Abstract
Broadband seismometers, distinguished by their large dynamic range and wide bandwidth, have seen increasingly widespread application in earthquake early warning systems and seismological research in recent years. A quantitative investigation into the discrepancies in background noise Power Spectral Density (PSD) recorded by co-located broadband seismometers, operating with and without protective enclosures, is of substantial importance for enhancing the data quality and improving the utilization efficiency of these instruments. This paper utilizes co-located observational data from seismographic instruments (equipped with enclosures) and early warning sensors (without enclosures), installed at earthquake early warning reference stations in the Inner Mongolia region, to quantitatively investigate the noise reduction characteristics of seismometer enclosures across various frequency points, under different spatio-temporal conditions, for different components, and in diverse observational settings. The results demonstrate that subsequent to the installation of seismometer enclosures: Within the low-frequency band of 0.02-0.05 Hz, the enclosures effectively mitigate temperature fluctuations and airflow disturbances, thereby suppressing background noise. The efficacy of this suppression exhibits dependencies on both component orientation and frequency; specifically, the suppression of horizontal noise components exceeds that of the vertical component, with this noise-reducing effect becoming increasingly prominent at longer periods. The mean difference for the East-West component is 3.5 dB (median: 1 dB), while the mean difference for the vertical component is 2.2 dB. This characteristic is consistently corroborated by amplitude-squared coherence analyses performed on teleseismic event data (with the difference between the two components being approximately 0.2). Furthermore, surface-based installations benefit more significantly from such noise reduction than those situated in vaults or caves, a difference potentially attributable to the inherently greater thermal stability of subterranean environments. In the primary microseism band (0.05-0.1 Hz), the enclosures provide a discernible noise reduction effect, suggesting that the sources of primary microseisms are not solely oceanic in origin but are also modulated to some extent by the local environment proximal to the seismometer. Conversely, in the secondary microseism band (0.1-0.5 Hz) and the high-frequency band (0.5-40 Hz), the enclosures offer essentially no discernible noise reduction.