Multi-parameter selection curves for machine-assisted annual layer interpretations of the WAIS-Divide core

Affiliation(s) PI Project period Funded by
DEES McGwire, Kenneth C 01/01/2012 - 12/31/2012 National Science Foundation

Project Description

This project extends a novel "selection curve" method was developed to interpret annual layers in the West Antarctic Ice Sheet (WAIS) Divide ice core based on dielectric properties (DEP) to make use of multiple chemical measurements. The selection curve algorithm uses a spline curve whose shape selects successive annual peaks in plots of ice core parameters. Using 50 meters of manually interpreted calibration data, the selection curve method matched a manual interpretation throughout a 1200 m dataset to within 1% of the total count and 2% RMSE. In this proposal, the selection curve method will be enhanced to incorporate multiple chemical measurements derived from high-resolution CFA measurements of the WAIS Divide ice core. This enhancement will include a number of indicators for the quality of annual peak selections. For each chemical species, the statistical distribution of values for these quality measures will be calculated in order to establish the significance of deviations from mean values. Instead of attempting to merge annual signals of different chemical species into a composite response variable, this new multi-parameter version of the selection curve algorithm will apply a multi-objective numerical optimization routine to each annual pick. Proceeding along the core, each annual pick will maximize the consistency and quality of peaks across all measurements. Stored values for peak quality measures will flag questionable annual picks for manual verification. The position and shape of annual peaks for each chemical will be characterized, and depth-dependent changes in the relative timing and variability of these peaks will be examined to determine to what degree this seasonal-scale variation correlates with known changes in climate. The intellectual merit of this project is many-fold. The majority of annually-resolved dating that has been published for long ice core records has relied on manual counting techniques. The proposed method would provide an objective, repeatable method for multi-parameter counting to support the development of the final WAIS Divide annually-resolved timescale. This capability would improve the speed and end-to-end consistency of timescale development at WAIS Divide and in future studies. The proposed method provides quality measures for each annual pick that will allow prioritized consideration of regions in the ice core that are difficult to interpret. The method also provides quantitative information regarding differences in the seasonal timing and variability of peaks for different chemical species that may relate to seasonal variation in accumulation rates, or flag suspected problems such as flipped cores. This ability to extract ice core information at sub-annual scales has the potential to greatly enrich the understanding of past climate regimes. By improving ice core science, this work will have broader impacts that benefit society. Ice core science addresses critically important questions related to global warming, abrupt climate change, and biogeochemical cycling that should inform policy makers. By integrating an objective, repeatable multi-parameter counting tool into the final timescale development strategy at WAIS Divide, the proposed effort would reduce potential undue criticism of subjectivity in ice core interpretations and climate change science. The proposed analytical software will be made available to the broad community, improving scientific infrastructure, and could be used for non-ice core applications as well. This proposal does not require field work in the Antarctic.

This website uses cookies
We use cookies to personalise content, provide social media features and  analyse our traffic. We also share information about your use of our site with our social media and analytics partners who may combine it with other information that you’ve provided to them or that they’ve collected from your use of their services. You consent to our cookies if you continue to use our website. DRI Privacy Policy >>