[MITgcm-support] 回复: 回复: 回复: 转发: The obvious difference between ERA5 and NCEP-R1 forcing

Matthew Mazloff mmazloff at ucsd.edu
Wed Mar 18 12:49:31 EDT 2020


Hi Mike

Are you getting spurious features at the open boundaries?
Are you using a restoring/sponge layer?
Do you have the bathymetry in the normal direction flat at the boundaries?
Where is the issue showing up - the corners?

This sounds more like an OBCS issue than a surface forcing issue, and your extra energy from hourly wind is just making your OBCS problem worse. 

Matt


> On Mar 18, 2020, at 5:00 AM, Leming Van <ifanliming at outlook.com> wrote:
> 
> Hi, everyone
> 
> I have found the cause of the difference. 
> 
> The MITgcm open boundaries are forced by monthly HYCOM data, while the forcing period of surface wind field is one hour, so the problem turns up. 
> 
> Now I uesd daily surface wind field, it works OK for me. Daily OB forcing field and hourly surface wind field may be suitable too.
> 
> Thanks for your attention and help.
> 
> 
> Cheers,
> -Mike
> 发件人: MITgcm-support <mitgcm-support-bounces at mitgcm.org <mailto:mitgcm-support-bounces at mitgcm.org>> 代表 Stanislav Martyanov <martyanov.sd at gmail.com <mailto:martyanov.sd at gmail.com>>
> 发送时间: 2020年3月17日 15:59
> 收件人: mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org> <mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org>>
> 主题: Re: [MITgcm-support] 回复: 回复: 转发: The obvious difference between ERA5 and NCEP-R1 forcing
>  
> Hello, Mike!
> >>But I am a bit puzzled. The GLOBAL_REANALYSIS_PHY_001_030 you
> >>mentioned should be a product of the NEMO model.
> >>Is this the open boundary forcing data you use?
> 
> Oh, yes, it was a misprint - I looked in the another file... Of course, the ERA5 meteorological forcing was taken from https://cds.climate.copernicus.eu/ <https://cds.climate.copernicus.eu/>, while the open boundary conditions for the ocean model - from the  GLOBAL_REANALYSIS_PHY_001_030.
> 
> By the way, concerning ERA5 units, some care should be taken. I'm sure you know all of this, as you have already posted, but for the users new to ERA5 reanalysis dataset some details might be useful:
> 
> The ERA5's field 'Mean total precipitation rate' is in [kg m-2 s-1], and dividing it by 1000 kg m-3 makes the field suitable for the MITgcm's EXF package, while ERA5's  'Total precipitation' is in [m]. Also, 'Mean surface downward long-wave and short-wave  radiation fluxes' are in [W m-2], while 'Surface solar and thermal radiation downwards' is in [J m-2]. The difference is the multiplier related to the time period between adjacent model records. So, information presented in the Overview table at the ERA5 page should be carefully read. It is not only about units, but also about the ERA5's convention for vertical fluxes' direction.
> 
> I hope you will figure out the couse of difference in your results. Please, let us know when you succeed, the topic is interesting and important indeed.
> 
> Regards,
> Stanislav
> 
> 
> вт, 17 мар. 2020 г. в 08:58, Leming Van <ifanliming at outlook.com <mailto:ifanliming at outlook.com>>:
> Hi, 
> 
> I am grateful to everyone for your reply, and here I will respond to your suggestions.
> 
> 1.for Dr. Stanislav Martyanov =========================================
> 
> Thank you for sharing the meteorological forcing variables used in the ERA5 dataset, which has benefited me a lot. 
> 
> But I am a bit puzzled. The GLOBAL_REANALYSIS_PHY_001_030 you mentioned should be a product of the NEMO model. Is this the open boundary forcing data you use?
> 
> 
> 2.for Estanislao ===================================================
> 
> Thank you for your reminding. Before using the data, it is very important to know the unit of the variable, just like the unit of precipitation in ERA5 is kg ・ m-2 s-1, and the unit of the corresponding variable in MITgcm-EXF is m ・ s-1. In this case, you need to divide by the density of fresh water for conversion. 
