Skip to main content
eScholarship
Open Access Publications from the University of California

Improving a Priori Regional Climate Model Estimates of Greenland Ice Sheet Surface Mass Loss Through Assimilation of Measured Ice Surface Temperatures

  • Author(s): Navari, Mahdi
  • Advisor(s): Margulis, Steven Adam
  • et al.
Abstract

The Greenland ice sheet has been the focus of climate studies due to its considerable impact on sea level rise. Accurate estimates of surface mass balance components - including precipitation, runoff, and evaporation - over the Greenland ice sheet would contribute to understanding the cause of the ice sheet’s recent changes (i.e., increase in melt amount and duration, thickening of ice sheet interior, thinning at the ice sheet margins) and help to forecast future changes. Deterministic approaches provide a general trend of the surface mass fluxes, but they cannot characterize the uncertainty of estimates. The data assimilation method developed in this dissertation aimed to optimally merge the satellite-derived ice surface temperature into a snow/ice model while taking into account the uncertainty of input variables. Satellite-derived ice surface temperatures were used to improve the estimates of the Greenland ice sheet surface mass fluxes.

Three studies were conducted on the Greenland ice sheet. The goal of the first study was to provide a proof of concept of the proposed methodology. A set of observing system simulation experiments was performed to retrieve the true surface mass fluxes of the Greenland ice sheet. The data assimilation framework was able to reduce the RMSE of the prior estimates of runoff, sublimation/evaporation, surface condensation, and surface mass loss fluxes by 61%, 64%, 76%, and 62%, respectively, over the nominal prior estimates from the regional climate model. In the second study, satellite-derived ice surface temperatures were assimilated into a snow/ice model. The results show that the data assimilation framework was capable of retrieving ice surface temperatures with a mean spatial RMSE of 0.3 K which was 69% less than that of the prior estimate without conditioning on satellite-derived ice surface measurements. Evaluation of surface mass fluxes is a critical part of the study; however, it is limited by the spare amount of independent data sets. Several data sets were used to investigate the feasibility of verification of results. It was found that predicted melt duration is in agreement with melt duration from passive microwave measurements; however, more efforts are needed to further verify the results. In the third study, the feasibility of microwave radiance assimilation was investigated by characterizing the error and uncertainty in predicted passive microwave brightness temperature from the radiative transfer model. We found significant uncertainty between the predicted measurement and satellite-derived passive microwave brightness temperature due to error in snow states, coarse resolution of the passive microwave and also an imperfect coupled snow/ice and radiative transfer model. Based on our findings, radiance assimilation requires more accurate snow grain size parameterization to take into account temporal and spatial variability of snow grain size. Furthermore, coarse resolution of both passive microwave brightness temperature and snow/ice model and attribute uncertainties of both predicted and measured brightness temperature make the radiance assimilation unattractive. This research demonstrates that ice surface temperature measurements have valuable information that can be extracted by a data assimilation technique to improve the estimates of the Greenland ice sheet surface mass fluxes.

Main Content
Current View