Abstract: LZDM is one of the 50 primary stations of the International Monitoring System (IMS), and the capacity to monitor microseismic activity in Lanzhou and its surrounding areas has been significantly enhanced by LZDM. Welch average periodogram method was used to estimate the background noise level of LZDM. The continuous waveform data recorded by nine elements of LZDM during 2018—2022 were systematically processed, and two methods, Power Spectral Density and Power Spectral Probability Density Function, were used to analyze the background noise in different time dimensions. The characteristics of noise in the frequency domain and cyclical variation were analyzed, and the noise sources were surveyed. Finally, suggestions on data processing were put forward considering the actual situation of Lanzhou Array. The research result shows that the power spectral density of long-period noise for nine elements showed significant annual variation with periodicity: It decreased from May to September and increased from September to May of the next year. The power spectral densities of short-period noise for nine elements were relatively high and showed significant diurnal and annual variation. The noise level near 3 Hz reached the peak and could be suppressed by narrowband filtering. Because A0 and B2 were affected by road traffic, the PSD values between 7 and 20 Hz were relatively high for the two elements, which reached -100 dB and -110 dB respectively, and could be suppressed by low-pass filtering or beamforming techniques. There were more instrument malfunctions with B4 compared to the other elements, and its data could be excluded when necessary. The research results of this paper could provide important references for the data processing and operation and maintenance of Lanzhou array.

Key words: Lanzhou array, Background noise, Power spectral density, Power spectral, Probability density function