A kilometer-thick ice sheet covers more than 98 % of Antarctica's surface. Therefore, the historical evolution, geological structure, and tectonic activity underneath the Antarctic ice sheet are in large part not well known. Continuous, year-round seismic recordings provide a remedy to overcome the direct inaccessibility of the Antarctic continent. The recordings of local, regional, and teleseismic earthquakes have been used in various studies. Thus, our present knowledge of the structure of the Earth's mantle, lithosphere, and crustal structure underneath the Antarctic ice cover is based on these records (e.g., Knopoff and Vane, 1978; Danesi and Morelli, 2001; Ritzwoller et al., 2001; Lawrence et al., 2006; Janik et al., 2014; An et al., 2015; Lough et al., 2018). However, there is only little seismic activity originating from the Antarctic plate itself due to a low level of tectonic activity (Sykes, 1978). Moreover, the low seismic activity is paired with a sparse distribution of seismic instrumentation in Antarctica (particularly in East Antarctica; Fig. 1a) and is thus difficult to verify. Only a few long-term seismic observatories exist. Most of them are constrained to the coastal region and in direct vicinity to research infrastructure (Fig. 1).

Many seismic experiments, both those using active source and those using passive sources, require numerous seismic stations. The deployments often have to be done in remote and difficult to access areas with very limited power supply and servicing infrastructure. In particular, a long-term AC power supply is not available in most cases. For areas with moderate climatic conditions or when batteries can easily be changed or replaced in regular intervals, this is not a problem. Furthermore, state-of-the-art solar panels provide sufficient electrical power for efficiently charging batteries, enabling instrument operation throughout almost the entire year. In polar regions, however, significant challenges arise in terms of the geographical setting, remoteness, and extreme weather conditions, which require a sophisticated power supply design. First, long periods of the dark polar winter with no sunlight available make it necessary to install an additional power supply. If sufficient backup power cannot be realized, it has to be taken into account that data acquisition will stop at some point during polar winter. Second, due to the low temperatures and high discharge of the batteries, it may occur that data acquisition cannot resume when sufficient sunlight is available after the winter break. Additionally, almost all types of batteries show reduced performance at low temperatures and show a substantially reduced effective capacity. Therefore, energy-efficient and for low-temperature-adapted renewable systems are required for a sustainable operation in polar environments (Tin et al., 2010).

A major step in this direction was realized with the large international POLENET project (Polar Earth Observing Network) within the activities of the International Polar Year (IPY) 2007–2009. In this project, a large number of seismic and GPS instruments were installed in remote sites in Antarctica for several years. The equipment required for POLENET was developed by IRIS (Incorporated Research Institutions for Seismology) PASSCAL (Portable Array Seismic Studies of the Continental Lithosphere) with a focus on a cold-resistant power and communication system that is easy to install and that can withstand Antarctic weather conditions. In this project, large-scale temporary coverage of West Antarctica up to the Transantarctic Mountains, as well as central parts of East Antarctica, was realized for the first time (Fig. 1a).

To ensure the continuous extension of seismic coverage in the polar regions, it is essential to find new solutions to the same problems again and again. For polar seismology, it is therefore important to build on previous experience (e.g., from IRIS PASSCAL) to optimize the use of self-sufficient seismometer stations and to find flexible solutions for the different survey areas and deployment lengths. The specifications of the stations must also be realizable with the available resources and be based on long-term scientific goals.

In this article, we describe the concept of an in-house-developed mobile and self-sufficient seismological broadband station designed for the extreme demands of the Antarctic ice sheet. A focus lies on (i) the compact modular design and conception of an energy supply to operate under extreme temperatures between -20 and -40 ∘C (with slightly warmer temperatures in summer and colder temperatures in winter on average) and (ii) to present strategies to get the system through the sunless polar winter. Our layout makes the system suitable for long-term operations over several years without regular maintenance and for shorter surveys. Some of the concepts presented here are already in use at numerous seismological stations in western Dronning Maud Land (East Antarctica) operated by the geophysical observatory of Neumayer III Station (Fig. 1). Other concepts presented here represent extensions for current and future projects.

The Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI), has been operating a local seismographic network for more than 2 decades in the vicinity of its permanent base Neumayer III Station (Fig. 1b). One permanent seismometer is located inside the geophysical observatory close to the station (VNA1), and two other permanent seismometers (VNA2 and VNA3) are deployed at ice rises at approximately 45 and 85 km distance off Neumayer III Station (Eckstaller et al., 2007). The data quality of VNA2 and VNA3 is substantially better than VNA1 data because the latter is stationed on the ice shelf and the former two are stationed on grounded ice (Fig. 1b). Data are continuously transmitted to the base in near-real-time quality via high-speed terrestrial data radio. This local network is supplemented by six offline remote semi-permanent seismic stations (Fig. 1b; Novolazevskaya, NVL; Kohnen Station, KOHN; Svea Station, SVEA; Forstefjell Nunatak, DS4; Weigel Nunatak, WEI; and Utpostane, UPST). In addition, several temporary single mobile seismic stations or arrays have been deployed in the vicinity of Neumayer III Station for testing purposes and geophysical surveys.

The permanent and mobile temporary seismological stations of the regional AWI seismographic network are located in different glaciological regimes in western Dronning Maud Land and thus are affected by different snow accumulation rates. None of our stations is located in an ablation area. Snow accumulation on the plateau (e.g., KOHN at Kohnen station; Fig. 1b) ranges between 15 and 20 cm yr-1. By contrast, we observe several meters per year of snow accumulation at the coastal stations (e.g., 3 m yr-1 at VNA3). Depending on the local snow accumulation, the components of the seismological stations, as well as the solar panels or masts, must be relocated to the ice surface, otherwise, they will be buried by the snow over a longer period. This action is mandatory once a year for VNA3 on Sörasen and every 3–5 years for the stations on the plateau. Some stations (e.g., UPST, SVEA, WEI; Fig. 1b) are located on nunataks where we observe neither significant snow accumulation nor ablation.

