Most Jurassic reservoirs in the Middle East contain bulk bitumen. Core data from the Middle Jurassic Uwainat reservoir in the Dukhan field of Qatar show that bulk bitumen predominantly occurs in grainstones, and to a lesser degree, in bioclastic packstone and wackestone intervals. Bitumen fills interparticle and intraparticle porosities, hairline fractures, and stylolites. The thickness of the bitumen varies, generally in response to the distribution of the porous-permeable lithologies. For this study a learning dataset was constructed, from which numerical estimations were made to isolate petrophysical characteristics of bitumen. Then qualitative estimates were made to predict the occurrence of bitumen. Data from cores and thin sections were integrated with porosity and resistivity logs of twenty cored wells to develop a technique for determining the presence of bulk bitumen in the uncored wells. As this methodology is largely qualitative, a binary flag system was employed to indicate bitumen. The technique involves cross-plotting of bulk density versus thermal neutron porosity, across zones where bulk bitumen has been identified in cores. A transform of the form y=abxxc relates the neutron porosity to the bulk density. This transform is then applied to the neutron data to arrive at a calculated bulk density, theoretically predicting a bulk density response to bitumen. When the difference between the calculated and the actual bulk density is within ± 0.03 gm/cc, a positive flag is generated (suggesting that bulk bitumen is present). Using cut-offs based on calculated bulk density values and the microresistivity log further refines this flag. This method is useful in detecting bulk bitumen in uncored wells, when bulk bitumen forms in intervals more than 2 feet thick. The zones identified as having bulk bitumen in cores were correctly identified by the log data. This technique has about 80 percent reliability when the rock interval containing bulk bitumen is more than 10 feet thick. The approach however, has limitations where bitumen occurs in low porosity rocks, as it becomes difficult to differentiate resistivity increases due to reduction in porosity from those due to bitumen filling the pore spaces. The agreement between core and log requires an accurate depth matching between core and logs.


In the Dukhan field of Qatar the Middle Jurassic Uwainat reservoir consists of lime grainstones, packstones and wackestones (Figures 1 and 2). These lithofacies are arranged in a series of geological layers that have consistent thickness and extent. The Uwainat carbonates were deposited in an open marine platform environment as overall regressive sequences. Hydrocarbon accumulations in the Uwainat reservoir occur in a relatively small, oil-saturated pool in the northern sector, and a small pool of non-associated gas in the southern sector of the Dukhan field (Figure 11). The Uwainat reservoir rock contains significant amounts of bitumen that forms localized flow barriers and impacts hydrocarbon recovery.

Bitumen has also been recognized and identified in cores from the Jurassic Arab-D reservoir of the Dukhan field. A study performed by Robertson Research International Limited (1990) concluded that in the Arab-D reservoir of Dukhan, bitumen concentration within the major bitumen mat vary from one percent to greater than 15 percent bulk volume (5 to 95 percent pore volume). Bitumen is more abundant in the northern portion of the Dukhan field. In the southern part of the field, bitumen is more disseminated and less abundant, both in bulk and pore volume. Where bitumen is present it completely reduces the reservoir quality (porosity/permeability) and acts as significant flow barrier in many crestal wells. This can constitute a problem in any future enhanced oil recovery, such as water flooding.

Bitumen precipitates from migrating or trapped hydrocarbons in both carbonate and sandstone reservoirs (Lomando, 1992). Bulk bitumen can form through several processes including thermal alteration, deasphalting and bacterial degradation (Tissot and Welte, 1978; Hunt, 1979). Numerous terms are commonly used in the oil industry for bulk reservoir bitumen. They include solid hydrocarbon, pyrobitumen, dead oil, tar mats, solid bitumen and black and asphaltic sands (Lomando, 1992). In this paper we will use the bulk bitumen term. We present an approach that integrates core data with open hole density, neutron, and resistivity logs of cored wells, in order to identify bulk bitumen in uncored wells. We believe that open hole logging can be effective in determining the presence of bitumen.

The occurence of bitumen in the Uwainat reservoir

Conventional cores from twenty wells representing the Uwainat reservoir were visually examined and described for the occurrence of bitumen. Core data show that bitumen predominantly occurs in the grainstone and to a lesser degree in the packstones and wackestones. Bulk bitumen completely fills interparticle and intraparticle pores, hairline fractures and stylolites (Figures 3 to 5). Whereas, patchy bitumen only partially fills or lines pore spaces. Measured porosity over the bitumen intervals ranges from 0 to 15 percent and the permeability range is from 0 to 10 mD.

