Discrepancy in ERA5 100m Offshore Wind Speeds

44 views
Skip to first unread message

Umed Paliwal

unread,
Aug 13, 2021, 10:28:00 PMAug 13
to openmod-i...@googlegroups.com
Hi All,

I am trying to do some analysis on offshore wind potential in the US and China. I am using 100m wind speeds from ERA5. I found a very weird pattern in the offshore wind speeds, the wind speeds always drop significantly from 9AM to 10 AM and 9 PM to 10 PM. This happens in all offshore locations whether the lat lon is on the coast of US, China, Japan and India. I have pulled out the data for 6 of the offshore points from across the globe and plotted the hourly wind speeds and diurnal mean in this tableau visualisation. You can see this weird pattern clearly in diurnal means and on most days in the hourly 8760 data.  This pattern doesn't show up when using the onshore wind speed data. Has anyone else noticed this? 

Any leads would be highly appreciated.

Best,
Umed   


Donna Peng

unread,
Aug 14, 2021, 2:42:05 AMAug 14
to Umed Paliwal, openmod-i...@googlegroups.com
Hi Umed,

I have come across this before in this paper:

The most relevant section is below.

4.1 Issue in ERA5 diurnal cycles

The author discovered an issue with the diurnal cycles of wind speeds in ERA5. The wind speeds at 10 m, 100 m and also on lower model levels, exhibit gaps at 10:00 UTC and following hours and similarly but in a lesser extent at 22:00 UTC. The wind deficit at 10:00 UTC may be larger than 0.5 m s−1 on average. 

The issue was reported to ECMWF which acknowledged the problem, which is linked to the assimilation process, affects mostly low latitude oceanic regions but also Europe and North America, and is not possible to correct (ERA5 data documentation2020). The issue appears at both 10:00 and 22:00 UTC but is more visible in convective atmospheric conditions thus appearing mostly at 10:00 UTC, during the day, over continental France.

To address this issue, ERA5 forecasts (denoted “fc” hereafter) were downloaded and studied in addition to the usual analysis data (denoted “an”). The forecasts do not exhibit the issue at 10:00 and 22:00 UTC but there may be jumps at 07:00 and 19:00 UTC when moving from one forecast stream onto the next. These jumps are small or not visible over continental France.

Best,
Donna 

On 14 Aug 2021, at 03:28, Umed Paliwal <paliwa...@gmail.com> wrote:


--
You received this message because you are subscribed to the Google Groups "openmod initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openmod-initiat...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/CANMsCvC9UxR%3Df1Rp4M%2BWecqboMFQwsUj5o2EGVcH7-tRwhs3Lg%40mail.gmail.com.

Matteo De Felice

unread,
Aug 14, 2021, 7:10:09 AMAug 14
to openmod initiative
Dear Umed,
as Donna pointed out, this is a well-known issue due to the assimilation cycles. It's quite annoying and the only solution is using the forecasts rather than the analysis,
Best regards,

vel...@gmail.com

unread,
Aug 15, 2021, 1:33:02 AMAug 15
to Matteo De Felice, openmod initiative

Dear All,

 

I am doing some PhD work on offshore CAES for WT and using the Sesame DNVGL software.

 

I know this has nothing to do with this trail but since I am seeing the group so helpful I though of promoting maybe you can guide me to some similar groups about my studies .. especially if the Sesame DNV software is being used.

 

Apologies for the intrusion but maybe you can help ..

 

Thanks,

Peter Vella

Johannes Hampp

unread,
Aug 16, 2021, 5:43:03 AMAug 16
to Matteo De Felice, openmod initiative
Dear Donna, Matteo, all,

Thanks for pointing out this issue and the reference.
Can either of you elaborate on which product you mean with "ERA5
forecast" / where we can find information on this dataset?
It seems to elude at least my online search queries.

Thanks & Best,
Johannes


Best regards,
Johannes Hampp (he/him)

Justus Liebig University Giessen (JLU)
Center for international Development and Environmental Research (ZEU)

