More than 1200 high-resolution CTD temperature-salinity profiles, collected during 1989-1999 from the Hawaii Ocean Time-series deep-water station ALOHA (22°45N, 158°W), were analyzed to determine surface mixed layer properties. Two methods of defining the mixed layer were used. One determines a nearly-isothermal surface layer by the depth where potential temperature is 0.5°C less than the surface value (MLDT). The other defines a nearly constant density layer by the depth where the anomaly of potential density is 0.14 kg m-3 less than the surface (MLDr). This value was chosen to be equivalent to a 0.5°C change when salinity is constant.
Often, these estimates of mixed layer depth during individual cruises are nearly identical. However, during some cruises, the MLDr is more than 60 m shallower than MLDT. The times when large difference occur are during or shortly after winter storms, when MLDT is deep. This result is likely sensitive to the phasing of heat flux, freshwater flux and wind work on the upper ocean during frontal passages associated with winter storms.
The average MLDT=65 m, while the average MLDr=56 m, the difference being due to salinity stratification. This result suggests that nearly a 20% error in the subtropical North Pacific Ocean mean mixed layer depth will be made in models which do not include salinity. Such models cannot properly distribute momentum and heat in the upper ocean on annual and longer times scales, let alone from month to month. Even models which include salinity will require accurate, high frequency rainfall estimates to correctly compute surface mixed layer depths. The impact of related model errors is largest during the winter months, when strong air-sea exchanges of heat, freshwater and momentum occur, and when subduction may sequester anomalies until the next winter or for longer periods. Models of North Pacific decadal variability will have to accurately model the hydrological cycle, or properly parameterize its influence on ocean heat storage and circulation.