Worldwide shale-oil reserves : towards a global approach based on the principles of Petroleum System and the Petroleum System Yield ★

9 Abstract – Global inventory of shale-oil resources and reserves are far from being complete even in mature basins which have been intensively drilled and produced and in which the main parameters of the regional or local oil-prone source rocks are known. But even in these cases, difficulties still occur for deriving reserves from resources: reaching a plausible recovery factor is actually a complex task because of the lack of production history in many shale-oil ventures. This exercise is in progress in several institutions (EIA, USGS, AAPG) or private oil and gas companies on a basin-by-basin basis in order to estimate the global potential. This analytical method is very useful and accurate but also very time consuming. In the last EIA report in 2013 “only” 95 basins had been surveyed whereas for example, no Middle-East or Caspian basins have been taken into account. In order to accelerate the process and to reach an order of magnitude of worldwide shale-oil reserves, we propose hereafter a method based on the Petroleum System principle as defined by Demaison and Huizinga (Demaison G and Huizinga B. 1991. Genetic classification of Petroleum Systems. AAPG Bulletin 75 (10): 1626–1643) and more precisely on the Petroleum System Yield (PSY) defined as the ratio (at a source-rock drainage area scale) between the accumulated hydrocarbons in conventional traps (HCA) and hydrocarbons generated by the mature parts of the source-rock (HCG). By knowing the initial oil reserves worldwide we can first derive the global HCA and then the HCG. Using a proxy for amount of the migrated oil from the source-rocks to the trap, one can obtain the retained accumulations within the shales and then their reserves by using assumptions about a possible average recovery factor for shale-oil. As a definition of shale-oil or more precisely LTO (light tight oil), we will follow Jarvie (Jarvie D. 2012. Shale resource systems for oil & gas: part 2 –Shale Oil Resources Systems. In: Breyer J, ed. Shale Reservoirs. AAPG, Memoir 97, pp. 89–119) stating that “shale-oil is oil stored in organic rich intervals (the source rock itself) or migrated into juxtaposed organic lean intervals”. According to several institutes or companies, the worldwide initial recoverable oil reserves should reach around 3000Gbo, taking into account the already produced oil (1000Gbo) and the “Yet to Find” oil (500Gbo). Following a review of more than 50 basins within different geodynamical contexts, the world average PSY value is around 5% except for very special Extra Heavy Oils (EHO) belts like the Orinoco or Alberta foreland basins where PSY can reach 50% (!) because large part of the migrated oils have been trapped and preserved and not destroyed by oxidation as it is so often the case. This 50% PSY figure is here considered as a good proxy for the global amount of expelled and migrated oil as compared to the HCG. Confirmation of such figures can also be achieved when studying the ratio of S1 (in-place hydrocarbon) versus S2 (potential hydrocarbons to be produced) of some source rocks in Rock-Eval laboratory measurements. Using 3000Gbo as worldwide oil reserves and assuming a quite optimistic average recovery factor of 40%, the corresponding HCA is close to 7500Gbo and HCG (=HCA/PSY) close to 150 000Gbo. Assuming a 50% expulsion (migration) factor, we obtain that 75 000Gbo is trapped in source-rocks worldwide which corresponds to the shale-oil resources. To derive the (recoverable) reserves from these resources, one needs to estimate an average recovery factor (RF). Main parameters for determining recovery factors are

reasonable values of porosity and saturation which is difficult to obtain in these extremely fine-grained, tight unconventional reservoirs associated with sampling and laboratories technical workflows which vary significantly.However, new logging technologies (NMR) as well as SEM images reveal that the main effective porosity in oil-shales is created, thanks to maturity increase, within the organic matter itself.Accordingly, porosity is increasing with Total Organic Carbon (TOC) and paradoxically with... burial!Moreover, porosity has never been water bearing, is mainly oil-wet and therefore oil saturation is very high, measured and calculated between 75 and 90%.Indirect validation of such high figures can be obtained when looking at the first vertical producing wells in the Bakken LTO before hydraulic fracturing started which show a very low water-cut (between 1 and 4%) up to a cumulative oil production of 300 Kbo.One can therefore assume that the highest RF values of around 10% should be used, as proposed by several researchers.Accordingly, the worldwide un-risked shale-oil reserves should be around 7500 Gbo.However, a high risk factor should be applied to some subsurface pitfalls (basins with mainly dispersed type III kerogen source-rocks or source rocks located in the gas window) and to many surface hurdles caused by human activities (farming, housing, transportation lines, etc...) which can hamper developments of shale-oil production.Assuming that only shale-oil basins in (semi) desert conditions (i.e., mainly parts of Middle East, Kazakstan, West Siberia, North Africa, West China, West Argentina, West USA and Canada, Mexico and Australia) will be developed, a probability factor of 20% can be used.Accordingly, the global shale-oil reserves could reach 1500 Gbo which is half the initial conventional reserves and could therefore double the present conventional oil remaining reserves.

