Permalink : http://hdl.handle.net/2115/45455

KEYWORDS : Svalbard;ice core;ion chemistry;climate change;statistical analysis;diffusion;percolation

We examine the impact that post-depositional change has on ion concentrations in ice cores that suffer limited seasonal melting. We show that the impact in the case of at least one Svalbard ice core is limited to decreasing resolution of signals to about 3 years - a similar accuracy as the best dating can usually provide. We model various effects such as percolation and washout of ions, and show how they can be used to construct useful ice core proxies. The sulfate profile shows how post depositional effects lead to magnesium and nitrate being excellent predictors under normal conditions for sulfate - indicating a difference in the properties of acidic and neutral salt sulfate components. The approach leads to an objective method of detecting volcanic signals in sulfate, even when the signals are much smaller than the sulfate variability caused by other factors. We emphasis the importance of correct statistical analysis the interpretation of ion profiles, taking into account how measurement errors and the distribution of ion chromatography data affects how statistical modelling should be done.