Extraction of data from HDF at different pressure level - r

Question

I am trying to extract a variable named Data Fields/OzoneTropColumn at a point location (lon=40, lat=34) at different pressure level (825.40198, 681.29102, 464.16000, 316.22699 hPa) from multiple hdf files
library(raster)
library(ncdf4)
library(RNetCDF)
# read file
nc <- nc_open("E:/Ozone/test1.nc")
list_col1 <- as.list(list.files("E:/Ozone/", pattern = "*.hdf",
full.names = TRUE))
> attributes(nc$var) #using a single hdf file to check its variables
$names
[1] "Data Fields/Latitude" "Data Fields/Longitude"
[3] "Data Fields/O3" "Data Fields/O3DataCount"
[5] "Data Fields/O3Maximum" "Data Fields/O3Minimum"
[7] "Data Fields/O3StdDeviation" "Data Fields/OzoneTropColumn"
[9] "Data Fields/Pressure" "Data Fields/TotColDensDataCount"
[11] "Data Fields/TotColDensMaximum" "Data Fields/TotColDensMinimum"
[13] "Data Fields/TotColDensStdDeviation" "Data Fields/TotalColumnDensity"
[15] "HDFEOS INFORMATION/StructMetadata.0" "HDFEOS INFORMATION/coremetadata"
> pres <- ncvar_get(nc, "Data Fields/Pressure") #looking at pressure level from single file of hdf
> pres
[1] 825.40198 681.29102 464.16000 316.22699 215.44400 146.77901 100.00000 68.12950 46.41580 31.62290
[11] 21.54430 14.67800 10.00000 6.81291 4.64160
ncin <- raster::stack(list_col1,
varname = "Data Fields/OzoneTropColumn",
ncdf=TRUE)
#cannot extract using the following code
o3 <- ncvar_get(list_col1,attributes(list_col1$var)$names[9])
"Error in ncvar_get(list_col1, attributes(list_col1$var)$names[9]) :
first argument (nc) is not of class ncdf4!"
#tried to extract pressure levels
> prsr <- raster::stack(list_col1,varname = "Data Fields/Pressure",ncdf=TRUE)
"Error in h(simpleError(msg, call)) :
error in evaluating the argument 'x' in selecting a method for function 'stack': varname: Data Fields/Pressure does not exist in the file. Select one from:
Data Fields/O3, Data Fields/O3DataCount, Data Fields/O3Maximum, Data Fields/O3Minimum, Data Fields/O3StdDeviation, Data Fields/OzoneTropColumn, Data Fields/TotColDensDataCount, Data Fields/TotColDensMaximum, Data Fields/TotColDensMinimum, Data Fields/TotColDensStdDeviation, Data Fields/TotalColumnDensity"
#tried using index
#Point location can also be written as below 1 deg by 1 deg resolution
lonIdx <- which(lon >32 & lon <36)
latIdx <- which(lat >38 & lat <42)
presIdx <- which(pres >= 400 & pres <= 900)
#also tried
# Option 2 -- subset using array indexing
o3 <- ncvar_get(list_col1,'Data Fields/OzoneTropColumn')
"Error in ncvar_get(list_col1, "Data Fields/OzoneTropColumn") :
first argument (nc) is not of class ncdf4!"
extract2 <- o3[lonIdx, latIdx, presIdx, ]
How to I extract these values vertically at each pressure level ? (SM=Some value)
I would like the output in following way at location (lon=40, lat=34):
Pressure 1 2 3 4 5 .... 10
825.40198 SM1 SM2 SM3 SM4 SM5... SM10
681.29102 SM11 SM12
464.16000
316.22699 SM.. SM.. SM.. SM.. SM.. SM..
Appreciate any help.
Thank you

Answer 1

This might be an issue with how netcdf4 and raster name each of the layers in the file. And perhaps some confusion with trying to create a multilayer object from multiple ncdf at once.
I would do the following, using only raster: Load a single netCDF, using stack() or brick(). This will load the file as a multilayer object in R. Use names() to identify what is the name of the Ozone layer according to the raster package.
firstraster <- stack("E:/Ozone/test1.nc")
names(firstraster)
Once you find out the name, you can just execute a reading of all objects as stack(), extract the information on points of interest, without even assembling all layers in a single stack.
Ozonelayername <- "put name here"
files <- list.files("E:/Ozone/", pattern = "*.hdf", full.names = TRUE)
stacklist <- lapply(files, stack)
Ozonelayerlist <- lapply(stacklist, "[[", "Ozonelayername")
The line above will output a list of rasters objects (not stacks or bricks, just plain rasters), with only the layer you want.
Now we just need to execute an extract on each of these layers. sapply() will format that neatly in a matrix for us.
pointsofinterest <- expand.grid(32:36,38:42)
values <- sapply(Ozonelayerlist, extract, pointsofinterest)
I can test it, since I do not have the data, but I assume this would work.