Home

Report on ECMWF User Survey

1. 70 of 141 29 06 05 17 21 ECMWFE User Survey Sorry I did not mean that I am not satisfied I just don t use it very often well I m speechless not applicable Web Services In this section we would like to find out about your use of the web services available at www ecmwf int How often do you access the ECMWF web site Each Respondent could choose only ONE of the following options Response Total of Total Respondents Daily su Ey 13 Several times a week 9 22 Once a week 16 Es 19 Occasionally 184 SS jas Never a ca Other please specify e 0 Total Responses 406 0 20 40 60 80 e Other please specify now and then sometimes Overall how useful do you find the ECMWF web site Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very useful 162 i 41 Useful 217 ii 55 Neither useful nor useless s i 4 Useless 2 1 Very useless o 0 Total Responses 396 0 20 40 60 80 Your comments pE free data are very limited and the data policy is too rescrited both for research people and public people 71 of 141 29 06 05 17 21 ECMWF User Survey used to be very bad and has been greatly improved in the past couple of years I found it quite useful for checking how to run jobs etc excellent site in all ways You appear to be limited in receiving information on ensemble products for
2. I have not used the service however I know very well of its existence Both User Support and Call Desk are always very helpful and if they can t answer my question pass it quickly to someone who can The Support resolved quickly my difficulties to run my jobs Sometimes I write with some difficulties my email in Shakspeare language and the use support have no difficulties for answer but her mother tongue is the same as me it s a pity not very satisfied of our contact point Dominique Lucas and his helpful support Areas for improvement In which of the following areas would you like to see an improvement to the service we offer Please comment Please tick each area where you want to see improvements and provide further details comments in the text boxes Each Respondent could choose ANY of the following options Response Total of Total of Total Respondents General purpose server General purpose server ecgate General purpose server ecgate 55 2 High Performance Computing Facility 21 HPCF Meteorological archive system MARS 7 26 Services offered on the website 53 Co 19 Information documentation available 72 26 on the website User Services User Support Call Desk 8 3 111 of 141 29 06 05 17 21 ECMWE User Survey Total Responses 281 0 20 40 60 80 General purpose server ecgate Each Respondent could choose ANY of the following options Response Tota
3. ECMWF User Survey The ease of access larger and more complete data accessibility for non MS Nat Met Serv more frequent updating More forecast products available to certificate access Availability of the daily forecast for every user from day 1 and not from day 3 easy way of finding information Higher resolution model output searching available data documentation on model ECMWF courses and seminars Verification of weekly and seasonal forecast More information on forecasts to less than 72 hours Searching forecast and EPSgramas a better interface to current weather forecast information and long range forecasts A seamless forecast view different tasks its fine as it is Better index and clearer links Forecasting fields in GRID code Diversity of software in Fortran to be free If I earn money for each time Having more time in my actual work to dedicate into work with fields and observations perhaps in the coming future Funding and time Forecast products that compete in timeliness and resolution with those of the national weather services If T d use the ECFS and the HPCF more frequently I d probably use more of the documentation More detailed forecast data similar to the output provided to Reading University each Friday A paper newsletter with details I receive too much electronic mail to the extent that I now ignore all except that which appears relevant A newsletter is more portable than a PC
4. ECMWF product are used to force the RCM and to validate the results At the moment my main interest in ECMWF data is in retrieving wind fields for trajectory calculations using a modified version of FLEXPART A Stohl that takes into account diffusive dispersion to study mixing properties Access to short term forecasts is particularly important implementation of satellite products specific studies of the tropopause weather forecaster during holiday season ECMWFE Operational Model data and ERA 40 data in the context of glacier mass balance modelling in the low latitudes Analysis of interannual variability of climate and predictability studies We retrieve ECMWF data to run our Particle transport model to simulate atmospheric transport processes Modelling tropospheric aerosol dynamics at the Global and Regional scale by global Chemistry Transport models remote sensing applications We are working with climatic extremes middle atmosphere variability ENSO variability And we have been using ERA40 reanalysis data Postproceso estadistico de datos del EPS del ECMWF para obtener predicciones de variables meteorologicas en la superficie T2m T2d U10 V10 etc Unix and Linux system administrator and user support maintenance of ECMWE software and tools installed in my systems Climate variability Was supporting and developing Metview but have since moved onto a new job developing web services In FORMOST the Monthly forecast
5. RSS feed or a feature like WeatherFox availability of frequent reports on the operational aspects of the models problems performance verification results error tracking results upgrades etc which can help to better interpret model outputs larger online data access Mostly needs of work I don t spend too much time in the Internet outside work Documentation easier way of GRIB data decoding more detailed analysis of forecasts e g forecasts for each country in Europe If I would be able to retrieve large amount of data several years or at least a year from MARS archive like from Data Server 89 of 141 29 06 05 17 21 ECMWF User Survey nothing in particular when I think the info I need is on the ECMWF web site I start searching More forecast products available online Higher facility of navigation Discussion on interesting events as they happen i e particularly good or bad IFS performance Having access to some informations without logging in My most frequent use is for documentation Navigation there is not always straightforward pdf documents do not always display well in my browsers Perhaps the search facilities can be reviewed and pdf more often converted to html reanalysis data ERA 40 at 0 5 X 0 5 resolution more examples about retrieval and processing if a graphical interface giving the short term forecasts I will visit the web site everyday Documentation about training course and weather forec
