Skip to content

2 degree DATM ensemble for CESM experiments (1998 thru 2010)

Jump to bottom
johnsonb-ucar edited this page Jul 7, 2020 · 4 revisions

CAM4 ensemble reanalysis using DART

Publication

A comprehensive description of this dataset has been published as:

Raeder, K., J. L. Anderson, N. Collins, T. J. Hoar, J. E. Kay, P. H. Lauritzen, and R. Pincus, 2012: DART/CAM: An Ensemble Data Assimilation System for CESM Atmospheric Models. J. Climate, 25, 6304-6317 (DOI: 10.1175/JCLI-D-11-00395.1).

Overview

DAReS created a global 80-member ensemble reanalysis using the finite-volume Community Atmosphere Model. This dataset was generated to provide ensembles of atmospheric forcing (from CAM4) to ocean assimilations (using POP) and has also been used to force land model assimilations using CLM4.0.

CAM Version

4.0.1

Levels

26 vertical levels

Resolution

Nominal 2-degree grid

Observations

All observations that were used in the NCEP/NCAR Reanalysis

Output Frequency

6-hourly resolution

Date Range

01-Dec-1997 through 31-Dec-2010

The full dataset is many Tb's of data and is stored on the NCAR/UCAR High Performance Storage System (HPSS). The data atmosphere surface fields can be used to force other components of the Community Earth System Model (CESM) are approximately 1.5 Tb of data and are available via NCAR's Research Data Archive.

The POP_force dataset was generated to provide ensembles of atmospheric forcing (from CAM4) to ocean assimilations (using POP). They have also been used to force land model assimilations (using CLM4.0). There are several data products, which are described in detail below:

  1. State space output from DART
  2. Observation-Space Datasets and Diagnostics
  3. CAM initial files, and CLM, CICE, and DART restart files
  4. "Stream files", which are the data source from which the coupler imported atmospheric forcing for use by other components in CCSM (circa 2011)

Overall Structure of Data Set

All of the data sets are on NCAR's HPSS in hpss:/home/raeder/RAEDER/DAI/POP_force. This location will be shortened to {hpss} for the remainder of the document.

There are several separate assimilations named POPyy, where yy = 12,15,74,79,84,89, or 94. 12 and 15 both refer to multiyear assimilations that jointly cover 1997/12/1 - 2010/12/31. 74,79,84,89,94 refer to years 1974, 1979, etc.

In each POPyy directory there is a series of directories named obs_####, where #### refers to the day within the assimilation series (padded with leading zeros).

  • 0001,...,0365 for assimilations 74,79,84,89,94
  • 0001,...,4779 for assimilations 12,15, as given by the DOY Table (1997-2010) below.

In each obs_#### directory there is a DART diagnostics file named diagnostics.tar.gz and some have CAM+DART restart file sets.

1. State-Space Diagnostics

obs_####/diagnostics.tar.gz has DART output: Prior_Diag.nc; Inflated prior ensemble of state vectors ('copy's 3,...,82) at hours 6Z, 12Z, 18Z, 24Z. Also contains the ensemble mean (copy 1) and spread(2), and inflation(83) and inflation standard deviation(84). The state vector is PS, T, US, VS, Q, CLDLIQ, and CLDICE. Posterior_Diag.nc; Same, but for not-inflated posterior ensemble.

2. Observation-Space Datasets and Diagnostics

Observation-Space Data

obs_####/diagnostics.tar.gz also has DART output obs_seq.final; The observations available for assimilation and the ensemble of estimates of those obs. These files can be plotted using DART's Matlab scripts. The observations consist of T, U, and V from the NCEP BUFR files: radiosondes, ACARS, AIRCRAFT, and satellite drift winds (no surface obs). After 2006 they also used radio occultation observations from COSMIC GPS satellites.

Observation-Space Diagnostics

Time series and vertical profiles of statistics of the ensemble estimates of the observations can be found in {hpss}/POP15/YYYY_Mon_obs_diag.gztar where YYYY = 2000,...,2010 and Mon = Jan,Feb,...,Dec. These were generated by ${DARTROOT}/diagnostics/matlab/plot_rmse_xxx_evolution.m and ${DARTROOT}/diagnostics/matlab/plot_rmse_xxx_profile.m with 'totalspread' and 'bias' passed in the argument lists. The statistics are calculated for the northern and southern extratropics, the tropics, and North America.

