MERRA-2 Notes, Questions, and Discussion

Created by mapr8844 on - Updated on 09/07/2021 12:19

Documentation:

It is best to consider data set developers documentation first, when questions arise.

MERRA-2: http://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

MERRA: http://gmao.gsfc.nasa.gov/research/merra/file_specifications.php

Notes:

  • For users: Most users will want to use assimilation fields and not analysis. The MERRA File Specification Document spells out the differences. The analysis states are not adjusted by the model, and are derived from the data assimilation system. These may be better for use as regional model lateral boundary conditions.
  • When you download data, there are streams (year ranges for different runs). For ftp access, you need to change the stream number in the files you download according to the following:   1980-1991: 100, 1992-2000: 200, 2001-2010: 300, 2011-present: 400.  In the case that data needs to be re-run with scientific changes, these numbers will be incremented (e.g. 401)

          Step by step directions are available for one type of download.

 

Citing MERRA-2 data has 2 steps for a full citation: 

First pick the correct variable here: https://disc.gsfc.nasa.gov/datasets?page=1&project=MERRA-2&keywords=%22MERRA-2%22

When you click on the correct variable, it will take you to a second webpage with tabs that you can click that include: (1) documentation papers you need to cite, and (2) the correct variable citation information. (And you apparently need both types of citation). For example, here is a link to the page for 3d assimilated, 3-hourly, model level data: https://disc.gsfc.nasa.gov/datasets/M2I3NVASM_V5.12.4/summary?keywords=%22MERRA-2%22

 

 

General good contact: data center or the GMAO.

Data Download questions should go to the GES DISC help email:

gsfc-help-disc@lists.nasa.gov

Science questions regarding MERRA data be emailed to

merra-questions@lists.nasa.gov

When contacting these emails, provide specific information and links to where you have attempted the data downloads. It is also very important to familiarize yourself with the existing documentation.

MERRA-2: http://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

MERRA: http://gmao.gsfc.nasa.gov/research/merra/file_specifications.php

 

 

Previous comments:

Re: MERRA-2 Notes, Questions, and Discussion

Submitted by kishorekumargrandhi on Thu, 06/02/2016 - 02:07. 

Is it possible to download MERRA 2 data as a single file for whole period for one pressure level and for one variable?

Re: MERRA-2 Notes, Questions, and Discussion

Submitted by michael.bosilovich on Thu, 06/02/2016 - 08:45. 

The data are not prepared that way. However, the subsetters will allow you to collect a pressure level with time dependent files that can be easily concatenated at your end.

Mike Bosilovich (NASA/GMAO)

assimilation of surface T, q 

Submitted by Justin Schoof (not verified) on Tue, 05/31/2016 - 12:43. 

Does MERRA-2 assimilated 2-m observations of air temperature and specific humidity. I've scoured the available references and can't find a clear answer. Thanks in advance.

Re: assimilation of surface T, q 

Submitted by michael.bosilovich on Thu, 06/02/2016 - 08:43. 

Neither MERRA or MERRA-2 assimilate surface meteorology station data over land, specifically T2m or q2m. The radiosondes do provide some low level observations, but well above the surface layer.

Mike Bosilovich (NASA/GMAO)

Re: assimilation of surface T, q 

Submitted by gilbert.p.compo on Tue, 05/31/2016 - 13:12. 

Justin, I suggest emailing your question to merra-questions@lists.nasa.gov You should be specific about whether you are interested in over land, over ocean, or both. best wishes,

Gilbert P. Compo

University of Colorado/CIRES

NOAA/ESRL/Physical Sciences Division

http://www.esrl.noaa.gov/psd/people/gilbert.p.compo

Access to data

Submitted by Johanna Yepes (not verified) on Thu, 05/26/2016 - 10:49. 

Hi, I downloaded MERRA 6-hourly data of Specific humidity, Temperature and Wind and I cannot read them. I've tried to read them from Panoply and Matlab and I couldn't have access to the data. Two years ago I downloaded similar data and I could read them, but now I cannot, Should have it suffered any change the format? or any idea what is happening? Thank you.

Re: Access to data

Submitted by michael.bosilovich on Fri, 05/27/2016 - 07:43. 

Hi Johanna,

A couple things here. It's very hard to answer your question, since thre is not much information about how you downloaded the data and what exactly the problem may be. First thing to do is to check the file sizes are appropriate. very small files may indicate a corrupted download. Also, this page is not the most efficient for getting a response from the data center or the GMAO. This type of question should go to the GES DISC help page, as it pertains to accessibility of the data.

gsfc-help-disc@lists.nasa.gov

Science questions regarding MERRA data can go here

merra-questions@lists.nasa.gov

When contacting these emails, provide specific information and links to where you have attempted the data downloads. It is also very important to familiarize yourself with the existing documentation.

MERRA-2: http://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

MERRA: http://gmao.gsfc.nasa.gov/research/merra/file_specifications.php

In particular, the File Specification Document, will tell you everything you need to know about our native data formats and the grid structures (these may be manipulated by GES-DISC utilities though).

