The work, supported by MAFF, aims to improve the usefulness and accuracy of forecasts. Accuracy is determined from routine comparisons between model predictions and surges derived from tide gauge measurements. Apparent forecast "errors" can thus arise from inconsistencies in the treatment of model results and tide gauge data. One inconsistency relating to the treatment of seasonal variations was investigated. Tidal analysis of observed water levels attributes this to the annual tide, Sa, whereas in reality it contains a substantial non-tidal component due to variation in storminess, correctly interpreted by the model as part of the surge. Analysis of results from a 12 month hindcast run of the 12 km operational model produced the distribution of the weather generated contribution to Sa shown in Figure 10a. The magnitude is substantial in some areas and when accounted for correctly (Figure 10b) reduces the discrepancy between model and observed surge by O(10 cm).
Purchase by the Met. Office of a new supercomputer, a massively parallel CRAY T3E, and adoption of a new operational environment for the Central Forecasting Office required revision of operational models and procedures for both mainframe and workstation based systems, which are well advanced.
Wave-tide-surge models
Model), an EU MAST 2 project, ended in February 1997 having successfully established the first such coupled model. Work on coupled wave-hydrodynamic models showed that interactions are significant in shallow coastal regions and areas with strong currents.
Experiments with the POL WAM+3D barotropic model showed some effect of waves on current profiles. Wave influence on surface stress was less important than anticipated whereas their influence on bottom stress was more important.
Figure 10
a) distribution of the weather generated contribution to Sa from the operational model, and b) time series at Aberdeen of observed, model and adjusted model surge.
Last updated: 12th May 1998. Please send comments to A.Lane@pol.ac.uk