Resolving the Discrepancy


Abdalati said, “We didn’t regard the discrepancy in findings as an error to be corrected. Rather, we needed to understand why the discrepancy existed and to learn what information was contained in those different results.”

Scott Luthcke, of NASA’s Goddard Space Flight Center Planetary Geodynamics Laboratory, set out to determine if the discrepancy was a result of analytical assumptions and numerical manipulations. Knowing that many factors can affect GRACE’s raw gravity measurements, Luthcke embarked upon his own reprocessing of the data that provided greater spatial and temporal resolution than the original monthly averaged GRACE gravity field maps that were provided to Velicogna. His “new” technique was actually a modern update of the approach used by Apollo missions to map the moon’s gravity field. With this updated system for processing GRACE data, Luthcke was able to isolate geographic areas by drainage system and by elevation. He also aggregated the data into shorter time (ten-day) increments.

Using these finer resolutions of space and time on GRACE data, Luthcke calculated a net annual ice loss for the Greenland ice sheet of about 101 gigatons per year between July 2003 and July 2005. He also duplicated Velicogna’s analysis, and he got results similar to hers. This suggested that the discrepancy was due neither to GRACE data nor to the scientists’ analytical models, but, rather, to how GRACE data had originally been processed. Luthcke’s effort confirmed that GRACE data (like many data sources) are extremely sensitive to the methodology used to process and convert the raw data into values ready for research calculations. Luthcke’s study also confirmed that GRACE data are very reliable and well-suited for monitoring Earth’s great ice sheets.

Velicogna herself pinpointed two other important reasons why her estimates of the ice loss were larger than estimates made by other scientists. The first is model approach and methodology. Studies using GRACE data convert gravitational force to mass, whereas studies using data from laser and radar altimeters such as ICESat or the European Remote Sensing (ERS) satellites convert changes in surface elevation to mass. Different sets of assumptions are required by each of these different data sources; each approach brings its own inherent sources of error and approximation into the calculations.

The second and potentially greater source of the discrepancy is time. Different studies used data from different years. Scientists now know there was a substantial jump in ice loss from 2002 through 2004 compared to previous years. The timing of the GRACE data was such that Velicogna’s study coincided with the beginning of a period of increased ice loss on Greenland.

Researchers Mass Change (GT/year) Method Time span
Krabill et al. 2000 -47 Aircraft Surveys 1994–1999
Velicogna et al. 2006 -200 to -260 GRACE 2002–2006
Luthcke et al. 2007* -145 to -175 GRACE 2003–2006
Zwally et al. 2007* -80 to -100 ICESat 2003–2005

Further, two years of data is a rather short time period to be the basis for reliable model comparisons. As data for additional years has become available all the NASA research community’s estimates are converging toward an average annual net loss of ice mass loss of 150 to 180 gigatons per year. For example, Byron Tapley, an Aerospace Engineering Professor at University of Texas at Austin (2006), estimated total ice sheet loss at 209 gigatons per year. Eric Rignot, of NASA’s JPL (2006), calculated a net loss of 162 gigatons for 2005. Preliminary numbers produced by Jay Zwally (also released in 2006) using data from ICESat ranged from 80 to 100 gigatons of annual ice loss. Today, Velicogna’s new calculations adding data for 2005 and 2006 to her previous estimates show ice loss at just over 210 gigatons per year. When Luthcke and his colleagues updated their calculations with GRACE data collected through June 2006, they calculated a loss of 161 gigatons per year.


Scientists using different sources of data and different ways of processing raw data have generated a range of estimates for the mass balance of the Greenland Ice Sheet. Despite some differences, all the results indicate that the ice sheet has lost hundreds of gigatons of mass in recent years. (*Estimates based on unpublished data.)

Map of elevation change on the Greenland ice sheet.

While some differences remain in current estimates, Abdalati says the real story is not the discrepancies between these studies, it is the degree of their agreement. Every study over the past seven years—using a variety of assumptions, data sources, and study methodologies—has come to the same conclusion: Greenland is now losing significant amounts of ice each year.

“While differences in these studies still exist,” Abdalati concludes, “collectively, they very convincingly paint a picture of the Greenland Ice Sheet as having been close to balance in the 1990s, contributing a small amount to sea level, but becoming significantly out of balance and losing a substantial amount of ice to the sea in the last several years.”

Today, the questions facing Abdalati and his colleagues have evolved from whether the ice sheet is growing or shrinking to what will the ice sheet do in the future? Will its rate of shrinking accelerate? Slow down? Reverse? Remain the same?

Velicogna marvels at her and her colleagues’ ability to address such questions today using the new capabilities of GRACE and ICESat. “Putting all these instruments’ data together is giving us the full picture,” she states. “Now we can look at things that are happening and see what is really going on. To me, that is exciting and really cool.”

Velicogna points out that for a long time scientists thought that models would be the way they would figure out what was happening on the ice sheets. “But the models aren’t good enough now to predict the changes we are observing,” she observes. “The physics of the ice sheets aren’t fully understood, and so while those models are good, nature is more complex.”

Abdalati says he is convinced that the Greenland Ice Sheet will continue to shrink at a significant and, perhaps, accelerating rate. Already he and his colleagues have new studies underway in which they are investigating the sensitivity of the ice sheet to rising temperatures and the specific mechanisms by which the ice sheet responds to increased warmth. “As we look at Greenland through the eyes of these remarkable new satellite missions,” he says, “the ice sheet continues to surprise us.”


