Towards a combined estimation of Greenland’s ice sheet mass balance using GRACE and ICESat data

Towards a combined estimation of Greenland’s ice sheet mass balance using GRACE and ICESat data

Slobbe, D.C.

Ditmar, P. (mentor)
Lindenbergh, R.C. (mentor)

Aerospace Engineering

Department of Earth Observation and Space Systems (DEOS)

Geomatics

2007-11-01

The Greenland ice sheet is sensitive to climate change. Global heating is expected to result in ice mass losses that will contribute to global sea level rise. For this reason monitoring Greenland’s ice mass balance is of utmost importance. Data of both the Ice, Cloud, and land Elevation Satellite (ICESat) laser altimetry mission and the Gravity Recovery and Climate Experiment (GRACE) gravity mission are used to create two independent estimates of Greenland’s ice sheet mass balance over the full measurement period of about 2003 until 2007. For ICESat data, a processing strategy is developed that uses the elevation differences of geometrically overlapping footprints of both crossing and repeated tracks. The dataset is cleaned using quality flags defined by the Geoscience Laser Altimeter System (GLAS) science team. (The GLAS is the sole scientific instrument on ICESat). The cleaned dataset reveals some strong, spatially correlated signals that are shown to be related to physical phenomena like melting glaciers. On the other hand, strong correlation is also visible between the observed elevation differences and the combined effect of roughness and surface slope. Different processing strategies applied to different sets of laser campaigns are used to convert the observed temporal elevation differences to mass changes for 6 different drainage systems, further divided into a region above and below 2000 meter elevation. Here all available laser campaigns are used and outliers are removed using N-sigma thresholding. Both a uniform and non-uniform weighting scheme,used to estimate the elevation changes with respect to a reference epoch, is evaluated. The non-uniform weighting scheme is developed to account for the influence of roughness and surface slope, but it turns out that also signals of interest are sometimes suppressed. In order to obtain our final estimates based on ICESat data, the uniform weighting scheme is applied. For the whole of Greenland the estimated mass change rate is equal to -142.6 Gton/year. This value can be mainly attributed to strong mass losses in the region below 2000 meter elevation. On the other hand we show that for different processing strategies this value ranges between approximately 0 and -200 Gton/year. In general, the obtained results confirm trends discovered by other authors who use altimetry. Differences can be explained by different time spans of the used datasets, but mainly by differences in sampling of the data in the region below 2000 meter. Furthermore, GRACE based monthly variations of the Earth’s gravity field as processed by CNES, CSR, DEOS and GFZ are used to estimate the mass change rate for North and South Greenland. Here, both a Gaussian filter, for half-widths of 300, 500 and 800 km, and a Wiener filter is used. It turns out that the Gaussian filter with a half-width of 500 km has the best performance. The final estimates obtained after application of this filter for the different GRACE solutions range between -60.9 and -93.1 Gton/year. The differences in estimates among different GRACE solutions can be mainly explained by differences in processing strategies used by the processing centers to obtain the monthly gravity fields. Only for the DEOS solutions, these differences can also be attributed to the different time span of this dataset. In any case the estimates are low compared with recently published GRACE estimates, which can be explained by an unaccounted leakage effect in our estimates. The unaccounted leakage effect also mainly explains the differences between estimates based on ICESat and GRACE data. Due to their global coverage and high temporal resolution both the ICESat and GRACE mission have improved the estimations of Greenland’s ice sheet mass balance. Further improvements are possible when both datasets are combined. Hence an attempt is made for a joint inversion of both datasets. Depending on the used GRACE solution the estimated combined mass change rates range between -114.3 and -124.7 Gton/year.

ICESat
GRACE
Greenland
ice sheet
mass balance

http://resolver.tudelft.nl/uuid:0d315b1b-84a4-4e03-8b65-95353e04a598

Student theses

master thesis

(c) 2007 Slobbe, D.C.