SPlus Scipts for Paleoecology

     

    Explanation and Warning!!

    I work with lake sediments and often find myself wishing for a statistical procedure(s) to explore and summarize my data more fully and to examine specific hypotheses. The latter often revolve around question of presence/absence of a significant changepoint within a series and/or presence of a trend throughout series. Like many paleo data sets, my data often suffer from incomplete sampling and/or irregular sampling with respect to time.

    Below you will find some scripts (SPlus 6.0) that I have worked on in past 3-4 years. Most have not undergone peer review. I am neither a statistician or programmer so if you use these scripts beware. They may contain programming errors and/or violate some basic statistical principles, and I can't provide any support for them nor any guarantees that they do what they are supposed to do. I am making them available in hope that they may prove of some utility. If you find them to be useful (or find errors/make changes) I would welcome any suggestions or corrections.

  • Two sample comparison based on user-defined changepoint. ChangeTest1.ssc
  • Identification of 1-many local changepoints using technique originally described by Webster (1973)
  • Do biplots for variables not sampled at same timepoint. Takes two series and creates a biplot by locating points in second series that within some -- user specified-- amount of time of points within first series.
  • Do a geospatial equivalent of cross-correlation function. rundcf3 script calls dcf5 script.
  • Do a Lomb periodogram on a single series and assess statistical significance of peaks based on permuting (with replacement) the series. Note that files are part of a zipped archive and all need to be loaded and compiled.

 

References

Kirchner, J.W. & Weil, A. (2000) Delayed biological recovery from extinctions throughout the fossil record. Nature, 404, 177-180.

Webster, R. (1973) Automatic soil-boundary location from transect data. Mathematical Geology, 5, 27-37.