To investigate SeaWinds on QuikSCAT's backscatter behavior (taken from JPL/PO.DAAC L2A product), two regions of high altitude snow/ice over Greenland and Antarctica, as well as one over the South American tropical rain forest, were chosen. After browsing the data over somewhat larger sectors, these regions were restricted to the following limits :
- Greenland region : latitude [71 N, 73 N], longitude [319 E, 325 E]
- Tropical forest : latitude [5 S, 2.5 S], longitude [285 E, 299.5 E]
- Antarctic sector : latitude [82 S, 80 S], longitude [68 E, 82 E]
For each region, and separately for the inner and outer beam, one or two relative orbit numbers (absolute orbit number modulo 57) were chosen, covering the region at least partially. The figures showing the measurement count indicate that the drift in the satellite orbit modifies notably this number as time passes, the swaths entering or leaving the region partially.
The plots of mean backscatter over the regions (linear or dB) show a smooth evolution in time as a function of absolute orbit number. The values of the coefficients of a least-squares fit of the general form :
sigma0 = at+b
are given in the figures. Measurements of a given beam for all relative orbits considered are used in the fit. The units of t are 100 days, and the origin of t is shifted to correspond to the beginning of each plot. Values of the parameter a depend not only on instrument drift, but also, and essentially, on the seasonal evolution of backscatter and, unfortunately, orbit drift, which changes the ground sampling over the given regions.
The standard deviations of the individual sigma0 measurements over a given region, plotted as a function of orbit number, are computed for each orbit in linear form then transformed to dB. The values observed, around -6 dB, are a combination of instrument noise and geographical fluctuations of backscatter over the given area.
The plot of relative orbit number as a function of absolute orbit number is there only to emphasize that the same relative orbits are used presently. They will have to be modified if the orbit drift becomes too large.