The long and short of faults

The first thing I have to do here is once again apologise for neglecting this blog. I have a number of excuses for this that I won't bother to go into.

Almost a month ago now, I wrote a short post on a paper that has been accepted for publication by the Journal of Structural Geology, and said I would write a bit more "probably next week". Better late than never.

The paper is entitled "Structural geology and 4D evolution of a half-graben: new digital outcrop modelling techniques applied to the Nukhul half-graben, Suez rift, Egypt" and the authors are myself, Dave Hodgetts, Frank Rarity, Rob Gawthorpe and Ian Sharp. It looks at a well-exposed extensional half-graben (the Wadi Nukhul half-graben) in the Suez rift, Egypt. We have collected a large amount of terrestrial LIDAR data from this half-graben, and used it to accurately map the structure and stratigraphy of the study area. By combining those data with our structural and sedimentological analysis, we can get an idea of how the fault system evolved over time, by looking at how sedimentation responds to the evolving structure.

There have been a number of studies looking at the evolution of normal fault systems, most of which have used seismic data or conventional field data and have looked at relatively simple fault systems. Seismic data, particularly 3D seismic, is great for looking at these questions because it has perfect 3D coverage. However, the maximum resolution tends to be on the order of 20 to 40 m. In the Wadi Nukhul area, the stratigraphic units we are interested in looking at are quite thin (up to 100 m or so), and more importantly the variations in thickness are quite subtle. Thickness variations in the strata that were laid down during active faulting tell you about the fault activity, so we need to be able to map out those thickness variations as accurately as we can.

Essentially, there are two hypotheses as to how normal fault systems develop. The first is that a large number of small faults initiate. Depending on their spatial relationship to neighbouring faults, these might die out, or they might grow in size (and displacement) and link to neighbouring faults. So there is a progression from a large number of low-displacement faults (in the "rift initiation" phase to a small number of large displacement faults, which have formed by the linkage of fault segments that were initially isolated from each other (in the "rift climax" phase). So faults increase their length gradually through time by growing laterally and linking to other faults. The second hypothesis is that faults establish their length very early in their history, and then accrue displacement through time without significantly increasing their length. This seems more likely to be applicable to reactivated faults. There is more on this in Morley et al. 2007.

Here's a simplified map of the study area:


For this summary, we're interested in the Nukhul fault. This is the main fault that controls deposition of syn-rift strata (Abu Zenima and Nukhul formations) in its hanging-wall. You can see that in plan view the fault is segmented: different portions of the fault have differing strike. Now, here's a graph showing how throw (vertical displacement) varies along the fault, taken from a horizon that we mapped on the LIDAR data :

There are a couple of things to notice here. Firstly, the maximum throw is approximately 1 km and occurs where the Nukhul fault intersects with the Baba-Markha fault to the south; throw then decreases towards the fault tip towards the north. Superimposed on that larger pattern, it can be seen that where the fault strike changes, there are minima in throw values, while in the centre of the fault segments there are maxima in throw values.

The interpretation of this is that the fault segments we can see at the present day, which have differing strike orientations, were initially isolated from each other. The variations in throw are then a kind of hangover from when the fault segments were isolated: the throw minima represent palaeo-fault tips at which displacement was zero before the fault segments became linked.

So what was the timing of linkage of the fault segments? It is difficult to know this unequivocally, but three lines of evidence suggest that the linkage was early, around the break between deposition of the Abu Zenima and Nukhul formations (which is locally marked by an angular unconformity). Firstly, the Abu Zenima Formation seems to have been deposited along the entire Nukhul fault. In the map above, it can be seen that the Abu Zenima is exposed at the linkage point between segments 2 and 3 of the fault. Secondly, the sedimentary facies in the areas around linkage points is much coarser than elsewhere in late Abu Zenima and early Nukhul time. That suggests that the linkage points were also sediment input points at those times. Thirdly, there is active faulting close to the linkage points prior to deposition of the Nukhul Formation. This suggests that those areas were subsiding rapidly at that time, perhaps because of the linkage of the fault segments. If this is correct, then the linkage of fault segments occurred within the first 2.5 million years of rifting (and perhaps much sooner). The Nukhul fault therefore seems to be an example of a normal fault that established its length fairly early in the rift history. This would make sense, because the orientations of some of the fault segments suggest that they were inherited from regional basement structures.

There is plenty more in the paper, but I'll post a link to a PDF when it becomes available.