The record of seismic, hydroacoustic, and infrasonic waves is essential to detect, identify, and localize sources of both natural and anthropogenic origin. To guarantee traceability and inter-station comparability, as well as an estimation of the measurement uncertainties leading to a better monitoring of natural disasters and environmental aspects, suitable measurement standards and reliable calibration procedures of sensors, especially in the low-frequency range down to 0.01 Hz, are required. Most of all with regard to the design goal of the Comprehensive Nuclear-Test-Ban Treaty Organisation’s International Monitoring System, which requires the stations to be operational nearly 100% of the time, the on-site calibration during operation is of special importance. The purpose of this paper is to identify suitable excitation sources and elaborate necessary requirements for on-site calibrations. We give an extensive literature review of a large variety of anthropogenic and natural sources of seismic, hydroacoustic, and infrasonic waves, describe their most prominent features regarding signal and spectral characteristics, explicitly highlight some source examples, and evaluate the reviewed sources with respect to requirements for on-site calibrations such as frequency bandwidth, signal properties as well as the applicability in terms of cost–benefit. According to our assessment, earthquakes stand out across all three waveform technologies as a good natural excitation signal meeting the majority of the requirements. Furthermore, microseisms and microbaroms allow a calibration at very low frequencies. We also find that in each waveform technique man-made controlled sources such as drop weights or air guns are in good agreement with the required properties, although limitations may arise regarding the practicability. Using these sources, procedures will be established allowing calibration without record interrupting, thereby improving data quality and the identification of treaty-related events.

<p><span>As part of the joint research project "Metrology for low-frequency sound and vibration - 19ENV03 Infra-AUV" laboratory calibration methods for seismometers and microbarometers in the low frequency range down to 0.01 Hz have been developed. These procedures provide the possibility of traceable on-site calibration during operation for field sensors of the Comprehensive Nuclear-Test-Ban Treaty Organization’s (CTBTO) International Monitoring System (IMS). The traceable calibration allows for accurate amplitude and phase information as well as for an assignment of uncertainties in amplitude and phase. Thereby, data quality and the identification of treaty-relevant events is improved. The on-site calibration procedure requires a reference sensor with a precise and traceable response function which is provided by the newly developed laboratory calibration methods, as well as the record of sufficient coherent excitation signals within the relevant frequency range. The reference sensors can be installed as transfer standards co-located to the operational IMS station sensors without disturbing their regular measurements for treaty validation purposes.</span></p><p><span>At IMS stations PS19 and IS26 in Germany we performed on-site calibration tests with both seismometers and microbarometers calibrated in the laboratories at PTB and CEA, respectively, using signals from different natural and anthropogenic excitation sources. Following the approach of Gabrielson (2011) with modifications from Charbit et al. (2015) and Green et al. (2021), the gain ratio between the station sensor under test and the reference sensor is calculated. By multiplying the gain ratio with the precise frequency response of the reference, the frequency response function for both magnitude and phase of the station sensors including site-specific factors such as the wind noise reduction system or possible effects of pre-amplifiers and data loggers are determined.</span></p><p><span>We present calibration results derived from the comparison of IMS station sensors with the laboratory-calibrated instruments along with the nominal responses. The results show agreement with deviations of less than 5% from the nominal response function for frequencies below 10 Hz for all components. The traceable determination of the response for the individual components in detail improves the sensor quality; subsequently waveform amplitudes can be estimated correctly.</span></p>

