The adaptive observation system based on ETKF(Ensemble Transform Kalman Filter) has been applied to identify sensitive observation area in Jianghuai heavy rain, typhoon, freezing rain disaster and etc.This paper considers the function of humidity factor, increases the information of atmosphere in low troposphere, and develops an optimization scheme to make the system more suitable for high impact weather in China.Selecting summer heavy rain and winter snow around Beijing as high impact weather examples, four groups of test schemes are designed for identifying observation sensitive area.This paper investigates the quality of observation sensitive area and evaluates the results of analysis and forecast.Results show that the optimized adaptive observation system is the best scheme, which has significantly improved the quality of observation sensitive region.The humidity factor can make the strongest observation sensitive area more concentrative, which is more obvious in the heavy rain than in the snow.The information of low-level atmosphere is helpful in identification of the strongest observation sensitive area.Target observations in the sensitive area have positive contribution for the quality of analysis and short-term forecast.