Two observations about low-frequency signals on accelerograms from the october 15, 1979 imperial valley,california earthquake

Accelerograms from the Imperial Valley have been integrated and baseline corrected by a simple procedure which keeps track of the signal that is removed at long periods. Specifically, acceleration is integrated once to obtain velocity, a linear baseline is subtracted from velocity, the velocity is integrated to displacement, and low order terms of a trigonometric expansion of displacement are subtracted. It is observed that even though this procedure is not suitable for routine processing, one obtains displacement traces which usually resemble the results obtained by standard methods. It is also observed that amplitudes of the trigonometric series expansions are consistent with a simple earthquake source theory for frequencies greater than 1/12 Hz, suggesting that the records are dominated by signal at those frequencies.