This repo contains the necessary data and R codes to reproduce the analysis of the publication titled: 'Applying Bayesian spatiotemporal models to fisheries bycatch in the Canadian Arctic' published in the Canadian Journal of Fisheries and Aquatic Sciences, 2015, 72(2): 186-197, 10.1139/cjfas-2014-0159 available at: http://www.nrcresearchpress.com/doi/abs/10.1139/cjfas-2014-0159#.VRalgEZTTZg
This directory contains two folders Data and Rcodes.
Datagridpred.txt- Grid (4 x 4 km) used in predictionfackdata.txt- Greenland shark bycatch data (counts)etopo1_bedrock.xyz- NOAA ETOPO1 Bathymetry of Baffin Bay from https://www.ngdc.noaa.gov/mgg/global/global.htmlEEZ- Folder containing the shapefile for the Canadian Northwest Atlantic Exclusive Economic Zone (EEZ)
RcodesAllmodelsDICs.R- R codes to run all spatiotemporal model tested (See Table 2 and 3 in paper)Suppl_Material.R- R codes to fit the best model, predict bycatch hotspots, and create most figures included in the paper.
As per data sharing agreement with Fisheries and Ocean Canada, fackdata.txt is not the original dataset used in the paper, but is similar - results will be slightly different.
For further information on R-INLA, please refer to their website: http://www.r-inla.org/ Information on the hurdle and mixture likelihood used in our modeling are available: http://www.math.ntnu.no/inla/r-inla.org/doc/likelihood/zeroinflated.pdf