Lastly, since you are reading data with third party software, I will mention that there is no guarantee that our help emails will be able to provide assistance with those software. Best is to refer to their documentation and help lines.

I hope this helps

Mike Bosilovich (NASA/GMAO)

Re: Access to data

Submitted by Cathy.Smith@noaa.gov on Thu, 05/26/2016 - 11:17. 

Hi- The most likely the reason you can't read the data is that it is netCDF4 and not netCDF3. the files now end in nc4 and if you do a ncdump on a file from the ftp submitter NASA has you get, for example, ncdump -k svc_MERRA2_100.instM_2d_asm_Nx.198001.nc4 netCDF-4 classic model Both Panoply and MATLAB can read netCDF4 but you may need to upgrade your netCDF library. I checked some MERRA files that I have and I believe there were netCDF3. Cathy

Re: Access to data

Submitted by gilbert.p.compo on Thu, 05/26/2016 - 11:01. 

Dr. Yepes, Would you send the precise path you used for the download, as well as the name of the resulting file? Additionally, if you can, please send the md5sum or checksum, so that one can verify the file integrity. Are you on a Unix (Linux) system? best wishes,

Gilbert P. Compo

University of Colorado/CIRES

NOAA/ESRL/Physical Sciences Division

http://www.esrl.noaa.gov/psd/people/gilbert.p.compo

Re: Access to data

Submitted by Johanna Yepes (not verified) on Fri, 05/27/2016 - 11:05. 

Dear Dr. Compo, I used the following path http://disc.sci.gsfc.nasa.gov/daac-bin/FTPSubset.pl and I got the file named: MERRA300.prod.assim.inst6_3d_ana_Np.20160115.SUB.nc I got from cheksum calculator: 8295317E44DE1615EC88CA9D46838FF7. Probably the file is corrupted because is size is: 1kb. Thank you.

Re: Access to data

Submitted by gilbert.p.compo on Fri, 05/27/2016 - 13:30. 

Dear Johanna, Please do let us know how it goes, after trying to the data download again and contacting the email address mentioned above by Mike.

Gilbert P. Compo

University of Colorado/CIRES

NOAA/ESRL/Physical Sciences Division

http://www.esrl.noaa.gov/psd/people/gilbert.p.compo

Ashwin (not verified)

Fri, 02/02/2024 - 09:03

Black Carbon Angstrom parameter [470-870 nm] Black Carbon Column Mass Density Black Carbon column u-wind mass flux Black Carbon column v-wind mass flux Black Carbon Extinction AOT [550 nm] Black Carbon Scattering AOT [550 nm] Black Carbon Surface Mass Concentration DMS Column Mass Density __ENSEMBLE__ DMS Surface Mass Concentration __ENSEMBLE__ Dust Angstrom parameter [470-870 nm] Dust Column Mass Density Dust Column Mass Density - PM 2.5 Dust column u-wind mass flux Dust column v-wind mass flux Dust Extinction AOT [550 nm] Dust Extinction AOT [550 nm] - PM 2.5 Dust Scattering AOT [550 nm] Dust Scattering AOT [550 nm] - PM 2.5 Dust Surface Mass Concentration Dust Surface Mass Concentration - PM 2.5 Organic Carbon Angstrom parameter [470-870 nm] __ENSEMBLE__ Organic Carbon Column Mass Density __ENSEMBLE__ Organic Carbon column u-wind mass flux __ENSEMBLE__ Organic Carbon column v-wind mass flux __ENSEMBLE__ Organic Carbon Extinction AOT [550 nm] __ENSEMBLE__ Organic Carbon Scattering AOT [550 nm] __ENSEMBLE__ Organic Carbon Surface Mass Concentration __ENSEMBLE__ Sea Salt Angstrom parameter [470-870 nm] Sea Salt Column Mass Density Sea Salt Column Mass Density - PM 2.5 Sea Salt column u-wind mass flux Sea Salt column v-wind mass flux Sea Salt Extinction AOT [550 nm] Sea Salt Extinction AOT [550 nm] - PM 2.5 Sea Salt Scattering AOT [550 nm] Sea Salt Scattering AOT [550 nm] - PM 2.5 Sea Salt Surface Mass Concentration Sea Salt Surface Mass Concentration - PM 2.5 SO2 Column Mass Density __ENSEMBLE__ SO2 Surface Mass Concentration __ENSEMBLE__ SO4 Angstrom parameter [470-870 nm] __ENSEMBLE__ SO4 Column Mass Density __ENSEMBLE__ SO4 column u-wind mass flux __ENSEMBLE__ SO4 column v-wind mass flux __ENSEMBLE__ SO4 Extinction AOT [550 nm] __ENSEMBLE__ SO4 Scattering AOT [550 nm] __ENSEMBLE__ SO4 Surface Mass Concentration __ENSEMBLE__ Total Aerosol Angstrom parameter [470-870 nm] Total Aerosol Extinction AOT [550 nm] Total Aerosol Scattering AOT [550 nm] 

MEERA2 data have above parameters    I want AOD for my analysis and so How I can calculate AOD??

