Sonobuoy-based velocity functions for sediment thickness calculation in the deep Canada Basin of the Arctic Ocean

The deep Canada Basin, which occupies much of the western (Amerasian) Arctic Ocean, is one of the most unexplored and difficult areas on Earth for marine seismic acquisition due to permanent sea ice cover. It extends northward from the Alaskan and Canadian margins for ~500 km and is characterized by a remarkably flat sea floor at 3.6-3.8 km below sea level (bsl) and covers ~500,000 km^2. A set of 143 sonobuoy records were collected during 2007-2010 over the Canada Basin by US-Canada collaborative expeditions. The sonobuoys were deployed along short streamer multi-channel seismic (MCS) lines for estimating seismic velocities in the sediments. A spatial coverage of sonobuoys at every ~80 km on MCS lines provides data for robust average empirical time-depth conversion functions.Sediments are nearly sub-horizontal and sub-parallel, mostly undisturbed with maximum two-way travel time (TWTT)thicknesses of ~5 s, but not greater than ~2.5-3.0 s TWTT closer to the Alpha Ridge and the Chukchi Borderland. Semblance velocity analysis has been applied to the sonobuoy records to determine sediment thickness and depth using the Dix equation. Assuming flat-lying sediments, the normal-moveout (NMO) velocity is approximately equal to the root-mean-square (RMS) average velocity from the sea surface to the reflection horizon. A dip of 5 degrees effects velocity less than 0.5%. Semblance velocity analyses were completed for 128 of 143 sonobuoy records in the deep Canada Basin; 25 records were excluded from later analysis due to bad quality or location on the slope. Results were also converted to interval velocities and depths.The RMS velocities were consistently picked from clear high-coherency events to yield only increasing interval velocities.Picking stopped at or above the bright reflection interpreted as basement on the MCS. Velocities were not picked on dipping reflections surfaces, along the margins where the seafloor was dipping, nor for sedimentary horizons characterized by rough morphology. Rather, horizons above or below these surfaces were picked. RMS velocity analysis yielded 3 patterns. 1) the lower RMS velocity group 1, consisting of 54 sonobuoys, occurs in the southern and central basin; Thickness = 0.17047*t^2 + 0.81173*t + 0.002564; R^2>0.995 (1) where thickness is in km; t is TWTT, below sea floor in s. 2) the faster RMS velocity group 3, consisting of 43 sonobuoys, is mainly located in the northern Canada Basin; Thickness = 0.26007*t^2 + 0.74357*t + 0.0078034; R^2>0.995 (2) 3) 10 sonobuoy records show RMS velocities between group 1 and 3 were grouped into a 'transition' group. Residuals of data for the derived functions have a normal distribution, 95% of data deviates from average for about 7 % (two standard deviations), the mean of normalized residuals are close to zero.Function (2) produces thicker sediments from ~25 m at t=1s to ~400 m at t=2.5s in relation to function (1). Velocities within most of the basin are laterally quite uniform for broad areas. Higher RMS velocities in the northern Canada Basin may be explained by thinning or pinching out some of the lower velocity units. It is possible that the deeper sediments in the northern area are affected by basalt intrusions or sills within sedimentary section; or some other variations in lithology of sediments.