> 
> About the unit of long and short wave radiation flux in ERA5, I'm sure that it is W ・ m-2, and there should be no problem in the unit of other variables.
> 
> 
> 3.for Raf =======================================================
> 
> Thank you for your experience. The package I use is EXF. I simply looked at the code of EXF and found that it is calling exf_bulkformulae.F for related calculations. 
> 
> C     us  - mean wind speed (m/s)     at height hu (m)
> C     th  - mean air temperature (K)  at height ht (m)
> C     qh  - mean air humidity (kg/kg) at height hq (m)
> 
> As for the definition of various variables in height, exf_readparms. F has made relevant definitions, and I have checked that it is correct.
> 
>    zref           =       10.000  _d 0
>    hu             =       10.000  _d 0
>    ht             =        2.000  _d 0
> 
> 
> 4.for Hong =======================================================
> 
> Thank you for your checking. I'm sorry I didn't make it clear. I put net short-wave radiation flux (nswrs) in swfluxfile, since swdown will be ignored when using swdown and swflux at the same time, so I removed swdownfile.
> 
> (PID.TID 0000.0001) ** WARNING ** EXF_CHECK: "swdown" field is loaded from file but not used within pkg/exf
> 
> 
> Next ============================================================
> It seems that you don't have the same problem as me. Next, I plan to use ERA5 and NCEP data at the same time to find the reason for the difference between the results of model by the method of exclusion.
> 
> 
> 
> Best,
> -Mike
> 发件人: MITgcm-support <mitgcm-support-bounces at mitgcm.org <mailto:mitgcm-support-bounces at mitgcm.org>> 代表 Raphael Dussin <raphael.dussin at gmail.com <mailto:raphael.dussin at gmail.com>>
> 发送时间: 2020年3月16日 20:24
> 收件人: mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org> <mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org>>
> 主题: Re: [MITgcm-support] 回复: 转发: The obvious difference between ERA5 and NCEP-R1 forcing
>  
> Have you checked the heights for the atmospheric variables are correctly set?
> you should have something like this in data.blk
> 
> zref = 10.0
> zwd = 10.0
> zth = 2.0    
> 
> I've seen a lot of runs going off the rails because of this kind of problem:
> If the height is not set right, the fluxes are not either and you end up with large biases.
> 
> hope it helps,
> Raf
> 
> 
> 
> 
> Raf
> 
> 
> 
> Le lun. 16 mars 2020 à 04:49, Stanislav Martyanov <martyanov.sd at gmail.com <mailto:martyanov.sd at gmail.com>> a écrit :
> Hello, Mike!
> 
> I have recently downloaded the ERA5 reanalysis data (daily fields,  GLOBAL_REANALYSIS_PHY_001_030 as labeled in Copernicus network) in order to construct the meteorological forcing for a regional ocean model (Kara Sea) based on the MITgcm. I chose the following fields:
> 
> %           10m u-component of wind     
> %           10m v-component of wind      
> %           2m dewpoint temperature     
> %           2m temperature              
> %           Mean total precipitation rate
> %           Mean sea level pressure
> %           Mean surface downward long-wave  radiation flux
> %           Mean surface downward short-wave radiation flux 
> 
> In order to compute the specific humidity, as advised in the ERA5 documentation, I used formulas given in IFS Documentation CY46R1 - part-iv-physical-processes (2m dewpoint temperature and Mean sea level pressure are used there).
> 
> Regards,
> 
> Dr. Stanislav Martyanov,
> Shirshov Institute of Oceanology,
> Russia
> 
> 
> пн, 16 мар. 2020 г. в 09:27, Leming Van <ifanliming at outlook.com <mailto:ifanliming at outlook.com>>:
> Hi, Matt
> 
> Thank you for your advice. I'm sorry that I didn't provide enough information.
