This algorithm is developed for annual layer counting in paleoclimate archives. The algorithm is based on the principles of statistical inference of hidden states in semi-Markov processes. States, and their associated confidence intervals, are inferred by the Forward-Backward algorithm. The EM (Expectation-Maximization) algorithm is used to find the optimal set of layer parameters for each data batch. Confidence intervals do not account for the uncertainty in estimation of layer parameters. If (absolute) tiepoints are given, the algorithm is run between these, while assuming constant annual layer signals between each pair. If no tiepoints, the algorithm is run batch-wise down the core, with a slight overlap between consecutive batches. See Winstrup (2011) and Winstrup et al. (2012) for further documentation.
Developed by Mai Winstrup. Contact: email@example.com
When using this script please cite: Winstrup et al., An automated approach for annual layer counting in ice cores, Clim. Past. 8, 1881-1895, 2012. Also please provide release date of the algorithm: 7/7-2015
StratiCounter is written in Matlab, and thus running the script requires a Matlab license.
Released versions of the program can be downloaded from: https://github.com/maiwinstrup/StratiCounter/releases
We always recommend using the latest release.
The software package includes a user manual describing how to set up and run the program, and how to interpret the results.
Or did you encounter a bug?
For questions and inquiries, send me an e-mail at firstname.lastname@example.org describing the issue, and I'll get back to you as soon as I can.
For filing a bug report, please use Github's issue tracking feature: https://github.com/maiwinstrup/StratiCounter/issues/new
Bug reports should at least contain the following information:
- Copy of error message
- Steps to reproduce the issue
- Expected result and what actually happens
- Version of StratiCounter
- Matlab version and operating system
Filed bug reports that do not contain the above information will not be acted upon.