ERA-Interim: Notes, Questions, and Discussion

Created by mapr8844 on - Updated on 09/01/2020 19:08




Direct dataset questions to ECMWF User Support

Note: the following three comments have been added to this page due to them not migrating during our upgrade:


Re: ERA-Interim: Notes, Questions, and Discussion

Submitted by Natalie T (not verified) on Tue, 07/12/2016 - 13:14. 

Dear all, Is acceleration due to gravity already accounted for in the vertical integral of northward/eastward water vapour flux? I didn't find mention of anything of the sort in the documentation, but didn't want to repeat the action in my analysis if it has already been done. Thank you in advance.

Re: ERA-Interim: Notes, Questions, and Discussion

Submitted by Namaoui (not verified) on Tue, 07/12/2016 - 01:58. 

Hello, Iam Beginner in Era interim , I want to know how download a frid of Era interim ? Regards

ERA-Interim: Questions

Submitted by Johanna Yepes (not verified) on Fri, 05/13/2016 - 09:33. 

Dear Era-Interim staff, I would like to ask about the units of the runoff data. I downloaded the monthly means of runoff and I can see that the units are meters but I don't know the time units for that, those are meters per day or meters per month? Thank you.

YEONWOO CHOI (not verified)

Thu, 10/22/2020 - 14:39

Dear, Sir/Madam

I have question about the surface downward UV radiation data (i.e., UV-B).

I found that the range of UV values extracted from ERA-INT and ERA5 is different, even though their units are the same (W/m2).

For example, in Jan 1979, global UV-B derived from ERA5 ranges from 0 to 45, but that derived from ERA-INT ranges from 20 to 130.

What is the exact difference between ERA-INT and ERA5 data on the value of the UVB?







thank you for your interest in ERA5 and ERA-Interim.

The difference between ERA5 and ERA-Interim is that for fluxes (and other accumulated quantities) is that for ERA5 accumulations are since the previous post-processing step (hourly for ERA5 and 3-hourly for its uncertainty estimate) while it is since from the start of the originating forecast for ERA-Interim.

More info can be found at:


Wed, 04/22/2020 - 09:18

How do we process ERA-Interim data in Matlab and obtain the plot for any of the parameters. Have downloaded the NetCDF file and was cable of reading the file but unable to plot the data. I am a Research Student working on Trend analysis of Total column Water Vapor.


ERA-Interim provides the parameter "Total column water vapour" (code 137.128). In addition, there is the parameter "Total column water" (code 136.128). Both of these parameters can be downloaded from the webpage at sub-daily temporal resolution. They are also available as monthly means, see the menu option "Monthly means of daily means" on the left hand side of the above page.



lais schmalfuss (not verified)

Fri, 12/07/2018 - 06:33


Somebody could help me with this?

I downloaded data from the Interim ERA and forgot to rename it.
Now I know the data years, but I do not know which months each data represents.



Dear Lais,

If you retrieved GRIB data, the date information (or century, year, month, day), time and length of the forecast, for forecast data, is contained within the GRIB header of each message. If you retrieved netCDF data, one of the coordinates of the file, is time since the reference date/time.



Thu, 07/12/2018 - 15:21

I used to get the last date of the monthly ERA-Interim reanalysis by going to a ERAInterim monthly webpage and noting the last month listed. Now I don't see a last month. How can I find the last month available? Is there a page that has it? Tx.

Dear Cathy,

The page you are looking for is the page for daily data:

The data are updated on a monthly basis, so if the daily data is available for a month, then the monthly mean data will be available too.



I see the last date again. But I have had no luck getting the (monthly) data via a python script for the last 2 months or so. The code was working and suddenly just hangs. I suspect it's an overloaded server. I was able to get a file via the web interface though it took a while. That's not a great option for me but I could try to make it work. 


We have looked at this and we can download data. Could you give us more details about what you are trying to obtain and how?

The easiest way to do this might be to create an issue on the C3S website: go to, click "login to the C3S Enquiry Portal", then "Any question" and fill in the form.

When the problem is resolved, we can post a summary here.



Kyle Itterly (not verified)

Fri, 05/25/2018 - 12:01


I have a question and a quick note to those looking at the diurnal cycle of accumulated forecast parameters in general.

