ERA52ARL Converter

Post questions and find resources to convert meteorological data into a format HYSPLIT can read.
Post Reply
rssv
Posts: 1
Joined: September 29th, 2020, 12:36 pm
Registered HYSPLIT User: Yes

ERA52ARL Converter

Post by rssv »

I am working on converting ECMWF ERA5 meteorological datasets to the HYSPLIT packed-ARL format. The goal is to create ERA5-based datasets that are comparable to the archived GFSQ-based datasets (in terms of variables provided) so that the same stability, dispersion, and deposition models can be used with either dataset.

The data2arl/era42arl utility, README, and corresponding example namelist/config file were immensely helpful in selecting and mapping the requisite variables, as was the “Meterology/Help/ARL Data Format” section of the HYSPLIT User’s guide and the ECMWF ERA5 data documentation web page <https://confluence.ecmwf.int/display/CK ... umentation>. However, I encountered a few issues and have a few questions regarding the era52arl.cfg input and overall variable mapping.

Issues:
  • ERA5 provides the total cloud cover (tcc/TCLD) as a fraction, while HYSPLIT expects it to be input as a percentage. The unit conversion factor should be 100 instead of 1.
  • The latent heat flux ARL variable should be “LHTF” instead of “'LTHF'”.
  • Upon comparing the (converted) GFSQ and ERA5 values for sensible and latent heat fluxes (SHTF, LHTF), it appears that GFSQ and ERA use different sign conventions (the ERA5 documentation states that “the ECMWF convention for vertical fluxes is positive downwards”). To address this, the unit conversion factors for SHTF and LHTF should be negative. There appears to be HYSPLIT logic that flips the sign if the model is labeled as “ETA” or “EDAS”, but not “ERA5”.
Questions
  • Is there available documentation for the GFS output variables? The web pages <https://www.nco.ncep.noaa.gov/pmb/produ ... f000.shtml>, <https://www.nco.ncep.noaa.gov/pmb/produ ... f003.shtml>, and <https://www.nco.ncep.noaa.gov/pmb/docs/ ... e4-1.shtml> summarize the grib contents, but a reference that provides more context would help resolve the vertical-flux sign convention issue.
  • The terrain height and surface roughness are provided to HYSPLIT via ASCII files in the bdyfiles directory. Do the variables in the met file take precedence over the variables in the ASCII file?
  • The file metinp.f, which contains the HYSPLIT subroutine which reads in the meteorological data, does not read in surface specific humidity or surface dew point temperature variables (SPH2 or DP2M) – it only reads in surface relative humidity (RH2M). By contrast both relative humidity and specific humidity are read for the atmosphere-level variables. Is this a possible oversight? Since GFSQ provides SPH2 and ERA5 DP2M, it would seem that the surface humidity values are not used for either met dataset.
  • The arl-224 technical document describes how the surface momentum fluxes are used to calculate the friction velocity (and then the boundary layer stability). Is it true then that if the met data provides the friction velocity (which both the GFSQ and ERA5 datasets do), the surface momentum fluxes are not necessary for HYSPLIT calculations (ERA5 provides turbulent surface stresses but not surface momentum fluxes).
  • Just a comment, the era52arl code matches variables in the grib files using both the shortname and the “indicatorOfParameter” parameters. The “indicatorOfParameter” value is different from the “paramId” value (listed on the main ERA5 documentation page) for some parameters (particularly for the friction velocity, zust) .
Below is a modified era52arl.cfg file:

Code: Select all

&SETUP
 numatm = 6
 atmgrb = 'z'            ,   't'   ,   'u'    ,   'v'   ,   'w'   ,   'r'    ,   
 atmcat =      129       ,      130,       131,      132,      135,       157,   
 atmnum =      129       ,      130,       131,      132,      135,       157,   
 atmcnv =     1.01972E-01,     1.0 ,      1.0 ,      1.0,     0.01,      1.0 ,  
 atmarl = 'HGTS'         ,   'TEMP',   'UWND' ,   'VWND',   'WWND',   'RELH' ,   
numsfc = 15
 sfcgrb = '2t'           ,   '10v' ,   '10u'  ,   'tcc' ,   'sp'  ,   '2d'   ,   'blh' ,   'cape',   'z'        ,   'tp'  ,   'sshf'       ,   'ssrd'       ,   'slhf'       ,   'zust',   'fsr' ,   
 sfccat =   167          ,      166,     165  ,     164 ,    134  ,    168   ,    159  ,    59   ,     129      ,    228  ,    146         ,    169         ,    147         ,    3    ,    244  ,   
 sfcnum =   167          ,      166,     165  ,     164 ,    134  ,    168   ,    159  ,    59   ,     129      ,    228  ,    146         ,    169         ,    147         ,    3    ,    244  ,   
 sfccnv =   1.0          ,    1.0  ,    1.0   ,   100.0 ,    0.01 ,   1.0    ,    1.0  ,    1.0  ,   1.01972E-01,    1.0  ,    -2.77778E-04,    +2.77778E-04,    -2.77778E-04,    1.0  ,    1.0  ,   
 sfcarl = 'T02M'         ,   'V10M',   'U10M' ,   'TCLD',   'PRSS',   'DP2M' ,   'PBLH',   'CAPE',   'SHGT'     ,   'TPP1',   'SHTF'       ,   'DSWF'       ,   'LHTF'       ,   'USTR',   'RGHS',  
 numlev = 37
 plev   = 1000, 975, 950, 925, 900, 875, 850, 825, 800, 775, 750, 700, 650, 600, 550, 500, 450, 400, 350, 300, 250, 225, 200, 175, 150, 125, 100, 70,  50, 30, 20, 10, 7, 5, 3, 2, 1, 
/
Thanks!
alicec
Posts: 281
Joined: February 8th, 2016, 12:56 pm
Registered HYSPLIT User: Yes

Re: ERA52ARL Converter

Post by alicec »

Thanks for pointing out some inconsistencies and errors.

- The correct conversion for TCLD is 100. It is only used in the GEM part of HYSPLIT.
- Yes HYSPLIT convention is that upward sensible heat flux is positive so it looks like the conversion factor here should be negative.

Yes, HYSPLIT will use terrain height and surface roughness from the met data first. If that is unavailable, it will use the values from the ASCII files in bdyfiles directory.

HYSPLIT currently uses relative or specific humidity from the first model layer rather than the surface value. So it does not use RH2M even though it will read it in.

Yes, the surface momentum fluxes are not needed if friction velocity (U*) is provided. HYSPLIT will estimate momentum fluxes from U*.

for zust the indicatorOfParameter is 3 which is correctly noted. https://apps.ecmwf.int/codes/grib/param-db?id=228003. Could you please specify which others you think are incorrect.
rssv
Posts: 1
Joined: September 29th, 2020, 12:36 pm
Registered HYSPLIT User: Yes

Re: ERA52ARL Converter

Post by rssv »

Thank you for your reply and clarifying answers.

The indicatorOfParameter values are correct. I was originally looking at the paramId values instead (because these are listed on the main ERA5 documentation page) and was thrown off because paramId and indicatorOfParameter do not always correspond.
Post Reply