This page provides data and code for Brandt et al. Continental-scale nutrient and contaminant delivery by Pacific salmon.
All figures and tables in the manuscript can be recreated by running the
R scripts in the code folder. The scripts are named in order (e.g., 1,
2, 3…), so that script 2 will not work without running script 1 first,
and so forth.
The scripts provide the following:
1) Import_data.R: Imports and wrangles contaminant and nutrient data
2) Model contaminants and nutrients.R: Imports data from script 1 and fits hierarchical Bayesian models estimating the concentrations of four nutrients and four contaminants in Pacific Salmon tissues. Prior predictive simulation is also conducted and justified.
3) Model salmon time series.R: Imports and wrangles data for salmon escapement abundance and salmon individual mass. Multiplies mass*abundance for each collection to generate salmon escapement biomass. Fits Bayesian Generalized Additive Models to estimate the escapement of salmon biomass across time, species, and regions.
4a) Prior Predictive Checks.R: Conducts prior predictive checks on the models fit in steps 2 and 3.
4b) Posterior Predictive Checks.R: Conducts posterior predictive checks on the models fit in steps 2 and 3.
5) Stack Posteriors.R: Combines posteriors from the separate models for contaminants and nutrients into a single data file for ease of use later.
6) Model Contaminants x Salmon.R: Multiplies the posteriors for contaminants/nutrients by the posteriors for salmon escapement to generate a distribution of contaminant and nutrient escapement.
7a) Wrangle Plot Data.R: Wrangles the posteriors to make underlying data for the figures in the main text and in the Extended Data.
7b) Make Plots.R: Uses the wrangled data from 7a to make the figures in the main text and in the Extended Data.
8) Sensitivity Analysis.R: Conducts a sensitivity analysis to determine the relative contributions of salmon abundance, salmon individual mass, and contaminant/nutrient concentrations to variation in contaminant/nutrient transport.
9) Summary Tabls.R: Computes summary statistics for the main text, Extended Data, and Supplementary Information
10) Check influence of 1980 Bering Sockeye outlier.R: Compares salmon escapement models that include or don’t include 1980 data on Bering Sockeye to check for influence of large escapement in that year.
11) data release.R: Wrangles data for USGS data release.
12) sanity checks.R: Various checks of the data compared to models. e.g., checks that the modeled escapement overlaps with raw data, and that raw estimates of contaminant/nutrient transport also match with modeled estimates, etc.
13) Biotransport per fish.R: Computes biotransport per individual fish.
14) re-run code 2) with weighted regression.R: Compare contaminant and nutrient models that account for variance/mean relationship with the main models from code 2.
Two additional scripts are included:
functions.R: Contains custom functions
simulate_salmon_data.R. Demonstrates method to randomize the salmon data for this release.
R version 4.4.0 (2024-04-24 ucrt) – “Puppy Cup” Copyright (C) 2024 The R Foundation for Statistical Computing Platform: x86_64-w64-mingw32/x64