 |
 |
Daily wind fields are produced using a heavily weighted temporal average. These products are on a 0.5° grid, for each day scatterometer observations were available over part of the globe. The scatterometer winds, and these gridded winds, are calibrated to a height of 10 m. This approach bins the satellite swath observations (in 0.5° bins) without any additional smoothing in space. The data set should be considered research quality.
The effective sampling time is non-homogeneous in space and time, but is typically between one and three days, with a peak at two days. A characteristic averaging time can be estimated by applying the averaging technique to the absolute value of the difference in time between each observation and 12Z on the day to which the wind field applies. This characteristic time is roughly half the effective sampling period. The probability distribution of this characteristic time for a similar 1° resolution product shows the effectiveness of this averaging technique. Data sets with homogeneous sampling characteristics require much larger spatial and temporal bins for averaging. This technique captures relatively rapid and small scale changes that might othe rwise be missed.
NetCDF libraries are required to access the data, and can be obtained from UNIDATA.