Characterizing Buoy Wind Speed Error in Extreme Conditions through a Comparison with Scatterometers and ERA5 Reanalysis

Buoys are a common source of validation data for remote sensing over the open ocean and are one of only a small number of in situ sources in this data sparse region. Previous studies have shown the presence of a low wind speed bias for buoys in extremely high wind conditions, but the data for the high wind speed ranges are limited. Therefore, it is important to test the validity of using buoys as an in situ source for satellite calibration in high winds and high seas using a relatively long record of buoy winds. Sub-setting scatterometer and buoy wind speed differences using wave parameters provides insight into the conditions under which wind speeds diverge between these wind sources. This study uses the Royal Netherlands Meteorological Institute ASCAT and NASA QuikSCAT Ocean Vector Wind datasets to compare with buoy winds processed through the Global Telecommunications System and European Centre for Medium-Range Weather Forecasts 5th Generation Reanalysis wave data from 1999 to 2018. The scatterometer, buoy and reanalysis data are triple collocated in space by less than 25 km, and time less than 30 minutes. A probability distribution function analysis with wind speed differences sub-divided by wind speed ranges and wave characteristics is performed to test for the dependence of wind speed differences on sea state. Wind speed differences are further binned by anemometer height to investigate how physical buoy characteristics affect these differences. To account for buoy wind speed differences due to different stability conditions in the boundary layer, buoy winds at different anemometer heights are converted to 10 m equivalent neutral winds (U10EN) which is the variable to which scatterometer winds are calibrated. Comparisons show a difference between the high wind speed calibrations of QuikSCAT and ASCAT where QuikSCAT winds exceed buoy U10EN by nearly 4 m s-1 and ASCAT winds exceed buoy U10EN by 0.65 m s-1 on average in the 20 to 25 m s-1 buoy U10EN range. However, the PDFs of wind speed differences (Buoy U10EN − Scatterometer) binned by wind speed ranges vary as a function of ERA5 significant wave height and buoy anemometer height. Buoy U10EN associated with low anemometers is typically lower than buoy U10EN from higher anemometers for the same significant wave height and scatterometer wind speed bins and these differences increase as significant wave height increases. This pattern is particularly apparent when significant wave heights exceeding 4 m are combined with wind speeds exceeding 12 m s-1. Therefore, it is important to consider modification of buoy winds by the sea state in extreme conditions prior to their application to validate remotely sensed winds.