Keywords: nesting, coastal modelling, performance assessment, hydrodynamics.

The development and application of multi-scale modelling techniques is well-established in oceanographic modelling. These techniques are particularly important for modelling the coastal zone where the greater influence of the coastline and sea bed requires a higher resolution than the deep ocean. Nesting techniques are used to develop multi-scale models where structured grids are employed. Nesting results in the generation of errors at the nested boundary interface which can propagate into the nested domain leading to a deterioration in accuracy. A critical component of nested model development then is the assessment of model performance. This paper presents details of a unique protocol for the assessment of model performance (PAP).

Tidally-driven flows vary in both space and time. The accuracy of a tidal flow model will therefore also vary in both space and time and any approach to performance assessment should reflect this. To date, no assessment method has been developed to simultaneously capture the spatially- and temporally-varying nature of tidally-driven flows. The PAP enables simultaneous analysis of both the spatial and the temporal errors in a nested model solution. The protocol is based on the mathematical quantification of error in a nested solution relative to a reference solution using a novel error parameter: the tidally-averaged relative error field. This error for a particular grid cell quantifies the temporal error at that location. The spatial variation in error is simultaneously captured when the complete tidally-averaged relative error field is graphed as an error distribution plot for the area of interest. The tidally-averaged relative error field can be further processed to calculate a single figure quantification of errors in the nested solution - the domain-averaged relative error. This unique error parameter combines both the spatial and temporal variations in model performance.

The PAP was applied to an idealised harbour with tidal forcing and was found to be both efficient and effective. The protocol enables fast and accurate quantification of the model error and has a number of applications. It can be used for the development and testing of nested modelling techniques. The effectiveness of a nesting technique can be measured by the accuracy of the nested solution. Reduction of the domain-averaged relative error for a nested domain will indicate a more effective nesting technique. The PAP can also be used to aid the selection of suitable boundaries for a nested domain. Application of the PAP to the low resolution model in which the nesting is to be implemented allows easy identification of areas of high model accuracy. Nested boundaries should be located in these areas to ensure that boundary data errors are low. This application is particularly important for operational modelling where nesting is commonly used and both accuracy and computational time are important.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £130 +P&P)