Data collected by the C band ERS-2 wind scatterometer (EScat), the Ku band ADEOS-1 NASA scatterometer (NSCAT), and the Ku band SeaWinds on QuikScat (QSCAT) satellite instruments are used to illustrate spatiotemporal variability in snow accumulation on the Greenland ice sheet. Microwave radar backscatter images of Greenland are derived using the scatterometer image reconstruction (SIR) method at 3-day intervals over the periods 1991--1998 and 1996--1997 for EScat and NSCAT, respectively. The backscatter coefficient &sgr;¿ normalized to 40¿ incidence, A, and gradient in backscatter, B, in the range 20¿--60¿ are compared with historical snow accumulation data and recent measurements made in the Program for Arctic Regional Climate Assessment (PARCA) shallow snow pits. Empirical relationships derived from these comparisons reveal different exponential relationships between C and Ku band A values and dry snow zone mean annual accumulation, Q. Frequency difference images between overlapping scatterometer images suggest that C band data are more sensitive to snow layering and buried inhomogeneities, whereas Ku band data are more sensitive to volume scattering from recently accumulated snow. Direct comparisons between NSCAT B values and in situ Q measurements show a linear relationship between ln(Q) and B, with a negative rank correlation of R=-0.8. The root-mean-square residual in fitting regression line equation ln(Q)=3.08-17.83B to the data is 0.05-m snow water equivalent. This simple Ku band empirical relationship is exploited to investigate decadal changes in dry snow zone accumulation between Seasat (1978) and NSCAT (1996). Additional comparisons between NSCAT and recent QSCAT (1999) data reveal significant upslope shifts in the dry snow line along the southwestern flank of the ice sheet. Recent acceleration in the increase in intensity of scattering is observed in the percolation zone, suggesting increased melting between 2000-and 3000-m elevation in the southern half of the ice sheet. ¿ 2001 American Geophysical Union