The thickness of the bitumen varies, generally in response to the distribution and thickness of the originally porous-permeable lithologies, in particular grainstones. The bulk bitumen can reach up to 25 feet in thickness. The bulk bitumen represents the maturation products of tarry and asphaltic residues formed by the alteration of an oil accumulation (Robertson Research Report, 1990).

Core-log calibration

We have made an attempt to identify bitumen-bearing intervals of uncored wells using a combination of open hole density, neutron and resistivity log data across the Uwainat reservoir in the Dukhan field. Data from cores and thin sections were integrated with porosity log data to develop an initial neutron-porosity to bulk density transform or learning curve (Figure 6a). This transform, along with resistivity data, was then used to develop a series of flags that predicted the presence of bitumen in the uncored wells. Bitumen data of the 20 cored and 21 uncored wells were then used to map the occurrence and distribution of bitumen in the Uwainat reservoirs. The main principles used to identify bitumen from log data are discussed below.

Porosity Logs

Both the density and neutron tool responses are affected, to some degree, by the presence of bitumen. The density tool (which is a pad device), and the neutron tool (which is run eccentralized) are additionally affected by the borehole wall conditions. This borehole effect is compensated, to some extent, by the presence of two detectors: (1) the near detector response compensates the measured density for borehole effects; and (2) the ratio of the two detectors compensates the neutron for borehole effects.

The presence of bitumen in the formation has the effect of replacing connate water with a material, which may have a higher density than that of water. In a previous in-house Qatar Petroleum study of bitumen in the Uwainat reservoir, the density of bitumen was seen to be as high as 1.5 to 1.7 gm/cc. Where the density of bitumen is greater than the density of water, the effect of bitumen in a formation increases the measured bulk density and therefore lowers the computed porosity. Such effects are predominant across zones where bulk bitumen is present. The neutron tool responds to the amount of hydrogen in the formation and hence is designed to measure the ‘Hydrogen Index’ rather than simply the formation porosity. In the presence of bitumen, there is a marked reduction in the amount of hydrogen detected, which leads to a reduction in the neutron porosity reading. The overall effect of the two tool responses in the calculation of porosity is one of reduction in effective porosity. A decrease in the actual pore size and/or thin bed effects can mask the detection of bitumen using this technique.

Resistivity Logs

The microresistivity log has a vertical resolution in the order of 2 to 3 inches. Because it is a pad device, its readings are those of the near borehole region. The deep resistivity has a much coarser vertical resolution of about 2 feet.

Bulk bitumen fills all the pore space. It replaces all fluids and prevents their movement through the pores. Nor is bitumen displaced by drilling fluid. As a result, both the deep-reading resistivity and shallow-reading microresistivity tools read high values. If the concentration of bitumen occurs in a thin bed, then the deep resistivity may not respond adequately to it. Hence, it was decided to rely solely on the microresistivity log to attempt identification of bitumen.


Zones were selected where bulk bitumen was detected from the cores. In these zones a cross-plot of bulk density versus thermal neutron porosity was generated (Figure 6a). Cores identified as having patchy bitumen in cores were excluded from this plot. This regression was of the form y =abxxc where a, b, and c represent the coefficients of the equation; x represents the neutron porosity and y the bulk density. This plot (Figure 6a) shows good correlation with a high regression coefficient (greater than 0.9). A predicted bulk density response to bitumen was estimated using the above transform from neutron data. The results yielded an absolute value for the majority of the data comparison (predicted versus actual RHOB) at less than 0.03 gm/cc (Figure 6b). Therefore, when the difference between the calculated bulk density and the actual bulk density was within ± 0.03 gm/cc, it was flagged with a value of 2, indicating the presence of bitumen. In addition, a visual inspection of the density log data across zones identified as containing bulk bitumen in core, reveals that the density values mostly range between 2.4 and 2.58 gm/cc (Figure 7). If the bulk density values were outside of the above range, the previously calculated flag was assigned a value of 0, indicating absence of bitumen in the rock.

After detailed examination of the log data across the zones identified as containing bulk bitumen from cores, a cut-off of 20 ohm-m on the microresistivity log was used. So for microresistivity values larger than 20 ohm-m, the previously calculated bitumen flag retained its value, or else it was set to 0. Thus this flag would have a value of 2 where bulk bitumen was predicted from log data, and values of 0 elsewhere. Similarly, a flag was created for the core data having values of 2 across bitumen zones and a value of 0 elsewhere. The absolute value of the difference of the two flags (core flag–log flag) was calculated producing agreement/non-agreement flags (effectively reverse flag values) and plotted as a histogram (Figure 8a). Where the log-derived predictions agree with the core data, this histogram has a value of 0, otherwise it has a value of 2.

