When working with ultra-low-frequency (ULF) magnetic datasets, as with most geophysical time-series data, it is important to be able to distinguish between cultural signals, internal instrument noise, and natural external signals with their induced telluric fields. This distinction is commonly attempted using simultaneously recorded data from a spatially remote reference site. Here, instead, we compared data recorded by two systems with different instrumental characteristics at the same location over the same time period. We collocated two independent ULF magnetic systems, one from the QuakeFinder network and the other from the United States Geological Survey (USGS)-Stanford network, in order to cross-compare their data, characterize data reproducibility, and characterize signal origin. In addition, we used simultaneous measurements at a remote geomagnetic observatory to distinguish global atmospheric signals from local cultural signals. We demonstrated that the QuakeFinder and USGS-Stanford systems have excellent coherence, despite their different sensors and digitizers. Rare instances of isolated signals recorded by only one system or only one sensor indicate that caution is needed when attributing specific recorded signal features to specific origins.

The study of normal geomagnetic background fields is the premise and basis for obtaining local geomagnetic anomalies caused by changes in underground pressure during gas injection and extraction in an underground gas storage (UGS) area. Measurements of the total magnetic field in an array consisting of sixty-three stations were acquired at a UGS and its surrounding areas. We completed field measurements at the stations using proton precession magnetometers (PPMs) with a sensitivity of 0.15 nT @ 1 Hz in April 2017 and March 2018. The geomagnetic background field model of the UGS area is calculated using the Taylor polynomial fitting method based on the data of all measured stations. A fourth-order Taylor polynomial model is appropriate for the UGS area according to a comprehensive comparative analysis of the mean square error and boundary distortion effect. The geomagnetic field distribution in the core area of the UGS can be better described by enlarging the network of repeat stations. The results provide a reference model for extracting the variations in the local magnetic field (LMF) caused by the underground pressure variations in the UGS area.

We improved the surface spline method to make the centers of the base functions fixed. This improvement makes it meaningful to analyze the time series of coefficients. Based on the revised surface spline (RSS) method and 2382 geomagnetic data measured in mainland China, we built Chinese Geomagnetic Reference Field 2020 (CGRF2020) to describe the spatial distribution of the geomagnetic field in most of China and surrounding areas. The root mean square residual is 0.1318 degrees for D, 0.0857 degrees for I, 93.70 nT for F, 73.94 nT for X, 65.02 nT for Y, and 97.82 nT for Z. The values of F, X, Y, and Z are about 20 nT less than those by the SCH method and IGRF13. We also preliminarily analyzed the correlation between geomagnetic anomalies and geological structures.

As a basic technological product describing the geomagnetic spatial distribution, the national geomagnetic chart should choose an appropriate method to describe the geomagnetic field on standard epoch and its secular variations for the coming 5 years truly. We calculated the geomagnetic data at 1119 sites and 36 observatories in China by using the spherical cap harmonic (SCH) method, and obtained the space distribution of the normal and the anomalous geomagnetic fields in China for 2005.0 and the SCH model of geomagnetic secular variations in China for 2005~2010. The results showed that the errors of the geomagnetic field by IGRF could be ~ 5ʹ for the declination D and the inclination I and ~100 nT for the total intensity F. Because the geomagnetic survey data with new data and better resolution were used in the calculation, the geomagnetic chart for 2005.0 by the SCH method can better describe the space distribution of the geomagnetic field than the IGRF. This method can be stably used in compiling the China Geomagnetic Chart.

Using the data of the geomagnetic diurnal averages during 2000~2004 at 34 observatories in China, we calculated the mean annual change rates on the quiet days and all days, and the fitting annual change rates on the quiet days. The Taylor polynomial model and the spline model of the geomagnetic secular variation in China for epoch 2000.0~2005.0 were established based on the above annual change rates. The similarity and difference between the Taylor polynomial model, the spline model and the secular variation model of International Geomagnetic Reference Field (IGRF‐SV) were analyzed and discussed. The results show that the shape and the value of the Taylor polynomial model and the spline model for the geomagnetic secular variation in China are quite consistent, and their long‐term trends are also consistent with IGRF‐SV. However, the geomagnetic secular variation in China has some regional characteristics.