mailto: johanne...@zeu.uni-giessen.de

Office 110
Senckenbergstr. 3
DE-35392 Giessen
https://uni-giessen.de/zeu

Am 14/08/2021 um 13:10 schrieb Matteo De Felice:
> Dear Umed,
> as Donna pointed out, this is a well-known issue due to the assimilation
> cycles. It's quite annoying and the only solution is using the forecasts
> rather than the analysis,
> Best regards,
>
> On Saturday, August 14, 2021 at 8:42:05 AM UTC+2 peng...@gmail.com wrote:
>
>> Hi Umed,
>>
>> I have come across this before in this paper:
>> https://asr.copernicus.org/articles/17/63/2020/
>>
>> The most relevant section is below.
>> 4.1 Issue in ERA5 diurnal cycles
>>
>> The author discovered an issue with the diurnal cycles of wind speeds in
>> ERA5. The wind speeds at 10 m, 100 m and also on lower model levels,
>> exhibit gaps at 10:00 UTC and following hours and similarly but in a lesser
>> extent at 22:00 UTC. The wind deficit at 10:00 UTC may be larger than 0.5 m 
>> s−1 on average.
>>
>> The issue was reported to ECMWF which acknowledged the problem, which is
>> linked to the assimilation process, affects mostly low latitude oceanic
>> regions but also Europe and North America, and is not possible to correct
>> (ERA5 data documentation
>> <https://asr.copernicus.org/articles/17/63/2020/#bib1.bibx9>, 2020
>> <https://asr.copernicus.org/articles/17/63/2020/#bib1.bibx9>). The issue
>> appears at both 10:00 and 22:00 UTC but is more visible in convective
>> atmospheric conditions thus appearing mostly at 10:00 UTC, during the day,
>> over continental France.
>>
>> To address this issue, ERA5 forecasts (denoted “fc” hereafter) were
>> downloaded and studied in addition to the usual analysis data
>> (denoted “an”). The forecasts do not exhibit the issue at 10:00 and 22:00 
>> UTC but there may be jumps at 07:00 and 19:00 UTC when moving from one
>> forecast stream onto the next. These jumps are small or not visible over
>> continental France.
>> Best,
>> Donna
>>
>> On 14 Aug 2021, at 03:28, Umed Paliwal <paliwa...@gmail.com> wrote:
>>
>> 
>>
>> Hi All,
>>
>> I am trying to do some analysis on offshore wind potential in the US and
>> China. I am using 100m wind speeds from ERA5. I found a very weird pattern
>> in the offshore wind speeds, the wind speeds always drop significantly from
>> 9AM to 10 AM and 9 PM to 10 PM. This happens in all offshore locations
>> whether the lat lon is on the coast of US, China, Japan and India. I have
>> pulled out the data for 6 of the offshore points from across the globe and
>> plotted the hourly wind speeds and diurnal mean in this tableau
>> visualisation
>> <https://public.tableau.com/app/profile/umed8052/viz/ERA5100mWindSpeedERA5/DiurnalWindSpeed?publish=yes>.
>> You can see this weird pattern clearly in diurnal means and on most days in
>> the hourly 8760 data. This pattern doesn't show up when using the onshore
>> wind speed data. Has anyone else noticed this?
>>
>> Any leads would be highly appreciated.
>>
>> Best,
>> Umed
>>
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "openmod initiative" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to openmod-initiat...@googlegroups.com.
>> To view this discussion on the web, visit
>> https://groups.google.com/d/msgid/openmod-initiative/CANMsCvC9UxR%3Df1Rp4M%2BWecqboMFQwsUj5o2EGVcH7-tRwhs3Lg%40mail.gmail.com
>> <https://groups.google.com/d/msgid/openmod-initiative/CANMsCvC9UxR%3Df1Rp4M%2BWecqboMFQwsUj5o2EGVcH7-tRwhs3Lg%40mail.gmail.com?utm_medium=email&utm_source=footer>
>> .
>>
>>
>

ki...@fias.uni-frankfurt.de

unread,
Aug 16, 2021, 5:57:47 AMAug 16
to Johannes Hampp, Matteo De Felice, openmod initiative
Dear Johannes,
most ERA5 variables (among them 100m wind speeds) come as analysis and
forecast, for the differences see this page
https://confluence.ecmwf.int/pages/viewpage.action?pageId=85402030
and how to access, as well.
Best
Alex
> https://groups.google.com/d/msgid/openmod-initiative/35a02c24-3740-24b8-69ab-9c3506c167c7%40zeu.uni-giessen.de.
>


Matteo De Felice

unread,
Aug 16, 2021, 9:35:03 AMAug 16
to openmod initiative
As specified in the "Known issues" page of ERA5 (https://confluence.ecmwf.int/display/CKB/ERA5%3A+data+documentation#ERA5:datadocumentation-Knownissues): 
" The forecast near surface winds are available from MARS, see the section, Data organisation and how to download ERA5."

AFAIK, the forecasts (stream='fc') are available only on MARS. You can start from here: https://apps.ecmwf.int/data-catalogues/era5/?stream=oper&levtype=sfc&expver=1&month=jan&year=1985&type=fc&class=ea

Good data hunting! 

Umed Paliwal

unread,
Aug 16, 2021, 3:48:10 PMAug 16
to Matteo De Felice, openmod initiative
Thanks everyone for your input !!!

Best,
Umed



--
Umed Paliwal
Master of Public Policy Candidate, 2019
Goldman School of Public Policy | University of California, Berkeley
Reply all
Reply to author
Forward
0 new messages