6. Some analyses to use as initial conditions Daily archives to be used as boundary fields in limited area models is the main interest Tropical storms trajectories TOGA data Only need for checking Metview COSMO LEPS Bologna IT Member state projects prepIFS in my case observations MAP SOP ERA Depending on what kind of observations I can retrieve that might be interesting Occasionnally data from other centres wind field for Global Atmosphere Watch Network Other Map reanalyses my own files on ecfs More specific types of data Please specify the importance of the data type or tick Not relevant e Analysis Each Respondent could choose only ONE of the following options Response Total of Total Respondents High eS 87 Not relevant E Total Responses 359 0 20 40 60 80 e Deterministic forecast Each Respondent could choose only ONE of the following options Response Total of Total Respondents 56 of 141 29 06 05 17 21 ECMWFE User Survey High 221 67 17 mE Low 54 eS 16 Not relevant 56 E Total Responses 331 0 20 40 60 80 BBE e Ensemble forecast Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 120 39 Not relevant 112 Total Responses 305 0 20 40 60 80 e Monthly forecast Each Respondent could choose only ONE of the following options Response Total of Total Respon
7. 3 EE 15 Low 39 FE 18 Not relevant 147 i 67 Total Responses 219 0 20 40 60 80 E e Not necessary for your work i e serial programming is adequate for your needs Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 127 51 Low Fai 18 Not relevant 6 an 31 43 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 247 0 20 40 60 80 e Difficult to access Each Respondent could choose only ONE of the following options Response Total of Total Respondents High lo i 4 Low 3s zy 17 Not relevant 163 P 79 Total Responses 207 0 20 40 60 80 BE e Difficult to use Each Respondent could choose only ONE of the following options Response Total of Total Respondents High h2 6 Low 3 ey 16 Not relevant 157 i 7 _ 78 Total Responses 202 0 20 40 60 80 e Lack of training Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 39 Es 18 Low 47 eS 22 BE e Lack of adequate assistance Each Respondent could choose only ONE of the following options Response Total of Total Respondents High s E 4 Low m Ea 16 Not relevant 160 30 Total Responses 201 0 20 40 60 80 44 of 141 29 06 05 17 21 ECMWFE User Survey e Lack of debugging tools Each Respondent could choose only ONE of
8. Response Total of Total Respondents Very satisfied 14 i 69 Satisfied 80 SS 28 108 of 141 29 06 05 17 21 BE ECMWFE User Survey Neither satisfied nor dissatisfied s 3 Dissatisfied o 0 Very dissatisfied o 0 Total Responses 282 0 20 40 60 80 e Clarity ability to explain Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 13 i 61 Satisfied 9 Ss 35 Neither satisfied nor dissatisfied u i 4 Dissatisfied oO 0 Very dissatisfied o 0 Total Responses 283 0 20 40 60 80 pE e Accessibility availability Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 166 M 59 Satisfied 8 eS 35 Neither satisfied nor dissatisfied e al 5 Dissatisfied 2 1 Very dissatisfied o 0 Total Responses 280 0 20 40 60 80 pE e Response to special requests Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 155 iii 57 Satisfied s SSS 30 Neither satisfied nor dissatisfied 32 Ea 12 Dissatisfied 2 1 Very dissatisfied o 0 Total Responses 270 0 20 40 60 80 HBE 109 of 141 29 06 05 17 21 ECMWE User Survey e Understanding of your local environment Each Respondent could choose only ONE of the following options Response Total
9. The telnet connection is very often slow and thus interactive activity is quite boring from time to time This may be not a Centre s problem and rather a bad priority given to telnet vs ftp for example A personal experience Before using a local ecaccess server using direct ecaccess ecmwf int access I thought the X connections were not allowed It was a pity since I made some work otherwise I could have done In other words a user may think erroneously after tests not giving the waited result First of all the messages should be as informative as possible It is a difficult matter my need only administrative tasks in interactive mode I have not used it enough to have an opinion How easy do you find ecgate to use Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very Easy 76 ae 20 Easy 204 52 Neither easy nor difficult 9 24 Difficult e 4 Very Difficult h 1 Total Responses 389 0 20 40 60 80 e Your comments BDE Baffling to start with straightforward once I got the hang of it Your documentation is among the best I ve ever seen and support during training courses has been excellent I find the access to ECMWF data very restricted and I access the data as little as possible For example not I haven t used it for some time and I know it will take me some effort to get back onto using it x windows access is a nuisance but I understa
10. l 2 Unhelpful o 0 Very unhelpful oO 0 Total Responses 295 0 20 40 60 80 e Your comments BE I am amazed by how good is the service with so many customers This refers to help over a year ago always got a quick and friendly response which is appreciated the support is always very helpful fast In case of absence of a certain person very fast other person are answering the questions So good that they even pre empted the failure of a job in msjobs because it had the wrong file permissions and they emailed me before it was due to run so that I could change it in time We had many contact in the past with Dr Lucas and more recently with Dr Modigliani in both cases they were extremely professional helpful patient and kind in helping us to solve our problems and last but not least thanks to their help we effectively SOLVED our problems Have they written any FAQs I contact Carsten Maass and am very happy with his support and help I feel they have always done their their best to help me Even when my prob were beyond their knowledge or responsibilities Staying so friendly and helpful as you are We find all user support services very helpful and would hope that the present very hight standard would be maintained into the future I am most impressed with the responses timely accurate usually solving my problem in no time Chapeau Special thanks to Norbert Kreitz for his help and patience It is ve
