Using a statistical correlation based on an ensemble forecast, observations and the sensitive region of a heavy rainfall that occurred around Beijing on July 23 2009 are analyzed. An ensemble of 30 initial samples is constructed using WRFDA (Weather Research and Forecasting Data Assimilation) and then a 12-h forecast is produced on the basis of the initial samples. To determine the sensitive variables and their regions, the sensitivity of the accumulated precipitation in the verification region (Beijing area) to the basic variables of the initial conditions is studied using the method described above. A preliminary determination indicates that the sensitive variables are the water vapor and temperature, and their respective sensitive regions are southwest and northeast of Beijing. An analysis of the real atmosphere found that the sensitive variables and their sensitive areas have a clear physical meaning. Moreover, data assimilation of an observing system simulation experiment is used to verify the sensitive region. It is shown that assimilating the sensitive variable water vapor in its sensitive region can improve the precipitation forecast accuracy; assimilating the sensitive variable temperature in its sensitive region can also improve the forecast quality, indicating the correctness of the sensitive variables and sensitive areas. Assimilating the sensitive variable water vapor in its sensitive region while assimilating the sensitive variable temperature in its sensitive region can have a greater positive effect on the forecast quality, illustrating that the effects of temperature and water vapor yields the greatest improvement in the rainfall prediction accuracy. Therefore, the method of statistical correlation can effectively identify the sensitive variables and their sensitive regions. This study can help to identify the observed sensitive region of rainfall in the Beijing area.