Geoscience Reference
In-Depth Information
5.10 Key Points from Chapter 5
•
Fully three-dimensional models for surface and subsurface flow processes have become computation-
ally feasible but most current physically based distributed rainfall-runoff models still discretise the
catchment into lower dimension subsystems. Such a discretisation leads to only approximate repre-
sentations of the processes and may lead to numerical problems in some cases.
•
The effects of heterogeneities of soil properties, preferential flows in the soil and irregularities of
surface flows are not generally well represented in the current generation of models. Small scale
heterogeneity in the unsaturated zone suggests that new grid or element scale representations might
be needed.
•
The widely used SHE model is an example of a model based on grid elements, using one-dimensional
finite difference solutions for channel reaches and the unsaturated zone in each grid element, and two-
dimensional solutions in plan for the saturated zone and overland flow processes. It includes sediment
transport and water quality components.
•
The InHM model is an example of a fully 3D subsurface model, coupled to a 2D surface flow model
that has been tested in applications of detailed experimental data. It has also been extended to predict
sediment transport. HydroGeoSphere is another 3D modelling system that simulates flow, sediment
and water quality variables.
•
In some circumstances it may be possible to use simpler solutions based on the kinematic wave equation
for both surface and subsurface flows.
•
All distributed models require effective parameter values to be specified at the scale of the calculation
elements that may be different from values measured in the field. Distributed predictions mean that
distributed data can be used in model calibration but evaluation of this type of model may be difficult
due to differences in scale of predictions and measurements and the fact that the initial and boundary
conditions for the model cannot normally be specified sufficiently accurately.
•
A small number of studies have investigated the uncertainties associated with the predictions of dis-
tributed models, based only on prior estimates of parameter values or on conditioning of prior ranges
on the predictions of observed data. In both cases, the models do not do always provide acceptable
simulations of all variables, particularly internal state states of the system.
•
It has been argued that the use of distributed physics-based models is the best way of doing hydrological
science. There are still theoretical problems that need to be overcome in dealing with heterogeneity and
preferential flows in this type of model but the problem of parameter identification, particularly for the
subsurface, will be even greater and significant progress will undoubtedly depend on the development
of improved measurement techniques.
