Modelling Catchment and Fluvial Processes and their Interactions - Environmental Modelling: Finding Simplicity in Complexity - page 189

greater constraint. For simplicity, natural timesteps are
always preferable to artificial ones (day and night sepa-
rated at sunrise and sunset instead being aggregated into a
whole day; rainfall rate based on timesteps that reflect the
real passage of rainfall intensities instead of hourly lumps
of rainfall - see Wainwright and Parsons (2002) on how
getting temporal variability in rainfall right also translates
into the correct representation of spatial variability.

of partial contributing area concepts into hydrologi-
cal models, our understanding of subsurface quickflow
mechanisms through macropore pipe networks and
shallow subsurface quickflow is much less advanced.
This lack of progress is partly the result of the difficulty
in measuring the variation in the physical properties
that control these processes at the catchment scale and
partly the result of the relatively recent recognition of
subsurface quickflow as hydrologically important.
• Hillslope-channel coupling . Modelling catchments is
about modelling the hydrology of hillslopes and of
channels. Many of the processes that determine chan-
nelization are still poorly understood and the coupling
of hillslopes to channels is an area in which new insights
are being made but further research is required to help
improve catchment models of streamflow, of the storm
hydrograph and of flooding. In particular, although
progress is being made in producing improved models
of (dis-)connected surface flow across the transition
from hillslopes to floodplains to channels, there is less
progress on the effects of subsurface flows, which can
be demonstrated to be significant at point and reach
scales (Ibrahim et al ., 2010). Although detailed models
have been applied to the floodplain-channel interface,
reproducing interesting behaviours of field conditions
(e.g. Cloke et al ., 2006), they tend to be carried out
without considering the hillslope context, at least in
part due to the computational overheads.
• Non-rainfall precipitation . There is still a relative dearth
of modelling efforts focused upon the hydrology of non-
rainfall inputs such as snow (Bloeschl, 1999) and occult
precipitation (Bruijnzeel et al ., 2010) at the catchment
scale. These inputs are significant in many catchments
but are more difficult to measure and to model than
rainfall.
• Tropical lowlands and tropical mountains . The hydro-
logical impacts of land use and cover change in the
tropics is much discussed in the literature but there
are very few studies that apply modelling to better
understand these systems (Mulligan et al ., 2010b) and
fewer still that combine modelling with intensive field-
monitoring programmes (Chappell et al ., 1998). As a
result there is still much confusion about the impli-
cations of land use change in these environments and
myths abound (Bruijnzeel, 1989; Calder, 1999; Mulligan
et al ., 2010b).
• Hydrological Connectivity . A key characteristic of
catchment-scale studies is the lateral connectivity
between patches that results from the existence surface

11.3.3 Simplifyingprocesscomplexity
There are many areas of hydrology where our under-
standing of processes is basic but still sufficient to develop
models but there are still areas in which the complexity
of hydrological processes is so great, or the information
so little, that we do not understand sufficient of the
processes to develop reliable models. This complexity of
processes is separate from the issues related to spatial and
temporal variation and the lack of data available to repre-
sent them as outlined above. Some of the areas in which
there is still much progress to be made are outlined below.

• The hydrology of sparse vegetation . Though techniques
for modelling the interception, drainage and evapotran-
spiration from forest canopies are now well established
there are still difficulties in understanding the role of
canopy properties in determining the partitioning of
water between the various fluxes. These difficulties are
particularly clear for nonclosed canopies or sparse vege-
tation where the impact of surface roughness is less well
known and the parameterization of evapotranspiration
is much more difficult. Furthermore the separation of
vegetation into patches may change evapotranspiration
loads in complex ways (see Veen et al ., 1996). Patch-
iness is not just important for evapotranspiration but
also affects the generation and propagation of runoff,
and sediment especially in arid and semi-arid environ-
ments (see Dunkerly and Brown, 1995, Wainwright and
Bracken, 2011, and the discussion in Chapter 10).

Subsurface quickflow mechanisms . The parameteriza-
tion of saturated hydraulic conductivity ( K sat ) at scales
greater than a few hundred cm 3 remains a major obsta-
cle to progress in subsurface hydrological modelling,
particularly given the importance of macropores in pro-
viding a mechanism for subsurface quickflow in many
environments (Elsenbeer and Vertessy, 2000; Uchida
et al ., 2001). Though our understanding of the mech-
anisms of runoff generation through Hortonian and
saturation-excess mechanisms has improved consider-
ably in recent years with the extensive incorporation

•