Introduction
A conference on source rocks cannot ignore the potential of oil production from source rocks, so called shale-oil, which since a few years surprisingly makes up a significant part of US liquid oil production, increasing between 2010 and 2014 from zero to 3 MMbpd.This is not an insignificant part of world oil production which has reached 94 MMbpd on average in 2014 and should reach 97 MMbpd in 2015 (Fig. 1).
In this article, we try to estimate the world-wide oil reserves from source rocks using the notion of generation potential of source rocks which is relatively well understood, as well as their retention capacity (or non-expulsion) which is less understood.Various empirical and geochemical approaches from the basin to the laboratory scale will be used to determine the retention potential of source rocks.We will also try to estimate a recovery factor for these retained fluids using analogues of petrographical, geomechanical and fluid characteristics from US shales.A recovery factor is difficult to know because little production data are available beyond the short production history of these unconventional reservoirs.Nevertheless, their production demonstrates a surprising production resilience due to the large number of wells and the possibility of refracking but also due to the better than thought petrophysical properties.
Unconventional oil (or LTO = light tight oil) is defined here in accordance with Jarvie (2012) as oil that is generated by rock formations rich in organic matter that are oilmature and trapped either within the source rock layers or in its immediate vicinity (not more than a few meters of migration within such an extremely low permeability rock matrix).These oils did not experience secondary migration and stay in these compact reservoirs and therefore are not limited by a water-oil contact.

Oil in source rocks: how is it retained
A first estimate of the quantity of oil trapped in source rocks has been done by P.K. Meyer in a short article in the Oil and Gas Journal in May 2012.He makes an intriguing and almost shocking statement that within a source rock, retained oil makes up about 80% of generated oil.When comparing the recovery factor of oil from source rocks and conventional reservoirs, P.K. Meyer estimates that the amount of recoverable oil (or oil reserves) is eight times greater in source rocks than in conventional reservoirs!This is very different from what the conventional wisdom of petroleum geology has taught us since the 1980s on both sides of the Atlantic (see Magoon, 1988;Demaison and Huizinga, 1991;Perrodon, 1995).The seminal article by W. England in the famous AAPG Memoir 60 (1994) did not even mention HC retention, if it is not synonymous with "loss to primary migration" which he considered about 15% of the generated oil (Fig. 2).In his model, the majority of oil (60%) is indeed lost during migration whereas only a small amount (8%) is trapped in commercial (conventional) accumulations, either structurally or stratigraphically.This percentage also corresponds to the sum of all conventionally trapped hydrocarbons in producing reservoirs (HCA).Yet another way to calculate this percentage is to estimate it from the petroleum potential of basins (HCG), combining amount of organic matter originally contained in source rocks and their state of maturity.
For a given source rock and a given drainage area, HCG can be calculated according to Demaison and Huizinga (1991) using the following formula: HCG = h Â (S1 þ S2) Â Density Â TR Â Drainage area Parameters S2 (producible hydrocarbons or hydrogen index) and S1 (free in situ hydrocarbons or hydrocarbon index) are expressed in mg HC/g TOC.S2 corresponds to total hydrocarbons that could be generated from a source rock if all its potential is realized (i.e., its transformation ratio TR = 100%) while S1 corresponds to hydrocarbons present in the source rock that had already been generated but are retained in the source rock.S1 is oftentimes underestimated because a large part is lost before its actual measurement, as already and perfectly stated by Price and LeFever (1992).Nevertheless, S1 can be a used as a proxy for a retention coefficient.
