Abstract: A signal-based anomaly detection method with the Box-Cox transformation technique are adopted in an effort to effectively extract anomalous information from multiple data sources, and to improve the identification of water vapor transport anomalies of flood-causing storms (FCSs). The combined approach is applied to the Huaihe River Basin (HRB) to studying the anomalous characteristics of 850 hPa water vapor flux for the years of 2003, 2005 and 2007. The results show that during FCS periods, the 850 hPa meridional flux of water vapor between longitude 105° E and longitude 125° E is likely to form a dipole pattern with the HRB as its center and the positive anomaly on the south and the negative on the north. The importance of water vapor transport channels varies with FCSs as revealed by the 850 hPa zonal transport anomalies of water vapor at lower latitudes.