Please explain with details.
 

Tertsea Igbawua (not verified)

Sat, 11/11/2023 - 00:46

I want to download the Monthly Aerosol Mixing Ratio for monthly analysis PM1 and PM10, but all I can see is the hourly version "Instantaneous Three-Dimensional Collections inst3_3d_aer_Nv (M2I3NVAER): Aerosol Mixing Ratio"

Is there a link that I can locate the monthly version of the same data set?

Omar Ramos Pérez (not verified)

Sun, 09/03/2023 - 23:41

Hi!

I'm utilizing the dataset, M2T1NXAER of MERRA-2. I have a query: What is the meaning of single-level in this dataset (M2T1NXAER)?  

Thanks for helping, and best regards, Omar Ramos 

Adnan (not verified)

Sun, 05/14/2023 - 01:43

Hi

Can I extract the MERRA-2 point data of AOD to validate with ground station point data at one of the AERONET stations? Is it acceptable in the research community for publication?  

Gopika (not verified)

Mon, 08/15/2022 - 07:43

Hi, 

I am currently trying to estimate Radiative Forcing due to different aerosol types, for which the basic parameters such as AOD, SSA, and Asymmetry parameters are the essential components. However, MERRA-2 straight away gives AOD and SSA parameters for 5 aerosol types including Sea Salt, Black Carbon, Organic Carbon, Dust, and Sulfate.

But I have not come across any information about how to estimate the Asymmetry Parameter for these 5 aerosol types globally. In this case, can anyone suggest how I can get the Asymmetry Parameter datasets across the globe? Or else, if the information about the Asymmetry Parameters datasets (globally) has already been provided somewhere that I have not come across, can you please share the same?

Thanks in advance

For properly computing aerosol radiative forcing you need PROFILES of extinction, ssa and asymmetry parameter, something that would need to be computed from the aerosol mixing ratio files (aer_Nv) and our lookup tables found at https://portal.nccs.nasa.gov/datashare/iesa/aerosol/AerosolOptics/. There you also find the full phase function (and phase matrix) for each aerosol type.

Now, if you would like to use the vertically integrated aerosol optical properties and prescribe some vertical structure, do it at your own peril. To answer your question, the asymmtry parameter due to all aerosols is computed as a weighted average, where the weights are the scattering optical depth, e.g., DUSCATAU, SSSCATAU, etc. Notice also, that the single scattering albedo should be computed as the ratio of TOTSCATAU / TOTEXTTAU.

June (not verified)

Mon, 05/16/2022 - 01:17

Hi,

I'm a student learning to work with MERRA2 datasets. However, I have a trouble with understanding some of values in a file.

Can you help me understand about the variable in MERRA2_tavgM_3d_odt_Np dataset? The set contains variables such as ozone tendency, and I can't fully understand them. I tried to find a document and guess through units, but that was not crystal clear for me.

Does ozone tendency mean ozone trend? and does odd oxygen due to moist processes mean the value only considers  odd oxygen related to a moist? Same with turbulences and chemistry.

How can I work with the dataset to make a comparison between ozone dataset which has a ppmv unit?

Thanks in advance,

June

June (not verified)

Mon, 05/16/2022 - 01:16

Hi,

I'm a student learning to work with MERRA2 datasets. However, I have a trouble with understanding some of values in a file.

Can you help me understand about the variable in MERRA2_tavgM_3d_odt_Np dataset? The set contains variables such as ozone tendency, and I can't fully understand them. I tried to find a document and guess through units, but that was not crystal clear for me.

Does ozone tendency mean ozone trend? and does odd oxygen due to moist processes mean the value only considers  odd oxygen related to a moist? Same with turbulences and chemistry.

How can I work with the dataset to make a comparison between ozone dataset which has a ppmv unit?

Thanks in advance,

June

Sarun (not verified)

Tue, 02/02/2021 - 01:26

I downloaded tavg1_2d_aer_Nx: Aerosol Diagnostics data which is the product of MERRA2 GMI in order to get PM2.5 data. My country of interest is Thailand. I found that many pixels contain extremely high value of PM2.5 in the range of 1e-06 - 2e-06 kg m-3, which is equivalent to 1000-2000 microgram m-3. Do I misunderstand anything? Or, any other conversion factors needed in this case?

For more information, I tried to check each component of PM2.5 following the equation PM2.5 = 1.375*SO4 + 1.8*OC + BC + DS +SS, and found that OC was also extremely high.

Best regards

I have same problem about reanalysis data for my study area Sulaimani-Iraq , specialy for finding variable , is the equation resalte have the same MERRA-2 resalt ? how we can change the unit for micro gram because if we change it it has really diffeance between MERRA-2 data and ground measured data.

how can we found pm10 and pm1

Muhammad Umar (not verified)

Wed, 12/30/2020 - 09:41

Sir

In MERRA, there are hourly data (24 layers in monthly layer) in monthly files, what does it means? And how to get monthly total evaporation in millimeters?

