Description of each field in the psobs_posterior.txt file for 20CRv3
Fields separated by single space
Field 1, len. 19 (pos.1-19)
Datetime & Observation ID
YYYYMMDDHH+{9 digit ISPD unique ID suffix}
Field 2, len. 30 (pos.21:50)
Station name
From source record or station library table
Field 3, len. 12 (pos 53-64)
Station ID
Represents marine call sign or other identifier when present
Field 4, len. 2 (pos 67-68)
Observation type
ps = pressure
Field 5, len. 3 (pos 70-72)
NCEP observation type
Missing: 999
180 Marine Observation Data
181 Station Observation Data
183 Station Observation only reporting sea level pressure
193 Digitized mean sea level pressure bogus
194 Old Weather
195 Weather Detective
3x0 Synoptic (0,6,12,18UTC) Central Pressure from a tropical cyclone best track dataset
3x1-3x5 Synoptic (0,6,12,18UTC) Central Pressure for a category 1-5 tropical cyclone from a tropical cyclone best track dataset
4x0 Non-synoptic Central Pressure from a tropical cyclone best track dataset
5x0 Bogus Central pressure for a tropical depression derived from tropical cyclone best track wind dataset.
5x1-585 Bogus Central Pressure for a category 1-5 tropical cyclone derived from a tropical cyclone best track wind dataset.
Second digits of the NCEP observation type code for tropical cyclones are designated for the regional codes based on the basin/sub-basins classifications by National Climatic Data Center Global Tropical Cyclone Stewardship.
1. Eastern North Pacific (sub-basin: Central Pacific)
2. Eastern North Pacific
3. North Atlantic
4. North Indian
5. South Indian
6. South Pacific (sub-basin: Eastern Australia)
7. South Pacific
8. West Pacific
Field 6, len. 6 (pos 75:80)
Longitude, in deg. E
Min: 0.00, Max: 359.99
Field 7, len. 6 (pos 82-87)
Latitude, in deg. N
Min: -90.00, Max: 90.00
Field 8, len. 7 (pos 89-95)
Observed station pressure or sea level pressure
Field 9, len. 5 (pos 98-102)
Time offset of observation time from analysis time
Field 10, len. 7 (pos 104-110)
Observed station pressure or sea level pressure (same as field 8)
Field 11, len. 1 (pos 112)
Skipped flag
1 – observation was skipped, 0 – observation was assimilated
Field 12, len. 10 (pos 114-123)
Observation bias
For station data, this is the difference between the observation and first guess averaged over the past sixty days. (Approx.) zero for marine observations. The 1851-1870 ship bias correction is not included in this field. (Bias is subtracted from observation for bias-corrected value.)
Field 13, len. 10 (pos 125-134)
Obfit_prior
Observation minus first guess
Field 14, len. 10 (pos 136-145)
Obfit_post
Observation minus analysis
Field 15, len. 10 (pos 147-156)
Obsprd_prior
Variance of first guess ensemble interpolated to observation time and location
Field 16, len. 10 (pos 158-167)
Obsprd_post
Variance analysis ensemble interpolated to observation time and location
Field 17, len. 7 (pos 169-175)
Oberrvar_orig_out
Original observation error variance
Field 18, len. 7 (pos 177-183)
Oberrvar_out
Observation error variance if modified by analysis
Field 19, len. 7 (pos 185-191)
Oberrvaruse
Observation error variance used in analysis (incl. nonlinear QC modification) Note values of 99.9999 are the “new QC”: instead of discarding the obs, we just set the observation error to be very large.
Field 20, len. 8 (pos 193-200)
Paoverpb_save
H(Pa)H^T/H(Pb)H^T; analysis ensemble covariance at the space-time location of the observation, divided by forecast ensemble covariance at the space-time location of the observation. Used in calculation of localization length scale.
Field 21, len 4 (pos 202-205)
Prob. of a gross error
Field 22, len 7 (pos 207-213)
Localization length scale (in km)
Field 23, len 7 (pos 215-221)
Localization length in thousands of km
Field 24, len 1 (pos 223)
QC failure flag
0 if passed quality control check, 1 if failed
v2c posterior observations files
Hello,
Thank you for the clear description of the v3 obs key. I am also using the v2c posterior observations files, I was wondering if you have a similar key please? I can work out what most of the columns are, but would like to know which column is the assimilation flag please? - columns 15-19 all have zeros and ones in them so are potential candidates for this?
Thanks and best wishes,
Julie
Hi Julie, can you be more…
Hi Julie,
can you be more specific which v2c posterior observation files you are using ... are they the ISPD observation records with 20CR feedback, the assimilation text files, or something else? Send me an example file name and where you downloaded it, and we should be able to help you out.
- Tom Cram
NCAR Research Data Archive
Assimilation text format
Hi Julie,
the document 'ISPD_quick_assimilated_ascii_format.pdf', available at https://rda.ucar.edu/datasets/ds131.2/#!docs describes the fields.
- Tom
Hi Tom, Thanks for…
Hi Tom,
Thanks for getting back to me, I think they're the assimilation text files. Philip Brohan kindly downloaded them for me, and put them on a link for me to download. As it looks like this is his private filespace, I'll e-mail you one of the files.
The filenames are prebufrobs_assim_yymmddhr text files.
Best wishes,
Julie