资源环境科技发展态势分析平台

Spatial pattern of meteorological droughts is a critical element for investigating drought condition. Empirical orthogonal function (EOF) and rotated EOF (REOF) methods are widely used to discriminate dominant patterns. Previous works focused on case applications but left the associated physical meanings unexplored. This study employs the EOF, wavelet analysis, and stepwise regression methods to distinguish spatio-temporal patterns of meteorological droughts with pattern interpretation. For this purpose, monthly precipitation data were used from 73 meteorological stations in the Poyang Lake basin of China for the period of 1960-2007. Standardized Precipitation Index (SPI) was applied to quantify meteorological droughts. Our results identified four leading EOF/REOF patterns at the basin scale. Seasonal SPI series showed more dominant periods than monthly SPI series, indicating periodic complexity of regional droughts. The dominant periods coincided with those of the EOF/REOF time coefficients. The four leading EOF/REOFs were significantly related to the dominant time scales of 2, 3, 6, 11, 16, and 17 years, except for the fifth EOF/REOF. Stepwise regression analysis further indicated that the four leading EOF/REOFs be likely the superposition of SPI variation from inter-annual and inter-decadal scales. The SPI series at different time scales (1-36 months) showed significant correlations with climate teleconnection modes, confirming the existence of identified spatio-temporal pattern. The Nino3.4 and Southern Oscillation were the primary teleconnection modes that influenced the variability of meteorological droughts over the basin. The Pacific Decadal Oscillation and the North Atlantic Oscillation also contributed to the regional dry/wet alternation. The findings provide important implications for understanding spatial and temporal variation of regional droughts.