 

Thanks

Umar

Shuai Yin (not verified)

Tue, 12/01/2020 - 04:06

Hi,

I am thinking of using MERRA-2 data to estimate the surface PM1 and PM10 concentration. But I found two different equations. 

First one,

PM1 = (1.375*SO4 + BCphobic + BCphilic + OCphobic + OCphilic + 0.7 * DU001 + SS01 + SS002) * AIRDENS

PM10 = (1.375*SO4 + BCphobic + BCphilic + OCphobic + OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * AIRDENS,

which is from the website of Goddard Space Flight Center (https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/FAQ/)

Second one, 

PM10 = (1.375*SO4 + BCphobic + BCphilic + 1.8*OCphobic + 1.8*OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * delp/g, which is from a discussion of this website.

We can notice that the second equation multiply the OC aerosols by 1.8 and the first one does not. So I would like to know which one is correct?

On the other hand, I also cannot find a paper including these equations. So can you provide a reference that we can cite these equations?

Thanks.

 

Shuai Yin

 

Shuai Yin (not verified)

Tue, 12/01/2020 - 04:00

Hi,

I am thinking of using MERRA-2 data to estimate the surface PM1 and PM10 concentration. But I found two different equations. 

First one,

PM1 = (1.375*SO4 + BCphobic + BCphilic + OCphobic + OCphilic + 0.7 * DU001 + SS01 + SS002) * AIRDENS

PM10 = (1.375*SO4 + BCphobic + BCphilic + OCphobic + OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * AIRDENS,

which is from the website of Goddard Space Flight Center (https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/FAQ/)

Second one, 

PM10 = (1.375*SO4 + BCphobic + BCphilic + 1.8*OCphobic + 1.8*OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * delp/g, which is from a discussion of this website.

We can notice that the second equation multiply the OC aerosols by a factor of 1.8 and the first one does not. So I would like to know which one is correct?

On the other hand, I also cannot find a paper including these equations. So can you provide a reference that we can cite these equations?

Thanks.

 

Shuai Yin

 

 

pritha pande

Wed, 07/15/2020 - 07:25

 

Hi,

 

I have a query regarding the way 1)  PM10 is calculated using MERRA-2 aerosol species i.e,  BC, OC, SS, SU, and DU.

equation given - PM10 = (1.375*SO4 + BCphobic + BCphilic + 1.8*OCphobic + 1.8*OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * delp/g

I want to know where the values of the factors you are multiplying different components with

(like SO4 with 1.375, OC with 1.8, DU with 0.74) come from?

 

 

2) Is there a way with which we can ad nitrate particulate concentration if we know ammonia gas concentration

in the atmosphere?

 

3) How can we calculate PM concentration with the provided the species concentration? Just in the same manner as PM10 is calculated, but including all the bins?

 

Posted on behalf of Arlindo da Silva, NASA:

I have a query regarding the way 1)  PM10 is calculated using MERRA-2 aerosol species i.e,  BC, OC, SS, SU, and DU. equation given - PM10 = (1.375*SO4 + BCphobic + BCphilic + 1.8*OCphobic + 1.8*OCphilic + DU001 + DU002 + DU003 + 0.74 * DU004 + SS01 + SS002 + SS003 + SS004) * delp/g. I want to know where the values of the factors you are multiplying different components with(like SO4 with 1.375, OC with 1.8, DU with 0.74) come from?

ANSWER:  Factor for dust takes in consideration the size distribution for that bin.  Since the GOCART tracer is the mass of the sulfate ion, we multiply it here by a factor of 1.375. MERRA-2 simulates OC aerosol, and the particulate organic matter (POM) is estimated from modeled OC multiplied by a factor (molecular weight per carbon weight ratio) that takes into account contributions from other elements associated with the organic matter. This factor varies spatially and temporally with values between 1.2 and 2.6 (Malm et al., 2011). A constant value of 1.8 is applied in MERRA-2.

Malm, W.C., Schichtel, B.A., Pitchford, M.L., 2011. Uncertainties in PM2.5 gravimetric and speciation measurements and what we can learn from them. J. Air Waste Manag. Assoc. 61 (11), 1131e1149.

 

2) Is there a way with which we can ad nitrate particulate concentration if we know ammonia gas concentration in the atmosphere?

 

No, MERRA-2 did not include Nitrates. Nitrates have been included in the latest version of our system, though.

3) How can we calculate PM concentration with the provided the species concentration? Just in the same manner as PM10 is calculated, but including all the bins?

 

See below about the size distribution for each bin. If you are not concerned about size, just do: PM = (1.375*SO4 + BCphobic + BCphilic + 1.8*OCphobic + 1.8*OCphilic + DU001 + DU002 + DU003 + DU004 + DU005 + SS01 + SS002 + SS003 + SS004 + SS005) * delp/g

Reeta (not verified)

Mon, 09/16/2019 - 09:22

I have a query regarding bin size of aerosol species.

 

I am using the mass-mixing ratio of 3-d, 3-hourly, instantaneous aerosol and ozone for my research.

species. I wanted to use the bin which has species' diameter less than 2.5 micrometers as I'll be extracting data for PM2.5.