> 
> I downloaded the ERA5 dataset at https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=form <https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=form>
> Copernicus Climate Data Store | Copernicus Climate Data Store <https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=form>
> CDS Menu. Home; Search; Datasets; Applications; Toolbox; FAQ; Live; Copernicus Climate Data Store
> cds.climate.copernicus.eu <http://cds.climate.copernicus.eu/>
> 
> And I selected precip[mean_total_precipitation_rate. kg m-2 s-1], atemp[2m_temperature. K], swdown[mean_surface_downward_short_wave_radiation_flux. W m-2], lwdown[mean_surface_downward_long_wave_radiation_flux. W m-2], uwind[10m_u_component_of_wind.  m s-1], vwind[10m_v_component_of_wind.  m s-1], because ERA5 does not provide aqh directly, I used air pressure[surface_pressure  Pa] and dewpoint temperature[2m_dewpoint_temperature  K] to calculate aqh.
> 
> Since the data is hourly, I only downloaded the area of interest(bigger than South China Sea), and besides, the data is zonally aligned from 35N to 5S, I flipped it to 5S-35N. 
> 
> I compared the ERA5 with NCEP, and I don't think there is much difference except for the short-wave radiation. Please check the attachment.
> 
> I only have one question, do you choose the same ERA5 variables like me? If not, could you tell me which variables you choose?
> 
> Please let me know if any other files are needed for diagnosis. Thanks in advance.
> 
> 
> -Mike
> 
> 
> Here is the setting of my data.exf
> 
> # $Header: /u/gcmpack/MITgcm_contrib/eh3/llc/ecco-godae/input_50lev/data.exf,v 1.1 2007/03/06 19:13:51 heimbach Exp $
> #
> # *********************
> # External Forcing Data
> # *********************
>  &EXF_NML_01
>  useAtmWind        = .TRUE.,
> # rotateStressOnAgrid = .TRUE.,
>  exf_iprec         = 32,
>  exf_yftype        = 'RL',
>  useExfYearlyFields = .false.,
>  twoDigitYear      = .false.,
>  useExfCheckRange  =.false.,
>  &
> 
>  &EXF_NML_02
>  climsstfile        = 'woa13_sst_monthly_globe_relax.box',
>  climsssfile        = 'woa13_sss_monthly_globe_relax.box',
>  precipfile         = 'era5_prate_2002-2005_hourly.box',
>  atempfile          = 'era5_air_2002-2005_hourly.box',
>  aqhfile            = 'era5_shum_2002-2005_hourly.box',
>  swdownfile         = '',
>  swfluxfile         = 'era5_nswrs_2002-2005_hourly.box',
>  lwfluxfile         = '',
>  lwdownfile         = 'era5_dlwrf_2002-2005_hourly.box',
>  uwindfile          = 'era5_uwnd_2002-2005_hourly.box',
>  vwindfile          = 'era5_vwnd_2002-2005_hourly.box',
>  runofffile         = 'run-off.bin_1x1',
> # 
>  climsststartdate1  = 20020101,
>  climsststartdate2  = 000000,
>  climsstperiod      = -12,
>  climsssstartdate1  = 20020101,
>  climsssstartdate2  = 000000,
>  climsssperiod      = -12,
>  runoffstartdate1   = 20020101,
>  runoffstartdate2   = 000000,
>  runoffperiod       = -12,
> # 
>  precipstartdate1=20020101,
>  precipstartdate2=000000,
>  precipperiod=3600.0,
>  atempstartdate1=20020101,
>  atempstartdate2=000000,
>  atempperiod=3600.0,
>  aqhstartdate1=20020101,
>  aqhstartdate2=000000,
>  aqhperiod=3600.0,
