Building Drought Indices Based On AMSR-E Data

The world is experiencing the climate warming. Extreme drought caused by climate warming has become one of the major issues of worldwide concern. It was estimated that because of the drought, more than6-8billion U.S. dollars has lost every year, far more than it caused by any other meteorological disaster.Therefore,real-time drought monitoring study has long-term significance on the research of climate change, drought defense, water resources management and protection issues. Now, there are two types of drought index, both of which are widely used at home and abroad:one is the traditional drought index that is based on the single observation point to be built. The monitoring result of this index is single and not very accurate. The other index relies on the modern satellite remote sensing technology. This index has the characteristic of wide-rage and high-precision monitoring,especiallythe widely-used passive microwave remote sensing information. AMSR-E(The Advanced Scanning Radiometer for EOS) is one kind of the passive microwave remote sensing, which is used for soil moisture retrieval. However, so far there is no mature drought indices based on AMSR-E. Therefore building drought indices based on AMSR-E has great practical significance on the dynamic agricultural drought monitoring.On the basis of antecedent work, this research selects the hilly area of central Sichuan basin as the study case. Based on the10bands of vertical and horizontal polarization AMSR-E passive microwave remote sensing data from2006to2010, by the way of addition, subtraction, multiplication and division, we have built125kinds of candidate drought indices. Then, in order to select the better ones for the drought research in hilly area of central Sichuan Basin, the candidate indices are compared with the measured soil moisture, daily precipitation, the standardized precipitation index (SPI) and AMSR-E soil moisture product. And the main conclusions are as follows:1. In this paper, by calculating and programming the measured soil moisture data with the125kinds of candidate drought indices in the hilly area, we can get each grid point values of correlation coefficient with time in the hilly area of central Sichuan. Among them the following eight indices have generally good correlation with the measured soil moisture: DI23(TBv06.9-TBh36.5), DI31(TBvO6.9/TBh36.5), DI74(TBv10.7/TBv18.7), DI75(TBv10.7/TBh36.5), DI92(TBv18.7/TBvO6.9), DI96(TBvl8.7/TBv10.7), DI101(TBh36.5/TBvO6.9), DI105(TBh36.5/TBv10.7). DI23, DI31, DI74andDI75show obvious negative correlation, namely bigger the value is, more arid it is. But DI92,96, DI101, and DI105in general perform out relatively good positive correlations.Hence, the eight drought indices are better for the drought research in hilly area of central Sichuan basin.2. By calculating and analysing the optimal eight drought indices with the daily precipitation and SPI in the seven measured sites, we find that DI23, DI31, DI74and DI75show negative correlation and DI92,96, DI101, DI105in general show positive correlations. From the time series of the optimal eight drought indices, it can beseen that the optimal eight drought indices can better reflect the spring, summer, and winter drought every year. Especially DI74, DI92, DI96can better reflect the spring and winter drought. Although in some years they are not satisfactory, such as the summer winter drought in2006and2007, their variation trends are in agreement with daily precipitation and SPI. They can more sufficiently reflect the drought conditions in each site. So the three drought indices are more applicable to the drought research in the hilly area.3. The overall trend of the three optimal drought indices (DI74, DI92and DI96) on the spatial distribution is basically the same and the degree of aridity from the southeast to the northwest region gradually deepened. The spatial distribution of the three optimal drought indices is roughly the same with the precipitation anomaly percentage, SPI, and soil relative humidity, especially in the northeast and southeast of the study area. But those in the northwest and southwest are quite different in scope and distribution of drought.4. The AMSR-E soil moisture product has a poor temporal correlation with the daily precipitation and SPI. Meanwhile, the spatial distribution of the AMSR-E soil moisture product is quite different with the precipitation anomaly percentage, SPI, and relative soil moisture. So the AMSR-E soil moisture product is not applicable to the drought research in the study area. Comparing the three optimal drought indices with the AMSR-E soil moisture product, the calculation of the three available drought indices are simpler, and the variation trends and the spatial distribution are more identical in the hilly area. Generally speaking, the three optimal drought indices are more applicable to the drought research in the hilly area of central Sichuan basin.