| The phenomenon of regional unbalance of income and expenses or demand of moisture content, and water shortage affects the human survival and economic development, known as drought. American Meteorological Society (AMS) divided drought into meteorological drought, agriculture drought, hydrologic drought and social economy drought. Agriculture serves as the first industry of China, this kind of drought occurs will have an enormous effect on the economic development of our country. Utilize advanced and effective method of drought monitoring has a great significance to the development of both people’s life and social economic. Traditional drought monitoring on the statistical data of meteorological sites basis does not have the characteristics of timely and accurate, on the contrary, satellite remote sensing has the characteristics of macroscopic, comprehensive, real-time, etc., which can be able to acquire a wide range of regional drought conditions timely, accurately and comprehensively, that facilitate promoting agricultural production, and guarantee the food security and regional sustainable.The study area of this article is the Xinjiang Uygur Autonomous Region which is the largest cotton and excellent melon and fruit base among the nation, but is located in the arid and semi-arid climate zones, such climate environment intensifies the development of salinization, and obstructs agricultural production. Land surface temperature (LST) and normalized difference vegetation index (NDVI) data have the meaning of revelation to drought conditions, and certain limitation will be existing If the data are used individually, while via on behalf of green vegetation in response to drought stress on the environment to reveal the information of soil moisture, and indicate whether vegetation is affected by drought by the combination of LST and NDVI. Based on these two factors, build a two dimensional feature space, and calculate the temperature-vegetation drought index (TVDI), a combination of two parameters respectively having their own ecological significance.The study uses Moderate Resolution Imaging Spectrometer (MODIS) remote sensing data, treat 10cm,20cm soil relative humidity data from meteorological site as verification data, based on the theory of feature space of LST-NDVI as foundation, we built the monthly LST-NDVI feature space of vegetation growing season from March to August of the year from 2004 to 2013, and calculated the monthly TVDI drought tendency chart. This article made full use of the highly advantage of time resolution of MODIS data, and its extensiveness area drought monitoring in accordance with space resolution, we got the drought distribution and variation conditions of Xinjiang from space and time aspects. The mainly study methods and results are as follows:(1) From 2004 to 2013, we made five years as a period of time, maximum and minimum synthetic method were respectively used to extract the maximum and minimum surface temperature between the same vegetation index from March to August, use these data to form a new feature space boundary, and finally acquired synthesis dry and wet edges of feature space by linear fitting, and got the synthesis LST-NDVI feature space of five years. According to the analysis of single time frame and the slope and intercept of the synthetic feature space boundaries, we determined the instability of the original boundary, the floating of slope and intercept of each month in different years is obviously, and synthesis feature space is more stable.(2) Divided TVDI obtained from calculating into five degrees, that are no drought (0-0.67), light drought (0.67-0.74), medium drought (0.74-0.8), heavy drought (0.8-0.86) and extra heavy drought (0.86-1). According to TVDI degree distribution figure, which shows that years of drought conditions distribution trend is consistent in Xinjiang, heavy and extra heavy drought mainly distributed in the two basins of the province, (the Tarim and Junggar basin), the drought severity alleviate gradually extending outward along the basin, major drought level are no drought, light drought in basin edge and along the mountain area, heavy and extra heavy drought are distributed in gobi and desert. However in March, because the surface temperature to too low, and the vegetation is at the early growth stage, the effect of monitoring by TVDI is not ideal, including many void value pixel point. As time goes on, temperature arise and vegetation growth, void value pixel point decrease, drought area has the trend to expand.(3) From the analysis of relevancy of TVDI index respectively with surface temperature and vegetation index of August,2013, we found relevancy of TVDI with surface temperature reached 0989, while relevancy of TVDI with vegetation index on reached 0.409, which is obviously lower than that with surface temperature, it is visible vegetation index has limited ability to reflect drought condition, while TVDI is more sensitive with the change of surface temperature, using surface temperature to assess drought is more desirable.(4) Using 10cm and 20cm soil relative humidity data from 21-24 meteorological site to verify.10cm soil relative humidity data represent negative correlation with TVDI, the absolute value of correlation index is more than 0.617, i.e. the higher the TVDI is, the lower the soil relative humidity.20cm soil relative humidity data represent no obvious correlation, which reflects TVDI is reliable for drought monitoring, and 10cm soil relative humidity can reflect drought conditions in study area better than 20cm soil relative humidity.(5) From the changing curve of the proportion that drought area occupy in the whole study area from 2004 to 2013, the proportion of drought area shows fluctuated change from each month and varies between 33% and 67%. According to the trend line, however, the variation trend turns to be stable which keeping around 59%. Heavy and extra heavy drought area proportion and drought area present the same change trend. Their peak and valley values mostly are in the same month. The variation range of heavy and extra heavy drought area proportion is between 2% and 33%, and from the trend line shows a slight increase from 16% to 20%. The decline in 2009 was due to low temperatures continue in spring and summer in Xinjiang and rise in 2011 was because of sustained high temperatures, floods and hailstones in spring and summer.(6) The difference value of monthly TVDI and average TVDI of each month reveals the TVDI did not dramatically change in study area in the decade, and dramatically change occurs only in some specific areas. The most obvious changing regions are the edge of Tarim basin, Southern Xinjiang and along the Tianshan area. Region located in the heavy drought throughout the year, TVDI rarely ups and downs, whereas oasis regions where vegetation grow more lush are the TVDI ups and downs severely regions, and also the drought monitoring key areas need concern.(7) Spring (March to May) and Summer (June to August), based on average TVDI, drought conditions change obviously, no drought area in summer is significantly lower than that of spring, heavy and extra heavy area is far higher than that of spring, TVDI seasonal variation is mainly due to the rise of the summer surface temperature and the better growth of vegetation, this result once again confirms that surface temperature has great importance on TVDI drought monitoring. |
PDF Full Text Request