You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.


Small-scale variability in sea surface heights and surface winds: Implications for errors in ocean models and observations

Alexey Kaplan(1) , Mark Cane(1) , and Dake Chen(1)

(1) Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY 10964, United States


Variability in nature exists on all spatial and temporal scales, including those smaller than the resolution of model and observational data sets. Imperfect parameterization of the small-scale variability (SSV) in models and incomplete sampling of it by observational systems creates model and observational error on the resolved scales of variability. The SSV in sea surface height was found to play a major role in defining the error pattern of wind-forced ocean simulation and satellite altimetry assimilation products. The statistical modeling of the SSV in sea surface height suggests that in the tropical Pacific the major portion of this variability can be explained as a dynamical ocean model response to the SSV (and error) in the wind. Areas of high error which are not associated with local wind SSV are those of high shear and current instabilities in the ocean. Most GCMs underestimate the wind-driven sea surface height SSV even if driven by wind forcing with well-represented SSV (possibly because of dissipation schemes that overdamp small scales) and, as a result, underestimate variability on signal scales as well. Not only magnitude, but also decorrelation scales of the wind error are crucial for determining the error in the ocean response. Data assimilation procedures usually interpret observed data as if they could be expressed in terms of the averages over model grid box areas despite in reality the observations are either pointwise values (for in situ data) or averages over certain footprints (for remote sensing data). Therefore the difference between observations and model values ought to reflect the influence of the small-scale variability (SSV) of the observed physical field, because this variability is getting averaged differently by the model grid and by the observational system. Knowledge of the statistical details of the SSV, e.g.its standard deviations and temporal-to-spatial SSV ratios, helps to model data error better.


Workshop poster


                 Last modified: 07.10.03