Scatterometers estimate the relative atmosphere-ocean motion at spatially high resolution and provide accurate inertial-scale ocean wind forcing information, which is crucial for many ocean, atmosphere and climate applications. An empirical scatterometer ocean stress (SOS) product is estimated and validated using available statistical information. A triple collocation dataset of scatterometer, moored buoy and numerical weather prediction (NWP) observations together with two commonly used surface layer (SL) models are used to characterize the SOS. First, a comparison between the two SL models is performed. Although their roughness length and the stability parameterizations differ somewhat, the two models show little differences in terms of stress estimation. A triple collocation exercise is then conducted to assess the errors of both the observations and the SL models. The results show that the uncertainty in the NWP dataset is generally larger than in the buoy and scatterometer wind/stress datasets, but depending on the spatial scales of interest. The triple collocation analysis also shows that scatterometer winds are as close to real winds as to neutral winds, provided that we use the appropriate scaling. An explanation for this duality is that the small stability effects found in the analysis are masked by the uncertainty in SL models and their inputs. The triple collocation analysis shows that scatterometer winds can be straightforwardly and reliably transformed to wind stress. This opens the door for the development of wind stress swath (level 2) and gridded (level 3) products for the Advanced Scatterometer (ASCAT) onboard MetOp and for further geophysical development.
M Portabella, A Stoffelen, A Verhoef, J Verspeek, J Vogelzang. Towards a high-resolution gridded ocean forcing
2007, 2007, EUMETSAT