  • Abdalati, W. and K. Steffen. (2001). Greenland ice sheet melt extent: 1979-1999. Journal of Geophysical Research,106, (D24), 33983-33988. doi:10.1029/2001JD900181.
  • Herring, D. (2005). Time on the shelf. NASA Earth Observatory. Accessed: August 16, 2007.
  • Howat, I., Joughin, I., Tulaczyk, S., and Gogineni, S. (2005). Rapid retreat and acceleration of the Helheim glacier, East Greenland. Geophysical Research Letters, 32(10), L22502, doi:10.1029/2005GL024737.
  • Johanessen, O., Khvorostovsky, K., Miles, M., and Bobylev, L. (2005). Recent ice-sheet growth in the interior of Greenland. Science,310(5750), 1013-1016. doi: 10.1126/science.1115356.
  • Krabill, W., Abdalati, W., Frederick, E., Manizade, S., Martin, C., Sonntag, J., Swift, R., Thomas, R., Wright, W., and Yungel, J. (2000). Greenland Ice Sheet: high-elevation balance and peripheral thinning. Science, 289(5478), 428-430. doi: 10.1126/science.289.5478.428.
  • Krabill, W., Frederick, E., Manizade, S., Martin, C., Sonntag, J., Swift, R., Thomas, R., Wright, W., and Yungel, J. (1999). Rapid thinning of parts of the Southern Greenland Ice Sheet. Science, 283(5407), 1522-1524. doi: 10.1126/science.283.5407.1522.
  • Krabill, W., Hanna, E., Huybrechts, P., Abdalati, W., Cappelen, J., Csatho, B., Frederick, E., Manzinade, S., Martin, C., Sonntag, J., Swift, R., Thomas, R., and Yungel, J. (2004). Greenland Ice Sheet: Increased coastal thinning. Geophysical Research Letters, 31, L24402, doi:10.1029/2004GL021533.
  • Luthcke, S., Zwally, H., Abdalati, W., Rowlands, D., Ray, R., Nerem, R., Lemoine, F., McCarthy, J., and Chinn, D. (2006). Recent Greenland ice mass loss by drainage system from satellite gravity observations. Science, 314(5803), 1286-1289. doi: 10.1126/science.1130776.
  • Luthcke, S., Zwally, H., Rowlands, D., Abdalati, W., Nerem, R., Ray, R., Lemoine, F., and Chinn, D. (2006). Recent ice sheet mass change observations from GRACE mascon solutions. EOS Transactions AGU,87(52), Fall Meeting Supplement, G12A-07.
  • Mosley-Thompson, E., Readinger, C., Craigmile, P., Thompson, L., and Calder, C. (2005). Regional sensitivity of Greenland precipitation to NAO variability. Geophysical Research Letters,32, L24707, doi:10.1029/2005GL024776.
  • Rignot, E., and Kanagaratnam, P. (2006). Changes in the velocity structure of the Greenland Ice Sheet. Science311(5763), 986-990. doi: 10.1126/science.1121381.
  • Rignot, E, and Thomas, R. (2002). Mass balance of the polar ice sheets. Science297(5586), 1502-1506. doi: 10.1126/science.1073888.
  • Rothrock, D., Yu, Y., and Maykut. G. (1999). Thinning of the Arctic sea-ice cover. Geophysical Research Letters, 26(33), 3469-3472. doi: 10.1029/1999GL010863.
  • Serreze, M., Maslanik, J., Scambos, T., Fetterer, F., Stroeve, J., Knowles, K., Fowler, C., Drobot, S., Barry, R, and Haran, T. (2003). A record minimum Arctic sea ice extent and area in 2002. Geophysical Research Letters, 30(3), 1110. doi: 10.1029/2002GL016406.
  • Velicogna, I., and Wahr, J. (2005). Ice mass balance in Greenland from GRACE. Geophysical Research Letters, 32(18), L18505. doi: 10.1029/2005GL023955.
  • Wahr, J., and Velicogna, I. (2003). What might GRACE contribute to studies of post glacial rebound? Space Science Reviews, 108(1-2), 319-330. doi: 10.1023/A:1026183526762.
  • Weir, J. (2003). Searching for Atlantic rhythms: winter weather and the North Atlantic Oscillation. NASA Earth Observatory. Accessed: August 16, 2007.
  • Zwally, H., Abdalati, W., Herring, T., Larson, K., Saba, J., and Steffen, K. (2002). Surface melt-induced acceleration of Greenland Ice Sheet flow. Science,297(5579), 218-222. doi: 10.1126/science.1072708.
  • Zwally, H., Giovinetto, M., Li, J., Cornejo, H., Beckley, M., Brenner, A., Saba, J., and Yi, D. (2005). Mass changes of the Greenland and Antarctic ice sheets and shelves and contributions to sea-level rise:1992-2002. Journal of Glaciology, 51(175), 509-527.

Using ICESat data, scientists calculated the change in thickness of the Greenland Ice Sheet between 2003 and 2006. Increased melting and faster glacial flow lowered the surface along the margins of the ice sheet (brown), while increased snowfall thickened the ice in the interior of Greenland (blue). This pattern is consistent with model predictions of how global warming will affect the ice sheet. (NASA map by Robert Simmon, based on ICESat data.)