Geoscience Reference
In-Depth Information
infiltrated (Carmi and Berliner, 2008). Considerations of
this kind make it clear that, in real drylands, the effects
of seals cannot readily be isolated and that, instead, it is
more informative to seek a comprehensive understanding
of the multiple controls on infiltration, runoff production
and sediment transport processes. These multiple drivers
of infiltration, runoff and erosion include of course the in-
fluence of biota such as termites (e.g. Mando, Stroosnijder
and Brussaard, 1996). Where soils are cultivated, such as
some of the Spanish drylands (e.g. Ries and Hirt, 2008)
there are differences in seal properties, infiltrability and
erosion rates among fields of different ages and at local
scale between ridges and furrows within the fields. Thus,
the runoff and erosional response of any sizeable area
is the aggregate of many disparate local and microscale
effects of soil surface features. As Ries and Hirt (2008)
showed, enhanced runoff from sealed areas can result in
gully development further downslope, so that again we
can see that the consequences of seal formation cannot
be established solely from core or small plot tests on the
sealed surfaces themselves. Rather, their position and role
in the larger landscape must be considered. Finally, of
course, we can expect differing responses between small
and large rain events, and between single, isolated rain
events and multiple, closely spaced events. This intro-
duces great complexity in the temporal response to seal
formation, in parallel with the spatial complexity.
A formalised study of the effect of rain event arrival on
soil seals (Fohrer
et al.
, 1999) compared the effects of a
single event of 60 mm with the same total rain depth de-
livered in five separate events, each of 12 mm, separated
in time. In the 'multiple events' treatment, the infiltration
rate declined much more rapidly with time than during
the single event. Consequently, even relatively short, low
rain rate events on sealed soils could yield overland flow
if they followed a prior wetting event. The effects in a
particular case would depend on the extent of soil dry-
ing achieved between each event, with the possibility of
surface cracking, for example, modifying the processes in
later events. Carmi and Berliner (2008) reported similar
findings from a study of field plots containing sealed soils
at an arid site in Israel. Their results showed that the cor-
relation between runoff coefficients and rainfall rate was
higher for events closely spaced in time and weaker for
widely spaced events, reflecting soil moisture levels and,
in particular, suction gradients within the soil.
Studies of particular processes on sealed soil surfaces
can be used to explore the means by which sealing af-
fects soil dislodgment. For example, Slattery and Bryan
(1992) studied seal development and erosion in a labora-
tory flume using simulated rain. They found that splash
aggregates were broken down by raindrop impact, thus
providing a supply of splashable particles. As the surface
seal developed, however, and the disaggregated particles
were partly washed away and partly packed into the sur-
face seal, the rate of soil splash declined with time to a
stable, lower value. Similar findings have been reported
by others (e.g. Roth and Helming, 1992).
Seals are believed to reduce wind erosion, because of
the stable packing of particles into the seal, which reduces
the flux of abrasive grains passing over the soil surface.
This is a function of soil texture, however. On sandy soils
in Niger, it was shown that despite the formation of a
sieving (structural) seal, in which some fines are moving
downwards among the framework grains, leaving a friable
sandy surface, the eroded flux of fine sediment particles
was not supply-limited. Moreover, as in the case of runon
arriving from upslope, there can be sources of abrasive
sand particles upwind from areas that are sealed. In such
a situation, the saltating grains can progressively cause
a loss of structural integrity in soil seals, with a conse-
quent increase in the flux of eroded particles (e.g. Hupy,
2004). As with the case of raindrop impact on exposed
soils quantified earlier, the number of impacts of saltat-
ing grains with the soil surface is likely to be very large
indeed, on the order of 10
3
/cm
2
s in a typical cloud of
moving sand (Rice, Willetts and McEwan, 1996). Only
moderate wind conditions are needed for such progres-
sive impact damage to occur (Rice, McEwan and Mullins,
1999). Most studies of wind erosion on inorganic seals
are short-term experimental investigations made in tun-
nels with fan-driven airflow. In drylands, whether a seal
survives drought conditions and sand-blasting effects pre-
sumably depends upon the upwind sand supply and upon
the duration of dry and windy conditions.
Having considered something of the properties and sig-
nificance of inorganic soil seals, we now turn to the second
major class of dryland soil surface features, the biological
soil crusts.
7.5.5
Biological soil crusts
Where the dryland surface is neither eroding nor aggrad-
ing too quickly, the uppermost parts of the soil commonly
form the habitat for hardy varieties of nonvascular plant
and other organisms, including algae, lichens (the sym-
biosis of an alga and a fungus) and bryophytes (mosses,
liverworts and so on). These terricolous (soil-dwelling) or-
ganisms can form extensive coverings on the soil surface,
commonly referred to as 'biological soil crusts' (BSCs).
Communities of organisms can also colonise rock ma-
