Geoscience Reference
In-Depth Information
typically a loamy clay layer with very low conductivity and higher bulk density than
the overlying layers. However, in some cases during their evolution, fragipan horizons
became fractured into a polygonal columnar structure with a network of interconnected
vertical fissures, again, as a result of shrinkage of the clay particles; these cracks are
then believed to have been filled with more permeable soil material from above, greatly
facilitating water transport. The cracks are typically 10-20 cm wide. In another type of
preferential flow, the paths are initially established as instabilities or
fingers
at infiltrating
wetting fronts in coarse soils, when the infiltration rate is smaller than the saturated
conductivity. A crucial point, however, is that, once established, these paths usually
become permanent features of the profile, each time the soil isbeing rewetted (Glass
et al
., 1989); exceptions may occur when the soil has undergone complete drying out or
complete saturation, both of which are rare if not unlikely in nature. Figure 11.6 shows an
example of the initial growth of fingers observed in the laboratory. Such fingers are not
so obvious in the soil profile, but they become visible with dyes or other tracers. Other
aspects of the nature and originofthis type of preferential paths have been clarified (see
Selker
et al
., 1992; Liu
et al
., 1994a; b).
Although the existence of macropores has been known for a long time, the precise
nature of their contribution to the streamflow generation processes has been emerging
only gradually. A few examples follow of investigations inwhich macropores were
observed to play a major role.
In a small (0
022 km
2
) basin in east-central Honshu, Tanaka
et al
. (1981; 1988)
observed that more than 90% of the storm runoff came from below the ground surface
mainly through pipe flow; some saturation overland flow occurred over the gentler slopes
(
S
0
=
.
12) of the valley floor, when the rainfall exceeded 50 mm; the saturated area
varied somewhat in location and extent from storm to storm, but it never occupied more
than 4.5% of the total area (see Figures 11.7 and 11.8). No overland flow was ever
observed on the steep (
S
0
=
0
.
50) hillsides.
Ina0.47 ha forested catchment in Tennessee, Wilson
et al
. (1991) found that the
initial subsurface stormflow water in moderate to high intensity events consisted mainly
(
0
.
70%) of new, i.e. event water; they concluded from this that it had bypassed the
unsaturated soil matrix, inwhich the pre-event water was stored, via macropores without
ever reaching the water table. Later on, however, as the flow continued, the fraction of
old water gradually increased.
In a catchment under pasture in southern Australia, Smettem
et al
. (1991) and Leaney
et al
. (1993) observed that winter stormflow reaches the channel mainly through macrop-
ores, bypassing the soil matrix, and creating perched water table conditions immediately
around these pores. In summer, however, overland flow was found to be dominant; they
did not observe evidence of partial area sources, as only a negligible fraction of the
catchment was occupied by wetland.
On a steep forested hillslope with cedar and cypress in Ibarakiin east-central Honshu,
Tsuboyama
et al
. (1994) observed a dynamic system of macropores, which expanded
and conducted increasing amounts of water as antecedent conditions became wetter.
Continued studies on that same catchment (Noguchi
et al
. 1999; Sidle
et al
., 2001) led
>
