Plotting and/or exporting VIIRS SDR DNB data as GeoTIFF in R?

Question

I already asked this question @stackoverflow however, perhaps it fits better here, due to the spatial nature of VIIRS data and the hdf5-eos file format.

Im looking into how to handle .h5 (hdf-eos) VIIRS DNB raster data in R (or any other opensource software). I know that GeoTIFF products are available, for monthly composites, however I need daily data, that is why I need to work with the .h5.

My data is ordered and downloaded through https://www.class.ngdc.noaa.gov/saa/products/welcome

Preferably I want to be able to plot and create time series of the data in R, but if that is not possible, then I want to export it as GeoTIFF to handle it in ArcGIS or QGIS (which is maybe also necessary to work with it in R?).

So far I am able to read in the file with

hdf<- h5file("mypath/GDNBO-SVDNB_npp_d20171101_t0110370_e0116174_b31151_c20180208224859630066_nobc_ops.h5", mode = "r")

But I cannot figure out how to access the "radiance" band, let alone plot it. I have read the "h5" package documentation without any progress.

As another option I have tried to export the h5 file to geotiff with

sds <- get_subdatasets("subset_1_d20171101_t0110370.h5")
gdal_translate(sds[17], dst_dataset = "radiance.tif", overwrite=T) #subset 17 is the radiance band i need

however the file has lost its geolocation during the transformation, when the geotiff is opened in e.g. QGIS, it does not have any location or projection.

I understand that the geolocation is stored in separate bands (latitude, longitude) however I don't understand how I can access meta data on the corner coordinates and the projection, and then pass this on to gdal_translate() or a software like ArcGIS?

Anyone knows how to deal with these type of files in R?

I seek an example in R how to plot or export this data file.

Answer 1

@dtanon, did you find any solution? I'm also interested. You can use polargrid or satpy to reproject and convert to geotiff your data. Did you use gdal_translate? You can try something like this but I don't know if it is the best solution:

gdal_translate -of VRT  HDF5:<yourhdffile>://<pathofdataset> <output.vrt>

and then to reproject to your AOI:

gdalwarp -te <xmin>, <ymin>, <xmax>, <ymax> -tr <xres> <yres> -r <resample method> -s_srs <sourceEPSG> -t_srs EPSG:<destinationEPSG> <output.vrt> <outputf.tif>