Once HCG is determined for all source rocks in a basin, it can be compared with HCA mainly in basins that are already highly explored, with important historical production where accumulations (oil in place) are generally well-known.Such results have been published by Biteau et al. (2010) (Fig. 3) for more than 60 basins in various structural and geodynamical settings.The ratio of HCA/HCG or petroleum systems yield (PSY) of these basins is between 1 and 10%, with an average of less than 5% (Fig. 5).Only the Alberta and Orinoco basins show an exceptionally high PSY of 50%!Both these basins contain large amounts of bituminous oils (or Extra Heavy oils) at shallow depths that are preserved at the fringes of these basins after have migrated long distances and having been trapped by topographic driven hydrodynamics.Therefore, these two basins are rare cases where the amount of migrated oil is preserved while normally a large amount of migrated oil is lost, both for explorers and scientists alike.If this is the case, we can estimate the amount of migrated oil at 45-50% of HCG.This also means that the amount of retained oil must be the remainder, around 50%.With this number we are much closer to the number published by Meyer (2012) than the numbers conventionally used (15%).
Another way to determine the amount of retention in source rocks is to measure in the laboratory the amount of HCs present in source rocks.Several studies by Espitalié et al. (1985) (Fig. 4) or Jarvie (2012) indicate very high numbers of S1 (free hydrocarbons) close to 250 mg HC/g TOC and ratios of S1/S2 near 35% which is not far away from our hypothesis of 50%.
A complete geochemical analysis of the Lower and Upper Bakken shales (Zhang et al., 2013) shows that no significant petroleum expulsion is observed in the center of the Upper and Lower Bakken.This shows that retention is indeed very important in shales even if it is certainly less near the Middle   Bakken whose lithology is less shaly and therefore acts more like a drain towards the under-and overlying oil-rich shales.

Shale-oil reserves: how much recovery
As indicated in the introduction, the recovery factor of shale-oil is not well-known because it is difficult to measure the oil and water saturations of these highly compact rocks with very short production histories.We need first to determine its porosity if we want to estimate oil saturation and then try to quantify initial water content and its evolution during production.
While the compaction law indicates that porosity of the mineral matrix decreases during burial, the organic porosity which is located within the organic material is not a function of burial.Porosity measurements in core material and porosity visualizations based on MEB and SEM photographs show that this intra-kerogen porosity develops at the beginning of maturation and increases with maturity and therefore increases with burial.Therefore, thermal maturation of kerogen not only generates hydrocarbons but also creates porosity.Figure 5 from Cander (2012) shows the evolution with burial of the porosity of the Eagleford shale in Texas, an LTO of Upper Cretaceous age.Organic porosity reaches 5% around 5000 m depth where it is greater than the porosity of the mineral matrix.This porosity is a function of the amount of organic matter whose volume may reach 5-20% of the bulk rock volume for a TOC ranging from 2-8%, given a mineral matrix density of 2.6 g/cm 3 and an organic matter density of 1.1-1.2g/ cm 3 .The oil is mainly present within this intra-kerogen organic porosity where oil saturation must be very high because this porosity is created at a later stage, during oil generation and never underwent any saturation by connate water like every classical sedimentary reservoir rock.Possible water connectivity may only be created once hydraulic fracturing occurred.In addition, the organic material is oil-wet (Rylander et al., 2013;Passey et al., 2010).Laboratory data confirm the hypothesis that oil saturations of LTOs are high to very high and stay high during production, much higher than during production from conventional reservoirs.These latter are already aquifers before oil migrates and accumulates and therefore these reservoirs always contain a significant amount of residual water which increases during production because water is replacing the produced oil either coming from the active aquifer itself or from water injection.
The few oil saturation data published in the literature indicate values of 75-90% for the Bakken shale (Philips et al., 2007) whereas Zhang et al. give a mean residual oil saturation of 40%.Using an S1 of 10-15 mg HC/g rock for the Lower and Upper Bakken, an average porosity of 8%, an average TOC of 10% (which yields a bulk rock density of 2.6 g/cm 3 according to Passey et al., 2010) and an in situ HC density of 0.6 g/cm 3 , we can calculate oil saturations between 50-75%.This means that 1 g of rock corresponds to 0.38 cm 3 or to 0.03 cm 3 of pore volume or 0.018 g of HC/g rock.For S1 = 10 mg HC/g rock, oil saturation reaches about 55% while for S1 = 15 mg/g, oil saturation reaches 80%.