11. options Response Total of Total Respondents Very satisfied oT E 17 Satisfied 226 el 58 Neither satisfied nor dissatisfied n R 18 Dissatisfied 2s pe Te Very dissatisfied o 0 Total Responses 392 0 20 40 60 80 e Kase of finding information Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied K hay 8 Satisfied 186 47 Neither satisfied nor dissatisfied Ea ie 31 Dissatisfied a2 119 Very dissatisfied 7 l ja Total Responses 392 0 20 40 60 80 e Search facilities Each Respondent could choose only ONE of the following options 73 of 141 29 06 05 17 21 ECMWFE User Survey Response Total of Total Respondents Very satisfied 32 i 8 Satisfied 177 47 Neither satisfied nor dissatisfied 133 a 35 Dissatisfied 35 By 9 Very dissatisfied 3 1 Total Responses 380 0 20 40 60 80 pE e Accuracy of information Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 13 ii 29 Satisfied 222 iT 57 Neither satisfied nor dissatisfied 12 Dissatisfied 7 2 Very dissatisfied o 0 Total Responses 389 0 20 40 60 80 HEE Timeliness of information Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied n aor 20 Satisfied 2
12. where the data is converted to NetCDF format and used to calculate secondary fields PV RH visualisation case studies trajectory calculations with own code run mesoscale model driven by ECMWF data We use analyses and forecasts provided by ECMWF to drive our Lagrangian CTM in order to study transport and chemistry in the stratosphere and the tropopause region R amp D into use of SSTs from satellite borne infra red radiometers Involves use of ERA 40 fluxes SST wind speeds in ocean surface modelling data assimilation in simplified models regional climate modelling in past years EPS development in collaboration with ECMWF Atmospheric Boundary Layer studies LES and mesoscale modelling with the Meso NH model at a local cluster for tests and at the ECMWF for large runs ECMWF see above KNMI research and development of models especially for maritime meteorological applications ECMWF running singular vector sensitivity jobs and forecasts Modifying IFS code KNMI developing optimal model initial condition perturbations Studying strato troposphere interaction data management NWP research and development ECMWE ERA 40 retrievals Home Regional climate modelling For my Master Theses I am using ERA40 data some monthly fields above Europe and monthly SST in equatorial Pacific to study ENSO impact on Europe and Croatia Using RegCM we have a plan to implay ECMWP s seasonal forecast to get the new
13. 1 last time was over year ago 1 once a month or more if it needs 1 seldom colleague is doing it mainly 1 several times per day over periods of a few days but occasionally 1 so far occasionally but I will use ECMWE services often in the near future 1 until now occasionally from now several times a week 1 variable but up to several times week 1 varies widely at present occasionally but in the past also daily May change in future How do you connect to ECMWF Telnet to ecaccess ecmwf int via Internet Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 14 44 Sometimes 127 iii 36 Never o aaa 20 Total Responses 351 0 20 40 60 80 e Telnet to msaccess ecmwf int via RMDCN Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 6 Sometimes s mm en ae Never 199 a Total Responses 248 0 20 40 60 80 BBE e Telnet to a local ECaccess gateway Each Respondent could choose only ONE of the following options 25 of 141 29 06 05 17 21 ECMWFE User Survey Response Total of Total Respondents Always 50 19 Sometimes 0 26 Never 6 55 Total Responses 266 0 20 40 60 80 BOE e Telnet to a local MSaccess gateway Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always a E
14. 133 80 Total Responses 166 0 20 40 60 80 HDE e Windows 3 x Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 0 0 Sometimes 1 1 Never 147 99 Total Responses 148 0 20 40 60 80 E e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 0 0 Sometimes 6 E 5 30 of 141 29 06 05 17 21 ECMWFE User Survey Never 107 95 Total Responses 113 0 20 40 60 80 e Please specify any other desktop system 2 FreeBSD 2 Mac OS X 1 cygwin 1 dec alpha 1 HP will now be Linux XP 1 I think I use OSF1 unix 1 LINUX 1 mostly from XP 1 Nokia Symbian telephone Which scripting language do you regularly use in your work at ECMWF and or in your organisation e Select all that apply Each Respondent could choose ANY of the following options Response Total of Total Respondents C shell csh 183 44 Korn shell ksh 225 54 Tab C shell tesh ss 21 Bourne Again shell bash 153 37 Perl s 17 Python js 4 Tcl Tk js 4 None a 6 Other please specify u 3 Total Responses 418 0 20 40 60 80 e Other please specify 2 MATLAB 2 PHP 2 sh 1 awk 1 ferret 1 ncl 2 octave 2 posix shell IR ul afes 1 sed 1 tesh 1 VBS or Javascript Which high level programming lang
15. 3 Sometimes Ea ee 10 Never 200 T_T 87 Total Responses 231 0 20 40 60 80 pE e ssh to an ECaccess MSaccess gateway Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 0 _ 30 Sometimes s mm 23 Never 127 C 47 Total Responses 268 0 20 40 60 80 e ECaccess shell commands ECtools Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 27 ae 11 Sometimes o 29 Never 144 60 Total Responses 241 0 20 40 60 80 i PBE e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 18 a 11 Sometimes 13 a 8 26 of 141 29 06 05 17 21 ECMWFE User Survey Never 140 82 Total Responses 171 0 20 40 60 80 e Don t know Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 7 7 Sometimes 4 i 4 Never E 89 Total Responses 104 0 20 40 60 80 e Specify other ways to connect to ECMWF 10 Web access 4 telnet to local ECaccess gateway 3 web access to ecaccess ecmwf int 3 WebMARS 2 ftp to ECaccess 3 ftp to ecaccess ecmwf int 2 ftp trough local gateway 2 Internet 1 Daily transfer by ECMWF to JRC FTP site 1 ECtrans via our loca
