Master Thesis Defense by Jesper Ejlebæk Holm

Title: Firn-air Modelling in the Community Firn Model

Over the last glacial period, a series of 25 climate fluctuations were observed, characterized by rapid warming followed by a longer cooling period. One of these Dansgaard-Oescher (D-O) events happened around 11,500 years ago and heralded the end of the ice age and the beginning of the Holocene period. To simulate these events, studies use firn densification models coupled to air and heat diffusion in order to reconstruct paleotemperatures.

The Community Firn Model (CFM) is one such example, and it provides an open-source-modular framework to simulate physical processes in the firn. A 1-dimensional Lagrangian grid is used where density is explicitly calculated while diffusion is solved through an implicit Finite Volume Method (FVM). The project seeks to aid in the interpretation of these DO-events by conducting
two experiments to better understand the model’s behaviour under different conditions.
Initially, we examine the behaviour of the CFM with different densification schemes by testing a variety of “DO-like” events. It is observed that higher temperatures and lower accumulation rates lead to a shallower close-off depth. Secondly, noisy 𝛿15N data, computed from NGRIP, is inverted into noisy temperature data by way of Brent’s root-finding method. Then the sensitivity of the inverted temperatures is estimated by including uncertainty from sources such as 𝛿15N, surface density and diffusivity parameterization. We find that the magnitude of the 𝛿15N uncertainty is paramount for reliable interpretation of reconstructed temperatures.

Finally, the halfway time to a new equilibrium following a linear fluctuation is computed for a series of different amplitudes and duration. We show that for changes in amplitude, the halfway time tends to decrease, while the behaviour is more complex for the duration. Moreover, we estimate the uncertainty of the temperature gradient by calculating the excess of stable nitrogen isotopes compared to argon. This is used to compute the true equilibrium time for which the opposite behaviour is seen. Based on this the merits of the Halfway time versus Equilibrium time are discussed in the view of interpreting past climates.

Supervisors: Vasileios Gkinis and Michael Döring
Censor: Sebastian Bjerregaard Simonsen, DTU