Soil and Hillslope (Eco)Hydrology - Environmental Modelling: Finding Simplicity in Complexity - page 161

hydrophobic (e.g. MacDonald and Huffman, 2004), and
the post-fire establishment of plant propagules may be
affected by soil-water repellence and its effect on soil
wetness.
In this chapter, it is argued that an ecohydrologi-
cal approach, as defined above, is required to improve
understanding, and to develop useful models, of hills-
lope hydrological behaviour. A key theme is the idea that
hillslopes, generally, are likely to be complex adaptive
systems (CAS). In the context of ecosystems - and why
should hillslopes be considered as something other than
ecosystems? - Levin (1998) (as modified by Belyea and
Baird 2006) identifies the following defining properties
of CAS:

such as a run of drought years, although the response
of the hillslope system to events will also be condi-
tioned by the developmental history of the hillslope
(its ecological memory - see Section 10.4). Suscepti-
bility to external forcing-the inverse of resistance as
defined by Harrison (1979) - will depend on the veg-
etation present, its growth stage, its pattern, and how
it has interacted with the soil to produce subsurface
structures and patterns.

These attributes of CAS are considered in the following
three sections. Rather than deal with each attribute sep-
arately, they are discussed using a typology that reflects
current understanding of hillslope ecohydrology and a
desire to propose an agenda for hillslope ecohydrological
research that will underpin future model development.
Attention is given first (in Section 10.2) to work on dry-
land hillslopes, and to a lesser extent peatlands. Both
display striking vegetation patterns, both contain each
of the attributes noted above, and both have been the
subject of research by landscape ecologists and ecohy-
drologists in the last 10-15 years. As part of this research
effort, attempts have been made to model emergent veg-
etation patterns, and these models have necessarily been
ecohydrological in that they consider the two-way inter-
actions between plants and soil and also the movement
and concentration of resources such as nutrients within
the soil.
Section 10.3 considers the idea that patterns may be
present even when there is no obvious surface expression
of heterogeneity, and stresses the need for ecohydrologists
to search for such patterns when attempting to understand
hillslope hydrological behaviour. To help balance the ear-
lier emphasis on vegetation patterns, the focus in this
section is on the role of soil invertebrates - particularly
earthworms - in pattern formation. Section 10.4 consid-
ers the role of 'ecological memory' in hillslope ecohy-
drology and how an appreciation of ecological memory
may help improve understanding of hillslope hydrological
function and pattern persistence.
The concluding section is an attempt to 'stir the pot'
and to encourage hydrologists to look beyond their tra-
ditional disciplinary boundaries. We are all guilty of
disciplinary insularity or of paying lip-service to others'
ways of doing things; perhaps it is time for most hydrol-
ogists to recognize that hydrology has stagnated. What is
the appropriate response to such stagnation? Should we
become ecohydrologists?

(i) Sustained spatial heterogeneity or aggregation
In simple terms this means that a system is com-
posed of fundamental units that may aggregate to
form groups or patches which are (usually) main-
tained over time by selective autonomous processes,
unless external forcing (e.g. climate change) causes
a shift in ecosystem behaviour. Examples of selective
autonomous processes are given in Section 10.2.
(ii) Localized flows or transfers
The fundamental units interact by localized transfers
of resources and energy. These interactions lead to
larger scale (emergent) patterns that, in turn, deter-
mine the trajectory of the ecosystem's development
(see (iii)).
(iii) Self-organizing structure
The ecosystem has a hierarchical structure that
emerges from local interactions between compo-
nent (fundamental) units; the patterns that emerge
constrain further development of the system; thus
giving rise to cross-scale dependency.
(iv) Nonlinearity
Chance events, such as changes in the weather over
a period of years, can be amplified by positive feed-
back, leading to rapid change in system properties
and dependence on past events and development
(path dependency). It is known that hillslopes dis-
play nonlinearity in their short-term hydrological
response to, for example, rainstorms. However, in
the context of CAS, nonlinearity refers to apparent
threshold behaviour over longer timescales (years
and decades) whereby the characteristic hydrologi-
cal response of a hillslope, given a certain input of
rainfall, changes. This change may be due to events