Geology Reference
In-Depth Information
(Pierson, 1983). Anderson et al. (2008) used dyes (as have others) to
trace pipe networks exposed by excavation. Not surprisingly, the size
and connectivity of these shallow features was related to surface
catchments. The same is probably not true of deeper features.
Channelised fracture networks from the parent rock often persist
through the various stages of weathering. This type of preferential
ow
needs to be considered in investigation, hydrogeological modelling and
design. Such natural pipes probably follow original structural paths
(especially master joint or fault intersections), but may also be formed
by seepage pressure in weak saprolite or in super
cial soils such as
colluvium. They also develop at permeability contrasts (e.g. colluvium
overlying saprolite). More details are given in Hencher (2010).
Pipes are commonly seen associated withmany deep-seated failures in
weathered terrain (Hencher, 2006). It is implied that the development of
pipes at depth may be linked to early stages of progressive failure, as the
rock mass dilates and ground water exploits the dilating and deteriorat-
ing rockmass. Such deep pipes are probably distinct in origin frompipes
found in upper soil horizons. Fletcher (2004) reports that in
lled pipes
are sometimes encountered up to depths of 80m below present sea level,
and these must be associated with ancient lowstand levels.
3.6.4 Preferential
ow paths through rock
cult to investigate,
characterise or model in any real sense (Black, 2010). Interpretation of
even sophisticated test data is not straightforward, depending upon
whether one assumes that measured
Fracture
flow in rock is poorly understood and dif
flow volume into or away from a
test location is three-dimensional, planar (along a planar feature such
as might be assumed for a fault or major joint) or essentially linear
along a preferential channel. In simple terms, transmissivity depends
upon the aperture (degree of closure of rock walls) and roughness of
the discontinuity walls and lengths of fractures, together with their
intersections. The
cult to
identify. Dershowitz & LaPointe (1994) report how new oil wells
caused large drops in production to existing wells at distances of
several kilometres, within two days, whist other wells between them
were unaffected. Such behaviour could not be predicted without extre-
mely good knowledge of the fracture network and understanding of
potential connectivity, which is unlikely ever to be the case. To do so
with any hope of reasonable success would require a good under-
standing of the fracture origins and this is quite unlikely given the
current state of geological knowledge. Most attempts are simplistic,
extrapolating from super
flow paths are tortuous and extremely dif
cial, statistical and poorly constrained
observations in exposures or boreholes, to draw implications for the
rock mass at some distance. Thomas & La Pointe (1995) attempted to
discriminate between dry and
flowing fractures in the drift at Kiamichi
