The Future is the Past: Challenges with and the Scientific Value of Legacy Seismic Data

Abstract:

Since ancient times humans have been fascinated with earthquakes and sought ways to understand the movement of the ground. From the recording of the first distant earthquake in 1889, which marked the beginning of modern seismology, until the dawn of the digital era at the end of the 20th century, nearly all observations of ground movements were recorded on large sheets of paper. Today, seismologists are awash in terabytes upon terabytes of digital data which, however, cover only a few decades-worth of time. On the other hand, archives of analog seismic data stretch back nearly a century. When these legacy data are considered alongside their modern digital counterparts, the length of time for which seismic data are available more than quadruples. The longer time period allows for “unusual” or “once-in-a-lifetime” events such as above-ground nuclear tests or very large earthquakes to be studied, and moreover, allows for the exploration of processes or phenomena such as the earthquake cycle or the temporal variability associated with slow changes of the Earth.

Before exciting science can happen with these data, the analog records must first be scanned and digitized. This transforms them from deteriorating sheets of paper or film into digital timeseries which can be exploited in the same ways as we manipulate their modern counterparts. In addition, this unique legacy data set comes with its own challenges. For example, without GPS clocks, timing of the record is significantly less precise, and there may be artifacts in the time series due to the mechanical motion of analog seismometers. Furthermore, the analog storage media and how they are scanned may introduce more complications. We present methods that address some of these issues in order to facilitate conversion from analog to digital format and to prepare data for rigorous scientific analyses. Finally, initial results of storm-induced ground motion recorded in the 1930s and their relationship to hurricanes will be presented. These efforts mark first steps towards more widespread use of legacy seismic data, and opens an exciting new possibility that legacy seismic data can be used to test and quantify the hypothesis that the strength of storms has changed through the past century with potential implications regarding human-induced climate change.