Kolmogorov second-order structure functions (second moment of velocity differences) are used to characterize and compare the information contained in five scatterometer wind products. Three of the wind products were obtained using different processing methods applied to SeaWinds-on-QuikSCAT measurements and two from processing ASCAT-on-MetOp-A measurements. The analysis is carried out for rainy and dry regions in the tropical Pacific (nine regions between latitudes 10S and 10N and longitudes 140 and 260E) for the period November 2008 - October 2009. Both monthly and regionally averaged longitudinal and transverse structure functions are calculated using along-track winds. The following quantities were estimated from the structure functions (i) noise levels, (ii) turbulent kinetic energies, (iii) vorticity-to-divergence ratios and (iv) structure function slopes for the range 50 - 250 km. The five wind products show good qualitative agreement, but also important differences due to instrument design and processing. Estimates of noise level are sensitive to the method used. Fits to a symmetric quadratic yield noise levels that correlate well with rain-rate. These noise levels also show that SeaWinds median filter products have larger noise in the transverse component, while ASCAT products have larger noise in the longitudinal component. Fits to an asymmetric quadratic yields information about the strength of the filtering used to reduce noise in Level 1 processing; results imply that ASCAT products are over filtered. It is shown that structure functions can be used as a proxy for the cumulative variance. Estimates of the turbulent kinetic energy show that ASCAT is greater than (less than) or equal to SeaWinds in the divergent (shear) component. Ratios of the shear to divergent turbulent kinetic energy shows that the greatest differences between SeaWinds median filtered and ASCAT winds occur in the convectively active months of each region. Longitudinal (transverse) structure function slopes are steeper (shallower) for SeaWinds than for ASCAT. Slope ratios in most regions show that SeaWinds median filtered winds have steeper longitudinal structure functions, while ASCAT has steeper transverse structure functions. Results for the SeaWinds 2DVAR winds vary, sometimes closer to ASCAT and sometimes closer to the other SeaWinds products.
G King, J Vogelzang, A Stoffelen. Second-order structure function analysis of scatterometer winds over the Tropical Pacific
published, J. Geophys. Res., 2014, 120