Real Time Data from US East Coast
Nelson-Nordgren, Astatic Spring, Broadband Force Balance, Vertical (FVB) Seismometers

Measuring Solid Earth Tides

Accurately measuring Earth tide amplitudes with a broadband seismometer, on a basement floor in a residential neighborhood, is difficult. To keep temperature variations to an acceptable level one would have to dig a deep hole and place the seismometer at the bottom. I decided to try the experiment, without digging the hole, to see if it was possible to see any evidence at all of Earth tides.

The experiment started with a Yuma Seismometer with a spring constant error of ~245 ppm/C, and a sensor error of ~43 um/s/C/hr. This would mean that a 0.01C change would result in output error of 430 nm/s, a sizeable amount. Given that my basement varies about 0.5C/day in season, and season transitions can see about 1C to 2C per day, it was unknown if any evidence of Earth tides would be seen. I buried the Yuma in 200 pounds of sand, to increase the thermal mass and began the experiment. Note: Do not do this for horizontal measurements.

The experiment was conducted for 45 days. I stopped the experiment due to extreme temperation variations, (2 C/day) due to the transition from Summer to Winter. Figure 1 shows the data resampled to 1 SPS from 100 SPS in sdrmanip, period extended to 0.0005 Hz and low pass filtered at 0.008 Hz in Winsdr. A final low pass filter (Figure 2) at 0.0000555 Hz (5 hrs.) was applied in SigView to smooth out the waveform.

Figure 3 is a plot of the spectral components of the Yuma data. Components with a period of 24 hours and greater are assumed to be temperature noise from the diurnal temperature changes in the basement environment.

The next set of plots (Figures 4 and 5) show solid Earth tide displacements calculated for Princeton Junction, NJ, from an Earth tide simulator "solid.exe"(1) and its spectral components.

The close match in semi-diurnal spectral components between simulation and the Yuma seismometer is reasonable evidence that Earth tides are being detected in the seismometer.

(Fig.1) Yuma data showing peaks in October due to above average temperature changes.

(Fig.2) Smoothed Yuma data with slight evidence of a periodic component.

(Fig.3) Spectral components of Yuma data

(Fig.4) "solid.exe" simulation data for Princeton Junction, NJ

(Fig.5) Spectral response of "solid.exe" results

Summary of results:

Simulation Results


45 Day "solid.exe" Spectrum

45 Day Yuma Spectrum

Darwin symbol

Period (hrs.)

f (Hz)

Period (hrs.)

f (Hz)

S2 (12.000 hrs.)





M2 (12.421 hrs.)





N2 (12.658 Hrs.)





Temperature Noise





K1 (23.934 hrs.)





O1 (25.819 hrs.)






(1) "Program "solid.exe" by Dennis Milbert is based on a version of the dehanttideinelMJD.f source code provided by Professor V. Dehant. This code is an implementation of the solid earth tide computation found in section 7.1.2 of the IERS Conventions (2003), IERS Technical Note No. 32". See for additional info. "Solid" does not implement ocean loading, atmospheric loading, or deformation due to polar motion.