glmSTARMA (version 1.0.0)

sst: Sea Surface Temperature Anomalies in the Pacific

Description

Multivariate time series containing monthly sea surface temperature anomalies in the Pacific

Format

SST

A matrix with monthly sea surface temperature anomalies (rows = spatial locations, columns = time points).

W_directed

A list of (sparse) matrices containing spatial weight matrices. See details

1. Identity matrix.

2. north: weight matrix for northward direction

3. east: weight matrix for eastward direction

4. south: weight matrix for southward direction

5. west: weight matrix for westward direction

locations

data.frame containing the latitude and longitude of each spatial location.

Details

This dataset contains monthly sea surface temperature anomalies (in degree Celsius) at 25 spatial locations in the Pacific Ocean over a time period of 33 years (396 months) from January 1970 to December 2002.

The neighborhood matrices in the list W_directed for orders 2, 3, 4, and 5 correspond to neighbors with the location directly north, east, south, and west. Coefficients estimated using these matrices then reflect directed dependencies. Not all neighbors always exist at boundary locations of the observed area. In these cases, the corresponding rows of the weight matrices contain only zeros. These matrices are stored as objects of class 'dgCMatrix' from the 'Matrix' R package.

The dataset is not included directly in the package. Use load_data("sst") to download it.

References

• Wikle, C.K., Zammit-Mangion, A., and Cressie, N. (2019). Spatio-Temporal Statistics with R. Chapman & Hall/CRC, Boca Raton, FL.

• Cressie, N, and Wikle, C.K. (2011). Statistics for Spatio-Temporal Data. John Wiley & Sons, Incorporated.

Examples

# \donttest{
# Note: Complete examples take around 4 minutes to run #
# Requires the 'Matrix' package
if(requireNamespace("Matrix")){
 dat <- load_data("sst", directory = tempdir())
 SST <- dat$SST
 W_directed <- dat$W_directed
 locations <- dat$locations

 times <- seq(from = as.Date("1970-01-01"), to = as.Date("2002-12-01"), by = "m")
 times <- format(times, "%b %Y")
 covariates <- list(trend = SpatialConstant(seq(times) / length(times)),
                    longitude = TimeConstant(locations$lon / 360),
                    season_cos = SpatialConstant(cos(2 * pi / 12 * seq(times))),
                    season_sin = SpatialConstant(sin(2 * pi / 12 * seq(times))),
                    abs_lat_inc = TimeConstant(pmin(abs(locations$lat), 6) / 90),
                    abs_lat_dec = TimeConstant(pmax(abs(locations$lat) - 6, 0) / 90))

 fit <- glmstarma(SST, model = list(past_obs = 4, past_mean = 4), 
                  wlist = W_directed, wlist_past_mean = W_directed,
                  covariates = covariates, family = vnormal())

 fit2 <- dglmstarma(SST, mean_model = list(past_obs = 4, past_mean = 4), 
                   dispersion_model = list(past_obs = 4),
                   wlist = W_directed, mean_covariates = covariates, 
                   dispersion_covariates = covariates, mean_family = vnormal())
}
# }