projected data ingest

[EarthScientist]

we need an ingest strategy for the projected data John is interested in.

From Email with JWalsh: For the "future", the monthly output of the NMME (North American Multi-Model Ensemble), plots are accessible at: http://www.cpc.ncep.noaa.gov/products/NMME/monanom.shtml

while the data access link (grb files) for the NMME is

http://www.cpc.ncep.noaa.gov/products/NMME/data.html

Of the various models in the NMME ensemble, the one we would be most interested in is CFSv2

[EarthScientist]

This FTP site contains some (all?) of the NMME ensemble model prediction data John was mentioning in that email referenced above.

These are GRIidded Binay (GRIB) files and I am working with some ways to extract some metadata from them so that I can properly work with the data on the GDAL commandline.

There is a package called pygrib that gives some methods for reading and writing these files if we want to build some automation in that direction.

[EarthScientist]

CFS_V2 ensemble models CFS_V2 realtime_anom

[EarthScientist]

NMME ENSEMBLE RECIPE

download test file:

wget http://ftp.cpc.ncep.noaa.gov/NMME/EMs/prate.010100.CFSv2.ensmean.anom.1x1.grb

the work:

# regrid to greenwich-centered WGS84
gdalwarp -nomd -overwrite -multi -r bilinear -t_srs "EPSG:4326" -tr 1 1 -te -180 -90 180 90 -wo "SOURCE_EXTRA=10" prate.010100.CFSv2.ensmean.anom.1x1.grb prate_010100_CFSv2_ensmean_anom_1x1_greenwich.tif

# warp to 3572
gdalwarp -nomd -overwrite -multi -r bilinear -t_srs "EPSG:3572" -wo "SOURCE_EXTRA=10" prate_010100_CFSv2_ensmean_anom_1x1_greenwich.tif prate_010100_CFSv2_ensmean_anom_1x1_3572.tif

# crop it to the 40N
gdalwarp -overwrite -cutline "40N_epsg3572.shp" -crop_to_cutline -dstnodata -9999 prate_010100_CFSv2_ensmean_anom_1x1_3572.tif prate_010100_CFSv2_ensmean_anom_1x1_3572_40N.tif

[brucecrevensten]

I sent a note to the people referenced for questions on this data:

Hi,

For a project I'm working on at the University of Alaska, Fairbanks, I'm using data from here:

http://ftp.cpc.ncep.noaa.gov/NMME/realtime_anom/CFSv2/2016010800

An example of a specific file I'm using is this:

tmpsfc.2016010100.01.CFSv2.anom.avrg.1x1.grb

I have two questions about this data: What do the eight bands available within each .grb file represent? In the filename, what is the meaning of the 2-digit number XX in this location in the file name: tmpsfc.2016010100.XX.CFSv2.anom.avrg.1x1.grb ? Is there a metadata or other explanation page I missed which would outline this? If so, please send me a link.

[brucecrevensten]

Response from people responsible for this data:

Each grib file is one ensemble member, indicated by the 2-digit number in the filename. For CFSv2, there are 32 members in the real-time forecast, so you'll find files numbered 01-32 for each variable. The directory indicates the initial date, as does the filename (January 2016 in your example file.)

There are eight forecast leads in each file, starting with Lead 1, denoted in the metadata as "1-2mon ave". Lead 1 is the forecast for the month subsequent to the initial month, e.g. the Lead 1 forecast from January is for February. The metadata can be read using wgrib.

We're using the first ensemble & Lead 1 for the next-month anomalies at this time, can update later if needed.