simField: Simulate A continous 2D spatial field evolving over time

In nmmarquez/DRU5MR: Evaluate Point and Polygon Data Simulatneously

Description

Calculates a continous underlying spatial field for the Dominican Republic over time and across ages for probability of mortality for one of 7 age groups. The final observed field is a linear additive model as described in Wakefield 2017.

Usage

simField(M, rangeE = 0.8, sigmaE = 0.04, rhoAR = 0.9,
  sigmaRW = 0.001, ageB = arm::logit((7:1/1000) + 0.00025),
  deltaUrban = -0.1, offset = 1, max.edge = c(0.15, 0.3))

Arguments

M	Number of years
rangeE	spatial range
sigmaE	spatial variance
rhoAR	Yearly autoregression
sigmaRW	Variation of independent temporal component
ageB	age specfifc effects
deltaUrban	urban effect
offset	mesh creation parameter
max.edge	mesh creation parameter

Value

field object, contains multiple items. Spatial points data frame with raster values for transformed linear combination, and other values. A mesh that was used to create the latent field. The latent field itself. A bounding shape where all observations take place. A projection matrix from the mesh to the entire field of interest. The spde for the matern approximation.

Examples

## Not run: 
require(tidyverse)

fieldDR <- simField(4, rangeE = 1.3)
 
fieldDR$fqzDF %>%
    ggplot(aes(x=long, y=lat, fill=fqz)) +
    geom_raster() +
    facet_wrap(~year) +
    coord_equal() +
    theme_void() +
    scale_fill_distiller(palette = "Spectral")

## End(Not run)

nmmarquez/DRU5MR documentation built on Sept. 3, 2019, 5:25 a.m.