3.3. Instrumentation/Specifications

The parameters listed below describe the specifications of a complete observatory system for a one-second vector data set including recording environment, magnetometer, and data processing procedure. INTERMAGNET does not specify the type of magnetometer that must be used (although a list of instrument suppliers can be provided on request), however a 1s-IMO must meet the specifications listed below. These specifications are used by INTERMAGNET as a guideline for reviewing data quality when data are either submitted for an IMO application, or when definitive data are submitted for inclusion in the archive. A scalar magnetometer is mandatory as well as temperature measurements of the vector magnetometer electronics and sensor with a resolution of 0.1°C.

3.3.1. Definitive Data

Due to advances in instrumentation and observatory practice, the INTERMAGNET requirements for one-second definitive data are more stringent (and defined more comprehensively) than those for one-minute data described in Chapter 2. The parameters listed below describe the specifications of a complete observatory system for a one-second vector data set including recording environment, magnetometer, and data processing procedure. These specifications are designed such that definitive one-minute data may be derived from definitive one-second data by means of a suitable filter.

The absolute accuracy of one-second definitive data is specified below as a maximum offset error of ±2.5 nT. This is the maximum low frequency error between absolute observations. In general, the largest source of offset error is instrumental or thermal drift of the variometer, therefore the frequency of absolute observations can be determined by the observer based on the stability of the variometer, its operating environment and its baselines.

The definitive one-second data standard is described in more detail in INTERMAGNET Technical Note TN6.

3.3.2. Absolute Measurements

The requirements are the same as for 1 minute data. See Chapter 4

3.3.3. Vector Magnetometer

3.3.3.1. General Specifications

Time-stamp accuracy	0.01s centered on UTC second
Phase response	Maximum group delay ±0.01s
Maximum filter width	25 seconds
Dynamic Range	±4000nT High Latitude ±3000nT Mid/Equatorial Lat.
Data resolution	1pT
Pass band	DC to 0.2Hz
Thermal stability	0.25nT/°C
Long term stability	5nT/year

3.3.3.2. Pass Band Specifications [DC to 8mHz (120s)]

Noise level	100pT RMS over 10 min
Maximum offset error (cumulative error between absolute observations)	±2.5nT
Maximum component scaling & linearity error	1%
Maximum component orthogonality error	2mrad
Maximum Z-component verticality error	2mrad

3.3.3.3. Pass Band Specifications [8mHz (120s) to 0.2Hz]

Noise level	10pT/ Hz at 0.1Hz
Maximum gain/attenuation	3dB

3.3.3.4. Stop Band Specifications [0.5 Hz]

Minimum attenuation in the stop band (0.5 Hz)

50 dB

3.3.4. Scalar Magnetometer

Resolution	0.01nT
Accuracy	1nT
Sampling rate	0.033Hz (30 sec)

3.3.5. Auxiliary Measurements

Compulsory vector magnetometer electronics and sensor temperature measurements with a resolution of 0.1°C at a minimum sample period of one minute.

3.3.6. Recorder

An on-site recorder is necessary so data are not lost as a result of data transmission outages.

3.3.7. Transmission

Transmission by electronic means to a Geomagnetic Information Node (GIN) must be within 72 hours of acquisition or sooner (see Section 6.1.4).

3.3.8. Other

Data format	IAGA2002 or ImagCDFV1.10 (or later)
Definitive data	to be submitted for inclusion in the INTERMAGNET archive
Baseline data	each component to be submitted for inclusion in the INTERMAGNET archive
Filtering	to INTERMAGNET standard ( Section 3.4 )

For more information see Section 6.1.1 for Data Formats and Section 6.1.2 for Data Types.

The consensus of the scientific community survey was that one-second data should be accurately timestamped and the instruments should have linear phase response. Hence, a maximum time-stamp error has been specified and the phase response quoted in terms of a maximum group delay, which limits the non-linearity of the phase. Data samples may be time-shifted to correct for latency (e.g. instrument response and filter delay) provided that the system phase response is met. The quoted instrument ranges are inherited from the INTERMAGNET one-minute specification, while a data format resolution is specified to reduce quantization noise and a maximum filter width is set to minimize the time extent of the system response to a step input i.e. filter ringing.

Observatories moving from one-minute recordings to one-second recordings having definitive data quality will need to monitor over both a larger frequency band and, due to the spectrum of the natural magnetic field, a larger dynamic range. To meet these stringent measurement requirements while ensuring that the instrumentation standards are realistic, the pass band has been split into two bands with different specifications:

1. the existing INTERMAGNET one-minute data band (DC to 120s) and

2. the extended high frequency band (8mHz to 0.2Hz)

For the low frequency band (DC to 120s), there is a higher system noise level limit than for the high frequency band, but more constraints on parameters affecting the absolute accuracy, such as sensor orthogonality errors, scale and offset errors. The offset error is expressed as a maximum error from all sources (including instrument and thermal drift) between absolute observations.

In the high frequency band of the pass band (8mHz to 0.2Hz), the noise level is set at a lower level to ensure sufficient resolution of low amplitude fluctuations in the geomagnetic field that may occur in this frequency range. Since absolute signal amplitude is not as critical in this band as it is in the low frequency band, and to allow for instrument roll-off with sufficient attenuation in the stop band, the maximum signal gain/attenuation is specified at a less stringent 3dB in the high frequency band.

The stop band starts at the Nyquist frequency, allowing for a sufficiently wide transition band to set a high stop band attenuation. This is necessary to attenuate typical natural signals in order to avoid errors due to aliased signal in the pass band and meet the noise specification. However, this specification alone will not sufficiently attenuate large amplitude artificial signals, such as 50/60 Hz mains interference. Hence it is recommended to separately attenuate non-natural, large-amplitude signals above the Nyquist frequency.