Soil and Hillslope (Eco)Hydrology - Environmental Modelling: Finding Simplicity in Complexity - page 159

spatial disposition of the K values. Binley et al . (1989a)
ran their model multiple times, each with a different
spatial arrangement of K values (their model had over
3,000 computational nodes). The different arrangements
of K comprised values of K drawn randomly in each
case from the same PDF. The randomly drawn K values
were then placed in the model hillslope, sometimes ran-
domly in space and sometimes in arrangements according
to certain levels of spatial autocorrelation. For different
arrangements with the same nonzero spatial correlation,
Binley et al . found wide variation in model output, which
they suggested could be explained by whether or not the
base of the model hillslope was occupied by low- K or
high- K soil. Overall, water flow from the hillslope was
sensitive to the value of K at the hillslope base. Therefore,
even with overall similar values and arrangements of K
on the hillslope, the exact disposition of areas of low
and high K can have a large effect on water flow from
its base.
Despite the work of Binley et al . (1989a, b) and
similar more recent work by, for example, Baird et al .
(2009), many hydrologists still ignore spatial pattern of
hydrophysical parameters such as K when considering
hillslope hydrological response to rainfall, and almost all,
as noted by McDonnell et al . (2007), ignore the causes
of such patterns (see also Bracken and Croke, 2007).
Hydrologists have identified 'hydrological connectivity'
as a key attribute that can help explain the hydrological
behaviour of hillslopes ( cf. Bracken and Croke, 2007),
but, to date, appear to have failed to identify some of the
important factors that might enhance or reduce connec-
tivity. As noted by Bracken and Croke (2007), the term
'hydrological connectivity' lacks a single, standardized,
definition. Nevertheless, included in the term is the idea
of the degree to which water can flow downslope without
entering longer term (days to weeks) storage.
How readily water flows down a hillslope will depend
on many factors, including the infiltrability ( sensu Hillel,
1998) and the surface and subsurface pattern of water
stores and connections between those stores. Bracken
and Croke (2007) note the importance of vegetation
as a control on connectivity and cite work done in
semi-arid or dryland environments, including, inter alia ,
Puigdef abregas (2005) and Boer and Puigdef abregas
(2005) (see also Section 10.2). Surprisingly, however,
some leading groups working on connectivity in non-
dryland environments apparently fail - at least some
of the time - to appreciate the role of vegetation and
ecological processes more generally; examples include
papers such as that of Hopp and McDonnell (2009)

in which some of the physical causes of hydrologi-
cal connectivity are investigated but where vegetation
as a factor in hillslope hydrological response to rain-
fall is not considered. Somewhat bafflingly, Hopp and
McDonnell (2009) cite Bracken and Croke (2007) and
note the role of ecologists in understanding hydrological
connectivity but then fail to discuss or investigate the
role of ecohydrological processes in hillslope hydrological
behaviour. Additionally, they do not look at how the
plan variability of K (i.e., variability down and across a
hillslope) affects hillslope hydrological response to rain-
fall. Using a series of numerical experiments similar to
those of Binley et al . (1989a, b), Hopp and McDonnell
(2009) consider the effect of storm size, slope angle,
soil thickness, and bedrock permeability on hillslope
hydrological response. While many of their results are
intuitive, some are not and reflect a complex interplay
between parameters. Hopp and McDonnell (2009) con-
clude that such a complex interplay could be explained
using hydrological connectivity as a conceptualization
tool. They suggest that hillslopes can be divided into
dynamic areas of 'fill' and 'spill' and that connectiv-
ity between these is the main control on overall hillslope
hydrological behaviour. Although their findings are inter-
esting and show that the identification of subsurface
hydrological networks within the soil matrix may prove
fruitful in helping explain thresholds in the response
of some hillslopes to rainfall ( cf . Tetzlaff et al ., 2008),
it is worth quoting the authors (Hopp and McDon-
nell, p. 380):

We acknowledge that our representation of the hillslope
is a gross simplification of a system that has evolved
over long time spans as a result of interacting climatic,
geomorphological and biological forces. Nevertheless, we
are of the opinion that valuable insight into hillslope-
scale subsurface stormflow generation can be gained with
this simplified description of the hillslope.

It is hard to reconcile the sentiment of the second
sentence with the admission in the first. Additionally,
the findings of the paper seem, in part, to constitute an
appeal to uniqueness ('all hillslopes are different') when
it is becoming increasingly clear that many hillslopes are
not necessarily unique because they show similarities in
their vegetation patterns (see Sections 10.2 and 10.5) (for
a discussion of the apparent uniqueness problem, see
Beven 2000).
Other papers of which McDonnell and Hopp are co-
authors, such as Hopp et al . (2009) and McDonnell