3. CESM CAM Restart Datasets

Some obs_#### have 'restart sets' in 80xEM/batchN, EM = number of ensemble members/file, (80 members total) N = 1,2,... Each of these files contains the CAM initial files, and CLM, CICE, and DART restart files for EM ensemble members. The CAM initial files contain the posterior model state at the end of the day represented by ####. These are slightly out of balance due to the perturbations introduced by the last assimilation. An example script, hpss2restart.csh, is available for downloading and unpacking these files. You must have an account on the HPSS for this method.

4. 'Data Atmosphere' Stream Files for CESM Experiments

"Stream files", which are the data source from which the coupler imported atmospheric forcing for use by other components in CCSM (circa 2011). Each file contains 1 member and 1 year and 48 files exist. Up to 80 of these files could be made.

{hpss}/POPyy/Forcing_YYYY/ FV2deg_Cplr_out_single-POP15-EM.cpl.ha2x1d.gztar where YYYY = 2000,...,2010 and EM = 1,...,48.

These contain files: FV2deg_Cplr_out_single-POP15-EM.cpl.ha2x1davg.2000.nc: 365 time slots: daily FV2deg_Cplr_out_single-POP15-EM.cpl.ha2x1dx6h.2000.nc: 1460 time slots: 4/day

Those files were constructed from contents of {hpss}/POP15/cplr_forcing_YYYY_Mon.gztar where YYYY = 2000,...,2010 and Mon = Jan, Feb... which contain files:

6hourly/FV2deg_Cplr_out_single-POP15-EM.cpl.ha2x1dx6h.2000-4.nc daily/FV2deg_Cplr_out_single-POP15-EM.cpl.ha2x1davg.2000-4.nc

where EM = 1,...,80. These are intermediate (monthly) files between what was written by CAM/cpl (6 hourly) and what was needed by other component assimilations (yearly).

The original files can be found in:

{hpss}/POPyy/obs_####/H_cplr.ha2x1d.gz.tar where #### and yy are as above.

DOY Table (1997-2010)

year month/day of first,middle,last obs_seq #### of first,middle,last
1997 12/ 1, 12/16, 12/31 1 - 16 - 31
year month/day of first,middle,last obs_seq #### of first,middle,last
1998

1/ 1, 1/16, 1/31

 32 - 47 - 62
1998

2/ 1, 2/16, 2/28

 63 - 78 - 90
1998

3/ 1, 3/16, 3/31

 91 - 106 - 121
1998

4/ 1, 4/16, 4/30

 122 - 137 - 151
1998

5/ 1, 5/16, 5/31

 152 - 167 - 182
1998

6/ 1, 6/16, 6/30

 183 - 198 - 212
1998

7/ 1, 7/16, 7/31

 213 - 228 - 243
1998

8/ 1, 8/16, 8/31

 244 - 259 - 274
1998

9/ 1, 9/16, 9/30

 275 - 290 - 304
1998 10/ 1, 10/16, 10/31 305 - 320 - 335
1998 11/ 1, 11/16, 11/30 336 - 351 - 365
1998 12/ 1, 12/16, 12/31 366 - 381 - 396
year month/day of first,middle,last obs_seq #### of first,middle,last
1999

1/ 1, 1/16, 1/31

 397 - 412 - 427
1999

2/ 1, 2/16, 2/28

 428 - 443 - 455
1999

3/ 1, 3/16, 3/31

 456 - 471 - 486
1999

4/ 1, 4/16, 4/30

 487 - 502 - 516
1999

5/ 1, 5/16, 5/31

 517 - 532 - 547
1999

6/ 1, 6/16, 6/30

 548 - 563 - 577
1999

7/ 1, 7/16, 7/31

 578 - 593 - 608
1999

8/ 1, 8/16, 8/31

 609 - 624 - 639
1999

9/ 1, 9/16, 9/30

 640 - 655 - 669
1999

10/ 1, 10/16, 10/31

 670 - 685 - 700
1999

11/ 1, 11/16, 11/30

 701 - 716 - 730
1999

12/ 1, 12/16, 12/31

 731 - 746 - 761
year month/day of first,middle,last obs_seq #### of first,middle,last
2000

