PhD Defense Lars Nørtoft Reiter

Title: Statistical Ecology: An Ecological Archive in Narwhal Tusks

Abstract: 

The Arctic Ocean is home to many remarkable species, among them the narwhal, best known for its spiraled tusk. The mineralized tissues within these tusks provide a window not only into the life history of the whales, but also into the Arctic environment they inhabit. Because many of the elements in the tusks are present in minute quantities (parts per million), tracing them requires highly precise equipment, a task made possibleby Laser Ablation Inductively Coupled Mass Spectrometry (LA-ICP-MS). This thesis analyses 16 tusks collected from various locations in Greenland in connection with local
Inuit hunts. With LA-ICP-MS, we are able to extract several trace element profiles that reflect various aspects of narwhal ecology. For example, the elements barium and strontium showed clear annual cycles reflecting migration between wintering and summering grounds. This observation motivated the development of a time warping model - the subject of
chapter 2 - which allows us to translate spatial positions along the tusk to yearly estimates. Formally, we fit a sum of two sinusoids to the observed cyclic signals. These sinusoids represent annual cycles and winter/summer differences. We model the changing growth as the integral of a (stochastic) square-root diffusion process. This integrated process, which we call the growth process, allows us to parametrize measurements in terms of years rather than spatial positions on the tusk.
  In chapter 3, we use the time warping model to produce age estimates for all whales (tusks) based on the combined seasonal signals of barium and strontium. Each tusk was sectioned longitudinally and then along the cross section to produce smaller pieces compatible with laboratory equipment. The model was applied to each tusk piece, and the resulting signals were then recombined in chronological order. Sectioning can result in repeated growth layers across adjacent tusk pieces, thus also temporal overlaps of signals. We address this using a least-squares comparison of signals after the time warping procedure. Finally, using residual bootstrapping we provide probabilistic estimates of the age under the model.
  Chapter 4 concludes the thesis with an analysis of the whole collection of trace elements using the year estimates obtained from the time warping model. Here we study correlations between groups of elements, and also differences between elements found in tusks from West (a sample of 9) and East (a sample of 7) Greenland. We also estimate important life-history events, such as the time of weaning. Several elements show distinct patterns. For example, barium and strontium exhibit clear annual cycles consistent with migration, whereas magnesium, potassium and rubidium show no cyclicity but are instead enriched in the dentine of the tusks. We also find discover strong correlations within a large group of heavy metals. Arguably, our most important finding is the difference in median element concentrations between East and West Greenland, alongside shared temporal patterns observed for a large subset of elements across both regions.

Supervisor:

Professor Susanne Ditlevsen, University of Copenhagen
Senior Researcher Eva Garde, Greenland Institute of Natural Resources
Professor Mads Peter Heide-Jørgensen, Greenland Institute of Natural Resource

Assessment Committee:

Professor Helle Sørensen (chair), University of Copenhagen
Professor Jacob Carstensen, Aarhus University
Associate Professor Evgeny A. Podolskiy, Hokkaido Universit