Notes, Questions and Discussion

Created by esrl_admin on - Updated on 03/29/2022 08:53

Users have requested workspaces and comment areas for general reanalysis questions and for asking specific questions or posting comments about specific reanalysis datasets.

The links below are good places to post notes, ask questions, comment, or start a discussion about each dataset.

Or, leave a comment at the bottom of this page. 

Links to Notes, Questions, and Discussions for each reanalysis dataset: Overview Comparison Table Overview of each reanalysis


SPARC-Reanalysis Intercomparison Project (S-RIP) Comparison Paper and its Tables 

  • the paper includes several tables comparing assimilating models, resolution, major physical parameterizations, data input, and boundary conditions

Bennet Foli (not verified)

Sat, 01/02/2021 - 11:01


I would like to know if PIRATA buoy data are assimilated in NCEP and ERA5 reanalysis winds and if so which specific buoys (locations) and which parameters from the buoys are utilized. Can I also have some references to these.

Thank you.

Thank you for the reply.

The following are the station IDs and corresponding locations

GL_TS_MO_13010; % Lat 0N, Lon 0E
GL_TS_MO_15002; % Lat 0N, Lon 10W
GL_TS_MO_31006; % Lat 4N, Lon 23W
GL_TS_MO_13002; % Lat 20N, Lon 23W

Thank you.


I tried to estimate the 2m horizontal wind using the log wind profile with 10 m horizontal wind, and found it is quite different to the 2m horizontal wind provided in MERRA2 output. So, I am wondering how did you calculate the horizontal wind at 2m? Thanks.

Posted from Arlindo da Silva and Andrea Molod, both of NASA


From Dr. da Sliva-

From the glossary: U10M, U2M, V10M, V2M: The eastward and northward wind component at 10 m, and 2m in the surface layer, in m s -1 . These values are defined above displacement height (DISPH) and are dependent on the surface layer parameterization being used.


From Dr. Molod

hi there,

MERRA-2 used what we call the Helfand (Helfand and Schubert, 1995) Monin-Obukhov surface layer parameterization. The charnock coefficients (relating stress to speed over the ocean) were modified substantially from what is in the manuscript....  The 2-m and 10-m winds are “picked off” the non-dimensional vertical wind profiles that the local stability dictates. It will not use a neutral (log wind) profile.

Our surface layer thickness in MERRA-2 is approximately 60m, and so in M2 the 2-m and 10-m winds we output are essentially interpolated along the correct profile between 0 at the sfc and the model lowest level wind at 60 m.

If you are interested in repeating this calculation, it would be possible to approximate the profile using the Cu, u* and Ri outputs from MERRA-2 and getting the Psi values (non-dimensional vertical wind profile) from those.

I can send you the algebra or you can find it in the Helfand and Schubert paper in the appendix....

Let me know if there is any more information you would like.



Recently I'd like to use MERRA2 data to do some analysis. I read some papers and knew that most output variables were calculated by

parameterizations in the GEOS-5 AGCM model it used and assimilated observations. 

May I ask how they made the output variables physically consistent? there are so many parameterizations that seem not related together.




Hi Orrin,

The parameterizations in any GCM or atmospheric model generally depend on each other. There are prognostic state fields: T, Q, U, V (sometime liquid and ice water too) and say aerosols and chemical constituents. clouds are formed depending on the states, radiation is computed from the states and the clouds, and each then feed back into the states.  Land and ocean surfaces connect with the atmosphere, exchanging information through fluxes.  This is how the models are designed.  The code can be complicated in places. Best is to have an advisor who has experience with models and work with one. (without knowing your background)


Richard Yablonsky (not verified)

Fri, 05/12/2017 - 12:23

Hello... we are attempting to simulate the coastal flooding in Great Britain from the 1953 North Sea flood, and we are having trouble finding an atmospheric reanalysis product that provides reasonably accurate surface wind and pressure forcing for our simulations.  Our time period of interest is January 30, 1953 to February 2, 1953.  Any advice on which reanalysis product(s) would be most appropriate for our purposes would be much appreciated.

Dear Richard,

Which datasets have you already tried in this context? 

There are only 5 possibilities: 

20CR v2

20CR v2c




Links to all of these are available at 

For the ensemble-based 20CR v2 and v2c, you should use individual members, though using the ensemble mean for 1953 Great Britain should be alright.  CERA-20C also has an ensemble, and individual members would probably be better for your purposes. 

best wishes,


20CR2c has a decent qualitative representation of the storm - - I've not looked at the quantitative accuracy. I think David Hein, at the Met Office has has a regional model run covering the UK downscaled from one of the 20CR ensemble members. That might also be worth a look.

Fri, 04/07/2017 - 09:57

The WRIT page is a good start.

For specific countries, The World Bank has data by country but not to 2015. You might contact them and see if they have it available.

They use CRU data.

You could also look at the URL below  (they have different years/months available). NCEI has many datasets and could have it lesehwre. You can look/ask them.

good luck!


Tracy Carty (not verified)

Thu, 04/06/2017 - 08:26


I work for a leading development NGO.  I'm writing a paper on the East Africa drought and I'm trying to get the following information, or close to it based on what's available.  I'm not up to speed with the data sets or how to use them and hence this message is a call to see if anyone would be able to lend me a hand.  I'm hoping this is a quick job for someone who has the data to hand and knows what they are doing??

[Mean annual temperatures increased from 1960-2016 by X in Kenya, X in Ethiopia and X in Somalia][would also be interested for 2015 and first few months of 2017 if available]

FYI - I've also emailed to see if anyone can help with this.

Best wishes,



Dear Tracy,

If you go to the Web-based reanalysis intercomparison tools and select the WRIT Timeseries button, you can make these calculations yourself for regions that encompass your countries of interest. The JRA-55 Screen dataset is probably the best for your purposes, as it includes actual near-surface thermometer values and is fairly up to date. The JRA-55 itself does not directly include the near surface temperature observations (it does include the upper-air, which start near the surface). 

Enter your latitude and longitude box approximating each country and you will get a time series that includes statistics about how much the temperature has changed.

You can also use the WRIT Maps and zoom in on Africa, and show the difference between the current period and an earlier period.

Hope that this is helpful.

Best wishes,



Ashwin (not verified)

Thu, 09/01/2016 - 02:31

Does anyone have any code samples that are newer than this one - ?


I want to be able to read a reanalysis 2 file using fortran and netcdf .

Add new comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.