The approach correctly predicted the presence of solid bitumen in 50 percent of the cases. In order to test the sensitivity of the approach to the presence of relatively thin beds of bitumen, a difference histogram was constructed of only those core-identified-bitumen-zones that are thicker than 2 ft (Figure 8b), > 4 ft (Figure 8c), > 6 ft (Figure 8d), > 8 ft (Figure 8e) and > 10 ft (Figure 8f). The accuracy of the prediction of bitumen progressively improves from 55 percent to 77 percent. Figure 9 shows that the prediction of bitumen from logs becomes more reliable with increased bitumen thickness.

The match obtained in cored wells, which were part of the database, is generally satisfactory. A plot of the comparison between core and log derived results for wells, which had significant bitumen detected in core is shown in Figure 10. To further verify the reliability of the predictions, the above approach was tested on a cored ‘blind’ well that was not part of the database (well A, Figures 10 and The zones identified as having bulk bitumen in cores were correctly identified by the log data, adding confidence to the reliability of the technique. Much of the agreement between core and log methods rely on accurate depth matching between core and log.

The bitumen thickness of two reservoir layers was combined and mapped. Figure 11a shows the distribution of bitumen based on cored wells only, whereas Figure 11b shows the distribution of bitumen from cores and logs. The data from the logs has improved the map in the north of the Dukhan field, where there is a lack of cored wells.

Limitations of the Technique for Detecting Bitumen in uncored wells

The approach has limitations where bitumen occurs in low porosity rocks. It becomes difficult to differentiate the resistivity increase due to reduction in porosity, from the resistivity decrease related to bitumen filling the pore space. Also the method sometimes indicates bitumen near the boundaries of vertically alternating tight and porous reservoir units. This is most likely an artifact of differences in vertical resolutions of the input logs. What complicates matters further is that the core in some of the wells shows bulk bitumen in the tight portions of the reservoir (assuming accurate depth matching of core to log). Both the bulk density and microresistivity are pad devices; therefore hole conditions have a big influence on the reliability of the final answer.


The impact of reservoir bitumen on the Uwainat reservoir ranges from reducing effective porosity and permeability, and changing wettability, to influencing reservoir quality predictions. The core-log calibration method presented in this study has proved useful in predicting bulk bitumen in uncored wells (with bitumen intervals exceeding 2 feet in thickness). This technique has about 80 percent reliability when the rock intervals containing bulk bitumen reach over 10 feet thickness. The approach has limitations where bitumen occurs in low porosity rock types.

The economic implications of the Uwainat bitumen mapping will result in improved determination of the recovery factor, calculations of the net pay, volumetrics, original oil-in-place, and estimated ultimate recovery. Detailed characterization of thickness and lateral distribution of bitumen are vital for reservoir simulation and history matching. This information will also be considered in optimizing future well location of infill producers and water injectors, and the selection of the type of well (deviated, horizontal or multilateral). The results from the study will be incorporated to produce a geologically realistic reservoir model as input to the reservoir engineering simulation program.


We thank the management of Qatar Petroleum for allowing us to publish this paper. We also thank Mr. S. Abohadoud for laying out the examined cores and Mr. A. Al Nour for the preparation of thin section. The design and drafting of the final graphics was by Gulf Petrolink.


Varavur Venkitardr Shankar is a Senior Petrophysicist with EPSA/DPSA (Exploration and Development Production Sharing Agreement) section in Qatar Petroleum (QP). Before joining QP in 1997, he was employed with Schlumberger Geoquest as a Geoscientist in Abu Dhabi. He has a BSc in Mathematics and a HND in Computer Science from the Scottish Vocational Education Council.


Ali Miloud Said Trabelsi is a Senior Reservoir Geologist and a Team Leader, Qatar Petroleum (QP). Before joining QP in 1997, Ali served as an Exploration Geologist, Senior Reservoir Geologist and Sedimentologist with AA Production Oil & Gas, Inc. and 3-D Exploration, Inc. in Lubbock, Texas, and Esso (Exxon) Libya. Ali consulted for many oil and gas companies in Houston and Dallas, Texas. He received his MSc and PhD degrees in Petroleum Geology in 1990 from Texas Tech University in Lubbock. His professional interests include quantitative sedimentology, sequence stratigraphy, reservoir characterization and 3-D reservoir modeling.


Mirza Arshad Beg is a Reservoir Geologist with Qatar Petroleum (QP). Before joining QP in 1999, Mirza was employed as a Senior Reservoir and Production Geologist with Sarawak Shell Berhad, and as a Sedimentologist with Petronas Research and Scientific Services, Malaysia. Mirza received his PhD in Applied Sedimentology in 1990 from Strathclyde University, UK. His professional interests include quantitative sedimentology, sequence stratigraphy, reservoir characterization and 3-D reservoir modeling.