Margulis Research Group

Department of Civil and Environmental Engineering


Home PeopleResearch ThemesResearch ProjectsStudent/Postdoc OpportunitiesTeachingContact

Improving surface mass balance estimation of Greenland ice sheet through assimilation of multi-sensor satellite products and ground measurements into a regional climate model.

(Sponsor: NSF Arctic Natural Sciences/Office of Polar Programs; Collaborator/s: Marco Tedesco, CUNY-CCNY)

The Greenland Ice Sheet (GIS) plays a major role in Arctic climate and is a major consideration in projections of sea level rise. Diagnosing the surface mass balance of the GIS is a critical objective that continues to involve large uncertainties from errors in modeled precipitation and errors related to sub-grid-scale process representation. To date, there has been limited work in integrating remote sensing techniques and ground-based data with modeling. In this project, a data assimilation approach will provide a rigorous framework for merging these disparate sources of information in a consistent way (based on their associated uncertainty) to obtain an optimal surface mass balance posterior estimate comprising maps in space and time. This project will focus on implementing a synergistic modeling/observation framework with the final outcome being improved estimates of the mean and uncertainty of the surface states and fluxes associated with the surface mass balance. Broader impacts include a significant contribution to understanding of GIS influence on sea level rise.

Topographic map of Greenland Ice Sheet.

Conceptual picture (top panel) of Greenland Surface Mass Balance (SMB) and total mass balance (MB) and topography at the MAR modeling resolution (25 km).


Schematic of methodology used to estimate Greenland SMB. The remote sensing data streams include microwave brightness temperature, infrared LST, and albedo. The models used include the MAR regional atmospheric model and the CROCUS ice/snow model.

Preliminary results:

True realization from an observing system simulation experiment (OSSE) showing the a) precipitation, b) evaporation, an c) runoff for 2010.

Spatially averaged SMB terms for the 2010 true realization.


Estimation error relative to the true realization (RMSE) for the prior case and those with assimilation of different remote sensing inputs for each of the key model inputs: precipitation, longwave radiation, shortwave radiation, and air temperature (reflected in terms of positive degree days (PDD) and negative degree days (NDD). All assimilation cases show improvement over the prior, with data streams (e.g. LST) providing more information.

Datasets: All datasets generated as part of this project will be released publically upon publication and completion of the project per NSF data management guidelines. Some of the MAR data is already being released on the Co-PI's Marco Tedesco's website. All assimilation results will be available via ftp from the our server and/or at the NSIDC website. Please contact us about expected release dates, preliminary data availability and consult this site for ftp links when they become available.