Two types of models are used to examine the sea surface height (SSH) variability at the principle short time frequency bands in the Yellow and East China Seas. The models are a primitive equation numerical model and a statistical model based on TOPEX/POSEIDON altimeter data. In situ pressure gauge measurements at two mooring locations are first used to evaluate each model. In addition, a comparison of the numerical and statistical models is made through duplicate analysis of the response to the main extended empirical orthogonal function wind stress modes. In the northern Yellow Sea the wind stress, pressure gauge data, numerical model, and statistical model all indicate local spectral maxima in bands centered on the weekly and semiweekly periods. The coherence of the two models with the in situ data is significant down to 3-day periods at the northern mooring and down to 5-day periods at the southern mooring. The meridional wind stress variability in the semiweekly band is higher than the weekly band energy in the northern Yellow Sea and Bohai Bay and lower than the weekly band energy across the East China Sea. The SSH variability contained in the weekly and semiweekly bands is largest in the Bohai Bay and northern Yellow Sea with a region of high energy along the Chinese coast. The spatial structure of the weekly and semiweekly SSH variability is similar to the wind stress variability at these bands with semiweekly variability higher in the Bohai Bay and northern Yellow Sea and weekly variability dominant in the East China Sea. ¿ 1998 American Geophysical Union