model {
# Spatially structured multivariate normal likelihood
# exponential correlation function
height[1:N] ~ spatial.exp(mu[], x[], y[], tau, phi, kappa)
# disc correlation function
# height[1:N] ~ spatial.disc(mu[], x[], y[], tau, alpha)
for(i in 1:N) {
mu[i] <- beta
}
# Priors
beta ~ dflat()
tau ~ dgamma(0.001, 0.001)
sigma2 <- 1/tau
# priors for spatial.exp parameters
# prior range for correlation at min distance (0.2 x 50 ft) is 0.02 to 0.99
phi ~ dunif(0.05, 20)
# prior range for correlation at max distance (8.3 x 50 ft) is 0 to 0.66
kappa ~ dunif(0.05,1.95)
# priors for spatial.disc parameter
# prior range for correlation at min distance (0.2 x 50 ft) is 0.07 to 0.96
# alpha ~ dunif(0.25, 48)
# prior range for correlation at max distance (8.3 x 50 ft) is 0 to 0.63
# Spatial prediction
# Single site prediction
for(j in 1:M) {
height.pred[j] ~ spatial.unipred(beta, x.pred[j], y.pred[j], height[])
}
# Only use joint prediction for small subset of points, due to length of time it takes to run
for(j in 1:10) { mu.pred[j] <- beta }
height.pred.multi[1:10] ~ spatial.pred(mu.pred[], x.pred[1:10], y.pred[1:10], height[])
}
}