1/ 1, 1/16, 1/31

 762 - 777 - 792
2000

2/ 1, 2/16, 2/29

 793 - 808 - 821
2000

3/ 1, 3/16, 3/31

 822 - 837 - 852
2000

4/ 1, 4/16, 4/30

 853 - 868 - 882
2000

5/ 1, 5/16, 5/31

 883 - 898 - 913
2000

6/ 1, 6/16, 6/30

 914 - 929 - 943
2000

7/ 1, 7/16, 7/31

 944 - 959 - 974
2000

8/ 1, 8/16, 8/31

 975 - 990 - 1005
2000

9/ 1, 9/16, 9/30

 1006 - 1021 - 1035
2000

10/ 1, 10/16, 10/31

 1036 - 1051 - 1066
2000

11/ 1, 11/16, 11/30

 1067 - 1082 - 1096
2000

12/ 1, 12/16, 12/31

 1097 - 1112 - 1127
year month/day of first,middle,last obs_seq #### of first,middle,last
2001

1/ 1, 1/16, 1/31

 1128 - 1143 - 1158
2001

2/ 1, 2/16, 2/28

 1159 - 1174 - 1186
2001

3/ 1, 3/16, 3/31

 1187 - 1202 - 1217
2001

4/ 1, 4/16, 4/30

 1218 - 1233 - 1247
2001

5/ 1, 5/16, 5/31

 1248 - 1263 - 1278
2001

6/ 1, 6/16, 6/30

 1279 - 1294 - 1308
2001

7/ 1, 7/16, 7/31

 1309 - 1324 - 1339
2001

8/ 1, 8/16, 8/31

 1340 - 1355 - 1370
2001

9/ 1, 9/16, 9/30

 1371 - 1386 - 1400
2001

10/ 1, 10/16, 10/31

 1401 - 1416 - 1431
2001

11/ 1, 11/16, 11/30

 1432 - 1447 - 1461
2001

12/ 1, 12/16, 12/31

 1462 - 1477 - 1492
year month/day of first,middle,last obs_seq #### of first,middle,last
2002

1/ 1, 1/16, 1/31

 1493 - 1508 - 1523
2002

2/ 1, 2/16, 2/28

 1524 - 1539 - 1551
2002

3/ 1, 3/16, 3/31

 1552 - 1567 - 1582
2002

4/ 1, 4/16, 4/30

 1583 - 1598 - 1612
2002

5/ 1, 5/16, 5/31

 1613 - 1628 - 1643
2002

6/ 1, 6/16, 6/30

 1644 - 1659 - 1673
2002

7/ 1, 7/16, 7/31

 1674 - 1689 - 1704
2002

8/ 1, 8/16, 8/31

 1705 - 1720 - 1735
2002

9/ 1, 9/16, 9/30

 1736 - 1751 - 1765
2002

10/ 1, 10/16, 10/31

 1766 - 1781 - 1796
2002

11/ 1, 11/16, 11/30

 1797 - 1812 - 1826
2002

12/ 1, 12/16, 12/31

 1827 - 1842 - 1857
year month/day of first,middle,last obs_seq #### of first,middle,last
2003

1/ 1, 1/16, 1/31

 1858 - 1873 - 1888
2003

2/ 1, 2/16, 2/28

 1889 - 1904 - 1916
2003

3/ 1, 3/16, 3/31

 1917 - 1932 - 1947
2003

4/ 1, 4/16, 4/30

 1948 - 1963 - 1977
2003

5/ 1, 5/16, 5/31

 1978 - 1993 - 2008
2003

6/ 1, 6/16, 6/30

 2009 - 2024 - 2038
2003

7/ 1, 7/16, 7/31

 2039 - 2054 - 2069
2003

8/ 1, 8/16, 8/31

 2070 - 2085 - 2100
2003

9/ 1, 9/16, 9/30

 2101 - 2116 - 2130
2003

10/ 1, 10/16, 10/31

 2131 - 2146 - 2161
2003

11/ 1, 11/16, 11/30

 2162 - 2177 - 2191
2003

12/ 1, 12/16, 12/31

 2192 - 2207 - 2222
year month/day of first,middle,last obs_seq #### of first,middle,last
2004

1/ 1, 1/16, 1/31

 2223 - 2238 - 2253
2004

2/ 1, 2/16, 2/29

 2254 - 2269 - 2282
2004

3/ 1, 3/16, 3/31

 2283 - 2298 - 2313
2004

4/ 1, 4/16, 4/30

 2314 - 2329 - 2343
2004

5/ 1, 5/16, 5/31

 2344 - 2359 - 2374
2004

6/ 1, 6/16, 6/30

 2375 - 2390 - 2404
2004

7/ 1, 7/16, 7/31

 2405 - 2420 - 2435
2004

8/ 1, 8/16, 8/31

 2436 - 2451 - 2466
2004

9/ 1, 9/16, 9/30

 2467 - 2482 - 2496
2004

10/ 1, 10/16, 10/31

 2497 - 2512 - 2527
2004

11/ 1, 11/16, 11/30

 2528 - 2543 - 2557
2004

12/ 1, 12/16, 12/31

 2558 - 2573 - 2588
year month/day of first,middle,last obs_seq #### of first,middle,last
2005

