[R-sig-Geo] SpatialGridDataFrame to netCDF via ncdf on Windows

[Andy Bunn]

Hello, I have a series of a few dozen SpatialGridDataFrame objects that I would like to write as a single netCDF object - i.e., a 2D field (x,y) as a function of time (t) containing one time-varying variable (z). Given the example below I'd like to create one netCDF object containing t1 and t2. I do not have a netCDF driver in my gdalDrivers() list. Can this be done via ncdf?

Thanks, Andy



# example
require(sp);require(rgdal);require(ncdf)

nlong <- 20
long.size <- 1
long.min <- 0
long <- seq(long.min+long.size/2,long.min+long.size*nlong,by=long.size)

nlat <- 20
lat.size <- 1
lat.min <- 45
lat <- seq(lat.min+lat.size/2,lat.min+lat.size*nlat,by=lat.size)
prj.str <- 'GEOGCS["GCS_WGS_1984",DATUM["D_WGS_1984",SPHEROID["WGS_1984",6378137.0,298.257223563]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]]'

t1 <- rnorm(nlong*nlat)
t1.sp = data.frame(z = t1,
    xc = rep(long,times=nlat),
    yc = rep(lat,each=nlong))
coordinates(t1.sp) = ~xc+yc
proj4string(t1.sp)<- CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs +towgs84=0,0,0")
gridded(t1.sp) = TRUE
t1.sp = as(t1.sp, "SpatialGridDataFrame")

t2 <- rnorm(nlong*nlat)
t2.sp = data.frame(z = t2,
    xc = rep(long,times=nlat),
    yc = rep(lat,each=nlong))
coordinates(t2.sp) = ~xc+yc
proj4string(t2.sp)<- CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs +towgs84=0,0,0")
gridded(t2.sp) = TRUE
t2.sp = as(t2.sp, "SpatialGridDataFrame")

# write netCDF: 2D field (x,y) as a function of time (t) containing one time-varying variable (z)
# ??