Risk, Uncertainty and Priorities - Quantitative Models
439
atility. Then the fixed value is systematically increased, and the volatility again
calculated for a suite of runs. The average of the volatility values then provides a
measure of the contribution of uncertainty of one parameter irrespective of the
value of the other parameter. The procedure is then reversed so that the volatility, Vj' of each parameter on its own is calculated. The relative importance of the
jth parameter is then defined by
N
RI(i)=vj/2vi'
H
where N is the number of parameters being varied (2 in the present case).
(2)
We use the measures Pso, v, and RI in the next sub-section to illustrate excess
pressure development across the section of Fig. 8.
4.1.2
Specific Model Results
The present-day section of Fig. 8 can be interpreted (Lowrie 1994) as providing
the spatial domain of an approximately wedge-shaped salt mass currently intersecting locations A, Band C, and which recently intersected location Dj together with a roughly block-shaped slab of salt currently intersecting locations
E, F and G, and which recently intersected location H. As each separated salt
mass moved laterally, the sedimentary infill behind the salt masses, at locations D and H respectively, is apparent from Fig. 8. The exact thickness of the
salt at any location is not too well determined from Fig. 8 due to both uncertainty in the seismic velocity used to convert a two-way travel-time section
into a physical depth section, plus the difficulty of imaging the base of salt on
the seismic section.
For instance, Lowrie (1994) has estimated that the salt thickness at location A
could be as large as 1500 m, is likely to be around 900 m, and could be as thin as
600 m. Equally, the lateral speed of motion is not precisely inferred, ranging
from a low estimate of about 1 cm a-I, with a likely value of about 10 cm a - 1 , but
with a high value of possibly 20 cm a-I.
At each of the locations B-H, the corresponding salt thicknesses today are taken to be equally as uncertain fractionally as at location A, so that each of the salt
masses can range from extremes of slow moving, relatively thin, bodies to rapidly moving, thick bodies across the section. Originally we did consider omitting calculations for the salt mass intersecting wells E-F at the present day, because we thought it unlikely that the mass has moved southward with the shape
it currently enjoysj it would seem more likely that the two "tongues" originating
in the south of the salt mass are what is currently moving. However, we have included the calculation as though the salt mass kept its present -day shape because
we were unable to prove unequivocally that the tongues were indeed the only
moving components of the salt. In any event the comparison between the two
salt masses provides an interesting end-member comparison of the range of
439
atility. Then the fixed value is systematically increased, and the volatility again
calculated for a suite of runs. The average of the volatility values then provides a
measure of the contribution of uncertainty of one parameter irrespective of the
value of the other parameter. The procedure is then reversed so that the volatility, Vj' of each parameter on its own is calculated. The relative importance of the
jth parameter is then defined by
N
RI(i)=vj/2vi'
H
where N is the number of parameters being varied (2 in the present case).
(2)
We use the measures Pso, v, and RI in the next sub-section to illustrate excess
pressure development across the section of Fig. 8.
4.1.2
Specific Model Results
The present-day section of Fig. 8 can be interpreted (Lowrie 1994) as providing
the spatial domain of an approximately wedge-shaped salt mass currently intersecting locations A, Band C, and which recently intersected location Dj together with a roughly block-shaped slab of salt currently intersecting locations
E, F and G, and which recently intersected location H. As each separated salt
mass moved laterally, the sedimentary infill behind the salt masses, at locations D and H respectively, is apparent from Fig. 8. The exact thickness of the
salt at any location is not too well determined from Fig. 8 due to both uncertainty in the seismic velocity used to convert a two-way travel-time section
into a physical depth section, plus the difficulty of imaging the base of salt on
the seismic section.
For instance, Lowrie (1994) has estimated that the salt thickness at location A
could be as large as 1500 m, is likely to be around 900 m, and could be as thin as
600 m. Equally, the lateral speed of motion is not precisely inferred, ranging
from a low estimate of about 1 cm a-I, with a likely value of about 10 cm a - 1 , but
with a high value of possibly 20 cm a-I.
At each of the locations B-H, the corresponding salt thicknesses today are taken to be equally as uncertain fractionally as at location A, so that each of the salt
masses can range from extremes of slow moving, relatively thin, bodies to rapidly moving, thick bodies across the section. Originally we did consider omitting calculations for the salt mass intersecting wells E-F at the present day, because we thought it unlikely that the mass has moved southward with the shape
it currently enjoysj it would seem more likely that the two "tongues" originating
in the south of the salt mass are what is currently moving. However, we have included the calculation as though the salt mass kept its present -day shape because
we were unable to prove unequivocally that the tongues were indeed the only
moving components of the salt. In any event the comparison between the two
salt masses provides an interesting end-member comparison of the range of