A potentially good site for a broad band seismic station is on hard-rock and far away from cultural noise. However, in remote places, such as Baja California, theft or vandalism of the equipment is a great concern. Therefore the stations have been selected with security as a priority, sometimes at the cost of a higher ambient noise.
All stations are housed in a purpose built shelter, made of hollow concrete blocks (figure 1). Roof and floor consist of reinforced concrete. To avoid ambient noise as much as possible, a seismometer platform is constructed on an outcrop which is decoupled from the structure of the building. Extra attention has been paid to thermal insulation. The inside roof and walls are insulated with 5 cm Polystyrene plates, carefully glued together, forming the insulating inside cabin. Between the Polystyrene wall or roof, an air space was left of 4 cm. In a few places of the outside wall, small gaps are left open to ventilate the air. Outside walls are white and the roof has been painted with sun reflecting paint. A plastic tube guides the cables from the roof to the inside cabin. The entrance door is made of insulated polyester, and where possible located on the shadow side of the shelter
The seismometer should be oriented as close as possible along the geographical NS/EW direction. A magnetic compass measures the direction to the magnetic pole, but has to be corrected for the local declination. Additionally, the presence of metal in the ground or construction and even solar wind can strongly disturb the measurement. To avoid all this uncertainty, we used a gyro compass (figure 2) in combination with two handheld GPS receivers to measure the geographical direction. We used the following procedure.
Two people with both handheld GPS receiver, read the same latitude/longitude on a position P1. Both agree on either latitude or longitude to be kept fixed on their GPS receiver. One person starts walking for approximately 150 meters to a position P2, keeping latitude or longitude to the agreed value. Possibly the person at P1 has to move a few meters to hold the same value. When both persons receive the fixed value, P1 and P2 are marked. Depending on latitude or longitude, the line (P1, P2) represents an exact geographic East/West or North/South line. The gyro compass is exactly positioned along the line (P1, P2) and calibrated. Once the gyro compass has been set, the shelter can be entered to position the seismometer (figure 3).
The absolute position read on the GPS receiver is not important. Both receive the same satellites with the same error at the same time. The difference between both receivers will decrease the direction accuracy. In practice this is better than 2 meters. To fix ideas, with a 4 m accuracy the error in orientation is only 1.5 degr (atan 4m/150m) .
Most sites are located far away from any power supply, therefore solar panels have to be used as a power source. The total power consumption of a NARS station is only 20 Watt. Depending on the local situation, 3 or more solar panels of 55 Watt each are installed on the roof of the shelter (figure 1). An intelligent charge controller regulates the power and maintains the 2 x 210 Amp Hour solar batteries. With fully charged batteries and a total absence of sun, the station will operate for at least 5 days. Depending on the battery charge, the power monitor will shut down or reboot the data logger system (figure 4).
Thermal insulation has a big influence on the overall performance of the broad band seismometer. To achieve this thermal shielding, the sensor is totally covered in sand. To protect the connector of the sensor from dirt, we first placed it in a thin plastic bag (figure 5). After installation and positioning, a wooden box with a 5 cm polystyrene inside insulation is placed over the edge of the platform.
The box is carefully filled with dry sand and closed with a wooden lid, also insulated with polystyrene (figure 5, 6). This simple construction has three big advantages: (1) long thermal time constant for the sensor, (2) full access to the seismometer for installation, and (3) easy removal of the sand by just lifting up the box.
Although the seismometer has been isolated as good as possible, long period temperature and barometric variations are still visible in the seismic signal. We installed a temperature sensor inside the rod hole of each STS-2 and in one station a barometric pressure sensor. The pressure and temperature variations are recorded together with the seismometer signals and can be used for signal corrections.