Quantifying Uncertainty in Carbonate Reservoirs—Humma Marrat Reservoir, Partitioned Neutral Zone (PNZ), Saudi Arabia and Kuwait

Abstract

The Jurassic-age Humma Marrat carbonate reservoir was discovered in 1998. Eleven wells have been drilled to date including several horizontal completions. The gross reservoir interval is about 235 m (730 feet) thick. The reservoir produces from three intervals -Marrat A, Marrat C, and Marrat E. The partially dolomitized lowermost Marrat E interval contributes 75-85% of the total production from zones averaging 20-25% porosity and 10-100 mD permeability. Productive intervals in the Marrat A and C zones average 15-20% porosity and 0.5-2 mD permeability. The current estimated original oil in place is about 500 million bbls.A volumetric uncertainty look-back (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007) has allowed a historical assessment to be made for porosity and water saturation (Sw) uncertainty. This look-back based assessment of porosity and Sw uncertainty allows the impact of increasing quantity of data, changing analytical workflows, and updating interpretations to be examined. Based on the standard deviation or range of the available data at various times in the look-back period, the estimated porosity uncertainty was about ±2-2.5 porosity units (pu) after the first three wells were completed. After two additional wells and integration of core data, the estimated porosity uncertainty was reduced to about ±1.5-2 pu. Data from all eleven wells available at the end of the look-back period showed that the porosity uncertainty reduced slightly to about ±1-1.5 pu. However, a significantly lower estimate for the uncertainty is derived from the variation of the average porosity of the individual well averages after the first three wells were drilled. Using this parameter as a measure of uncertainty provides an uncertainty of ±0.5-1 pu early in field history (after three wells) and ±0.25-0.5 pu at the end of the look-back. Likewise, the uncertainty estimate for Sw decreased from ±15-20 saturation units (su) in 2000 to ±10 su in 2007 based on the standard deviation and range of the data available at each analysis date in the look-back period. Using the change in average Sw of the individual well averages, the uncertainty was ±5-10 su early in field history (after three wells) and ±2-4 su at the end of the look-back' significantly less than that derived from standard deviation or data range. A proposal is made to use the variation of the average of the individual well averages to define uncertainty in cases for which enough data (generally more than 3 wells) is available and there are no changes expected to analytical workflows (e.g. recalculating log derived porosity once core data is available). If changes to analytical workflows are anticipated a more conservative (e.g. larger) estimate for uncertainty should be used, perhaps based on the standard deviation or range of the available data.An original oil in place (OOIP) uncertainty look-back based on a consistent design of experiments-based approach (Meddaugh et al (2006a) is also presented along with a brief discussion of using a normalized uncertaint...

Reservoir Forecast Optimism – Impact of Geostatistics, Reservoir Modeling, Heterogeneity, and Uncertainty

“…The low slope for periods 7-9 suggests that the final wells of the delineation program added little new information (afterMeddaugh et al, 2009. Note that the increase in OOIP between analysis dates 2-4 is due largely to drilling "safe" wells that had significantly higher than average porosity.…”

Improving Reservoir Forecasts by Understanding the Relative Impacts of Sparse Data, Reservoir Modeling Workflow and Parameter Selection, and Human Bias

“…Reservoir data for a Middle East carbonate as given by , , and Meddaugh et al (2009) was used to generate Figure 12 which shows how the mid-case OOIP for the reservoir varied over time as additional data became available from new wells, as seismic data was reprocessed and reinterpreted, and as well log data was reinterpreted in light of the availability of core Figure 12 is unscaled, the OOIP variation between analysis date "1" (initial discovery well) and analysis date "6" (start of true appraisal as several of the early, post-discovery wells were drilled high on structure as "safe" producers) is quite large. Obviously, produced volume forecasts given data at analysis dates "1", "4", and "6" would be quite different, perhaps 10-20 RFUs or more.…”