Improved Drought Monitoring Model Based On Multi-source Remote Sensing Data

Drought is the most serious natural disaster affecting human life and social production.Although there are many indices which can monitor meteorological drought or agricultural drought,few drought indices that can accurately reflect the variation of vegetation growth caused by drought and timely detect the occurrence of agricultural drought.Drought is a complex and changeable natural disaster,short term drought can occur many times in a month.Short term drought may lead to grain output drops during the growth period of crops.Therefore,Monitoring drought timely and accurately is significant to agricultural production.This paper uses multi-source remote sensing data to construct a new drought monitoring index based on 10-days time scale by the use of principle component analysis,then the information of the 2000-2010 year drought in Jilin province was obtained.And use the linear spatial trend analysis method to analyze the spatial-temporal drought variations in Jilin province,the main contents of this study include the following aspects:(1)This study use the 500-m surface reflectance of MODIS09GA band 1(620-670nm),band 2(841-876nm)and band 7(2105-2155nm),then calculate the Normalized Difference Vegetation Index(NDVI)and Normalized Difference Water Index(NDWI),and the NDVI and NDWI based on 10-days time scale were calculated by the method of maximum value composite.(2)The TRMM3B42 contains a reliable dataset of precipitation rates(mm/h)over a day with a resolution of 0.25°× 0.25°latitude/longitude and was resampled to 500m by using the same projection as NDVI.The ten-day precipitation value is calculated as accumulating ten days precipitation,and used to calculate the Precipitation Condition Index(PCI).(3)Based on principle component analysis(PCA),NDVI,NDWI and PCI were used as principle components,then use the largest component as Modified Drought Index(MDI).(4)This paper has a correlation test between Modified Drought Index(MDI)and soil moisture at the depth of 10cm and 20cm to determine whether the MDI can eliminates or weakens the lag of agricultural drought monitoring.To determine whether the MDI index can accurately reflect the change of vegetation growth,there is also a time series analysis between Leaf Area Index(LAI)and MDI index,and the correlation analysis was made between the grain yield and the drought area with the MDI index respectively to determine the accuracy of the MDI index for monitoring the agricultural drought and the determination of the drought area.(5)In this paper,MDI index is used to analyze the drought trend in Jilin Province in 2000-2010 years.The linear regression method was applied to study the temporal and spatial variation of drought in Jilin Province in the past 11 years.