1/ 1, 1/16, 1/31

 2589 - 2604 - 2619
2005

2/ 1, 2/16, 2/28

 2620 - 2635 - 2647
2005

3/ 1, 3/16, 3/31

 2648 - 2663 - 2678
2005

4/ 1, 4/16, 4/30

 2679 - 2694 - 2708
2005

5/ 1, 5/16, 5/31

 2709 - 2724 - 2739
2005

6/ 1, 6/16, 6/30

 2740 - 2755 - 2769
2005

7/ 1, 7/16, 7/31

 2770 - 2785 - 2800
2005

8/ 1, 8/16, 8/31

 2801 - 2816 - 2831
2005

9/ 1, 9/16, 9/30

 2832 - 2847 - 2861
2005

10/ 1, 10/16, 10/31

 2862 - 2877 - 2892
2005

11/ 1, 11/16, 11/30

 2893 - 2908 - 2922
2005

12/ 1, 12/16, 12/31

 2923 - 2938 - 2953
year month/day of first,middle,last obs_seq #### of first,middle,last
Includ e GPS when it becomes available?
2006

1/ 1, 1/16, 1/31

 2954 - 2969 - 2984
2006

2/ 1, 2/16, 2/28

 2985 - 3000 - 3012
2006

3/ 1, 3/16, 3/31

 3013 - 3028 - 3043
2006

4/ 1, 4/16, 4/30

 3044 - 3059 - 3073
2006

5/ 1, 5/16, 5/31

 3074 - 3089 - 3104
2006

6/ 1, 6/16, 6/30

 3105 - 3120 - 3134
2006

7/ 1, 7/16, 7/31

 3135 - 3150 - 3165
2006

8/ 1, 8/16, 8/31

 3166 - 3181 - 3196
2006

9/ 1, 9/16, 9/30

 3197 - 3212 - 3226
2006

10/ 1, 10/16, 10/31

 3227 - 3242 - 3257
2006

11/ 1, 11/16, 11/30

 3258 - 3273 - 3287
2006

12/ 1, 12/16, 12/31

 3288 - 3303 - 3318
year month/day of first,middle,last obs_seq #### of first,middle,last
2007

1/ 1, 1/16, 1/31

 3319 - 3334 - 3349
2007

2/ 1, 2/16, 2/28

 3350 - 3365 - 3377
2007

3/ 1, 3/16, 3/31

 3378 - 3393 - 3408
2007

4/ 1, 4/16, 4/30

 3409 - 3424 - 3438
2007

5/ 1, 5/16, 5/31

 3439 - 3454 - 3469
2007

6/ 1, 6/16, 6/30

 3470 - 3485 - 3499
2007

7/ 1, 7/16, 7/31

 3500 - 3515 - 3530
2007

8/ 1, 8/16, 8/31

 3531 - 3546 - 3561
2007

9/ 1, 9/16, 9/30

 3562 - 3577 - 3591
2007

10/ 1, 10/16, 10/31

 3592 - 3607 - 3622
2007

11/ 1, 11/16, 11/30

 3623 - 3638 - 3652
2007

12/ 1, 12/16, 12/31

 3653 - 3668 - 3683
year month/day of first,middle,last obs_seq #### of first,middle,last
2008

1/ 1, 1/16, 1/31

 3684 - 3699 - 3714
2008

2/ 1, 2/16, 2/29

 3715 - 3730 - 3743
2008

3/ 1, 3/16, 3/31

 3744 - 3759 - 3774
2008

4/ 1, 4/16, 4/30

 3775 - 3790 - 3804
2008

5/ 1, 5/16, 5/31

 3805 - 3820 - 3835
2008

6/ 1, 6/16, 6/30

 3836 - 3851 - 3865
2008

7/ 1, 7/16, 7/31

 3866 - 3881 - 3896
2008

8/ 1, 8/16, 8/31

 3897 - 3912 - 3927
2008

9/ 1, 9/16, 9/30

 3928 - 3943 - 3957
2008

10/ 1, 10/16, 10/31

 3958 - 3973 - 3988
2008

11/ 1, 11/16, 11/30

 3989 - 4004 - 4018 POP15_badSST
2008

11/ 2, 11/16, 11/30

 3990 - 4004 - 4018 good SST
2008

12/ 1, 12/16, 12/31

 4019 - 4034 - 4049
year month/day of first,middle,last obs_seq #### of first,middle,last
2009