 

Please clarify which bin should I use for this purpose for dust, sea salt, black carbon, organic carbon and sulfate?

 

 

 

Kunal Bali (not verified)

Mon, 03/11/2019 - 04:56

could you please let me know the bin size for different aerosol species?

like dust have 5 bins size such as 0.1--1, 1--1.8, 1.8--3, 3--6, 6--10 so, how about

black carbon (2bin size), organic carbon (2 bin size), sulfate (4 bin size), sea salt (5 bin size).

could you please provide some information on that.  

 

* Dust
Dust aerosol is represented with 5 bins that correspond to dry size
ranges (in 1e-6m) and densities (kg m-3)
---------------------------------------------
bin            1      2     3     4      5
---------------------------------------------
radius         0.73   1.4   2.4   4.5    8.0
radius lower   0.1    1.0   1.8   3.0    6.0
radius upper   1.0    1.8   3.0   6.0   10.0
density        2500   2650  2650  2650  2650
---------------------------------------------

* Sea salt
Sea salt aerosol is represented with 5 bins that correspond to dry size
ranges (in 1e-6m) and densities (kg m-3)
---------------------------------------------
bin           1     2     3      4       5
---------------------------------------------
radius        0.079 0.316 1.119  2.818   7.772
radius lower  0.03  0.1   0.5    1.5     5.0
radius upper  0.1   0.5   1.5    5.0    10.0
density       2200  2200  2200   2200   2200
---------------------------------------------

* Organic carbon
OC aerosol is represented with two tracers/bin that correspond to
hydrophobic and (aged) hydrophilic particles. Dry size and density for
the two OC tracers are: 0.35e-6m and 1800 kg/m3.

* Black carbon
BC aerosol is represented with two tracers/bin that correspond to
hydrophobic and (aged) hydrophilic particles. Dry size and density for
the two BC tracers are: 0.35e-6m and 1800 kg/m3.

* Sulfate aerosol
Four tracers are used to represent the sulfate cycle in the atmosphere.
These correspond to DMS, SO2, SO4 and MSA species. Tracers 1, 2 and 4
are gas species and do not have associated sizes with them. The SO4
tracer (#3) corresponds to sulfate particles and has dry size and
density of 0.35e-6m and 1700 kg/m3.

 

Cristian Waimann (not verified)

Thu, 11/01/2018 - 13:37

Hi,

I wonder if you could give me information about the averaging of radiation data. For example, SWGDN at 01:00 is the average of the SWGDN between 00:00 and 01:00? Or is it centered at 01:00? I am having problems using SWGDN data on PVLIB package to estimate PV output power. 

Thanks!

Dear Cristian,

Have your tried sending this science question to

merra-questions@lists.nasa.gov

If that isn't responsive pretty quickly, please try posting here. I would expect that deep in the metadata the "averaging interval" and the start and end times for the average are given somewhere, but merra-questions is the best place to start.

Best wishes,

Kiswendsida Hy… (not verified)

Tue, 09/11/2018 - 06:17

Hi,

I want to download MERRA-2 radiation data: 

Surface net shortwave radiation

Surface net longwave radiation

Both sensible and latent fluxes.

For the two first one I've gone through the specification doc but it is in my sense not clear enough. Thus for shortwave radiation all these 03 parameters seem to be what I want:

-surface net downward shortwave flux SWGNT (I think this should be the one I need but I am not confident because of the "downward"),

-surface incoming shortwave flux SWGDN(what's the difference between the "downward" and the "incoming") and 

-surface net downward shortwave flux SWNETSRF in the vertically integrated collection (clearly what makes the difference between SWGNT and SWNETSRF?).

I would like then to know which one is the real surface NET SW radiation  (not only downward) so that I may be able to derive the "upward" one as well. 

 

For longwave radiation, similar confusions exist between LWGNT and  LWGNET (this time there is no GDN). The question here is then which one is the "net" one as well.

 

Regarding turbulent fluxes, I have opted for EFLUX (latent) and HFLUX (sensible). I hope I have done right.

 

Your inputs are very appreciated.

 

Regards

 

GUIGMA 

Firstly, downward in "net downward" give the direction of the radiative flux.  So, the value is the net flux, positive downward.

Some variables were duplicated across file collections for convenience.  I expect that SWGNT and SWNETSRF are the same values. The land model group likely named the SWNETSRF variable, while the atmospheric physics group named SWGNT. 

Ipsita (not verified)

Mon, 08/06/2018 - 04:38

Hi,

Can you please help me understand about the variable "relative humidity after moist"? What is that mean actually. Normally we find RH in percentage %, whereas here the unit is 1, whereas the data range varies from -1e+15 to 1e+15. How can I compare it with other RH values which follow the normal % unit?

Though I have gone through the documents, still this variable is not much clear to me.

Thanks in advance,

Ipsita

In most GMAO pressure level data collections, there is no extrapolation of the surface meteorology to P > PS. Essentially, pressure coordinate locations under the ground are set to undefined. If you mask out these values, you should find real values of RH. The RH is written as a fraction, so multiply by 100 for percent. Also, see this discussion of the MERRA pressure level missing data fields.

