Surface latent and sensible heat fluxes over the Pacific Sub-Arctic Ocean from saildrone observations and three global reanalysis products

Sea surface latent and sensible heat fluxes are crucial components of the air-sea energy exchanges that influence the upper-ocean heat content and the marine atmospheric boundary layer. Due to the limited availability of in situ observations, assessing their impact on Arctic weather and climate has mainly been done using data assimilation products and numerical model simulations. The accuracy of the surface fluxes in numerical models are, however, largely unvalidated. Recent deployments of saildrones, remotely piloted uncrewed surface vehicles, can help bridge this data gap of in situ observations. This study represents an initial effort to validate sea surface latent and sensible heat fluxes over the Pacific sub-Arctic open ocean from three commonly used global reanalysis products (NASA MERRA2, ECMWF ERA5, NOAA CFSR2) against observations by saildrones. In general, fluxes from these reanalysis products and saildrone observations agree well, except for CFSR2 sensible heat fluxes, which exhibit systematic negative biases. Sporadic, very large (greater than two observed standard deviations) discrepancies between fluxes from the reanalysis products and observations do occur. These substantial discrepancies in the reanalysis products primarily result from errors in temperature for sensible heat fluxes and errors in both humidity and wind speed for latent heat fluxes. The results from this study suggest that the sea surface latent and sensible heat fluxes from MERRA2 and ERA5 are reliable in representing the mean features of air-sea exchanges in the sub-Arctic region. Nonetheless, their reliability is limited when used for studies of high-frequency variability, such as synoptic weather events.