Petroleum Migration Rates and Distances

 In my training classes, I am often asked about the rate and distances of oil and gas migration. There seems to be much confusion in understanding how petroleum migrates, and the controlling factors. 

Migration Distance

Long distance migration are observed in many basins. Here is a list of basins I am familiar with:

• Athabasca field, Alberta basin, Canada > 700 km from the kitchen
• Orinoco oil field,  Venezuela, > 100 km  
• East Texas field, Texas, USA, > 100 km 
• Wolfcamp accumulations near outcrop in central Texas, Permian basin, > 200 km.
• Rubiales field, Llanos basin, Colombia > 100 km
• Mississippi oil trend in Kansas, Anadarko basin, > 500 km 
• Ghawar field, Saudi Arabia, > 200 km. 

All of these are from Foreland basins or continental sags which have marine depositional systems with good vertical seals, and the structure relief is too low to allow vertical migration. Another requirement for long distance migration seems to be the source rock - enough petroleum needs to be available to feed the migration chain.  

Below are two figures from the Anadarko basin in Oklahoma and Kansas.

Figure 1. oil and gas wells in the Anadarko basin. The accumulations in Kansas are 100s of kilometers north of the Woodford kitchen in southern Oklahoma. The oil production trend (a giant "river" of oil migration) continues further north into Nebraska.  

Figure 2. Map location of wells in the Anadarko basin, showing patterns of migration into Kansas and beyond. The orange outline is the mature kitchen of the Woodford shale, which is the main source rock for the oil in the basin. There is also production in Nebraska, Colorado that are not indicated here.

Large vertical distances are associated with very good source rock potential and the geological controls that limit lateral migration (such as high structure relief, faults, and salt), and may be then limited by the thickness of the sediment column. In the GoM basin discussed above, sea bottom seeps are very common, the source rock is 10 km below surface. 

Migration Rate

Below is a typical burial history diagram from deep water of Gulf of Mexico basin. It shows that the "oil window" is roughly 10 million years. The Tithonian source rock is excellent (~ 100 m thickness, TOC ~6%, HI ~600mg/g ) with an UEP of about 10 m³ of oil per m². Assuming a porosity of ~10%, this converts to a flux of 1 m³ per million years, and a flow rate of about 0.00001 m/year for primary migration. Imagine it taking a year to move from one clay sized pore to the next! The capillary number Ca, is about 10^-15, ten orders of magnitude too low for viscosity to have any effect, so no Darcy flow! Note that the GoM has one of the fastest burial rates in the world, and most basins will have rates 10 times slower.

Figure 2. Typical burial history from deep water of Gulf of Mexico basin. The main generation window occurs over about 10 million years.

Similarly we can infer rate of secondary migration or charging, again with an example of a large accumulation and fast generation rate.  A one billion barrel field filled over 10 million years equates to 100 barrels (16 m³) per year. Even if the charge is occurring over a 1 square meter area (which is unrealistically too small), the rate is still slower than typical glacier, and 4 orders of magnitude too slow for Darcy flow. 

A related question is whether oil generation is fast enough to create micro-fractures. If we take the 1 m³/my rate generated, and calculate how much volume is generated per unit rock volume, we come to 1/100/1000000 = 1e-8 m³/year per m³, or 0.01 cc/m³/year.  A drop is 005 cc, so one drop every 5 years. You see where I am going. 

Flow Mechanism

On a related subject, some authors have suggested that Darcy flow be used to explain/model the higher saturation in some source rocks. I think it is just a misunderstanding of Darcy behavior. If we examine the geology of these source rocks, such as the Eagle Ford shale, the higher saturation in these source rocks is a result of the source rock being sandwiched in between tight limestones which act as seals. The source rock needs to build up saturation (higher local capillary pressure) to exceed the seal capillary entry pressure for primary migration. In areas where the seals are not present, the saturation is very low. The Eagle Ford is underlain by the Woodbine sandstone in East Texas, where the Eagle ford is not a good unconventional target. The oil expelled downward into the Woodbine and migrated up dip to accumulate in the giant East Texas field. Vast majority of source rocks in the world, such as the well known CT & A source rocks in the Atlantic margins, the Kimmeridge clay formation in the North Sea, and the lacustrine source rocks in the Bohai basins are also excellent source rocks but have very low saturations shown by the fact that S1 Rock Eval extract correlates very well with TOC. It indicates that the residual oil is mostly adsorbed by the organic matter. 

Another argument against using Darcy model for unconventional settings, is that if Darcy (viscous, transient) flow is the mechanism, we would expect high saturation in source rocks in younger basins. On the contrary, most unconventional plays are in old basins, and some of which have had no deposition (and therefore HC generation) for over 100 million years.

Source rocks are not homogeneous, and there are always some inter-bedded capillary contrasts over the thickness of the source rock, so some saturation is often necessary for primary migration, and often higher in the middle due to having to overcome addition capillary barriers to reach the edge. Migration will be from high saturation toward low saturation, so both up and down from the middle of the source rock. It is not a viscosity (Darcy) effect as the capillary number is just orders of magnitudes too low. The capillary contrast within a source rock are much, much higher than buoyancy - so downward migration is pretty common. See this post on that.

Simple experiment/Analog for Migration

Here is something you can try or imagine. Take a bottle of water, walk outside and pore it on the street. It will flow down the street, may be for a few feet. Then it will stop. Right? Then take another bottle and pore it at the same spot, the water on the street will start moving again, a few feet further, and stops again.  The distance of migration is related to the amount of water you have (UEP of the source rock), and the rate is related to how fast you can pore it (rate of generation). If there is a big pot hole, it will not continue further down until it is filled. This is how I think of migration, the roughness of the street surface mimics a capillary system, with small pools and barriers between them, when additional water is added, it allows the small pools to connect and flow continues, and then when you stop adding water, the barriers between the pools will hold the water in place, and flow stops.