For observations of vertical‐component acceleration in the normal‐mode band (0.3–10 mHz), the detection sensitivity for signals from the Earth’s body can be improved to levels below the Peterson low‐noise model (PLNM). This is achieved by deterministic procedures that (at least partly) remove the accelerations originating from atmospheric mass fluctuations. The physical models used in such corrections are still too simple and fail at frequencies above 3 mHz. Anticipating improved atmospheric correction procedures, we explore the prospects of lowering the detection level. From recordings of excellent vertical‐component sensors operated under exceptional site conditions at the Black Forest Observatory, we select time windows of very low background signal, for which all of the contributing broadband seismometers showed their best performance. Streckeisen seismometers of type STS‐1, STS‐2, and STS‐6A, a Nanometrics Trillium T360, and the superconducting gravimeter (SG) SG056 manufactured by GWR Instruments take part in this comparison. Because of their low level of self‐noise, the STS‐1 and the SG056‐G1 benefit the most from a correction with the best currently available improved Bouguer plate model for atmospherically induced signals at frequencies below 1 mHz. As far as we know, this is the first case in which the background level of a broadband seismometer could be lowered below the PLNM. At signal periods beyond the normal‐mode band (investigated up to 12 hr), the gravimeters show the lowest level of self‐noise, directly followed by the STS‐6A. In the band from 0.3 to 10 mHz, the STS‐1 has the lowest level of self‐noise, which is at least 4 dB below the PLNM, directly followed by the T360 and the STS‐6A. Sensors of lower self‐noise than the currently manufactured STS‐6A or T360 are needed before improved atmospheric correction procedures lead to a significantly lower vertical‐component detection threshold.

中文翻译：

---

降低垂直分量长期检测水平的挑战和前景

对于正常模式波段 (0.3-10 mHz) 中垂直分量加速度的观测，地球体信号的检测灵敏度可以提高到低于彼得森低噪声模型 (PLNM) 的水平。这是通过确定性程序实现的，该程序（至少部分）消除了源自大气质量波动的加速度。此类校正中使用的物理模型仍然过于简单，并且在高于 3 mHz 的频率时失效。预计改进大气校正程序，我们探索降低检测水平的前景。从黑森林天文台在特殊现场条件下运行的优秀垂直分量传感器的记录中，我们选择了非常低背景信号的时间窗口，所有贡献的宽带地震仪都表现出最佳性能。STS-1、STS-2和STS-6A型Streckeisen地震仪、Nanometrics Trillium T360和GWR Instruments制造的超导重力仪（SG）SG056参与了这次比较。由于它们的低水平自噪声，STS-1 和 SG056-G1 从使用目前可用的最佳改进布格板模型的校正中受益最大，该模型用于频率低于 1 mHz 的大气感应信号。据我们所知，这是第一个可以将宽带地震仪的背景水平降低到 PLNM 以下的情​​况。在正常模式波段以外的信号周期（调查长达 12 小时），重力仪显示出最低水平的自噪声，紧随其后的是 STS-6A。在 0.3 到 10 mHz 的频段内，STS-1 的自噪声水平最低，至少比 PLNM 低 4 dB，紧随其后的是 T360 和 STS-6A。在改进的大气校正程序导致显着降低垂直分量检测阈值之前，需要比当前制造的 STS-6A 或 T360 具有更低自噪声的传感器。