Rich Neale’s Blocking Diagnostic Documentation

Evaluate blocking frequency by season as determined by the meridional gradient above a threshold value of daily 500mb height following D’Andrea et al (1998) and a 2D variant of the above where lat and lon daily 500-mb thresholds are considered Davini et al

Version & Contact info

Version 2.0, implemented in MDTF v3.0beta3
PI: Rich Neale, NCAR, rneale@ucar.edu
Contact: Dani Coleman, NCAR, bundy@ucar.edu

Functionality

Required programming language and libraries

  • Python 3

  • NCL (usual NCAR/CESM packages: contributed, gsn_code, gsn_csm)

Required model output variables

Geopotential height at 500 milibars or 3D field with air pressure available

Settings/Environment Variables

There are two sets of environment variables that can be set in the input jsonc file:

  1. What to compare input case(s) against:

    • observational products from ERAI & MERRA. Can be turned on or off together. Default: On

    • three CESM 5-member ensembles (CAM3, CAM4, CAM5) which can be turned on or off individually. Default: CAM5 only. (Note that the code runs with these but the html file does not yet respond to this. Therefore the POD will fail to link the figures if non-default options are used)

    Variable name

    Default

    MDTF_BLOCKING_OBS

    True

    MDTF_BLOCKING_CAM3

    False

    MDTF_BLOCKING_CAM4

    False

    MDTF_BLOCKING_CAM5

    True

  2. Variables to control reading or writing of digested data

    • MDTF_BLOCKING_READ_DIGESTED (Default True) If True, the POD looks for digested data for the obs & CAM ensembles (MDTF_BLOCKING_READ_DIGESTED = False). Digested means the data has been processed by this script into the block_time variable name and format.

      The POD is not currently capable of running with a mix of digested and undigested (other than the undigested MDTF input case). If the user desires this, it is recommended to run only the components that are not digested, and write out digested. Then re-running with MDTF_BLOCKING_READ_DIGESTED will work.

      It is not yet possible to run the MDTF input case with digested data, although it is possible to write it out

      OBS/CAM files are expected to be delineated with the file name & dir name as follows.
      Digested: obs_data/blocking/ERAI/ERAI.z500.day. digested .nc
      Undigested: obs_data/blocking/undigested/ERAI/ERAI.z500.day.nc

    • MDTF_BLOCKING_WRITE_DIGESTED (Default False) If True, whatever data was read in raw will be written out digested. This is how a user can make a new digested file, whether obs, a new case or ensemble, or a previous MDTF input case to compare against.

Code Overview

The POD is called by a python wrapper script, blocking_neale.py which calls blocking.ncl.

blocking.ncl summary:

  1. Read all case data from environmental variables and store it in all_cases structure

    all_cases = blocking_get_all_case_info() is a data structure defined in blocking_funcs.ncl which contains arrays of length number of cases (mdtf case + obs cases + comparison models) and contains the following components:

    • case_names (short names; for ensembles these are the group name repeated length ncase)

    • file_names (paths to files for each case (distinct for each ensemble member))

    • var_names (what the variables are named in the datasets)

    • years_run0, years_run1 (start and stop years for each case)

    • plot_colors hard-coded for consistency from run to run

  2. Sets up figures that need to be done before each season is processed To be done

  3. Loops over seasons (tested on ANN. DJF & JJA not tested)

    • Loops over files. File settings obtained by file_opts = extract_file_opts(all_cases,ifile)

      1. If MDTF_BLOCKING_READ_DIGESTED = True

        • Calls blocking_digested.ncl:open_digested() to open file_opts@file_name (set in blocking_funcs.ncl:set_and_check_file_names()) reads block_time = f_dig->$var_name$(idays,:)

        • else reads raw Z500 data using var_in_p = blocking_readfile_orig and computes the blocking index as block_time(lon) following Tibaldi and Molteni (1990)

      2. If MDTF_BLOCKING_WRITE_DIGESTED Optionally writes block_time out as digested data

      3. computes block_days as sum of all days that were blocked over the entire time period, as a function of longitude (as well as std for ensembles) block_std and block_freq to be blocked/all days.

      4. for season ANN (annual) only, calculate daily frequency block_freq_yr(ndoys,nlons) for Hovmüllers, smoothed by smth9_Wrap

    • Loops over ensemble groups for figures

      1. Store data in arrays block_freq_aves_p, block_freq_min_p/max_p

      2. Set more plotting resources res_m

      3. Individual (Hovmueller figures) plot_t(ifile) = gsn_csm_hov(wks_ens_page, lonPivot(block_freq_yr,270.), res_t)

      4. Combined seasonal figure (one figure with a line for each dataset) plot(ip) = gsn_csm_xy (wks,block_freq_aves_p&lon,block_freq_aves_p,res_m)

Figures

../_images/figure_freq_season.png
../_images/figure_anncycle.OBS.png
../_images/figure_anncycle.CAM5.png

References

1. Tibaldi and Molteni (1990): On the operational predictability of blocking. Tellus A: Dynamic Meteorology and Oceanography, 42 (3), 343-365, doi:10.3402/tellusa.v42i3.11882.

2. D’Andrea, F., Tibaldi, S., Blackburn, M. et al. (1998): Northern Hemisphere atmospheric blocking as simulated by 15 atmospheric general circulation models in the period 1979–1988. Climate Dynamics, 14, 385–407 doi:10.1007/s003820050230.

3. Davini, P., Cagnazzo, C., Neale, R., and Tribbia, J. (2012): Coupling between Greenland blocking and the North Atlantic Oscillation pattern, Geophys. Res. Lett., 39, L14701, doi:10.1029/2012GL052315.