APPENDIX 1.
Rank abundance curves for simulated fossil assemblages. For each of the 355 samples Gulf of Alaska data set (>63 um size fraction) used herein, we generated 100 simulated assemblages, composed of 300 specimens each, at the DCA-1 value that had been found for that sample in the original study. These simulations used the kernel density estimates and probabilities of occurrence estimated from the total 2021 data set. Each plot depicts all 48 species in the total data set, which are ordered along the horizontal axis according to the rank order abundance of species in that empirical sample. The proportional abundance (y-axis) of each species in the empirical sample is indicated by a red symbol, consisting of a circle with a red ‘x’ in the middle. A red vertical line indicates the cutoff between species that occurred in the sample and those species that were not present (had an abundance of 0) in that specific empirical sample. The simulated proportional abundances for each species are represented by a black box (made using the ‘boxplot’ function in R package ‘stats’) where the bold horizontal line is the median of the proportional abundance of each species across all 100 simulations performed at that DCA value. The top and bottom of the box represent the first and third quartiles (the 25% and 75% quantiles) of the simulated proportional abundance. The whiskers on each box extend to the location of the most extreme data point that falls within 1.5x the length of the box from the upper or lower edge of the box. Simulated proportional abundances that fall outside of the ‘whisker’ range are represented as black circles. In figures where the boxplots of simulated data overlap with the red symbols that indicate the empirical proportional abundances, the simulations produced a similar abundance distribution as the original samples. For some samples, some of the simulated data plots below the empirical data, which indicates where the simulated assemblages have lower proportional abundances relative to the observed abundances. In some samples, particularly for species to the right of the red vertical line that have an abundance of zero in the empirical sample, the simulated data plots above the empirical data, indicating where the simulated assemblages predicted higher proportional abundances relative to the observed abundance. Appendix 1 and 2 are available for download in one zipped file.
APPENDIX 2.
R code underlying the paleoAM framework and data analyses presented. All files are in a single zipped file directory. The top-level files of the directory include a series of programming scripts which, if run in the sequence indicated by the numbering in their file name, recreate the simulations, analyses and figures included in the manuscript. Subdirectories contain the scripts used for generating individual simulations, the new R functions created for paleoAM, and the data files used in our analyses. Appendix 1 and 2 are available for download in one zipped file.
