Skip to content
Marine Socio-Environmental Covariates
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
distmarket
humanpop
npp
reeflandarea
wave
.gitignore
README.md
msec.Rproj
utils.R

README.md

Marine Socio-Environmental Covariates

This repository contains the R source code to create the following summary statistics of environmental and anthropogenic variables on a common 2.5 arc-minute grid, for use in marine science research:

  • Net Primary Productivity (with correction applied to shallow locations)
  • Reef Area (within 15km and 200km) and Land Area (within 15km and 50km)
  • Wave Energy
  • Human Population within 20km and 55km for 7 years between 1990 and 2020
  • Distance to Market

The details for each variable's calculation can be found in the README.md file of the corresponding subfolder.

Citation

These data products were published in the following study:

  • Yeager, L.A., Marchand, P., Gill, D.A., Baum, J.K., and McPherson, J.M. (2017) Queryable global layers of environmental and anthropogenic variables for marine ecosystem studies. Ecology. In Press. doi: 10.1002/ecy.1884.

Please cite this article for any work that re-uses this code.

Web Application

A web application to extract these variables for specific points of interest can be found at: http://shiny.sesync.org/apps/msec.

Input Data Sources

The data products were created from the following publicly available data sources:

R Packages Required

  • The following tidyverse packages: dplyr, tidyr, stringr and lubridate;
  • the sp, rgdal, rgeos and raster packages to manipulate spatial objects;
  • the geosphere package to calculate geodesic distances; and
  • the waver package to calculate fetch and wave energy.

Acknowledgement

This work was supported by the National Socio-Environmental Synthesis Center (SESYNC) under funding received from the National Science Foundation DBI-1052875.

Miscellaneous notes

  • The utils.R script in the main folder contains a few functions to manipulate spatial objects, e.g. creating a rectangular buffer of minimum distance, converting vector and raster layers from a (-180, 180) to a (0, 360) longitude range. These functions are re-used in multiple parts of the project.

  • The land mask raster (land_final.grd), which serves as a grid template for each final layer, should be generated first via the script found in the reeflandarea subfolder.

  • As indicated in the various scripts, the most computationally-intensive steps were performed in parallel on a HPC cluster. More specifically, these calculations were run on a 20-node SLURM cluster at SESYNC, each node having 8 cores and 60 Gb of RAM.

  • All processing steps from the original data sources to the final products were performed in R, with the exception of the distance buffers around the Reefs at Risk coral reef map, which were computed in ArcGIS. This is due to the absence of efficient R functions to compute distance buffers in geographic (unprojected) coordinates.

  • While all raster calculations use the native .grd format from the R raster package, the final products were saved as NetCDF files with a compression level of 5, using the ncdf4 package.

You can’t perform that action at this time.