[R-sig-Geo] Selecting variable within nc file in raster() or brick()

[Michael Sumner]

There is a hidden argument stopIfNotEqualSpaced to use for ignoring
georeferencing for advanced use. Netcdf does not always translate into a
regular raster cleanly, if at all.

I will have a look if I get a chance.

M

On Sat, 10 Jan 2015 00:09 Agustin Lobo <alobolistas at gmail.com> wrote:

> hmm.. strange, I do not get the same behaviour for this file:
>
> nom5 <- "_NRTSM003D025A_ALL_2013.nc"
> info5 <- open.ncdf(file.path(ncdir,nom5))
> print(info5)
> [1] "file /media/alobo/FREECOM
> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_NRTSM003D025A_ALL_2013.nc has 4
> dimensions:"
> [1] "time   Size: 365"
> [1] "lat   Size: 44"
> [1] "lon   Size: 62"
> [1] "time_strlen   Size: 20"
> [1] "------------------------"
> [1] "file /media/alobo/FREECOM
> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_NRTSM003D025A_ALL_2013.nc has 4
> variables:"
> [1] "int L2_Points[lon,lat,time]  Longname:Number of Measures Missval:0"
> [1] "float SM[lon,lat,time]  Longname:Surface soil moisture Missval:-999"
> [1] "float SM_DQX[lon,lat,time]  Longname:Surface soil moisture_dqx
> Missval:-999"
> [1] "float VARIANCE_SM[lon,lat,time]  Longname:Variance of Surface
> soil moisture Missval:-999"
>
> nc5 <- brick(file.path(ncdir,nom5), varname="SM")
> Error in .rasterObjectFromCDF(x, type = objecttype, band = band, ...) :
>   cells are not equally spaced; you should extract values as points
>
> I doubt the message is true (those files are satellite images),
> although cannot tell for sure because I've not made the nc file
> myself.
> Also, the following works and does not complain about the spacing:
>
> nc5 <- get.var.ncdf(info5, "SM", verbose = F)
> r5 <- aperm(nc5[,ncol(nc5):1,],c(2,1,3))
> r5 <- brick(r5)
> plot(r5[[200]])
>
> I can use this as a workaround, but it would be cleaner using brick()
> directly.
>
> I've put the file here in case somebody would like to check:
> https://dl.dropboxusercontent.com/u/3180464/_NRTSM003D025A_ALL_2013.nc
>
> Agus
>
> On Fri, Jan 9, 2015 at 12:48 PM, Michael Sumner <mdsumner at gmail.com>
> wrote:
> > Use the varname argument to choose a specific variable. Try
> >
> > brick(file.path(ncdir,nom5))
> >
> > it will tell you what to do. The level argument provides a similar
> facility
> > for slicing from a 4d var, btw.
> >
> > Cheers, Mike
> >
> > On Fri, 9 Jan 2015 22:37 Agustin Lobo <alobolistas at gmail.com> wrote:
> >
> >> Hi!
> >>
> >> For nc (v3) files having only 1 field variable along time such as
> >> open.ncdf(file.path(ncdir,nom1))
> >> [1] "file /media/alobo/FREECOM
> >> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_DELSMH001DIBEA.nc has 4
> >> dimensions:"
> >> [1] "time   Size: 257"
> >> [1] "lat   Size: 1100"
> >> [1] "lon   Size: 1600"
> >> [1] "time_strlen   Size: 20"
> >> [1] "------------------------"
> >> [1] "file /media/alobo/FREECOM
> >> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_DELSMH001DIBEA.nc has 1 variables:"
> >> [1] "float HighRes_Soil_Moisture[lon,lat,time]  Longname:1 km Soil
> >> Moisture Missval:-999"
> >>
> >> I can simply use
> >> nc1 <- brick(file.path(ncdir,nom1))
> >>
> >> But, what in the case of having several variables in the nc file, i.e.:
> >> open.ncdf(file.path(ncdir,nom5))
> >> [1] "file /media/alobo/FREECOM
> >> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_NRTSM003D025A_ALL_2013.nc has 4
> >> dimensions:"
> >> [1] "time   Size: 365"
> >> [1] "lat   Size: 44"
> >> [1] "lon   Size: 62"
> >> [1] "time_strlen   Size: 20"
> >> [1] "------------------------"
> >> [1] "file /media/alobo/FREECOM
> >> HDD/BIGFILES/IMASCOARSESTOCK/SMOS/_NRTSM003D025A_ALL_2013.nc has 4
> >> variables:"
> >> [1] "int L2_Points[lon,lat,time]  Longname:Number of Measures Missval:0"
> >> [1] "float SM[lon,lat,time]  Longname:Surface soil moisture
> Missval:-999"
> >> [1] "float SM_DQX[lon,lat,time]  Longname:Surface soil moisture_dqx
> >> Missval:-999"
> >> [1] "float VARIANCE_SM[lon,lat,time]  Longname:Variance of Surface
> >> soil moisture Missval:-999"
> >>
> >> Is there a way to select the desired variable within brick() ?
> >> Or should I use  get.var.ncdf() and then transform to raster?
> >> (the only problem in this case is that I have to rotate etc, while
> >> brick() does it all in the first case)
> >>
> >> Thanks
> >>
> >> Agus
> >>
> >> _______________________________________________
> >> R-sig-Geo mailing list
> >> R-sig-Geo at r-project.org
> >> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> >>
> >
> >         [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > R-sig-Geo mailing list
> > R-sig-Geo at r-project.org
> > https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

	[[alternative HTML version deleted]]