I am a PhD student at the University of Nairobi, Kenya and I am using Black Carbon Mass Concentration data to check for its spatial-temporal characteristics over Eastern Africa region (28E,-12S,42E,6N). I would like to know MERRA-2.0/MERRA-2.1 Reanalysis Data is from which Satellite, can I get the satellite photograph, platform/instrument, methods it uses to generate data, whether the data is synthetic or observational one, possible sources of error and accuracy percentage.

Hi Daniel,

First, please read through the MERRA-2 documentation to get a sense of the data and its strenghts an dweaknesses.

https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

There is documentation on the aerosols. In short, emission sources are included, as well as assimilation of AVHRR and MODIS total AOD, which is then speciated and spread in the vertical column. There are also some papers in the Journal of Climate MERRA-2 psecial issue that may be of help.

http://journals.ametsoc.org/topic/merra-2

Mike

Boxuan Wong (not verified)

Sun, 04/15/2018 - 12:25

I got confused about these two terms. The mid_layer_heights apparently follow the topography. The geopotential heights, defined as the adjustment relative to the mean sea level, supposedly to be ~5 km over the Tibetan plateau at 1000-hPa. But from the MERRA2 "inst6_3d_ana_Np" data, when you look at the 1000 hPa level, the geopotential heights over the Tibetan plateau is about 100-200m. Why is that? 

 

Geopotential height is derived from other state variables. In these files, the heights are generated from extrapolation beneath the surface. This allows further diagnostics (such at 1000-500mb Thickness, which has a certain meteorological value, or for comparison with other reanalyses height fields). The extrapolated state fields are not included at pressure levels greater than the surface pressure, and marked undefined. So, if you would like to exclude the 1000mb geopotential heights beneath the surface, 1000mb temperature can be used as a mask.

If you are looking for height above ground level (AGL), that is not provided, but could be determined from model eta level coordinates. Also keep in mind that the surface pressure over the Tibetan plateau is on the order of 550mb.

I hope this helps.

David (not verified)

Tue, 03/27/2018 - 15:36

Hi, I need to know what variables (T, UV winds, surface pressure, etc.) MERRA2 assimilates from land met stations and moored buoy arrays such as TAO, TRITON and RAMA.

Also, does MERRA assimilates any variables measured by moored buoys offshore the Spanish Atlantic coast, operated by the Spanish Agency Puertos del Estado?

Any help would be much appreciated, thanks!

Christopher Frank (not verified)

Wed, 02/21/2018 - 04:14

Hi,

I was just looking into the wind speed profiles provided by MERRA-2. I merged the 3d field with the 10m and 50m wind speed and plottet the resulting profiles. Surprisingly, I found the 50m wind speed being exactly the same value as the lowest level 3d wind speed. The lowest level 3d wind speed is given at an height of about 65m. This is the lowest entry of the mid_level_heights variable.

Is this known? Do I have done something wrong?

Christopher

Dear Christopher, 

I am cutting and pasting a response that Mike Bosilovich previously posted. I think these suggestions will be your best bet on getting the most specific answers:

It's very hard to answer your question, since thre is not much information about how you downloaded the data and what exactly the problem may be. First thing to do is to check the file sizes are appropriate. very small files may indicate a corrupted download. Also, this page is not the most efficient for getting a response from the data center or the GMAO. This type of question should go to the GES DISC help page, as it pertains to accessibility of the data.

gsfc-help-disc@lists.nasa.gov

Science questions regarding MERRA data can go here

merra-questions@lists.nasa.gov

When contacting these emails, provide specific information and links to where you have attempted the data downloads. It is also very important to familiarize yourself with the existing documentation.

MERRA-2: http://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/
MERRA: http://gmao.gsfc.nasa.gov/research/merra/file_specifications.php

In particular, the File Specification Document, will tell you everything you need to know about our native data formats and the grid structures (these may be manipulated by GES-DISC utilities though).

 

Please let us here at reanalyses.org let us know if we can be of more help. Also, if you get an account, you can make a page that illustrates what you are seeing and provides the steps you took to get to your surprising result.  (See a page such as http://reanalyses.org/atmosphere/20cr-ensemble-mean-wind-speeds-appear-wrong for a good example of exchanging information about findings).

Best wishes,

gil

 

Hi Christopher

50m winds are interpolated from the lowest model level using Monin Obukov surface layer similarity theory. So it should be close to, but not exactly equal the lowest model level winds.  1 hourly lowest model level data are included with SLV data collection. (see the LML variables)

 

I hope that helps

 

Mike

Alemu Tadesse (not verified)

Wed, 04/19/2017 - 15:56

We were comparing MERRA 2 wind and temperature data to measured data in Mexico. The bias from MERRA2 was very large compared to METAR stations data. Any thoughts  ?

