ENSEMBLE SENSITIVITY ANALYSIS OF ICE FLOW SIMULATIONS WITH DIFFERENT PARAMETRIC MODEL UNCERTAINTIES

Open Access
Author:
Li, Wenjie
Graduate Program:
Meteorology
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
August 13, 2018
Committee Members:
  • Fuqing Zhang, Thesis Advisor
  • Richard B Alley, Thesis Advisor
  • Byron Richard Parizek, Thesis Advisor
  • Johannes Verlinde, Committee Member
Keywords:
  • ensemble sensitivity
  • ice flow
  • basal shear stress
Abstract:
Large ensembles of forward model simulations with well-constructed parametric model perturbations can provide useful insights to assessing different sources of uncertainties in ice- stream modeling. Ensemble sensitivity analysis of these simulations can further provide guidance to retrieve and improve these uncertain parameters through data assimilation, as well as to assess what potential observations will be more valuable for reducing such uncertainties.  In this thesis, we use a 2-d flowline model to study the ensemble sensitivity of ice-flow speed and ice thickness to the spatially varying basal friction profile, bottom topography and upstream flux. We found that ice surface velocity is highly correlated to the basal drag at the downstream end.  The results also show that a 5% uncertainty in the initial upstream mass flux will have a strong impact on the surface ice velocity simulation (their correlations are 0.8 or greater everywhere). This upstream flux uncertainty can be more dominant than the uncertainty due to the basal friction coefficient. Thus, previous retrieval results that did not consider uncertainty in the upstream flux may be overconfident.