-
Notifications
You must be signed in to change notification settings - Fork 17
Glossary
The following terms are used in the documentation (on GitHub and/or the R package help files):
Knots: The location of samples are provided in eastings-northings (or Latitude-Longitude), and all occur with a fixed sampling domain (area for the survey). Spatial variation in density is approximated as being piecewise constant by estimating spatial and spatiotemporal variation at a series of "knots". Each location in the sampling domain then has the value of the nearest "knot".
Polygons: The total spatial area in a sampling domain has piecewise constant density. The area that is closest to knot x (and which has constant density equal to the density at that knot) is refered to as the k-th "polygon". Each polygon is technically a voronoi tesselation of the surveys sampling domain.
Extrapolation grid: The analyst typically provides an "extrapolation grid" corresponding to the area and average value of covariates on sampling grid that has a fine resolution. The area for each "polygon" is then equal to the total area of all "extrapolation grid cells" that are closest to the knot corresponding to that polygon. Similarly, each polygon has as covariates the average of that covariate for each extrapolation grid cell corresponding to that polygon.
Covariates: There are two types of covariates used in the spatial delta-GLMM: spatial covariates, and catchability covariates. Catchability covariates are included to "control for" some confounding variable, while spatial covariates are used to improve prediction when "integrating across" spatial variation in density.
Spatial covariates: Spatial covariates are characteristics of each polygon, and are used as predictor variables when explaining spatial variation in density. Examples might include: bottom depth; bottom substrate; biogenic habitat.
Catchability covariates: Catchability covariates are characteristics of each sampling occasion. They are used as predictors of catch rates beyond what is explained by spatial variation in density. Examples might include: ocean condition; date within a season; experience of fishing operators.