search

mcglinnlab / shark-ray-div

Global patterns of shark and ray diversity
0 stars 1 forks source link

readme

shark-ray-div

Global patterns of shark and ray diversity

All raw polygon data for this project can be downloaded from zenodo here: https://zenodo.org/record/6321610#.YkEfVefMK3A

Naylor, Gavin. (2022). Global Dataset of Shark and Chimaera Ranges [Data set]. Zenodo. https://doi.org/10.5281/zenodo.6321610

Chlorophyll data was retrived from NASA Earth Observatory: https://neo.sci.gsfc.nasa.gov/view.php?datasetId=MY1DMM_CHLORA

All other environmental variables were retrieved from NOAA's World Ocean Database: https://www.nodc.noaa.gov/OC5/WOD/pr_wod.html

Hurlbert and Stegen (2014) simulation code: https://doi.org/10.5281/zenodo.5523198

Instructions for recreating analysis

  1. Simulation
    • Code: ./scripts/Analysis_workflow_local.R
      • Produces needed simulation results and adapted from Hurlbert and Stegen (2014). We specifically modified this to remove an if else which prevented the beta value for large trees from being computed. The sim IDs listed in which.sims corresponds to the following models:
        • 3465:3474 = Niche Conservatism Tropical Origin
        • 3565:3574 = Niche Conservatism Temperate Origin
        • 4065:4074 = Ecological Limits Tropical Origin
        • 4075:4084 = Ecological Limits Temperate Origin This script sources the following scripts and data files, all of which were copied from the above Zenodo link:
        • ./scripts/reg_calc_and_analysis.r
        • ./scripts/make.phylo.jimmy.fun.r
        • ./scripts/lat.grad.time.plot.r
        • ./scripts/clade.origin.corr.plot.r
        • ./scripts/clade.exmpl.figs.r
        • ./scripts/extinct.calc.r
        • ./scripts/unzipping_files.r
        • ./data/SENC_Master_Simulation_Matrix.csv
    • Code: ./scripts/Graphics.R
      • Creates the graphics for figure 1. It also creates two data frames, data/stats/cor_df_sims.Rdata and data/stats/p_df_sims.Rdata, which are used for calculating MSE.
  2. Shark analysis
    • Code: ./scripts/Initial_cleanup.R
      • Where all environmental rasters are read in, created, and cleaned
    • Code: ./scripts/Rasterize_polygons.R
      • Where range map polygons are read in, transformed into rasters, stacked, and summed to create richness maps and rasters of species richness for the global analysis, the subclade analysis, and an IUCN test case
      • Environmental rasters are also plotted here
    • Code: ./scripts/Phylogenetic_metrics.R
      • The whole tree is read in and cleaned, so duplicates and polytomies are removed
      • The mini_tree function creates subset trees for sub-analyses
      • The com_mat function creates community matrices to be used to calculate MRD
      • Rasters and maps are created for MRD
      • The beta statistic is calculated using the maxlik.betasplit function for the total analysis and the two subclade analyses
      • A ggtree object is created for the total shark tree to be used in the figure 2 graphic
    • Code: ./scripts/Data_analysis.R
      • All raster data is pulled into a single data frame, and the make_plot function creates regression plots for all variables
      • Figure 2 is generated here using the make_plot_fig2 function and grid.arrange
    • Code: ./scripts/Ecoregions.R
      • Realm polygons are pulled in and cleaned
      • All components of the analysis (rasterization, phylogenetic metrics, and data analysis/graphing) are repeated in this script for the Tropical Atlantic and Central Indo-Pacific realms
    • Code: ./scripts/Final_graphic.R
      • All correlation coefficients and beta values for the global analysis, subclade analyses, and ecoregion analyses are pulled into a list of data frames corresponding to scale resolution, along with the same statistics from the simulation
      • Data frames are similarly created for the upper and lower confidence interval of each statistic
    • Code: Obs_pred_plot.R
      • MSE is calculated for each analysis and hypothesis combination using the table_list of correlation coefficients and beta values and the error_list of confidence intervals
      • Figure 3 is created here along with other data visualizations describing how close each hypothesis is to fitting the real world data from each analysis