 |
| Titel |
Mass loss over the Greenland ice sheet from GRACE: A reappraisal |
| VerfasserIn |
C. Harig, F. J. Simons |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250070509
|
|
|
|
|
|
| Zusammenfassung |
| Where has Greenland been losing mass, and when? Has the mass loss been accelerating in the
second half of the last decade? We have developed a versatile new method to answer
questions like this and others, on a regional and global scale, and use it to reassess the
spatial and temporal behavior of Greenland’s mass balance using data from the
Gravity Recovery and Climate Experiment (GRACE) produced in the last decade.
Our method is validated with extensive simulation-based synthetic recovery tests,
and requires very few initial assumptions, e.g. there is no a priori subdivision of
the continent into geographic units over which the mass loss would commonly be
assumed to be going on at a constant rate, nor does it start with an explicit filtering
step. Monthly satellite gravity measurements from GRACE, produced since 2002,
have led to many estimates of the magnitude, spatial and temporal distribution of
mass change of Earth’s ice sheets. These estimates often disagree for a variety of
known reasons, such as the time span considered, however the reported errors for
equivalent mass estimates usually do not encompass estimates from other researchers.
We examine Level-2 GRACE data with the goal of determining what portion of
the gravity field can be considered signal and which can be considered noise. We
perform our analysis in both the spherical harmonic basis and in the spherical Slepian
basis, the latter providing spatiospectral localization of both mass change and noise
in our regions of interest, Greenland and Antarctica. After fitting and removing
secular, annual, and semiannual periodic functions, we use the time series residuals
to determine a time averaged noise covariance matrix, which unlike the Level-2
calibrated errors, has strong off-diagonal terms. These terms add contributions important
for determining the spatial variance and spatial covariance of the noise. With a
more robust estimate of signal to noise in the gravity measurements, we determine
what portion of the uncertainty in mass estimation is due to error inherent to the
data. We also determine, using Slepian functions, the spatial distribution of mass
change (and its variance) and how this change is spatially distributed in time at a
much improved resolution over previous methods. In 2003 and 2004, mass loss is
concentrated along the eastern coast of Greenland. In 2005 and 2006, mass loss is
reduced in the northeast while it is increased in the southeast. Meanwhile, mass loss
begins to increase along the northwest coast. From 2007 to 2010, mass loss further
increases in northwest Greenland while mass loss diminishes in the southeast coast
areas. During the period from 2002 to 2011 we find the overall total integrated mass
change rate over Greenland remains constant with very little observed acceleration. |
|
|
|
|
|
|