<p>Infrasound signals can be detected using a time-delay of arrival approach to derive the back azimuth and trace velocity of the coherent wave. For these calculations, it is necessary to have a calibrated measure of the pressure. Although the calibration of microbarometers can be performed in a laboratory setting with specific metrological means such as those developed by the CEA, it is much more difficult to determine the transfer function of the wind noise reduction systems (WNRS), designed to reduce the wind associated noise. In-situ calibration of these WNRS’s can be performed (as described by Gabrielson*) using a co-located reference sensor and comparing the response to that of the array sensor (considering only highly coherent signals) to determine the relative response of the WNRS. System defects, such as flooded pipes or blocked inlets, have significant impacts on the response, which in turn would influence the calculated infrasonic wave parameters. These defects can be characterized using in-situ calibration measurements. To demonstrate the importance of these measurements, experiments were undertaken at the infrasound station IS26, using a temporary detector whose defects on the WNRS can be produced. This will allow for the effects on real infrasound detections to be quantified and corrected using in-situ calibrations. Comparisons between models of these defects and experimental results allow for the characterization of their effects on infrasound parameter measurements and improvements of the models and WNRS designs.</p><p>* Thomas B. Gabrielson, “In situ calibration of atmospheric-infrasound sensors including the effects of wind-noise-reduction pipe systems”, The Journal of the Acoustical Society of America 130, 1154-1163 (2011) https://doi.org/10.1121/1.3613925</p>

<p>Transfer standards have an important role in calibration, in enabling traceability to be transferred across calibration facilities or physical locations. It is often the case in laboratory calibration, that the sensors best suited to achieving the optimum calibration accuracy, or measurement uncertainty, have different characteristics to those that need to be deployed in the field. Laboratory calibration techniques are often tailored to work with sensors of a specific type or form factor, and requirements on ruggedness and tolerance of a wide range of environmental conditions are usually of lesser importance under laboratory conditions. Conversely, a transfer standard is ideally suited to both laboratory and field environments. The Infra-AUV project is developing a complete calibration-chain solution to establish measurement traceability for IMS measurements, and consideration has therefore been given to the specification of suitable transfer standard sensors. The thought process behind this, and the infrasound sensors, hydrophones and seismometers ultimately proposed for designation as transfer standards will be presented, together with some characteristic calibration data.</p>

<p>At present, seismometers are not traceably calibrated. This means that their output sensitivity is not determined in a way that is traceable to the International Systems of Units (SI). The European research project <em>19ENV03 </em><em>Infra-AUV</em>, which is part of the EMPIR programme, develops methods and procedures to enable such traceable calibrations.</p> <p>In contrast to many other sensors, seismometers are operated stationary in their typical measurement application, i.e., they must not be moved after their deployment. Conventional calibration approaches which involve a laboratory calibration of the seismometer to be calibrated are therefore not feasible. For this reason, a new concept currently developed by different European partners within the <em>Infra-AUV</em> project proposes an on-site calibration scheme.</p> <p>For the on-site calibration, a reference seismometer is traceably calibrated to the SI in a laboratory. This reference is then used on-site to provide a secondary calibration of other seismometers, e.g. in a seismic station, using natural excitation sources [Schwardt et al., 2022, DOI: 10.1007/s10712-022-09713-4].</p> <p>The calibration of reference seismometers in the laboratory is carried out as a primary calibration. This means that the measured quantity (the velocity-proportional voltage output) is compared to a different quantity, in this case to a dynamic displacement measurement traced back to the units length and time, which can be measured very precisely by laser interferometry. In this calibration, the seismometer is excited with low-frequency mechanical vibrations generated by electrodynamic exciters. These calibrations must be performed for the horizontal and vertical axes. The frequency range of interest is from 20 Hz down to 0.01 Hz, depending on the seismometer under test. Either mono-frequency sinusoidal excitations of different frequencies are applied subsequently, or multiple frequencies are excited simultaneously using a multi-sine approach. The magnitudes and phases of both measured signals, the interferometric reference and the seismometer under test, are determined by using sine approximation algorithms or by applying a discrete Fourier transform (DFT).</p> <p>The results of the laboratory calibration, the transfer function of the reference seismometer, can then be derived from the ratios of the measured magnitudes and the differences of the phase angles for the different excitation frequencies. In addition, the associated measurement uncertainties are estimated and are part of the calibration result. Influences that may change the sensitivity of a seismometer, e.g., temperature effects, electromagnetic sensitivity, or ground stiffness need to be analysed and additionally taken into account for the uncertainty estimation.</p> <p>For the uncertainty of the on-site calibration, differences between the laboratory and the on-site environment also need to be taken into account. This includes, for example, aspects like typically different temperatures or different ground materials.</p>