The motivation for developing and optimizing mobile stations is to use them for temporary regional array studies in Dronning Maud Land. The ambition is to use a moving array of seismic stations to acquire data for one to two years and relocate the instruments after that to a new site. Our scientific interests will focus on the analysis of the regional tectonic seismicity associated with a potential neotectonic activity and the analysis of receiver functions to determine the Moho depths and eventually resolve major structural features in the upper mantle in this region.

Our requirements for the mobile seismometer stations must allow a rapid installation and be as modular and compact as possible to enable economic transport and fast deployment and recovery. A single mobile station comprises a solar panel (rack included), a seismometer with casing, one recorder box, and one to two battery boxes (Fig. 2). For the instrument boxes, we use Peli ISP2 CASES (Inter-Stacking Pattern Cases) boxes because they are waterproof and mechanically stable at low temperatures (Fig. 3). A single mobile station in its minimum configuration (one seismometer with casing, one seismic recorder, two AGM batteries, one solar charge and iridium controller, two solar panels mounted on one rack, one GPS, one iridium antenna, and all cables; Fig. 3a–e) weighs ∼ 140 kg and has ∼ 2–3 m3 of storage space.

We can equip our stations with two data logger types: three-channel Reftek RT-130 and six-channel Quanterra Q330S+ (or Quanterra Q330 + baler) recorders (Table 1). Both logger types can be deployed at permanent or temporary mobile seismic stations. The power drain of the recorders is approximately 50–83 mA and depends on the number of active channels, sample rate, and desired GPS-clock operation. The advantage of the Quanterras is the lower power consumption and the larger storage space. In addition, the Quanterra is easier and more versatile to configure (via a web page GUI from any computer) and has more modern interfaces. However, in contrast to the Reftek data loggers, they are also more expensive.

We commonly use Guralp CMG-3ESP and Kinemetrics Metrozet MBB2 (three-component) broadband seismometers with a lower corner period of 120 s, and in some cases we also use Lennartz LE-3D/20s seismometers. The only exception represents UPST station, where we have deployed a Streckeisen STS-2 and a small short period tripartite array. A relevant disadvantage of the Guralp seismometer for the mobile stations is that during transport the mass must be locked to prevent damage. In addition, the instrument must be manually leveled during installation. The advantage of the Metrozet MBB-2 seismometers is the compact design and the higher transport safety, as it is self-locking and able to center its mass automatically during operation. In addition, the power consumption is very low for an active sensor (20 mA) in comparison to the Guralp or Lennartz (50 mA) seismometers.

The solar-powered energy supply system consists of 100 W Solara S405M36 Ultra solar cells and a Morning Star SunSaver SS-MPPT-15L charge controller. Every seismic station is equipped with a state of health (SOH) transmitter that sends the station's operation status in regular intervals once a day via iridium satellite radio to AWI. For the Quanterra Q330 recorders, we use XEOS XI-202 controllers because they have an existing interface. However, this interface is not available for the newer Q330S+ recorders. For the RT-130 we use a custom-made iridium controller (SeiDL – Seismic Data Link) to influence all parameters and configurations. For example, it gives us the possibility to transmit data from additional environmental sensors, such as wind, temperature, solar radiation, current, and voltage (if available). This controller was developed by Arne Schwab (SchwaRTech, based near Bremen, Germany) and also uses the iridium short burst data (SBD) transmission technique. The average power drain of our iridium controllers is very low with 0.1 mA in sleep mode and 50 mA in transmission mode (2 min d-1). The overall power consumption of an entire single mobile station is 4 Ah d-1 and 1447 Ah yr-1. The calculation is based on a station design that comprises a Guralp CMG-3ESP seismometer and a Reftek-130 data logger. For the wiring of all devices, we have moved away from PVC insulated cables, as they are too brittle at low temperatures. Now we use almost exclusively more flexible cables with PE or PU insulation with improved UV and cold resistance. A complete list of the specifications of the instruments is provided in Table 1.

In addition to the concepts currently under development, we also have ideas for subsequent developments. Above all, we see much potential in optimizing battery management and input energy management. For example, a multiple-battery option would be desirable, in which individual batteries are charged step by step after the polar winter when the current flow is low so that a high voltage is available quickly. It would also be desirable to disconnect deeply discharged batteries from the overall system. In terms of input energy management, a variant is conceivable in which a wind generator is switched on exclusively in the polar winter. This would close the energy gap in the polar winter (with the acceptance of increased noise in the data) and generate no noise during the summer season while recording data. Another possibility to reduce the noise influence and material stress of the wind generators would be to switch on the wind generator only for a particular time when the total voltage of the batteries drops below a certain range.

We have presented a fast and easy to deploy modular, compact, mobile and self-sufficient seismometer station concept for the polar regions. Due to its modular design, it can be used in various ways, for example, for short-term deployment as an array over 1–2 years or as a long-lasting permanent station. The energy supply can be adapted as required using the modular cascading of battery boxes, wind generators, solar cells, or backup batteries, which enables optimum use of limited resources. The stations' modules are designed so that only the cables have to be connected in the field. Parts of the concepts presented here are already in use as part of the extended seismology network of the Neumayer III Station. Our system concept is not specifically limited to the application to seismology stations (except for noise suppression) and can also be extended by additional instruments with low power consumption (e.g., to monitor environmental parameters). Moreover, it is a suitable system for managing the power supply for all types of self-sufficient measuring systems in polar regions.