西南地区综合干旱监测模型构建与验证

在全球极端天气事件越来越多的大背景下,准确监测西南干旱对区域农业可持续发展具有重要的现实意义。该文选取降水距平百分率(percentage of precipitation anomaly index,Pa)、标准化降水指数(standard precipitation index,SPI)、相对湿润指数(relative moisture index,MI)等3种气象类干旱监测模型以及植被供水指数(vegetation water supply index,VWSI)与归一化植被指数(normalized differential vegetation index,NDVI)等2种遥感类干旱监测模型,并分别与实测土壤湿度作相关分析,在此基础上选取相关系数最高的相对湿润指数与归一化植被指数为自变量建立综合干旱监测指数(comprehensive drought monitoring index,DI)。结果表明,综合干旱指数与土壤水分实测值有较好的相关性,监测精度可达88.38%;在不同海拔高度内,综合干旱指数的拟合效果比单一指数效果更好,精度更高;在分析2009-2010年西南特大干旱旱情发展的时空演变过程中,综合干旱监测结果与实际干旱情况有较好的空间一致性,监测效果佳。研究成果为西南丘陵山区干旱监测提供了一种新的方法。

Construction and validation of comprehensive drought monitoring model in Southwest China

Abstract: Under the context of more frequent global extreme weather events, accurately monitoring the impact of drought on crop growth in Southwest China has important practical significance for the sustainable development of regional agriculture. Firstly, the study selected 3 types of meteorological drought monitoring models including the percentage of precipitation anomaly(Pa), the standardized precipitation index(SPI), and the relative moisture index (MI) and 2 types of remote sensing drought monitoring models including the vegetation supply water index(VSWI) and the normalized differential vegetation index(NDVI). Secondly, the correlation analysis between 3 meteorological drought monitoring indices, 2 remote sensing monitoring indices and soil relative moisture data was made by using a pixel-to-station paired correlation approach. Thirdly, MI and NDVI, which had the highest correlation coefficients with soil relative moisture, were selected to develop a comprehensive drought index(DI) finally. The new comprehensive DI makes full use of the complementary advantage of ground meteorological site data and remote sensing spatial data, and is suitable to the condition of complex underlying surface. The independent soil moisture data and observed actual drought disaster were used to test the reliability of model. The study result showed that in a month time scale, MI had a highest correlation coefficient of 0.477 with soil relative humidity among all the meteorological drought indices while NDVI had a higher correlation coefficient of 0.416 with soil relative humidity than VSWI. In addition, the correlation of the same type of drought monitoring indices was higher than the different type of drought monitoring indices. This demonstrated that different types of drought indices were complementary because they represented different drought information. Estimated DI had a good correlation with measured soil moisture with the r of 0.816 and the estimated average accuracy reached 88.38%, which was a high accuracy for drought monitoring in southwest area. Furthermore, DI performed better than the single index MI or NDVI in all altitudes, which suggested that DI based on multiple data sources was better than the index based on single data source in different altitudes. The spatial-temporal distribution of drought in 2009-2010 over the southwest region was analyzed according to the DI. The results of drought monitoring showed that the drought disaster area was mainly concentrated in Panxi region in southern Sichuan Province, most part of Yunnan Province and western Guizhou Province. The drought emerged from September 2009, increased gradually from October 2009 to February 2010, relieved gradually from March to May 2010 and terminated in June 2010. The temporal and spatial distribution of drought based on the drought monitoring model was consistent with the actual observed data, which showed DI had a good reliability to monitor the drought process in Southwest China. DI integrated the information of vegetation, rainfall, temperature and evapotranspiration and reduced the uncertainty of the single index inmonitoring drought. Therefore, DI could monitor drought more stably, continuously and truly compared to other drought monitoring indices. This work provides a new approach to monitor drought in Southwest China.