Extracting climate mechanisms from noisy data
Leader: John Moore Start of the project: 2001
Research group: Global change End of the project: Ongoing
The project is funded by Thule Institute, University of Oulu and Academy of Finland.
In general all data collected in Environmental Geosciences (and also in many other fields) can be classified as short, noisy and far from optimally sampled. This is unavoidable given the difficult logistical and financial constraints on collecting the data. However within the last decade spectacular progress has been made in the mathematics of just such types of problem – many of which are based on very general mathematical techniques. We have developed methods can be widely applied to many problems, not just in Global Change, but in other fields where data are sparse, the problem is ill-defined, and the dynamics essentially non-linear. In practice this is description of what we understand of the natural environment.
The main goals of the research are:
1. To develop methods to extract simple relationships from our archive of ice core proxies of climate, and relate those observations to other time series of the climate.
2. To examine causality relationships between different variables that measure climate.
3. To prove whether the simple oscillations, trends and relationships are statistically significant using our best knowledge of how the real-world behaves.
4. To make the methods we develop widely available to other researchers studying climate or similar time series.
Our research has focused to date on Singular Spectrum Analysis, Wavelet methods, and Phasing between time series. The time series we have worked on include: ice core isotope and chemistry records, large-scale atmospheric and oceanic circulation indices (SOI, AO, NAO, PDO etc), sunspot numbers, Hurricane and cyclone statistics, tide gauge sea level records, ground penetrating radar series
Key outcomes of the research so far have been:
• The development of an optimal method of determining the non-linear trend through a time series, and its associated confidence interval.
• Development of a popular website where the method of wavelet coherence is demonstrated and software provided in Matlab
• Extraction of novel proxy climate data: continentality and a chemical proxy for summer melting in Svalbard ice cores.
• Development of phase sensitive methods to extract causality relationships between time series: application to sunspot numbers and solar irradiance data, and tropical cyclones.
• Novel global and regional sea level reconstructions based on an optimal method of analyzing tide-gauge stations back to 1850.
PDFs of all these publications and others from the group are available here:
• Grinsted, A. J.C. Moore, V. Pohjola, T. Martma and E. Isaksson, 2006 Svalbard summer melting, continentality and sea ice extent from the Lomonosovfonna ice core, J. Geophys. Res., D07110, 10.1029/2005JD00649 pdf
• Jevrejeva, S., A. Grinsted, J. C. Moore, and S. Holgate, 2006 Nonlinear trends and multi-year cycles in sea level records. J. Geophys. Res., C09012, 10.1029/2005JC003229 pdf
• Moore, J.C., A. Grinsted and S. Jevrejeva 2006, Is there evidence for sunspot forcing of climate at multi-year and decadal periods? Geophys. Res. Lett. L17705, doi:10.1029/2006GL026501 pdf
• Moore J.C. and A. Grinsted, 2006, Singular spectrum analysis and envelope detection: methods of enhancing the utility of ground-penetrating radar data, J. Glaciol., 52 (176), 159-163 pdf
• Moore, J.C., Grinsted, A., & Jevrejeva, S. 2005 New tools for analyzing time series relationships and trends. EOS, 86, 226 & 232. pdf
• Grinsted, A. J. C. Moore, S. Jevrejeva 2004 Application of the cross wavelet transform and wavelet coherence to geophysical time series, Nonlin. Proc. Geophys. 11, 561-566 pdf
The Arctic Centre Modeling team are:
Senior Scientist John Moore
Arctic Centre, University of Lapland
john.moore @ ulapland.fi
Tel: + 358 16 341 2757 / +358-40-5007775
Scientist Aslak Grinsted (modeler)
Arctic Centre, University of Lapland
ag @ glaciology.net
Cross Wavelet and Wavelet Coherence website