View Source

Will the real-time file-naming convention stay the same?
Yes, there will be no changes to the file-naming convention, see File naming convention and format for real-time data.
Will the real-time dissemination schedule stay the same?
Yes, currently we are not expecting any changes to the Dissemination schedule.
Will changes made in our test configuration be carried over into production?
Yes, due to the nature of the upgrade, ECMWF will transfer all installed requirements in the test PREd to the production system on the day of the cycle implementation. During the RCP, any changes made in the production system should also be installed in the test system to ensure continuity.
Is it possible to have automatic dissemination of test files for 48R1?
Yes, but we recommend having a delay between the operational files and the test files. Interested users are invited to contact Data Support team via the Support Portal. Please note that test files are not guaranteed and may not arrive according to the schedule.
Please also note that this is not possible for Basic or Bronze service packs.
Will GRIB1 fields be impacted by the new CCSDS packing?
No, GRIB1 fields will not use the new CCSDS packing.
Will the ENS resolution be automatically increased to 9km?
In dissemination users will need to adjust their requirements or notify Data Support team via the Support Portal. If in MARS the resolution is not explicitly specified for 48r1 data, users will automatically get the higher resolution.
What will the file size increase be when the ENS becomes 9km?
Without the CCSDS packing, the file size of lat/lon data requesting the highest resolution will be 4x larger. This is due to twice the number of points of latitude and longitude.
Will the ENS-Ext have daily re-forecasts/hindcasts?
No, re-forecasts/hindcasts will be produced twice weekly as per the current configuration.
What will the ensemble mean (mean value derived from all the ensemble members) be calculated with?
With the implementation of 48r1 post-processed products (means, probabilities etc. ) from ENS (stream enfo) will continue to be derived from 51 members (50 perturbed forecasts + 1 control) whereas those products derived from ENS extended (stream eefo) will be processed from 101 members (100 perturbed forecasts + 1 control).
How can I specify to get ENS-Ext products for specific days?
This can be done with keyword 'use'. As before, users can specify use=monday/thursday, use=monday or use=thursday. Other weekdays e.g. 'tuesday' are not valid except for extended range weekly means. If keyword 'use' is not specified, the ENS-Ext products will be delivered daily.
Is it possible to stay with 50 members for the ENS-Extended?
Yes, in general users can choose any number between 1 and 100. However, commercial users can choose to purchase either 50 or 100 members. The Information Cost will be the same irrespective of the number of members, but the Volume Band for 100 members will be higher.
If I decide not to take all 100 members of ENS-Extended, will the first 50 members be a reasonable distribution?
Yes, the first 50 members will give a reasonable distribution as in the current system. As we have only 5 perturbed ocean analyses, ensemble members 1, 6, 11, 16, ... have the same ocean initial conditions. Therefore ensemble members 1 to 50 will give a good sample of the ocean initial conditions. Commercial users can request either 50 or 100 members.
How will the new ENS-Extended fields be charged?
For commercial users, the daily ENS-Ext fields will be charged at the current rate of twice per week for Information Cost. Users should be aware that adding the daily fields will increase the daily volume produced and thus the Volume Band may increase. Users who continue to adopt Monday and/or Thursday will pay the same.
What will the cost of our new data be after the implementation?
ECMWF will release an experimental version of the Product Requirements Catalogue in due course. With this, you will be able to determine the cost of datasets after the cycle change.
I am using reforecasts for calibration. What should I do after implementation of 48r1?
If you use past reforecasts e.g. for the calibration of Cycle 48r1 data, you need to make sure that these reforecasts are also from 48r1. As the reforecasts are produced two weeks in advance there are already reforecasts in MARS for dates after the 48r1 implementation which have been produced with the previous IFS Cycle, 47r3. To help with the migration, these reforecasts will be "backed up" as expver=2 and overwritten with those from the 48r1 e-suite so that all expver=1 reforecasts for dates after the implementation will be based on Cycle 48r1.
Let's illustrate that with the medium range model climate (M-climate) which is constructed from 9 consecutive re-forecast sets (covering a 5-week period), the middle one of which corresponds to the preceding Monday or Thursday that is closest to the actual ENS run date. With implementation on 27 June the model climate for 13 July will be the first which can be consistently computed entirely from reforecasts with expver=1 whereas the M-climate in the period 29 June to 10 July will need be computed from a mixture of expver 1 and 78.