gridConstruct {lgc}R Documentation

Construct a grid based on observation locations.

Description

gridConstruct constructs a grid.

gridFilter filters off unwanted grid points.

gridLocate finds the closest grid point for a set of locations.

gridFactor constructs a gridFactor object.

Usage

  gridConstruct(data,
    type = c("squareGrid", "triangularGrid", "scatterGrid"),
    filter = TRUE, wet = !wetEdges, wetEdges = FALSE,
    connected = TRUE, ordertol = 1, ...)

  gridFilter(grid, data, icesSquare = FALSE,
    nearestObs = Inf, wet = FALSE, connected = FALSE,
    ordertol = 0, wetEdges = FALSE, ...)

  gridLocate(grid, points)

  gridFactor(data, grid, ...)

Arguments

data

data.frame.

type

Neighborhood structure type.

filter

Call gridFilter after grid construction?

...
center

Optional list to control origo of grid.

km

Optional distance between neighboring grid points in km.

icesSquare

Remove grid points outside ICES squares in the data?

nearestObs

Remove grid points with closest data point greater than nearestObs.

wet

Remove grid points on land.

wetEdges

Alternative: Keep edges passing through water.

ordertol

Require at least ordertol neighbors to every grid point.

grid
points

Details

These functions help to construct the object required to build Gaussian Markov Random fields with the formula interface. Most situations can be handled by

  1. Creating a grid using grid <- gridConstruct(data).

  2. Building a gridFactor object using gridFactor(data,grid).

Construction of grids in practice generally involves 3 steps:

  1. gridConstruct - Construct the grid sufficiently fine and sufficiently large to contain all data points.

  2. gridFilter - Filter off unwanted grid points. For instance grid points on land or grid points too far away from the region of interest.

  3. gridLocate - For each data point locate the nearest grid point.

gridLocate performs a brute force search of closest gridpoint to each data point. The index of the closest grid point is returned.

Value

grid object

Filtered grid object

Examples

## Construct grid of North Sea and lookup points in the grid.

## Data
df <- data.frame(lon=c(0 ,1 ,5 ,6 ,2 ,2, 5, 1),
                 lat=c(56,60,55,57,54,58,55,60))

## Construct grid
gr <- gridConstruct(df,km=30,scale=1.3,filter=!FALSE)
plot(gr);points(df);map("worldHires",add=TRUE)

## grid factor
gf <- gridFactor(df,gr)
points(gr[gf,],col="red")

## Max distance to nearest grid point
max(dist.km(df,gr[gf,],outer=FALSE))

## 2.
gr <- gridConstruct(df,km=30,scale=1.3,nearestObs=100,connected=FALSE)

[Package lgc version 1.4 Index]