Study Of Drought Monitoring By Remote Sensing Based On MODIS Data In Henan Province

Drought, due to its high frequency, has always been a hot issue of global concern. Having a serious negative impact on environment, agriculture and people’s livelihood, drought has been restricting the sustainable development of China’s economy. Henan Province is located in the eastern part of China, and it is a typical drought-prone area. Therefore, it is of great practical significance to take advantage of the remote sensing technology in order to strengthen the investigation of the drought problems in Henan and to establish appropriate remote sensing monitoring models of drought.Taking Henan Province as the sample study area, this paper selected the data of MODIS surface reflectance, land surface temperature, and vegetation index product from the year of 2002 to 2010.Many years of soil moisture inversion model were constructed based on the apparent thermal inertia ATI and temperature vegetation drought index TVDI. The following conclusions has been reached:(1) When conducting a monthly analysis of the spatial and temporal distribution of relative humidity of Henan Province based on years of data of average relative soil humidity, soil moisture shows a strong seasonal pattern in time:The relative soil humidity is low from February to April; it rises after May; reaches its peak in September and then declines. Spatially speaking, the soil humidity greater than 10-20cm is presented more in the east than in the west. From January to June, soil humidity 50cm is presented more in the south than in the east. The soil humidity is relatively lower in the central and western part of the province every month. Vertically speaking, Most of the study area is mainly characterized by increasing relative soil humidity along with the increase of soil depth.(2) Years of data were divided to four seasons to establish models. Based on the apparent thermal inertia ATI, soil humidity inversion models were constructed for spring (March, April, and May) and winter (December, January, and February) data. After the completion of the model, analysis and calculation for each month and the whole seasons were done based on measures such as the measured average, the simulated average, the measured standard deviation, the simulated standard deviation, the root mean square error, the mean absolute error, and the mean relative error for each month. Throughout the spring, the average relative error for 10cm,20cm,50cm depth of soil layers were 21.276%,18.590%, and 19.114% respectively. Throughout the winter, the average relative error for 10cm,20cm,50cm depth of soil layers were 15.227%,14.939%, and 15.602% respectively.(3) Data of January, April, July, and October of the year 2012 were selected to represent winter, spring, summer, and fall. Models that were constructed based on the inversion data of the four seasons were applied to the relative soil humidity of these four months. Four graphs of inversion results were obtained from comprehensive comparison. Results that were obtained from the inversion for each month of data are generally consistent with the distribution of measured relative soil humidity. Of course, individual differences also exist, but that primarily come from variation errors caused by various sources such as the measured data, the ATI and TVDI calculation, and the inversion model. The results from the inversion model are generally in line with the measured data, which confirms and supports the credibly and the reliability of this study.(4) Nearly a decade of data monthly data, combined with the percentage of sunshine, precipitation, average temperature and other meteorological factors analysis of influence of soil humidity, and verifying the feasibility of this model, research shows that compared to the single-factor model, the correlation coefficient has risen, root mean square error decreased inversion of model accuracy. Each month break between seasons to establish soil humidity using 2002-2011 multiple-factor model, show that in January 2012, April, July, October 19 sites, soil humidity, and drought. Comparison of simulation with actual drought, only a limited number of sites and measured the levels of drought than one grade, and most sites results of numerical simulation with drought the same level.