16. GRID computing environment our performance will improve significantly indeed But I think this cannot be realized in a medium range time scale increase EPS post processing so that it is disseminated every 3h Extend GRIB libraries to code probabilistic and deterministic non ECMWE produced GRIB files Better tools or better available tools to decode from MARS to something that is understood by a Fortran program Faster Internet connection to transfer data to KNMI provide for a server in which all member states and others could cooperate in developing open source GNU style routines for Fortran90 and other languages that are useful essential for a meteorological researcher Make access to MARS archive easier i e user friendly Licenses for Matlab clearly state what variables can be downloaded from MARS reload from ECFS and MARS MARS less important needs some speed up large file support Nothing ECMWF has little impact on my productivity I don t know how much ECMWF can do because there are also other things that affect but more efficient timesharing possibilities when working from outside UK faster response time more reliable X connections I have no wishes MARS is still way to slow More service oriented access from client point of view example retrieve manipulate and download operational data from my application without access to ecgate or ECaccess o web access GRID oriented services ECMWF could
17. HPCF Each Respondent could choose ANY of the following options Response Total of Total Respondents High Performance Computing Facility Total Responses 21 0 20 40 60 80 e Your comments 3 more information training would be useful It would be nice if you could run longer jobs now the time limit appears to be something like 20 hours disc space provide tcsh or zsh More influence to remove experiments that failed I need user support now for doing this update software environment to the one on ecgate Please provide help for specific porting problem 113 of 141 29 06 05 17 21 ECMWE User Survey As cooperating states we would like to have a possible access for HPCF ectrans should remove any temporary files it has been creating LoadLeveler should be able to recognise a successful termination of ectrans SFTP should be accessible A little tricky to make changes and copy move things between archive and home documentation on the machine specific commands for new users Direct help to optimise my code if possible Also disk space ECMWF environment very different from local in that respect same remark Receive by email schedulinf of system session on HPCD and any modification on compilor and tools a faster machine Meteorological archive system MARS Each Respondent could choose ANY of the following options Response Total of Total Respondents Meteorological archi
18. IFS documentation 8 MARS manual 5 ERA 40 5 manuals publications gt manuals 5 Technical memoranda 4 Training course documentation 3 batch job examples 3 ECMWF Reports and Proceedings 3 GRIBEX 3 Introduction to computing facilities 3 MARS manual 3 on line 3 technical report 91 of 141 29 06 05 17 21 ECMWF User Survey 3 Workshop Seminar proceedings 2 computing services 2 Data assimilation 2 Dissemination documentation 2 EMOS lib encoding decoding 2 LoadLeveler 2 MARS documentation 2 prepIFS 2 User guide to ECMWF forecast products 1 Access to IFS source code 1 both on line and paper 1 Both paper and PS PDF format 1 BUFR User s Guide 1 Computing Manuals 1 courses lectures 1 daily dissemination products 1 decoding of MARS retrievals 1 DEMETER 1 Documentation about the best models for long range forecasting 1 Documentation and instructions about use of computing services 1 Documentation available on the net especially that describing different products and how to retrieve them 1 documentation from the computer user training course 1 documentation of the models 1 documentation on programming languages 1 ECMWE EPS 1 ECMWF lecture notes 1 ECMWF model user guide 1 field descriptions 1 Forecast and analysis products availability and the way in which a MARS retrieve command is generated 1 forecast products 1 FORTRAN Language Reference 1 gateway installation 1 Getting Started with MA
19. Response Total of Total Respondents Very satisfied 36 20 Satisfied a el 42 Neither satisfied nor dissatisfied 66 Ln 37 Dissatisfied oO 0 Very dissatisfied iz 1 Total Responses 177 0 20 40 60 80 e Your comments BDE 8 do can not use it Is it possible to increase the max number of jobs running on HPCD for a user Makes large experiments or fast testing possible will have to work on sufficient speed up runs very stable always disc space problem with a CTM input data and source code are deleted too quickly from disc Can t tell maybe we will use EC computers more often in the future We should try to get access Maybe you have good performance to run our model I m reflecting the general views of users who give me feedback when I say that we are very satisfied with the HPCF Only problem is disk space The capacity seems to be sufficient and runs go through efficiently I know the HPCF only from the computer user training From there i am very satisfied 41 of 141 29 06 05 17 21 ECMWE User Survey This facility is of the uttermost importance However to run very long jobs can be difficult I would appreciate to have tcsh available on HPCD Excellent support Umberto M and John G quota system unnecessarily restrictive when eg met office never uses all of its quota too little experience to make a judgement I haven t used it yet because P ve just got my account I am currently not
20. Sites So far as I ve needed or used ECMWF services until now MARS access and different aspects of Web services I am pleased with all of them Almost everything People there are very friendly and very patient with MS users quick and friendly response to problems and needs Pleasing Remote access to ECMWF very straightforward and reliable atmospheric model performance EPS included reanalisys archiving training courses fast data availability large data sets very recent data provide data for models 125 of 141 29 06 05 17 21 ECMWF User Survey Good archiving facilities I only access MARS data and have only done this occasionally Therefore I feel unable to answer these questions Meteorological training courses are really good Data Achieves The data you provide and Quality of Data efficient supply of huge amount of data sets Archiving and accessibility of the products Very capable experts working at the centre I feel that ECMWF provide a very good high performance computing facility including data storage I have always found the staff helpful and responsive to my problems queries with Data server and MARS archive Reliability Overall I am very satisfied Good computing facilities good service good information Efficiency and reliability in solving problems by ECMWE user support staff Documentation ECfile archive system access to operational archive It s too early for me to say u