>  swdownstartdate1=20020101,
>  swdownstartdate2=000000,
>  swdownperiod=3600.0,
>  lwfluxstartdate1=20020101,
>  lwfluxstartdate2=000000,
>  lwfluxperiod=3600.0,
>  swfluxstartdate1=20020101,
>  swfluxstartdate2=000000,
>  swfluxperiod=3600.0,
>  lwdownstartdate1=20020101,
>  lwdownstartdate2=000000,
>  lwdownperiod=3600.0,
>  uwindstartdate1=20020101,
>  uwindstartdate2=000000,
>  uwindperiod=3600.0,
>  vwindstartdate1=20020101,
>  vwindstartdate2=000000,
>  vwindperiod=3600.0,
>  hfluxstartdate1=20020101,
>  hfluxstartdate2=000000,
>  hfluxperiod=3600.0,
>  sfluxstartdate1=20020101,
>  sfluxstartdate2=000000,
>  sfluxperiod=3600.0,
>  ustressstartdate1=20020101,
>  ustressstartdate2=000000,
>  ustressperiod=3600.0,
>  vstressstartdate1=20020101,
>  vstressstartdate2=000000,
>  vstressperiod=3600.0,
>  &
> 
>  &EXF_NML_03
>  exf_inscal_precip=1.,
>  exf_offset_atemp=0.0,
>  exf_inscal_runoff  = 3.1710e-08,
>  &
> 
>  &EXF_NML_04
>  precip_lon0        = 95.0D0,
>  precip_lon_inc     = 0.25D0,
>  precip_lat0        = -5.0D0,
>  precip_lat_inc     = 160*0.25D0,
>  precip_nlon        = 160, 
>  precip_nlat        = 160,
> # 
>  atemp_lon0        = 95.0D0,
>  atemp_lon_inc     = 0.25D0,
>  atemp_lat0        = -5.0D0,
>  atemp_lat_inc     = 160*0.25D0,
>  atemp_nlon        = 160, 
>  atemp_nlat        = 160,
> # 
>  aqh_lon0        = 95.0D0,
>  aqh_lon_inc     = 0.25D0,
>  aqh_lat0        = -5.0D0,
>  aqh_lat_inc     = 160*0.25D0,
>  aqh_nlon        = 160, 
>  aqh_nlat        = 160,
> # 
>  swdown_lon0        = 95.0D0,
>  swdown_lon_inc     = 0.25D0,
>  swdown_lat0        = -5.0D0,
>  swdown_lat_inc     = 160*0.25D0,
>  swdown_nlon        = 160, 
>  swdown_nlat        = 160,
> # 
>  lwflux_lon0        = 95.0D0,
>  lwflux_lon_inc     = 0.25D0,
>  lwflux_lat0        = -5.0D0,
>  lwflux_lat_inc     = 160*0.25D0,
>  lwflux_nlon        = 160, 
>  lwflux_nlat        = 160,
> # 
>  swflux_lon0        = 95.0D0,
>  swflux_lon_inc     = 0.25D0,
>  swflux_lat0        = -5.0D0,
>  swflux_lat_inc     = 160*0.25D0,
>  swflux_nlon        = 160, 
>  swflux_nlat        = 160,
> # 
>  lwdown_lon0        = 95.0D0,
>  lwdown_lon_inc     = 0.25D0,
>  lwdown_lat0        = -5.0D0,
>  lwdown_lat_inc     = 160*0.25D0,
>  lwdown_nlon        = 160, 
>  lwdown_nlat        = 160,
> #
>  uwind_lon0        = 95.0D0,
>  uwind_lon_inc     = 0.25D0,
>  uwind_lat0        = -5.0D0,
>  uwind_lat_inc     = 160*0.25D0,
>  uwind_nlon        = 160,
>  uwind_nlat        = 160,
> # 
>  vwind_lon0        = 95.0D0,
>  vwind_lon_inc     = 0.25D0,
>  vwind_lat0        = -5.0D0,
>  vwind_lat_inc     = 160*0.25D0,
>  vwind_nlon        = 160,
>  vwind_nlat        = 160,
> #
>  ustress_lon0        = 95.0D0,
>  ustress_lon_inc     = 0.25D0,
>  ustress_lat0        = -5.0D0,
>  ustress_lat_inc     = 160*0.25D0,
>  ustress_nlon        = 160,
>  ustress_nlat        = 160,
> # 
>  vstress_lon0        = 95.0D0,
>  vstress_lon_inc     = 0.25D0,
>  vstress_lat0        = -5.0D0,
>  vstress_lat_inc     = 160*0.25D0,
>  vstress_nlon        = 160,
>  vstress_nlat        = 160, 
> #
>  hflux_lon0        = 95.0D0,
>  hflux_lon_inc     = 0.25D0,
