 |
| Titel |
Applications of EOF filters to GRACE |
| VerfasserIn |
E. Schrama, B. Wouters |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250026909
|
|
|
|
|
|
| Zusammenfassung |
| The Gravity Recovery and Climate Experiment (GRACE) mission is designed to map the
Earth’s gravity field at monthly intervals with a spatial resolution of about 250 to
500 km. The solutions provided by the GRACE science team demonstrate that a wide
variety of surface mass signals can be observed. For this paper we used the RL04
solution developed by the center of space research at the University of Texas at
Austin. We use all available monthly gravity solutions and represent the temporal
signal as a surface mass layer under the assumption of an elastic lithosphere.
GRACE gravity fields solutions are affected by noise which we attempt to suppress by
means of an Empirical Orthogonal Filtering technique based on Gaussian smoothed
surface mass data. We will discuss various implementations of the EOF filtering
technique that have enhanced the spatial resolution by which surface mass anomalies
can be retrieved from GRACE. Surface mass solutions with EOF algorithm 1 are based
upon a singular value decomposition of smoothed surface mass fields. A later version
of the EOF method, called algorithm 2, directly operates on the spherical harmonic
coefficient sets grouped by spherical harmonic order m.
We obtained surface mass field with EOF algorithm 1 and compared the vertical loading
displacements to a set of independent IGS station data acquired within the GRACE
observation window. An optimum was found at an EOF compression level of 3 and a
smoothing radius of 5 or 6.25 degrees so that the discrepancy between GRACE and GPS
vertical loading signals is around 1.9 mm for 59 IGS stations. In this we assume that
temporal gravity field changes occur on frequencies shorter than about 3 years because
of the way we preprocess the GPS data and GRACE surface mass estimates.
The performance of EOF algorithm 2 was evaluated by means of simulations which
demonstrate that it is feasible to obtain a spatial resolution of 250 km. We
use this method to estimate small scale surface mass trends for
Greenland and found a surface mass trend of 179+/-25 Gton/yr between 2003 and 2008.
Between 2003 and 2005 we observed 121+/-27 Gton/yr and between 2006-2008
204+/-25 Gton/yr. This method showed for the first time mass losses
above 2000m elevation. |
|
|
|
|
|
|