21. am not sure to collect the information I need Sometimes difficult to find information you need Not intuitive you know that a specific piece of information doc or forecast product is there because you ve found it once but you can t find it easily anymore the your room facility is very convenient Which areas do you access on the ECMWF web site e Each Respondent could choose ANY of the following options Response Total of Total Respondents Documentation 365 92 Forecast products 253 SSS 64 WebMARS 197 a 49 Your Room 126 EAE 32 Other please specify below La E 5 Total Responses 398 0 20 40 60 80 77 of 141 29 06 05 17 21 ECMWFE User Survey e Please specify other areas 3 publications 2 Employment 2 Epsgrams 2 general information 2 IFS source code 2 News amp Events 2 PrepIFS 2 Research documents 1 adverts for talks and courses 1 Archive computing Manuals 1 BUFR software 1 Calendar for meeting schedules and associated documentation 1 Committee Meetings 1 Courses 1 data availability 1 Dissemination requirements 1 Ectrans manager 1 ERA 40 1 Era40 Data 1 Info of Training courses etc 1 manuals 1 MARS manuals 1 Ordered programs 1 other research project descriptions 1 Reanalysis Docu 1 reanalysis products 1 RMDCN 1 Seminars Events 1 services research 1 sometimes links to ECPDS And to EMOS related software 1 special project s
22. and the multi model Seasonal forecast Glosea and System2 build the backbone for Met Office long range predictions At ECMWF the multi model ensemble forecasts are analysed and the information 16 of 141 29 06 05 17 21 ECMWF User Survey is compressed for further use at the Met Office Operational products are derived locally I am responsible for feeding our real time database with actual data from various sources That includes managing our dissemination requirements I also act as computing representative for FMI TAC representative Project Manager of many software projects related to computer operational issues and software development For example projects related database web web services system monitoring data communications At ECMWF Computing Representative Operational Products Visualisation ERA4O retrieval for dynamic downscaling At CHMI Head of NWP Group Development and operations of ALADIN Database Administrator in my organisation and MARS data retrieval at ECMWF Work with COSMO LEPS system to investigate mesoscale short term predictability of precipitation i e interaction of orography with precipitation Special project at the Centre The purpose of our work is to develop forecasting methods for the wind energy sector provide atmospheric forcings to a surface model provide atmospheric data e g boundary layer height for the interpretation of in situ measurements support publish diagnost
23. as ECMWF for example is also funded by mostly by public funds None that concern the ease of use once an account permission has been obtained There have been no questions about the use of ECMWE software BUFR library Metview Although I am quite satisfied with it there may be improvements possible particularly concerning the documentation Computing courses could be more specialised e g focusing on MPI and OpenMP only To obtain Passcode security ID one must wait a long period Maybe that is a little bit complicated and frustrating but I am aware that it can not be much improved due to all security purposes Indeed GRIB format is not very common e g visualisation with MATLAB is impossible It would be very useful to have the opportunity to download dynamics in HDF file format I use occasionally ECMWF computer without formation not enough time to do it For any information that I ve not given I could anwer you by mail Best regards just a suggestion ask the members states which forecast or verification products they would like to have on the web server the questionnaire was very long and it was difficult to me to give clear answers to some questions Congratulations for your questionnaire It s rather long to fill but it s worth to me As I mentioned for the time being I have a very comfortable work environment at Meteo France and little time to try other computing facilities However I remain
24. at ECMWF than I might have questions more frequently than monthly At the beginning once or twice once or twice Once or twice yet Who do you contact first in the case of a problem Each Respondent could choose only ONE of the following options Response Total of Total Respondents User Support 159 42 Call Desk Ea eal 8 Computing Representative 66 i 18 98 of 141 29 06 05 17 21 ECMWFE User Survey Colleague 80 EAS 21 The ECMWF specialist in charge directly 23 E 6 Other please specify 17 5 Total Responses 376 0 20 40 60 80 e Other please specify metoffice contact User support at FMI representative for Switzerland John Greenaway contact to our meteorological service through e mail C Maass national weather service myself ZAMG My regular contact at ECMWF I m not quite sure what his function at ECMWF is it depends on the subject Computing Representative User Support Call desk Either user support or the call desk depending on the urgency and nature of the problem Umberto M and John G John Greenaway Depends on the nature of the problem roughly in the order given above Metview team ECMWF specialist at Meteo France How important is it for you to receive advice in your own language Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very important 9 Ian 13 Quite important a eS 17 Not importa
