How accurately do we know interannual variations of surface mass balance and firn volume in Antarctica?

Knowing the interannual variations in the Antarctic ice sheet net snow accumulation, or surface mass balance (SMB), is essential for analyzing and interpreting present-day observations. For example, accumulation events like the one in East Antarctica in 2009 (Shepherd et al. 2012, Science, doi: 10.1126/science.1228102) challenge our ability to interpret observed decadal-scale trends in terms of long-term changes versus natural fluctuations. SMB variations cause changes in the firn density structure, which need to be accounted for when converting volume trends from satellite altimetry into mass trends. Recent assessments of SMB and firn volume variations mainly rely on atmospheric modeling and firn densification modeling (FDM). The modeling results need observational validation, which has been limited by now. Geodetic observations by satellite gravimetry and satellite altimetry reflect interannual firn mass and volume changes, among other signals like changes in ice flow dynamics. Therefore, these observations provide a means of validating modeling results over the observational period. We present comprehensive comparisons between interannual mass variations from SMB modeling by the regional atmospheric climate model RACMO2 and GRACE satellite gravimetry, and between interannual volume variations from ENVISAT radar altimetry (RA) and firn densification modeling (FDM). The RA-FDM comparison spans the spatial scales from 27km to the continental scale. The mass comparison refers to the regional (drainage basin) and continental scale. Overall, we find good agreement between the interannual variations described by the models and by the geodetic observations. This agreement proves our ability to track and understand SMB-related ice sheet variations from year to year. The assessment of differences between modeling and observations allows us to quantify upper bounds on the uncertainties of the results from either technique. Moreover, the joint analysis of the four techniques aids in attributing the remaining discrepancies to deficits in individual techniques. These analyses may guide further improvements in observing and modeling SMB-related ice sheet variations.