In order to assess the performance of current meters within and near the surf zone, data for biaxial electromagnetic current meters with spherical and open frame probe geometries were intercompared. One bottom-mounted flow meter of each type was deployed in a mean depth of 7.0 m for 17 days, and two sensors of each type were deployed in a mean depth of 2.0 m for 5 days. Sensors in the shallow deployment were frequently in the surf zone. Hourly averaged mean flows measured by different sensor types are highly correlated, averaging above 0.98. The largest difference between measured mean flows is constant bias, typically about 3.0 cm/s, which is roughly equal to the estimated accuracy of the sensor offset calibrations. Root-mean-square deviations from this constant bias are less than 2.0 cm/s, and are contributed to by errors in both gain calibration and sensor orientation. Comparisons of measured (surface gravity wave) oscillatory currents were made both between current meter types and with velocities inferred from the application of linear theory to pressure sensor data.

Correlations between time series of U&tgr;rms (the rms total oscillatory velocity for a 1-hour record) were all above 0.99 in 7.0-m depth and averaged 0.95 for the shallow deployment. The average U&tgr;rms ratio (over all hour-long records) was within 1.0¿0.07 for all current meter pairs in both deployments, which is consistent with the estimated 5% uncertainties in the flow meter gain calibration. Typical fluctuations of the U&tgr;rms ratio of any spherical and open frame sensor pair about its mean ratio, indicative of flow meter gain distortions probably associated with variations in the hydrodynamic environment, were less than 0.04 for any one deployment. Ratios of U&tgr;rms from both deployments taken together suggest that the open frame sensor overresponds, relative to the spherical probe, by about 5% at low (about 10.0 cm/s) total (mean-U&tgr;rms) speeds, and underresponds by about 5% at higher total speeds of about 75 cm/s. Relative to pressure data and linear theory, both flow meter types overrespond at low total speeds and underrespond at high total speeds.

We are, however, unable to determine whether these apparent gain deviations (of less than 10%) of the flow meters relative to pressure are associated with errors in the linear theory used to convert pressure to velocity, or with the response characteristics of the flow meters. Cross spectra between all sensors (including pressure) show high coherence and phase differences of a few degress, and they suggest that the response of both flow meters is only slightly frequency dependent. Various practical difficulties in accurately measuring the flow-induced microvolt potentials in electromagnetically noisy environments, with potentially interfering current meters, are discussed. ¿ American Geophysical Union 1988