Indeed, the water cut of production of LTO is very low, as shown for example in the first vertical wells drilled without hydraulic fracturing in the 1980s into the Bakken source rock of Devonian to Carboniferous age that have a ridiculously low water production of 1-4% even after 300 000 barrels of cumulative production from a single well.The water cut of horizontal wells (after frac water removal) starts at about 20% in 2008 and increases to about 30% in 2014 (see Fig. 6).The most recently drilled wells show an initial water cut of 35-40%, there is therefore a global increase of water cut during production of the field.This is probably due to the frequent fraccing operations which create ever increasing contact areas between fractures and the source rock matrix.But again, the increase in water cut during production remains moderate, increasing from 40 to 45% in two years.Saturation remains oil dominated and it is this low water fraction which is in our opinion a major reason for the resilience of production despite a drastic reduction in the number of on stream wells in 2014 due to the decrease in oil price (see Fig. 7).
The recovery factors used by the EIA in its report of 2013 which range between 2 and 8% for American shale oil should be questioned.They are given as a function of natural and induced fraccability and formation pressure but seem to totally neglect the oil saturation of the organic material.An average recovery factor of 10% seems more appropriate, given analytical laboratory results and well production data.This corresponds also to the higher value proposed by Charlez (2015).
4 From shale-oil resources to reserves: how and where?
If we accept a retention coefficient of 50% of generated HCs and estimate a recovery factor of 10% in unconventional LTO accumulations, we can calculate resources and world reserves of shale-oil.The in-place volumes of conventional oil and gas fields are fairly well-known except for the yet to find volumes: the initial world reserves are about 3000 Gbo of which 1000 have been produced and 500 are still to be discovered.Assuming a final recovery factor of 40%, the total in-place volume in accumulations is 3000/0.4= 7500 Gbo.If these accumulations (HCA) correspond to 10% of migrated HCs which themselves correspond to 50% of generated HC, then the total generated HC (HCG) is 7500/0.05= 150 000 Gbo. Since the retained oil corresponds to 50% of the generated oil, this means that oil generated and retained in source rocks (resources) equals to 75 000 Gbo (see Fig. 8).
Of course, not all source rocks of all the sedimentary basins in the world can be exploited as in the US shales (Vially, 2015) because they may have different geological  parameters (the nature of the source rock, particularly difficult to produce if the organic matter is dispersed as is often the case in Type III kerogen source rocks; the position of the oil window in source rocks of Type I and II kerogens; the mineralogy therefore the geomechanical properties of the source-rock that can be little or not frackable) or may be located in economically or geographically unfavorable areas for production such as highly populated areas.
Assuming that only the mature basins from a petroleum exploration point of view that are located in deserts or semideserts or in very sparsely populated areas will be exploited (such as Western Siberia, Kazakhstan, North Africa, Middle East, Western China, Andean foreland of Argentina, Australia, Western US or Canada, Appalachians of Eastern US), and assuming that only 20% of the world's source rocks (a conservative estimate) apply to these criteria, the risked resources in this case are 75 000 Â 0.2 = 15 000 Gbo. Using a recovery factor of 10% yields about 1500 Gb of shale-oil reserves.This number is about half of the total initial liquid reserves of conventional oil fields or the equivalent of today's oil reserves.This result appears to be very different from the number published by the EIA in its report on shale-oil from 2013 (335 Gb).However, the EIA number only covers 95 terrestrial basins outside the Middle East and the Caspian Sea areas which contain about half of the world's initial conventional reserves, and has been determined with a recovery factor of only 5%.By doubling the volume of source rock and using a recovery factor of 10% we arrive at a similar number (335 Â 4 = 1340 Gb).
Sedimentary basins with source rocks that are capable of retaining HCs are those that contain source rocks of kerogen Type II (possibly also Type I) that are not affected by faults or salt tectonics which offer vertical pathways for migrating HCs.Such basins correspond to intracontinental basins or foreland basins.Such basins have mainly horizontal migration and drainage patterns and sometimes exhibit even topographic driven hydrodynamic flow (Perrodon, 1995) exhibited on a diagram of initial recoverable reserves versus basin surface area (see Fig. 9).It is therefore not surprising that important unconventional shale-oil and shale gas discoveries of the last few years such as the Bakken, Utica, and Marcellus shales have been precisely made in this type of basins.
Further readings

Fig. 8 .
Fig. 8.A global approach to retention in Source Rocks.