22
Statistical Methods for Estimating Petroleum Resources
within the box indicates the median of the sample, whereas the short
vertical bar above the box covers the range occupied by three quarters of the data. The small squares and crosses outside the box indicate
relatively large values. The largest one or two values in each sample are
classifi ed as outliers, the magnitudes of which are relative to the values of the sample. The difference between the largest and second largest pools of the Cardium sandstone play shown in Figure 2.10 is much
greater than that of the other pools. Details for constructing a box plot
are discussed in Lee et al. (1999).
Correlation between Random Variables
Correlation between geological random variables (such as pool area,
net pay, recovery factor, reservoir pressure, and others) is also a common feature of geological populations. For example, the pool area
and net pay variables of the Zama reef play of the Western Canada
Sedimentary Basin exhibit a negative log–log association (Fig. 2.11A).
In other words, as the log pool area value increases, the log net pay value
decreases. In contrast, the pool area and net pay of the Beaverhill Lake
play (Fig. 2.11B) show a positive log–log association: As the pool area
value increases, the log net pay value increases. Correlation between
Figure 2.11. (A, B) Log–log associations for Western Canada Sedimentary Basin
plays showing a negative correlation between pool area and average net pay
variables for the Zama reef play (A), and a positive correlation between pool area
and average net pay for the Beaverhill Lake play (B).
Statistical Methods for Estimating Petroleum Resources
within the box indicates the median of the sample, whereas the short
vertical bar above the box covers the range occupied by three quarters of the data. The small squares and crosses outside the box indicate
relatively large values. The largest one or two values in each sample are
classifi ed as outliers, the magnitudes of which are relative to the values of the sample. The difference between the largest and second largest pools of the Cardium sandstone play shown in Figure 2.10 is much
greater than that of the other pools. Details for constructing a box plot
are discussed in Lee et al. (1999).
Correlation between Random Variables
Correlation between geological random variables (such as pool area,
net pay, recovery factor, reservoir pressure, and others) is also a common feature of geological populations. For example, the pool area
and net pay variables of the Zama reef play of the Western Canada
Sedimentary Basin exhibit a negative log–log association (Fig. 2.11A).
In other words, as the log pool area value increases, the log net pay value
decreases. In contrast, the pool area and net pay of the Beaverhill Lake
play (Fig. 2.11B) show a positive log–log association: As the pool area
value increases, the log net pay value increases. Correlation between
Figure 2.11. (A, B) Log–log associations for Western Canada Sedimentary Basin
plays showing a negative correlation between pool area and average net pay
variables for the Zama reef play (A), and a positive correlation between pool area
and average net pay for the Beaverhill Lake play (B).