25. at home a HDSL of 0 5 Mb for 39 euros a month Hope they fix that So I cannot tell about your web speed see above With the comment above statements cannot be made here on serious basis The Web design you set up perhaps one year ago is less handy than the old one It may be prettier but it is more difficult to find things Some products are very east to access like 10 day EPS forecasts but when outside companies request the ECMWF data I put them directly on to the web site and they invariable find it very difficult to navigate i e when they want several specific years of certain parameters Site navigation is an acquired skill But I m used to it and I know where to find what I need However novices sometimes find it difficult to navigate the site detailed information is sometimes hard to find Just an example I remind is the definition of the model levels The search did not easily bring me http www ecmwf int products data technical model_levels model_def_60 html structure and linking could be better Never had any problems with the site and it seems very professional and nice ECMWF web side is a very good source of information on my daily work The documentation about operational products and research products are difficult to find or many times out of date It is unacceptable that in the web site from the deterministic forecasts i can only see forecasts from 72 hours and on and only for three fields Retaining valuable informati
26. available in Met Office archive and EPS data research on ensemble seasonal forecasts Preparing boundary and initial conditions for a mesoscale meteorological model extracting data from MARS archive I provide user support for Graphics Applications Currently the Met Office only offers support in Metview and PV WAVE ECMWF Acquiring MARS data for models trajectories FMI Greenhouse and related gas measurements Educator in meteorology Organising training courses making learning material eg SatManu At SMHI I work at the maintenance department My title is Application manager so I mostly work with applications Very often these applications use ECMWF data GRIB BUFR etc I also have a responsibility to fetch and compile emoslib when a new version is available Verification of DWD Models and ECMWF precipitation forecasts over Germany Analysis of ERA15 40 for the study of the global and regional aspects of the water cycle Diagnostics of climate variability on interannual to decadal timescales forecasting meteorological modelling dynamic climate seasonal simulation Head of Operational Suite of C N M C A ITALY Postprocessing of numerical models and application s development Basically I retrieve ECMWEF data for carrying out research projects doctoral thesis and certain supervised works to be developed by students as part of their pre doctoral formation gaining ECMWF model results for regular model comparison with
27. bash as login shell a more detail information Es muy lento porque hay mucha gente conectada a ecgate Deberia haber varios portales ecgatel 2 3 como antes Improve bandwidth for data exchange between Met Office HQ Exeter and ecgate Pd wish I could use tesh automatic ftp to external computer in order to reduce the temporary disk usage for large data retrievals I hope the data sets would be decoded by wgrib as well I should learn how to use the decoding program for ECMWF data sets I only use it occasionally when I need it so I don t intend to use it more often Apart from the instability in X Iam happy Is it possible to find out which process in the chain is responsible for disconnecting just use ecgate for MARS retrieval no suggestions up to the moment It would be useful to describe the processes through more examples As I use your service rarely it is satisfactory for me I have not checked if there is a manual on your web site about how to use ecgate but if there is not it would be good to make one it would be nice that on ecgate the shell tsch could be available Access to the R statistical package would be nice http www r project org I am not as yet very strong in programming which is why currently I only use the web services 39 of 141 29 06 05 17 21 ECMWE User Survey More graphical less script unix style gadgets more EXAMPLES of SIMPLE retrievals a new MARS function to extract time series
28. be in m s or in mb hour or I simply want precipitation and only at the end want to know whether this is large scale and convective combined or whether they are added already It used to work very well But for some weeks already I have some problems Checking the availability of model fields for instance I get the response submitted but I never get an answer I have notified the problem the problem will be solved soon I guess great to generate the database call to be used in scripts later if it does work retrieval is too slow often This is a good facility which works well The only extra which would be useful is an automatic GRIB decoder which works on xp and Linux like ncep s wgrib I like the NetCDF interface I did it not use so I can nothing say but i think that WebMARS a useful thing I use webMARS only for a quick look at some data e g to plot a certain field These plots are rarely useful since I cannot modify choose a good scale neither can I modify the range of the data it should plot Would be nice to be able to define my own areas for display Isolines or contours need to be labelled better It would be great to be able to download files in NetCDF format for all type of data that are provided It takes much more time than run an retrieval by Metview 62 of 141 29 06 05 17 21 ECMWF User Survey I haven t found a description of the grid structures of the archived fields Maybe there is
29. comments made previously Complete EMOSLIB manual in one document More explicit documentation about how to use the ECACCESS software See earlier comments on Metview documentation comparison with others models It is difficult to find information on the website Despite the general good documentation I thing some points are still missing or not so well explained Relevant meteorological papers in PDF very strong good but there is always improvement Better documentation about current future plans for reanalyses Providing contact details for the specialists people in charge to various areas in particular for research related subjects Sometimes searching too long because the search engines produces quite a lot of links which are not always as relevant as is indicated more adequate and condensed information on the forecast system no PDFs Easier finding and browsing topics on the Website xcdp sms documentation is limited and old Better search facility and organisation of the information A clear and transparent quickstart manual see earlier statements in this questionnaire MARS documentation Updated Please give more technical information about previous experience of porting codes from other machines The first use of ECtrans was not easy without help from ECMWF support As indicated above it would be nice if one could easily find an overview of model changes on the web site MARS see previous questionnaire
