Soil and Hillslope (Eco)Hydrology - Environmental Modelling: Finding Simplicity in Complexity - page 169

Richards illustrates the latter with reference to the
work of Shakesby et al . (2006 and 2007) on post-fire
soils. On these soils, nests from Aphaenogaster longiceps
Smith were found to promote infiltration through the
hydrophobic top layer of the soil (produced by fire),
thus reducing overland flow and erosion. However, in
a similar way to earthworms, the effects of ants on soil
hydrological processes will depend on a range of factors
including ant species, the density of nests, and the type
of soil. Indeed, the presence of ants from the genus
Aphaenogaster may promote both soil horizonation and
soil homogenization and may be a cause of enhanced
erosion and reductions in soil erosion. As with studies of
earthworms, much remains to be learnt on the role of
ants as hydrological agents.
It is worth concluding this section with one of the rare
studies that has looked at spatial variability in macropores
at the scale of the hillslope: Holden (2009). In keep-
ing with the caution expressed by Beven and Germann
(1982), Holden (2009) distinguishes between macropores
that have a hydrological function and those that do not,
where hydrological function can be defined in terms of
water flow. So, water may flow in some macropores (func-
tional macropores) and not in others. Holden (2009) also
coins the term 'functioning macroporosity' to describe
the proportion of soil water flow taking place within
macropores. Holden (2009) looked at variations in func-
tional macroposity on six humid-temperate slopes with
three different soil types and found clear spatial variability
on all the studied slopes. He did not examine the causes
of the variability but is able to conclude that his results
contrast with the conceptual nine-unit land surface model
of Conacher and Dalrymple (1977) (cit. Holden, 2009)
in which it is suggested that macropores will be most
abundant immediately downslope of the hillslope divide.
Holden (2009) does suggest that his observed patterns of
functional macroporosity may be due to soil faunal activ-
ity but provides no further information on what sorts of
activity might explain his results. Therefore, the study is
useful in revealing pattern but needs to be followed up
by an investigation of the processes that might have led
to the pattern. This is a theme that has emerged from
the last two sections: even when pattern is measured, the
causes of the pattern and the scales at which these causes
operate have yet to be established. Only when they are can
we start to model their effects on hillslope hydrological
behaviour. For example, it is perhaps too easy to think
that the appropriate scale for representing faunal effects
in models is that of the earthworm burrow or single
antnest( cf . Smettem, 1986). It might make more sense

to think in terms of aggregations of burrows and nests
and how these enhance hydrological connectivity (see
Section 10.1) both vertically through the soil profile and
in plan across the hillslope. In addition, these aggregations
and their effects need to be considered alongside other
biological controls on soil hydrological processes. Scale
and process are closely linked, and only through field and
laboratory investigation and numerical exploration with
a range of model types will it be possible to identify the
ecohydrological rules of hillslopes (see Section 10.1).

10.4 Memories

Time present and time past
Are both perhaps present in time future,
( Burnt Norton ,No.1of The Four Quartets ,T.S.Eliot,1936).

It was noted in Section 10.1, that one of the defin-
ing attributes of CAS is that the behaviour of a system
shows path dependency: system behaviour in 'time future'
depends on its status in 'Time present and time past'.
What exactly does this mean? Perhaps the easiest way to
explain it is with reference to the ecological literature and
the literature on CA.
'Ecological memory' is often defined rather loosely
as the effect of past processes or ecological structures
(e.g. patterns of vegetation) on the current behaviour
and structure of an ecological system ( cf .Hendryand
McGlade, 1995; Peterson, 2002). For example, a wood-
land exhibits ecological memory if the location of current
tree-fall gaps is influenced by the location of past gaps
(Peterson, 2002). Thus, in systems in which ecological
memory is important, relatively short-term changes in
structure have longer term effects on ecosystem pro-
cesses. Memory and how it relates to spatial scale can be
illustrated with reference to a two-D cellular automaton
(CA). If we consider what happens between neighbouring
time levels (over single time steps), each individual cell
in the automaton will undergo transitions between states
according to a rule set based on the states of cells in its
neighbourhood. We may start with an initially random
arrangement of cell states at time level zero ( t

=

0), from
which a pattern starts to form from time level 1 ( t = 1),
onwards. The pattern at t

=

1 can be thought of as the ini-
tial condition of the next time step ( t = 2) and will affect
the transitions that take place in each cell during that time
step. We may think of this type of interaction as giving rise
to a weak memory effect. This type of memory fits within