Environmental Engineering Reference
In-Depth Information
extraction rate, a low extraction rate from the gas reservoir, and a similar amount of
vertical ground displacement in absolute value. The mean value of annual gas
production for this period is 3.63*10
8
Sm
3
, corresponding to a down-hole value of
3.17*10
6
m
3
, for an average pressure of 11 MPa and a temperature of 40°C. This has
to be compared with the above referenced annual water injection rate of 13*10
6
m
3
in Venice.
The measured surface subsidence for the period 1972 to 1977 over part of the
area is shown in Figure 5.25, together with the location of the productive wells,
location of Ravenna and the contour of the reservoir. The observed behavior is far
from being smooth, and a gradient of 1*10
-4
of vertical ground displacement can be
found. The observed ground movements do not appear to be directly affected by the
specific location of the pumping wells, but rather to a local heterogeneity of the
strata involved, details of which are not captured with usual subsidence models. In
addition it must be recalled that the exploited reservoir is located at an average
depth of about 1,800 m below the surface so that some smoothing effect of the
overburden may be expected. In Venice, where the injection of seawater is foreseen
at a depth of 600-800 m, we should expect much less attenuation for differential
vertical displacements.
By inspection of the distribution of productive wells in the Ravenna area and
accounting for the observed non-uniformity of subsidence, it is possible to argue
that a limited number of injection wells (12 for the area of Venice) located on a
circle with radius 10 km is not sufficient to guarantee limited vertical differential
movements.
Careful and detailed measurements are also available for the period 1977-1998,
when extraction of both water and gas was stopped and pressure recovery has been
measured in both the aquifers and reservoirs (see [SCH 09] for more detail).
More uniform surface settlements are now observed (see Figure 5.26), even
though some smoothing is introduced by data elaboration. The real picture is far
from being a smooth surface, however, and average vertical displacement gradients
larger than 5*10
-5
over distances between 1 and 2 km are observed. This does not
exclude greater local values. The largest surface movements and gradients are
observed close to the reservoir boundary. This can be explained by capillary effects,
e.g. [DEL 96].
Some conclusions can be drawn for the application of this uplifting technique.
The total lack of experience and the sensitivity of the problem when uplifting is
applied to areas with historical settlements mean that numerical models are the
unique tool for a reasonable forecast of future effects. At the same time, a careful
examination of well-documented subsidence histories and comparison with the