30. data archive Great archiving system It is the only place I know you can access to all of what was produced Operational aspects like system sessions notification reliability efficency and good network connection I am most pleased on the Call desk help A good source of atmospheric data Most of services available on line are very professional We are very satisfied with the services we are receiving from ECMWF The service we look for is limited with computation of backtrajectories and support from the user support desk when it becomes neccesary Reliability of the services Support ECMWF does an excellent job providing users the fest and anlys data Quality of model data 123 of 141 29 06 05 17 21 ECMWF User Survey The ECMWF services are handled very professionally and efficiently Good HPCF Excellent user support Reanalysis and observation data prepIFS provide good products courses facilities and information in meteorology and HPC Access and retrieval of data Web based access to MARS to quickly see what s available I am generally satisfied with ECMWF it is a high level and professional organisation ECMWF does almost everything very well I am very pleased with the hpc environment with the access to the operational forecasts and in particular with the support of ECMWF staff user support Metview MAGICS team and research staff resources availability and documentation MARS website Expertise excellen
31. domain but am not logging in I have a suspicion that the web site will act differently depending on which of the methods above is used for accessing it But I can not find any information on these differences I suppose since I am accessing it from a member state domain I have broad access to the web site It would be nice to know more about the consequences of these different access methods and also which parts of the web site are available for the general public Have tried to login in using this same error message since mid January 05 Can t call method uid when I try to view the EPS products and get the prompt to change the initial password at https www ecmwf int login using gt Your ECMWF user ID is tonymet Have followed instructions advised by user support This is VERY strange since we tried here again and it works perfectly So another test can you delete all the cookies in your browser then restart windows and try again Let s see if this makes any difference I still could not login when I followed this advice I both use certificate and SecurID At my Met Service Log in from your authorised Internet domain Otherwise Use your own ECMWF user certificate to log in or Use your userid and SecurID passcode I am coming from meteo fr but there is no explicit logging in Either access with rights of meteo fr or access via passcode would be useful to see the level of authorization of the different produc
32. find it easy enough with the SecurID It would be easier with a password From home I also use the web site sometimes directly and sometimes with the need for userid and passcode and authorised used domains Excellent Access from Home using User Cert also When logging in from home I need securID I also used the SecurID once in while also using our own user certificate I log in from both an authorised domain and my ECMWF user certificate Often try to see how far I get without log in use SecurID when necessary It however often happens that I have to log in with user id and passwd if I want to see the page but that might have s th to do with my own browser not how the page is protected I also log in using my userid and SecurID passcode Mostly from auth Inet domain met no At times when off work by use of SecurID Rarely works for any of login from authorised domain ECMWF User Certificate SecurID so tend not to use it This may be due to settings on my PC I log in from my authorised Internet domain too Sometimes I log in from the authorised domain too I log in from a variety of sites office home university institutes abroad we also use our userid and SecurID passcode In fact It is not clear if we have to register with passcode or not and we don t really know which sections are public or restricted 79 of 141 29 06 05 17 21 ECMWE User Survey I am using the web site from an authorised Internet
33. for one position not grid only extracts It would be useful to provide descriptions to explain all the possibilities on ecgate through examples Why are you not satisfied with the ecgate service I am not satisfied with the ecgate service because e g insufficient processing power insufficient disk space poor turnaround difficult interactive use difficult batch use bandwidth limitation unavailability of software etc The keys are not automatically set up Some shells use delete and some back space The arrow keys give strange symbols See previous question Especially define the problems of X windows My guess is that system does not allow opening too many X windows in our institute The first in the morning gets the X window possibility However there may be some improvements very lately Insufficient user disk space Primitive programming environment difficult interactive use it s really some kind of batch system and there are several layers between me and the data melbrowserlecaccessINQS IMARSldata The support resolved all my difficulties to run the jobs The users for France are all fr ones This is a too wide gathering in my opinion I used frtodcli and Dominique Lucas told me to abandon this way of doing Computing Services HPCF In this section we would like to find out about your use of the computing services provided by the High Performance Computing Facility HPCF How often do you use the Ce
