First arrival tomography-derived velocity models for Oneida Lake, New York

The data set contains velocity models derived from first arrival seismic tomography methods applied to a seismic survey collected for reflection seismic purposes. The original data set was collected as a reflection seismic survey and therefore the data set did not contain forward and reverse shots. Additionally, unlike most seismic tomography surveys, the seismic source and the hydrophones or receivers remained at a constant distance (e.g. the source was always the same distance from channel 1 or channel 144). The source and streamer (array of receivers) were towed behind the vessel and the source or air guns were fired every 6.25 meters. Shot gathers in SEG-Y format were imported into the Geogiga DW TOMO software in 10 shot increments with spacings of ~62.5 meters. This provided a spatially-dense data set for every line, with overlapping ray paths. This software uses ray tracing or the shortest path method in the forward modeling process (Moser, 1991). The forward model or initial velocity model is a sub-surface velocity model defined by the user. The intent of the tomography-derived velocity models was to independently confirm results from seismic reflection data, as well as provide information on the depth to basement in areas where reflection data failed to image the sub-surface stratigraphy. Therefore, a priori data from the reflection data set was limited in the tomography inversion process. The initial model contained a surface velocity of 1350 m/s and increased to 1600 m/s over a depth of 15 meters. Below 15 meters the model increased by 390 m/s every 1 meter to reach a velocity of 5500 meters at 25 meters depth. The same forward model was used for every line, and each line was subjected to 9 iterations with a maximum disturbance of 75%. The inversion process used an algorithm adapted from Toomey et al. (1994) which accounts for previous knowledge such as the prior iteration, smoothing parameters, and travel time uncertainties which are established by the user. Following careful parameter testing, the smoothing parameters chosen for use in the algorithm were 3.75 m for the vertical smoothing, 15.5 m for the horizontal smoothing, and a picking uncertainty of 1 ms. The velocity models are in ASCII text file format and each text file is named after the corresponding seismic reflection profile. In the text files, the first column is depth in meters, the second column is velocity in meters/second, the third column is UTM northing and the fourth is UTM easting (both in meters, UTM zone 18N). A total of 13 lines were processed using the first arrival tomography technique. It should be noted that data coverage or ray path density is much lower for the first few hundred meters of each line. Therefore, the beginning and end of each velocity model should be examined carefully and likely removed from the survey. The data files were generated as part of a project called P2C2: A High Resolution Paleoclimate Archive of Termination I in Oneida Lake and Glacial Lake Iroquois Sediments. Funding was provided through NSF grant EAR18-04460 to Syracuse University.