Modelling Catchment and Fluvial Processes and their Interactions - Environmental Modelling: Finding Simplicity in Complexity - page 185

over many other simple approaches by taking into account
the propensity for differences in catchment saturation in
accordance with the compound topographic index and
its effect on the generation of saturation overland flow.
These differences are represented statistically rather than
physically. In TOPMODEL, rainfall occurs and water
enters the soil along with water draining from other parts
of the catchment causing the groundwater table to rise.
When groundwater rises to the land surface, the area
becomes saturated and saturation overland flow occurs.
Water below the surface is assumed also to move downs-
lope as throughflow since the hydraulic gradient of the
saturated zone is approximated as the local topographic
slope. Total outflow is throughflow plus saturation and
infiltration-excess overland flow. More recent versions
have added further sophistication to processes such as
subsurface-flow routing (Beven and Freer, 2001) and a
spatially variable soil thickness (Saulnier et al ., 1997).
The Stanford Watershed model (SWM) is one of the
earliest and best known conceptual catchment models.
The model has an hourly soil-moisture budget and stor-
age and routing functions for the redistribution of water
entering the channels to provide catchment scale runoff
on a daily timestep (Crawford and Linsley, 1966; Viess-
man and Lewis, 1996). The model requires climate data
and some 34 parameters describing the physical prop-
erties of the catchment. More recent models have given
greater importance to the overarching control of the geo-
morphological properties of the catchment as exercised
through the influence of the Horton-Strahler stream-
order configuration. Furthermore it is envisaged that this
type of model can be applied in ungauged catchments
for the calculation of geomorphologically controlled unit
hydrographs (Gupta et al ., 1996; Schmidt et al ., 2000;
Yang et al ., 2001), particularly for large catchments. The
fractal-scaling properties of river basins help significantly
to simplify runoff modelling. Szilagyi and Parlange (1999)
describe a semi-distributed conceptual model in which
the catchment is conceptualized as a series of stores whose
dimensions are derived from the Horton-Strahler stream
order (see Figure 11.6). Overland flow fills these stores
and water is routed between them. A separate ground-
water model provides baseflow. The model requires the
calibration of only seven parameters on the basis of a year
long rainfall-runoff record.

and physiological ones (e.g. Hatton et al ., 1992) and with
erosion and sediment-transport models (De Roo, 1998;
Ewen et al ., 2000). Modelling scales have increased from
regional through to continental and global (Vorosmarty
et al ., 2000; Gosling and Arnell, 2010) in part to provide
interfaces with general circulation models of the atmo-
sphere for climate change (Stieglitz et al ., 1997; Sperna
Weiland et al ., 2010) and climate-change-impact (Bron-
stert et al ., 2002; Mulligan et al ., 2011) studies. These
larger scales and this deeper integration have been in
response to increasing pressures to make the models
address some of the most serious environmental prob-
lems that face governments and citizens, accelerated land
use and climate change. Since computer power has risen
exponentially in the last decades, the models (and the
modellers) have been able to keep up with these demands.
However, the gap between capability and parameterizabil-
ity (the C:P gap) continues to increase.
Recent years have also seen better integration of data
resources within flexible and intuitive databases, particu-
larly in GIS. More catchment models are designed to work
with data held in common GIS formats or indeed to run
within GIS software (particularly ESRI ArcMap, PCRas-
ter, GRASS and SAGAGIS). This trend has both facilitated
the spatialization of models and the potential for mod-
elling over larger catchments and has also prevented the
C:P gap from being even wider. The hydrological func-
tionality of GIS and the links between common spatial
hydrological models and GIS is reviewed by Ogden et al .
(2001). The advances in data availability and manip-
ulation within a GIS context that have facilitated this
integration of hydrology and GIS are reviewed in the
companion paper by Garbrecht et al . (2001). De Roo
(1998) outlines the state-of-the-art in dynamic spatial
modelling languages for the development and applica-
tion of hydrological models but reminds us that all of this
spatial detail does not necessarily mean better results.
In line with the greater emphasis on having the hydro-
logical models contribute to the solution of environmen-
tal problems and in line with the greater accessibility of
personal computers, catchment hydrological models have
also moved into the public domain. Decision-support
tools, which incorporate hydrological models for the
purposes of integrated catchment management (Walker
and Johnson, 1996; Martens and Di Biase, 1996), spatial
land-use planning (Greiner, 1996), or management of
ecosystem services (Mulligan and Burke, 2005; Mulligan
et al ., 2010a) are now common (see also Chapter 20).
These tools are often used for better understanding the
unforeseen consequences of human intervention in the

11.2.7 Recentdevelopments

Recent trends in catchment modelling have seen the inte-
gration of catchment hydrology models with ecological