The Importance of Internal Validation in the Assessment of Physically Based Distributed Models

As the use of physically based distributed models for solving environmental problems becomes more widespread it is important that we are satisfied with the manner in which they represent the processes operating within the modelled system. This paper seeks, with reference to three examples drawn from...

Full description

Saved in:
Bibliographic Details
Published in:Transactions - Institute of British Geographers (1965) 1995-01, Vol.20 (2), p.248-265
Main Authors: Fawcett, K. R., Anderson, M. G., Bates, P. D., J-P. Jordan, Bathhurst, J. C.
Format: Article
Language:English
Subjects:
Bgi / Prodig
Calibration
Geomorphology
Hydrological modeling
Hydrology
Methods and techniques
Modeling
Parametric models
Physical geography
Simulations
Soil water
Stormwater
Two dimensional modeling
Watersheds
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:As the use of physically based distributed models for solving environmental problems becomes more widespread it is important that we are satisfied with the manner in which they represent the processes operating within the modelled system. This paper seeks, with reference to three examples drawn from physical hydrology, to assess the utility of model internal validation to improve the analysis of distributed model behaviour. Obtaining a correct representation of physical process is particularly important in hydrology if the water movements simulated within a catchment will subsequently be used to drive simulations of solute and sediment transfers. We review current knowledge concerning catchment runoff processes to illustrate the range of physical mechanisms which physically based models may need to include. Examples of model applications of the SHE (Système Hydrologique Européen), the VSAS (Variable Source Area Simulator) and the RMA-2 models are then used to demonstrate the use of 'internal' validation approaches to identify the extent to which model simulations can correspond to physical reality. It is argued that validation using measurements of internal model output values should form an essential part of any model evaluation scheme.
ISSN:0020-2754
1475-5661
DOI:10.2307/622435