Optimizing dynamic downscaling in one-way nesting using a regional ocean model
•The Big-Brother Experiment may be used for downscaling ocean circulation.•Careful choices of grid, updating time, and domain can minimize errors.•The optimal grid size ratio was found for downscaling from OGCM to ORCM.•Frequently updating the lateral boundary conditions increases errors.•Buffer are...
Gespeichert in:
Veröffentlicht in: | Ocean modelling (Oxford) 2016-10, Vol.106, p.104-120 |
---|---|
Hauptverfasser: | , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | •The Big-Brother Experiment may be used for downscaling ocean circulation.•Careful choices of grid, updating time, and domain can minimize errors.•The optimal grid size ratio was found for downscaling from OGCM to ORCM.•Frequently updating the lateral boundary conditions increases errors.•Buffer areas inhibit boundary errors from propagating into the domain of interest.
Dynamical downscaling with nested regional oceanographic models has been demonstrated to be an effective approach for both operationally forecasted sea weather on regional scales and projections of future climate change and its impact on the ocean. However, when nesting procedures are carried out in dynamic downscaling from a larger-scale model or set of observations to a smaller scale, errors are unavoidable due to the differences in grid sizes and updating intervals. The present work assesses the impact of errors produced by nesting procedures on the downscaled results from Ocean Regional Circulation Models (ORCMs). Errors are identified and evaluated based on their sources and characteristics by employing the Big-Brother Experiment (BBE). The BBE uses the same model to produce both nesting and nested simulations; so it addresses those error sources separately (i.e., without combining the contributions of errors from different sources). Here, we focus on discussing errors resulting from the spatial grids’ differences, the updating times and the domain sizes. After the BBE was separately run for diverse cases, a Taylor diagram was used to analyze the results and recommend an optimal combination of grid size, updating period and domain sizes. Finally, suggested setups for the downscaling were evaluated by examining the spatial correlations of variables and the relative magnitudes of variances between the nested model and the original data. |
---|---|
ISSN: | 1463-5003 1463-5011 |
DOI: | 10.1016/j.ocemod.2016.09.009 |