>  hflux_lat0        = -5.0D0,
>  hflux_lat_inc     = 160*0.25D0,
>  hflux_nlon        = 160,
>  hflux_nlat        = 160,
> # 
>  sflux_lon0        = 95.0D0,
>  sflux_lon_inc     = 0.25D0,
>  sflux_lat0        = -5.0D0,
>  sflux_lat_inc     = 160*0.25D0,
>  sflux_nlon        = 160,
>  sflux_nlat        = 160,
> #
>  runoff_lon0        = 0.50D0,
>  runoff_lon_inc     = 1.0D0,
>  runoff_lat0        = -79.5D0,
>  runoff_lat_inc     = 159*1.0D0,
>  runoff_nlon        = 360,
>  runoff_nlat        = 160,
> #
>  climsst_lon0    = -179.875D0,
>  climsst_lon_inc = 0.25D0,
>  climsst_lat0    = -89.875D0,
>  climsst_lat_inc = 719*0.25D0,
>  climsst_nlon    = 1440,
>  climsst_nlat    = 720,
> #
>  climsss_lon0    = -179.875D0,
>  climsss_lon_inc = 0.25D0,
>  climsss_lat0    = -89.875D0,
>  climsss_lat_inc = 719*0.25D0,
>  climsss_nlon    = 1440,
>  climsss_nlat    = 720,
> # 
>  &
> 
>  &EXF_NML_OBCS 
>  obcsNstartdate1   = 20020101,
>  obcsNstartdate2   = 000000,
>  obcsNperiod       = 2628000.0,
>  obcsEstartdate1   = 20020101,
>  obcsEstartdate2   = 000000,
>  obcsEperiod       = 2628000.0,
>  obcsSstartdate1   = 20020101,
>  obcsSstartdate2   = 000000,
>  obcsSperiod       = 2628000.,
>  &
> 
> 发件人: MITgcm-support <mitgcm-support-bounces at mitgcm.org <mailto:mitgcm-support-bounces at mitgcm.org>> 代表 Matthew Mazloff <mmazloff at ucsd.edu <mailto:mmazloff at ucsd.edu>>
> 发送时间: 2020年3月16日 0:44
> 收件人: mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org> <mitgcm-support at mitgcm.org <mailto:mitgcm-support at mitgcm.org>>
> 主题: Re: [MITgcm-support] 转发: The obvious difference between ERA5 and NCEP-R1 forcing
>  
> Hi Mike
> 
> It is possible it could give worse results, but very unlikely. However its too hard for me to diagnose what is going on. There are many variables to your problem.
> 
> For ERA5 I have this in my data.exf:
>  &EXF_NML_03
>  exf_offset_atemp   = 273.15,
>  exf_inscal_swdown  = -1.0,
>  exf_inscal_lwdown  = -1.0,
> 
> Not sure how to help you, 
> Matt
> 
> 
> 
>> On Mar 15, 2020, at 5:41 AM, Leming Van <ifanliming at outlook.com <mailto:ifanliming at outlook.com>> wrote:
>> 
>> Hi, everyone
>> 
>> Recently I tried to change forcing dataset from NCEP-R1(6-hourly) to ERA5(hourly), but the results of them look so different. I compared the results of NCEP-R1 with the reanalysis data of HYCOM, and these two data look similar. 
>> 
>> Did I pick the wrong variables from ERA5? Or does this mean that higher temporal and spatial resolution atmospheric forcing ERA5 results in worse simulation results than lower resolution NCEP-R1?
>> 
>> I put the results at here, https://1drv.ms/u/s!AoS5GLJhmvDtjB5Ovnz63FidKeMo?e=c2hS7z <https://1drv.ms/u/s!AoS5GLJhmvDtjB5Ovnz63FidKeMo?e=c2hS7z>
>> Suggestions are welcome.
>> 
>> 
>> -Mike
>> 
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>
> <ncep&era5.jpg>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org <mailto:MITgcm-support at mitgcm.org>
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support <http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200318/a7dc0cd5/attachment-0001.html>


More information about the MITgcm-support mailing list