Abstract: It is shown from the observational data that there is a phenomenon of “memory” lag in response to the information of well level. It is different from general phase lag. The on-the-spot experiment in the well Lu-29 is also shown the phenomenon of “memory” lag in water level response to the change of air pressure. Using a model of the pressured aquifer with horizontal layers, the phenomenon of “memory”lag in water level response to the change of air pressure is theoretically explained. It is considered that the phenomenon is related to the transmissibility in the well aquifer. The lower the transmissibility of the aquifer is, the more obvious phenomenon of “memory” lag will be. It is difficult to remove the influence of memory” lag in the well water level by using the general multiple regression. In this paper, a new method to remove this influence is proposed. The observation data in the well Lu-03 are processed by using this method, and the result shows that the tidal coefficient of the well water level is 1. 8 mm/10 8, the lag time of the well water level to the earth tide is 1. 53 h, the barometric coefficient is 6. 4 mm/hPa and lag time is 1. 50 h. The standard deviation of the water level corrected by the new method evidently decreases. The corrected result is better.

Key words: Groundwater level, Well water level, “Memory” lag, Transmissibility, Response