Thanks,


Alemu

 

Dear Alemu,

Very interesting. I suggest directing your 

Science questions regarding MERRA data to

merra-questions@lists.nasa.gov

When contacting these emails, provide specific information and links to where you have attempted the data downloads. It is also very important to familiarize yourself with the existing documentation.

MERRA-2: http://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

Additionally, may I suggest creating a page here at reanalyses.org with figures that illustrates the issue? Then, the page can linked to in the message, and other users could comment on what you are showing.  If you don't have an account, direct your browser to http://reanalyses.org/user/register

and then you can create a page for posting a description of your findings and import a few figures. 

Please let me know if I can provide any additional information. 

best wishes,

gil

 

Gil made some good points, words to live by, as it were.  it's hard to answer this question since there are few specifics.  Is it temperature, wind or both that are biased? Which days, months season?  Which variables are you comparing to METAR?  Where is the bias exactly? Warm or cold, fast or slow?

We are working on some regional near surface temperature cold biases over desert surface types, somewhat prominent over the Sahara, but if it is surface type dependent, then it would seem Mexico might be affected similarly.  When we have results on that, I will try to remember to open a page on reanalysis.org for it.

I want to compare MERRA 2 wind data with land data for Pakistan, and I am unable to access/download MERRA 2 data for Pakistan under following required parameters:

MERRA 2 dataset required:
variables : u-component and v-component for wind
Temporal/Date range: from 2008-01-01 to 2012-01-01 within Grid 60, 23, 76, 39
Temporal Resolution = 3 hours
Spatial resolution =   0.25° x 0.25° or 0.5 ° x 0.5 ° or  1 ° x 1 °
pressure level= 1013.25hPa = 1atm

Any help to access this data.

Dear Faisal,

This is exactly the sort of detailed information that is needed, except you should specify what is longitude and what is latitude on your grid. I expect it is 60E, 23N to 76E, 39N, but it would be best to confirm. With that addition, I suggest you send your required parameters to gsfc-help-disc@lists.nasa.gov.   That is the data center, and they have staff who can address any issues when acquiring MERRA-2 data. In addition, in your message to them describe how you can handle the data. Can you use the netCDF format? Do you need comma delimited data? Are you expecting something else? 

best wishes,

gil

 

 

 

 

Hello Faisal Qureshi,

 

You can access the data for your requirements best with the MERRA-2 data subsetter:

https://disc.gsfc.nasa.gov/daac-bin/FTPSubset2.pl

You can select parameter, spatial and temporal subset. You can also select to regrid the data however since the data is 0.625 x 0.5 degree resolution I would only select a grid with a coarser resolution. 

To access the data you need to set up an Earthdata account (free) and add "NASA GESDISC DATA ARCHIVE" to the applications. The instructions are in the following links:

To set up your account and download data:

https://disc.gsfc.nasa.gov/data-access

to add the application to your account:

https://disc.gsfc.nasa.gov/earthdata-login

 

Please let us know if you have any further questions,

Dana Ostrenga

VK (not verified)

Tue, 03/07/2017 - 00:55

I am thinking of using MERRA2 for WRF but conducting research about this topic on the web has gotten me more confused.

Some webpages are indicating that you need to combine MERRA and another source of land data e.g. GLDAS

http://lundquistgroup.blogspot.com/2014/04/using-input-datasets-in-real-wrf-runs.html

On the other hand, there are numerous changes which directly use MERRA2 with out mentioning this requirement.

Can someone please explain the right way to configure MERRA2 data for use as WRF reanalyses dataset?

Thanks,

VK

Emanuele (not verified)

Fri, 01/27/2017 - 09:00

Dear Mike, 

I'm trying to extrapolate the roughness lengths from tavg1_2d_flx_Nx (M2T1NXFLX). I'm struggling to understand the datas, from the description the surface roughness should be in [m] but I can't figure how the value are so low. Could you give me some details or text me some references to better understand the values? Many thanks for that.

Just to let you know I'm assuming that a roughness length of 0.002 m is interpretable as a flat surface and a length of 0.8 as an high density metropolitan area.

Thanks a lot.

Emanuele  

Emanuele,

I can't say for sure if what you are seeing is low or not. The post does not include the values that you are seeing, or their location.

I picked a point in the central US (100W, 40N) and find z0h=z0m=0.0446857m

As far as I know now the documented units of meters for roughness are correct.

If your point is near water, you may have a predominantly water surface which would have small roughness.  Check the land fraction variable in the constants file.  If you are looking for high density metropolitan areas, keep in mind that the MERRA-2 resolition of ~1/2 degree would include values representative for the whole grid box.

Mike

KOSTAS PAPADOPOULOS (not verified)

Wed, 01/25/2017 - 07:24

I have just shifted from using MERRA reanalysis dataset to MERRA-2. My observation regards to the U50m, V50m winds from collection tavg1_2d_slv_Nx. The analysis area is Greece. I have noticed a large systematic change in the wind speed calculated for almost adjacent nodes of MERRA2 and MERRA. The "new" winds are about 25% lower than before. The correlation of the wind speeds are ~94% between the data sets but how is this large change explained.