1/ 1, 1/16, 1/31

 4050 - 4065 - 4080
2009

2/ 1, 2/16, 2/28

 4081 - 4096 - 4108
2009

3/ 1, 3/16, 3/31

 4109 - 4124 - 4139
2009

4/ 1, 4/16, 4/30

 4140 - 4155 - 4169
2009

5/ 1, 5/16, 5/31

 4170 - 4185 - 4200
2009

6/ 1, 6/16, 6/30

 4201 - 4216 - 4230
2009

7/ 1, 7/16, 7/31

 4231 - 4246 - 4261
2009

8/ 1, 8/16, 8/31

 4262 - 4277 - 4292
2009

9/ 1, 9/16, 9/30

 4293 - 4308 - 4322
2009

10/ 1, 10/16, 10/31

 4323 - 4338 - 4353
2009

11/ 1, 11/16, 11/30

 4354 - 4369 - 4383
2009

12/ 1, 12/16, 12/31

 4384 - 4399 - 4414
year month/day of first,middle,last obs_seq #### of first,middle,last
2010

1/ 1, 1/16, 1/31

 4415 - 4430 - 4445
2010

2/ 1, 2/16, 2/28

 4446 - 4461 - 4473
2010

3/ 1, 3/16, 3/31

 4474 - 4489 - 4504
2010

4/ 1, 4/16, 4/30

 4505 - 4520 - 4534
2010

5/ 1, 5/16, 5/31

 4535 - 4550 - 4565
2010

6/ 1, 6/16, 6/30

 4566 - 4581 - 4595
2010

7/ 1, 7/16, 7/31

 4596 - 4611 - 4626
2010

8/ 1, 8/16, 8/31

 4627 - 4642 - 4657
2010

9/ 1, 9/16, 9/30

 4658 - 4673 - 4687
2010

10/ 1, 10/16, 10/31

 4688 - 4703 - 4718
2010

11/ 1, 11/16, 11/30

 4719 - 4734 - 4748
2010

12/ 1, 12/16, 12/31

 4749 - 4764 - 4779

hpss2restart.csh

===========================================================================
hpss2restart.csh:
===========================================================================
#!/bin/csh

# script for copying 1 day/obs_seq of restart files from the HPSS.
# CAM, CLM, CICE and possibly filter_ic for each ensemble member are grouped together
# so that we can retrieve a subset of the ensemble members for a new experiment.
# Each batch has several ensemble member groups.


setenv LD_LIBRARY_PATH ${LD_LIBRARY_PATH}:/usr/local/dcs/lib
echo $LD_LIBRARY_PATH

# set  echo verbose

if ($#argv < 3) then
   echo "usage; from case/experiment/obs_seq directory; "
   echo "       hpss2restart.csh HSIpathname(no file name) batch1_number batchn_number uncompress(optnl)"
   echo "       untars ensemble members and puts them where they belong."
   echo "       Uncompresses CAM, CLM, and CICE files, if told to."
   echo "       The POP_force restart sets are not compressed."
   exit
endif

set ms_root = $1
set ms_dir = $ms_root

hsi ls -l $ms_root
if ($status == 0) then
   echo "files will be read from $ms_root/batch#"
else
   echo "$ms_root does not exist.  Check name and try again"
   exit
endif

set uncomp = false
if ($#argv == 4) set uncomp = true

set batchn = $3
set batch = $2

while($batch <= $batchn)

   echo "fetching ${ms_dir}/batch${batch}"
   if ($uncomp == true) then
      # msrcp ${ms_dir}/batch${batch}.cmp batch$batch
      hsi get batch$batch : ${ms_dir}/batch${batch}.cmp
   else
      # msrcp ${ms_dir}/batch$batch .
      hsi get batch$batch : ${ms_dir}/batch$batch
   endif

#  This makes CAM CLM CICE DART directories, as necessary, to place its files.
   echo "untarring"
   tar -x -f batch$batch
   rm batch$batch
   @ batch++
end

if ($uncomp == true) then
   gunzip -r CAM
   gunzip -r CLM
endif

echo "hsi get done; check local files "

exit
Clone this wiki locally