The impact of human factors and measurement obliquity when extracting geological slip-rate from seismically imaged normal faults.

Seismic reflection datasets can help unpick the long term (i.e., 100 kyr to Ma) slip history of active normal faults (e.g., Nicol et al., 2005). To constrain slip-rate from seismically imaged normal faults you measure the across-fault offset of stratigraphic markers of known age.  Ideally, this is undertaken parallel to the slip-vector, i.e., orthogonal to fault-strike. In many active systems this is not possible with only non-optimally orientated 2D surveys available. Here we assess how obliquity effects continuous and discontinuous fault properties (throw, heave, dip, displacement) extracted from normal faults imaged in a 3D seismic cube. We targeted ‘straight’ faults and extracted cut off data from sequentially oblique sample lines ranging from ±50˚, comparing oblique data to that extracted from optimally orientated lines. Additionally, repeat picks were undertaken on two horizons to investigate the relative importance of measurement obliquity and human error.

Oblique measurements showed different along-fault profiles compared to orientated sample lines. This causes some datasets to be statistically different, with >100 % difference occasionally observed. Continuous deformation is more prone to obliquity errors, with the measurement of an apparent dip causing heave, and therefore displacement and dip, to display large differences at high obliquity. The dip of horizons close to the fault and localised complexity at the sample location (e.g., nearby faults) are also important factors. Differences regularly exceed 20% at high obliquity and we therefore suggest obliquity should not exceed 15˚ and were this is not possible measurements are corrected using fault cut offs and local fault strike.

For repeats picks, the shape of along-fault profiles is similar; however, subtle differences exist. Variability depends on the fault and fault parameter, with greater differences observed for faults with low displacement. Several locations along the fault exceeded the population difference. This was locally associated with a particular dataset; however, trends rarely persisted along the whole fault. Factors influencing this include a) shallow folding close to the fault, b) localised complexity at the sample location, and/or c) poor imaging near the fault plane. Unexpectedly, no correlation between variability and obliquity was found. Overall, we suggest errors due to human factors could be ~10-15% for throw and 20-25% for heave, with higher errors possible.

If we consider the fault data extracted using 2D seismic lines across the Cape Egmont Fault by Nicol et al., 2005, >50% of the measurement points exceed our recommended maximum obliquity. Nicol et al. (2005) report that the maximum throw between the 3.2 and 3.7 Ma horizons as 1364 m, giving a throw rate of 0.0028 mm/yr. However, the 2D survey at this location has a measurement obliquity of 21˚. Considering our findings throw rate could range from 0.0022 to 0.0033 mm/yr due to obliquity and human factors. It is therefore important users are aware that spatio-temporal variations in slip-rate may be caused by geological controls, human errors, and measurement obliquity.

Nicol et al., 2005. Growth of a normal fault by the accumulation of slip over millions of years. J. Struct. Geol. 27 (2), 327-342.