 

Best regards

Kostas,

There were numerous model changes from MERRA to MERRA-2, and since 50m is within the lowest model layer, the model physics (especially Similarity profiles) would contribute significantly to the result. (to be sure you understand that a reanalysis is not an observation data set, but uses observations and models to produce the resulting data).  Aside from the model changes, this may also depend on the availability and assimilation of winds in the region.

I would suggest starting by reviewing the documentation available, which may provide more context fro the differences between MERRA and MERRA-2. https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/docs/

Next, I would suggest looking at the data provided that can contribute to wind speed. Both MERRA and MERRA-2 provide surface layer stability parameters, transfer coefficients and roughness lengths. These may show some systematic variation that explains the wind speed difference.

Another possibility is that the grid changed from MERRA to MERRA-2. Since you are in a region with both topography and coastlines, the grid points you are looking at may have different  geographical characteristics. You should check the constants file for the surface geopotential (altitude) and land/sea fraction. Both of which could dramatically influence surface winds.

Mike

Dear Mike,

Thank you for the prompt reply. Your comments were very helpful. I am aware of the grid point change between the two datasets; my comparison focused on nodes which were really close and the terrain was the same. If I have understood well, you imply that if the comparison is repeated with wind at 850hPa, then the results from the two datasets will converge. So what has changed is the parameterization of surface layer turbulence (scheme of Helfand and Schubert (1995) based on MO similarity vs Louis scheme. This would explain the noticeable deviation of near-surface winds.

Kostas

 

Kishore Kumar (not verified)

Wed, 10/05/2016 - 03:17

Hi 

I was trying to estimate trends in tropical middle atmosphere. The trends are very small compared to the earlier reports. I look back in the time series then I came to know that there was sudden shift in the winds around 2000. Does it something due to the input data?

I got doubt whether the MERRA is suitable for trend analysis or not.

 

kishore

 

Trends typically depend on the variable, and you don't say which variable or which reports.  Water vapor, for example, depends greatly on the availability of observations, which changes substantially over the typical satellite reanalysis period. All reanalyses will be subject to uncertainty there. Temperature of the lower troposphere (TLT), however, is fairly well observed, and the reanalyses tend toward the observed values reasonably well.  If you are talking about winds, then around 2000 MODIS AMV become available, and could affect time series analysis.

This isn't necessarily a problem for MERRA itself, any of the reanalyses, especially the full satellite observing system reanalyses (including MERRA-2, ERAI, JRA55 and CFSR) should be considered very carefully before assuming a physical forcing behind their trends.

Hi 

 

   I am looking on middle atmosphere temperature and horizontal wind trends over tropical latitudes. I used 72 level "inst6_3d_ana_Nv" product. Here with i am attaching link, https://sites.google.com/site/kishorekumargrandhi/merra-trends/Annualmean_temperature_8N.png, that shows annual mean temperature over 8 degree North. The annual mean temp. suddenly drop down after 2003. The temperature drop is very drastic. Is it real  or any kind artefact in the reanalysis?  

 

 

michael.bosilovich

Wed, 10/12/2016 - 13:31

In reply to by kishorekumargrandhi

Just to be clear, your first post states the question is about MERRA, but this is a MERRA-2 area of the www pages. The figure does not clear that difference, except the path says "merra-trends".

But to answer your question, that is not a trend in the figure, it is clearly a discontinuity, and so not to be trusted. My first suggestion in cases like this is double check your coding to create the figure and data. MERRA does not include monthly or annual data on the model vertical coordinate (Nv), so you likely created these data files on your own. So, it is worth double checking to be sure that the code has not introduced the discontinuity.

There are two possibilities, for a data issue in MERRA to occur at 2003. First, In the early part of the analysis, SSU radiances were assimilated with the GLATOVS radiative transfer  model, not the CRTM as AMSU data were assimilated. At the boundary between these, there is a problem, because the SSU were not handeled well by the older GLATOVS code.  The other less likely possibility is that AIRS is having some effect there. I only mention it because AIRS will be in full year of 2003.

In either case, MERRA is not the data you want to be using right now. MERRA-2 is available and has a newer CRTM that can handle SSU more seemlessly with AMSU.  There isn't an Nv collection for monthly temperature in MERRA-2, either.

 

To access MERRA data (for example const_2d_asm_Nx, inst6_3d_ana_Np, tavg1_2d_rad_Nx, tavg1_2d_slv_Nx), I developed an Excel Tool in VBA, which allows the download and the conversation to Excel format for selected subsets of this data.
Now I want to enhance this tool to MERRA-2. But this data are password protected and I don't know, how and to which address the login information has to be transmitted from the Excel-VBA procedure to the MERRA-2 Database. Exists a help Information for VBA programmers or generally for Software developers, how to solve the login procedure to access the MERRA-2 database from Software tools, developed for such purposes?
It's no Problem for me, to download the needed data in numerous single steps with the web browser, but this way of conversation by hand into the calculation tools in Excel is much too time consuming and hardly practicable.

Thanks for help an best regards, Matthias Popp    

Add new comment

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