This list describes the status of some feedback and requests put forward by users during the User Voice Corner feedback session at the ECMWF UEF meeting held in June 2019. Items are included here mainly to provide some useful clarification and/or an update where significant progress has already been made. Many more items are being worked on or considered for future action. References are made below to cycle 47r1; this new IFS model version went live on June 30th 2020. This update was delayed by the COVID outbreak. ECMWF is also working on feedback provided at UEF 2020, the intention is to provide a web page describing progress on those items early in 2021.
- Re-analyses and Re-forecasts - big demand
The ERA5 re-analysis data, from 1978 until 2019, went live in January 2019. Resolution is 31km compared to 80km for ERA-Interim. Late in 2019 ERA5 re-analysis data from 1950-1978 will also be released. Re-forecasts, which currently use ERA-Interim analyses as the basis for their initialisation, will start using ERA5 analyses for this purpose with cycle 46r1. (ERA5 versus ERA-Interim)
- Can 1 year of re-forecasts be made available immediately with each new cycle ?
This is technically not possible at present with current supercomputer resources. It is also important to stress that model changes between cycles are not usually that great, and so model characteristics do not commonly change that much. This means that old re-forecasts are not as redundant as they might appear.
- Provide 06Z and 18Z BC forecast runs, and 1-hourly IFS data where available
ECMWF tabled a motion to Council in late June 2018. seeking approval for these to be disseminated for a fee. This motion was passed and the data was first made available in October 2018. (licenses for ECMWF data)
- Can the IFS model issue issue of snow on the ground not melting quickly enough be fixed ?
This is an active area of research; a new 5 layer snow scheme has proven beneficial in addressing this and other near surface weather-related issues, although further tests are still needed before operational implementation becomes viable. (Model issues)
- Dissatisfaction in some quarters with the ECMWF method of computing CAPE
In response to this being repeatedly voiced, ECMWF has set up a short collaborative project with ESSL (the European Severe Storms Laboratory) to review computation of CAPE in the IFS and related usage in member and co-operating states. There will be some changes to the "max CAPE" and "max CAPE-shear" output in 46r1.
- Why are winds over mountains underestimated ?
The Verification team have been tasked with investigating and documenting this issue.
- Can ECMWF improve its visibility diagnostic ?
Dynamic aerosol-moisture interaction is an important missing link. Parallel work on the atmospheric composition version of the IFS (C-IFS) should in the long term (years) help address this issue. In the meantime we continue to investigate cases and document findings. Users should also note that some C-IFS output is now available in ecCharts which can help, for example, with the prediction of reduced visibility in duststorm/sandstorm situations. (ecCharts update that included CAMS data e.g. for dust)
- Winds wanted at more near-surface levels, for wind energy applications
The utility of providing data for other levels, in addition to the 100m level data already disseminated, has been studied; the conclusion was that "200m winds" were sufficiently different to justify inclusion, so these will be made available for HRES and ENS with cycle 46r1 (u,v). 50m winds provided little extra information, partly because of lack of ENS levels near the surface, and so will not be added.
- More aviation-related indices wanted
ECMWF has been researching turbulence and icing diagnostics, but these will not be introduced in 2019.
- Can ECMWF make ecCharts refresh faster?
This long-standing problem is being addressed via new cloud and map-drawing technologies (one component is similar to the tiles used by google maps). We plan to advertise to users a pre-operational version of this at the June 2019 UEF meeting.
- Request for variables on the "PV=1.5" level
Will be added in cycle 46r1: variables will mirror those already available for the PV=2 level.
- Users want more fire and drought-related products
Operational fire danger products are produced at ECMWF as part of the Copernicus Emergency Management Service, in the European Forest Fire Information System (EFFIS) . Drought-related products are not being actively developed at the current time, due to resource limitations. (EFFIS)
- Better MARS documentation wanted
ECMWF reviews regularly the MARS user documentation. A more substantial review of the MARS request language is planned for this year as part of consolidating the MARS and Dissemination request languages. (MARS user documentation)
- Faster MARS access wanted
- Rationalise computation of CAPE and other convective indices in various ways
For cycle 48r1 (late 2022?) preparations have been made to introduce several new CAPE-type diagnostics, following a collaboration with ESSL (the European Severe Storms Laboratory). In Cycle 47r1 (June 2020) the method of computation of convective inhibition (CIN) was markedly improved, and it was also made consistent with the CAPE computation level (for more details see here).
- 2m temperature biases, of different types, noted by many users
Following a cross-departmental project at ECMWF - USURF - biases in near surface parameters are now much better understood, and various initiatives are underway to reduce these (see e.g. here). ECMWF is also exploring, in an EU-funded project, how ecPoint post-processing can help with error-source identification and correction. Introduction of urban tiles (for heat island effects) is also part of ECMWF future plans.
- Requests for better visualisation of extended range and long range (=monthly & seasonal) output
Although this will not satisfy all requests, ECWMF has recently introduced a web-product-based CDF-intercomparison tool for weekly mean 2m temperature and weekly total precipitation, to see how handling by successive extended range forecasts compares, for a user-selected (clickable) site.
- Can we have more cluster-style products?
ECMWF has now moved from experimental to operational status a number of regime-related products, for both medium and extended ranges. These effectively cluster output according to standard regimes (such ad NAO+) and display the output in new and innovative ways.
- ecCharts is still slow!
ECWMF made some structural changes to ecCharts, focused on this particular problem, and whilst performance is still not perfect it seems that the number of users complaining has diminished as a result.ECMWF has increased capacity with new disk technology. Some data, for example re-analyses, can be accessed more quickly via another route - the CDS (Climate Data Store). Users are encouraged to use this channel wherever possible. (downloading ERA5 data through the CDS)