values less than adjacent years in individual months,during which vegetation growth was affected.(5)From the aspect of mathematical statistics,the correlation analysis between SDI and NDVI in the upper reaches of Fenhe River is carried out.The Pearson correlation coefficient method showed that the regional drought in March was significantly correlated with vegetation in May,June,July,August,November and December.The regional SDI in February,March,and April is mostly correlated with NDVI in April,May,June,July,and August.The regional SDI in October,November,and December also correlated well with the NDVI in September,October,November,and December.Regional drought in winter is related to vegetation in spring,summer and autumn,and regional drought in spring is related to vegetation in summer and autumn.The drought in summer and autumn is slightly correlated with vegetation,and the correlation between vegetation in autumn is better.The correlation between seasons is mostly positive,and there is a certain lag between regional drought and vegetation.According to the Spearman rank correlation method,there is a strong correlation between regional drought and vegetation in March,April,November,and December.The correlation between regional drought and vegetation in May,June,July,and August is generally small,mostly showing a positive correlation.There is more rain in the study area.There is a real correlation between regional drought in winter and vegetation in spring,summer and autumn.In summer,regional drought has a substantial correlation with vegetation,and is slightly correlated with vegetation in other seasons,and autumn vegetation has a strong correlation with regional drought.In the correlation result statistics,the correlation degree obtained by the two methods is more than 60% of the micro-correlation relationship.Compared with other scales,the correlation analysis between regional drought and vegetation on the monthly scale shows a better correlation.(6)In the wavelet analysis of regional drought and vegetation at different scales,monthly-scale regional SDI and NDVI have obvious periodic oscillations of about 12 months,and there is a linear positive correlation between regional drought and vegetation.Within the wavelet cone,there is a local negatively correlated oscillation region from 0 to 8 months.Seasonal regional SDI and NDVI have a positive correlation of 0-6 quarters,and there is a20-30 quarter cycle oscillation in the 40-100 th quarter.Annual drought and vegetation have periodic oscillations of about 6 years.(7)The GAMLSS model was used to analyze the trend of vegetation changes in the upper reaches of the Fenhe River with regional drought.When severe droughts and extreme droughts occur,NDVI remains at a low level.When regional droughts weaken,the NDVI under each guarantee rate shows a significant upward trend,and the maximum rises to 0.8.As the vegetation index continues to increase,the vegetation remains in the range of 0.5-0.6,and the vegetation becomes drier and drier as the drought degree changes.(8)The SWAT model is used to analyze the interaction between drought and land use/cover change in the upper reaches of the Fen River from the perspective of physical causes.With the decrease of water bodies and grassland in the upper reaches of the Fen River,and the increase of residential land,arable land,and woodland,the frequency of non-drought,light-drought,and moderate-drought decreased by about 27% in the monthly scale of regional droughts from 1992 to 2000,and severe and extreme droughts The frequency of occurrence has increased significantly.The decrease of water bodies and grassland,and the increase of residential land,arable land,and woodland affect the frequency and grade of drought in the basin,and aggravate the drought in the basin. |
PDF Full Text Request