NetCDF Variable Extraction – How to Extract at Specific Time and Depth in R

netcdfrraster

I have the following netcdf file containing 7 variables and 4 dimensions. 

[1] "file data.nc has 4 dimensions:"
[1] "depth   Size: 12"
[1] "time   Size: 37"
[1] "x   Size: 305"
[1] "y   Size: 432"
[1] "------------------------"
[1] "file data.nc has 7 variables:"
[1] "int stereographic[]  Longname:stereographic Missval:NA"
[1] "float longitude[x,y]  Longname:longitude Missval:1e+30"
[1] "float latitude[x,y]  Longname:latitude Missval:1e+30"
[1] "short chla[x,y,depth,time]  Longname:chla Missval:-32767"
[1] "short nitrat[x,y,depth,time]  Longname:nitrat Missval:-32767"
[1] "short phosphat[x,y,depth,time]  Longname:phosphat Missval:-32767"
[1] "short pbiomass[x,y,depth,time]  Longname:pbiomass Missval:-32767"

I would like to extract several months at a specific depth and convert them to a ASCII raster. I am able to use this command myraster<-raster("data.nc", varname="nitrat", lvar=1, band=2, level=20) but I am confused about which depth and time I am actually extracting, it is unclear to me how to use the parameters of the raster function to point to the specific time and depth.
Also I am not able to read a variable using the function get_att and I get this message

test<-open.ncdf("data.nc")
Error in ncatt_get(test, "time") : 
  Error, first passed argument must be an object of class ncdf4

I would like to be able to get a list of the different time and depth so I can use the correct parameters in the raster function.I have both "ncdf","ncdf4" packages installed.I am working in Windows with R 3.0

Best Answer

It might be easier to see what is going on when you create a RasterBrick (multiple layers), either representing depth, or time. The number of layers should make it clear which is is. the z-values too.

You could start with

b <- brick("data.nc", varname="nitrat")

and go from there.

The error in hour ncdf code is because you mix up the ncdf package (open.ncdf) with the ncdf4 package (ncatt_get). You need to pick one and stick with it.

Related Question