Let's use total precipitation as the example. I seem to be getting a few negative values for 3-hourly precipitation rate when subtracting the 3-hourly accumulated steps and dividing by the interval. It turns out that it's relatively rare and the non-negative values appear to match other results and to be of sufficient accuracy. Applying this same methodology to other accumulated parameters e.g. radiative fluxes appears to be fine. My theory is that bilinearly interpolating the accumulated precipitation (before subtraction of steps) to 1 degree by 1 degree (done by the WebAPI server) is the core of my issue. So, what would be the proper resolution and order of computation to ensure the 3-hourly precipitation rate is being calculated as intended?

My note to others looking at the diurnal cycle of ERA forecast parameters is to consider using the WebAPI so you can concatenate later forecast steps 6-18 hours from 00 and 12 UTC instead of using the first few steps 3-12, which can be affected by model spinup errors. Additionally, using forecast steps 6-18 hours provides a full diurnal cycle of model CAPE without erroneous values of zero in forecast step 3. This took me a while to figure out, and it turns out that the diurnal cycle of some parameters (e.g. OLR) appears to be less biased using the later steps.

Thanks in advance for any insight,


David Carvalho (not verified)

Tue, 03/27/2018 - 15:31

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

Also, does ERA-Interim 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!

Paul Berrisford (not verified)

Mon, 09/04/2017 - 11:37

Dear Andrew,

From the data, you could calculate the geopotential height on the model levels and interpolate the winds to 100gpm. This would give you the most accurate answer. For calculating geopotential, see equation 2.21 in Part III "Dynamics and numerical procedures" of the ERA-Interim model documentation:

To a first approximation the near surface model levels in ERA-Interim are at a constant height above the surface. However, that approximate height is proportional to temperature (in K), so the standard atmosphere gives slightly different results to those assuming a uniform scale height, see:

With the standard atmosphere, 100m is located between levels 58 and 57, though slightly closer to 57. So you could assume the standard atmosphere and interpolate between these two levels.

If you're happy with the discrepancy due to temperature, you could just use the winds at level 57, following the result from a 7km scale height.




Hi Paul,

Thanks for your reply, and sorry it's been awhile. I ended up doing some sensitivity tests for different model levels around 100m on the annual time scale (yearly means) for the geographic area of Texas. From what I can tell, there is no discernible difference between the different methods (using model level 57 versus interpolation of closest model levels). 

Thank you for your help.



Andrew Kumler (not verified)

Tue, 08/22/2017 - 16:42


I am currently doing research using ERA-Interim on hub-height winds (100 meters). ERA-Interim outputs data at model levels, one of these being model level 57 (1000 mb, or 100 m agl). This assumes a surface pressure of 1013.25 mb and scale height of the atmosphere of 7 km. Would this be an appropriate level to use to assess 100 meter winds over a large geographic area, or should alternative levels be used? The wind power law can be used to calculate 100 meter winds using winds at a lower heights as well. Or should I use geopotential heights to account for changes in pressure at 100 m agl? I assume changes in geopotential at 100 m would be fairly negligible over long periods of time.



Dear Cathy,

as Paul said, there is no automatic way of checking the latest available date. The ERA-Interim release cycle is described here:



Wed, 04/12/2017 - 08:52

How can I automatically find the last date of the ERA-inerim? If I request a date past the last date, the ecmwf perl module  fails with many errors. I can look at the webpage but then I can't update automatically.


Dear Cathy,

There is no automatic way of checking the latest available date. In order to find this date, you need to look at the ERA-Interim data webpage:



Lu Sun (not verified)

Thu, 11/17/2016 - 12:05

Dear Sir/Madam,

I am wondering how the CAPE in ERA-Interim data be calculated? Because I want to compared the results with other reanalysis data but most of them need us to calculate CAPE individually. Some important information such as: what level should be treated us the beginning point (1000mb? 1017mb? or the surface pressure?) Thank you for your replying.


Dear Lu Sun,

There is a description of the definition of CAPE in the ERA-Interim model documentation, Part IV: Physical processes:

In particular, look at equation 5.61 in Section 5.11 Diagnostics for post processing: CAPE, which is in Chapter 5 Convection.

The calculation is done from model (hybrid sigma/pressure) level data.



Add new comment

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