The field observations of englacial and subglacial processes will directly address unknowns in current modelling efforts, and will be directly integrated into our numerical modelling efforts. Our state-of-the-art 3-D dynamic glacier model, iSOSIA53, has been developed specifically to capture the dynamics of ice flow within mountain glaciers. iSOSIA is a second-order shallow-ice model that includes 3-D stress and velocity fields but is sufficiently economical to make large numbers of simulations at a high resolution (100-m grid spacing) over domains of tens of km2. We have recently developed this model to explicitly include the feedbacks between debris transport, ice flow and mass balance6, which enables us to realistically simulate glacier mass loss by surface lowering rather than making predictions based on clean ice glaciers as other studies have done. We have applied the latest version of iSOSIA to investigate the changes in volume and dynamics of Khumbu Glacier from the Little Ice Age through the present day to AD2200, finding that our simulations compared well with observations of contemporary velocities, surface elevation change and geodetic mass balance. However, the model has never been parameterised with real subsurface data, meaning the outputs have large uncertainty; by addressing this shortcoming we will make the first ever predictions of Himalayan glacier evolution based on real-world measurements and thus set a new standard in debris-covered glacier modelling.
In our approach, the distribution of ice and debris are simulated from spin-up, so the model can be initialised with minimal input data. The model has high computational efficiency, such that a simulation of a large glacier like the tongue of Khumbu (≤50 km2 ice area) can be run using 12 CPU cores in around 12 hours. This allows us to make multiple simulations and perform multi-parameter sensitivity experiments to test the effects on predicted glacier change of model parameterisations of englacial and subglacial properties. We propose running such sensitivity tests in the initial stages of the project to identify specific requirements for the modelling and tailor our field programme accordingly. The modelling strategy will therefore comprise three phases: (i) testing the sensitivity of the model to the parameters we are able to measure in the field to better formulate model descriptions of the englacial and subglacial processes operating on glaciers in the Everest region (and fine-tune our field data collection), (ii) introducing field data into the model to include processes such as hydrologically-driven variations in basal motion and updated boundary conditions such as englacial temperature, and tuning the model to e.g. englacial tilt and sediment transfer, (iii) running the upgraded version of the model to simulate the evolution of glaciers in the Everest region under IPCC CMIP5 climate scenarios (or more recent results if available) using published mass balance data and tuned to observations of present and past glacier extents.