Study Of The Agricultural Drought Monitoring By Remote Sensing Based On MODIS Data In Heilongjiang Province

Drought, which occurred frequently all over the world in recent years, seriously threatens the balance of natural eco-environment, and has great influence on the sustainable development of mankind, society and economy. So, it is of great significance to study drought monitoring, as well as to establish the suitable drought monitoring models based on remote sensing which have unmatchable advantages in information collecting and processing.In this study, the current situation and development tendency of drought monitoring based on remote sensing and soil moisture prediction were analyzed and the applying scopes and the characteristics of the related models were discussed. Based on the shortages of the two methods and the former investigator's study, energy index method was put forward in the study. The method considering surface temperature and reflection attributes was built with the soil moisture. Then the result of energy index method was compared with the result of the apparent soil inertia method and vegetation supply water index method. The conclusions are as follows:1) Apparent thermal inertia method was only suited to monitor soil moisture in low vegetation coverage condition and vegetation supply water index method in high vegetation coverage condition. Both of them were not stable in various periods. Energy index method was suited to monitor soil moisture with any level of vegetation cover and offered an effective method for monitoring soil moisture in moderate vegetation coverage condition.2) Energy index method was better than the other two methods in low and high vegetation coverage condition and the average fitting errors were all less than 13% in any level vegetation coverage condition. The error was 2.04% less than apparent thermal inertia method's in low vegetation coverage condition and 3.07% less than vegetation supply water index method in high vegetation coverage condition.3) The relative soil moisture with the depth of 10cm showed the highest correlation with apparent thermal inertia and the correlation between apparent thermal inertia and the relative soil moisture with the depth of 50cm was the lowest.The relative soil moisture with the depth of 50cm showed the highest correlation with vegetation supply water index and the correlation between vegetation supply water index and the relative soil moisture with the depth of 10cm was the lowest. Energy index method was suited to monitor soil moisture at any soil layer. The relative soil moisture with the depth of 50cm showed the highest correlation with energy index and the correlation between energy index and the relative soil moisture with the depth of 10cm was the lowest.4) Soil moisture retrieval model based on energy index method using several years data from 2003 to 2009 was used to monitor the soil moisture in Heilongjiang in 2010, and showed good results. Energy index model was suited to monitor soil moisture at any soil layer with any level of vegetation cover. The fitting error was 20.0% in low vegetation coverage condition, and 14.4% in moderate vegetation coverage condition,11.7% in high vegetation coverage condition. The average error was 15.4%. Besides, The average fitting errors became higher and higher as the change of time at any soil layer. The retrieval soil moisture was consistant with the actral situation when monitoring the drought in spring, summer and autumn. The monitoring results were best in sping and worst in autumn.5) The retrievals of soil moisture binary regression model based on energy index method were better than the energy index model at any soil layer to monitor the soil moisture in Heilongjiang in sping and summer,2010. So it could more precisely monitor the drought in sping and summer in Heilongjiang.6) Energy index method solved the disadvantages of other methods that couldn't monitor soil moisture with any level of vegetation cover. Furthermore, the monitoring results were good at any soil layer and the index could be calculated easily. So it could ensure somewhat the efficiency of monitoring soil moisture in Heilongjiang Province.