34. has the same data A nice thing that NOAA sites do is to allow the user to specify operations on data that are then carried out on the fly See for instance http www cdc noaa gov map time_plot This allows for anomaly maps and such to be made A service like this could be tagged on to existing ECMWF services I use netCDF formats instead of GRIB ones I am not thrilled with the GRIB format as it is less flexible than netCDF plus there are now many different programs that read netCDF but only a few that handle GRIB To me it is therefore important as I have already stated that the grib2netcdf translator be improved Finally many places are now using the 133 of 141 29 06 05 17 21 ECMWF User Survey openNDAP DODS framework to exchange climate data Data can then be accessed from within a user application in a sense similar to what Metview can do It would be nice if ECMWF could implement an DODS server for data most relevant to this open access I think ECMWF does it much better Computer resources information and help etc More complex They show more statistical climatic products I am not satisfied with CINECA I would like to use ECMWF in the future Huge amount of disc space for model results Poor by comparison ECMWF is easier to access and has clearer instructions Fine What they offer in addition is very much related to our needs concerning our model mainly efficiency improvement of code which we do not n
35. help with specific technical issues regarding parallel computing some training We need law for meteorology here in Montenegro 138 of 141 29 06 05 17 21 ECMWF User Survey provide data also in netCDF format i will think about it At the moment i don t know not so much since I am happy with my local resources now But if for example the use of your new IBM systems would have been more advertised and guided I may have chosen to continue to work at ECMWF computing resources make operational products on web available to public regional weather services of MS at the same level as for Nat Met Serv currently only WMO products are available which is sort of ridiculous offer me a job there So far almost all works well I will appreciate a lot to receive regularly an hard copy or digital format also link to a document via e mail results of the ECMWF overall activities like a few pages quarterly to be aware and to feel the scientific operational and computational developments of the centre The earlier the model output is available the better Simplify the data transfer from ECFS to my local computer system Providing access to HPC Improve connection bandwidth to ecgate be more relaxed on data issues Simplify the secured connection The SecurID card is a pain ECMWF does not affect that Access more easily to operational products in ASCII or GRIB with a routine in exe easily handing at the Northern Hemisphere leve
36. in areas for improvement below more meteorological products available free better link Easier navigation I do get frustrated when I can t find what I m looking for When I find it it s always very good though To have access to more real time forecast products like EPS monthly and seasonal forecasts I would like to find documentation more quickly Nothing else because I want to visit the ECMWF web site only because work what it is the current reason More freely available operational forecasts products I do not wish to visit the site more often The go there whenever I need to find documentation and information on your met products more forecast products zoomed maps for my region maps with 6 hour intervals up dated ERA 40 data set If I could ask advice understanding the usefulness to visit Easy access to forecasts change of subject if I would find more hot news 86 of 141 29 06 05 17 21 ECMWF User Survey higher resolution products if data availability would be more easy often I prefer another site like NCAR it depends what I m working with I m fully satisfied with the web site access to forecasts such like epsgrams More products like seasonal forecasts which aren t so often to find Easier way to animate manually step by step through user selected forecast products Regarding documentation and papers the possibility to export pdf Regarding forecasts the availability of more products
37. is used for private issues only easier navigation and structuring of the WEBsite so that you can easier find things availability of satellite data and recent weather reports analysis Games 87 of 141 29 06 05 17 21 ECMWF User Survey why would you want that I visit as often as needed and if I find what I was looking for I am satisfied If there were no problems when logging in An improvement would be to clearly state what s new on the website or at the centre I use it strictly for my work If my work necessitates more frequent visits I will visit it more often In general I m quite happy with the web site Daily forecasts More up to date forecasts easily available To access to your technical and scientific documents More information about news when logging on ECGATE weather information program of courses and seminars The frequency I visit the ECMWF web site has solely to do with the need for my work The more I need it for my work the more I will use it If I was able to find easily information when I need it For instance I have to go often to JPL and connect from there which I do using a web certificate Finding how to revoque the old and make a new certificate is a pain A good source of documentation about ECMWF products and non ECMWE products More access to forecast products my MS account is very limited I cannot get much real time operation products nor easy access to recent archives i e web
38. it for the first time you make your choice Then you accept the terms of use Then your choice is lost and you have to start again Very annoying Not all the data I was interested in was available when I tried to get it with this I only used this Service once or twice I was not aware of the existence of this service The downloading is fast and the server works quite well However I think that would be nice to have some online utilities to decode the GRIB files EMOSLIB isn t so easy to use As a GrAds user I will be happy with the following GRIB to GrAds set of utilities 1 automatic decoding from GRIB to GRADS 2 automatic build up of the GRads control file or option 2 the automatic creation of the index file idx and the control file ctl that allow to open a GRIB file using Grads incidentally this kind of GRads facilities are provided by NCEP for their reanalysis dataset helpful for downloading ERA40 fields This seems to be a service very similar to webMARS or have I missed something Why provide two such services Why not concentrate on one thing Initially however some of the documentation on time intervals forecast vs current etc were not very clear particularly in the NetCDF format Again I only used it briefly to see what data were available I just tried the system for ERA40 data Normally I don t have to use this system since data resolution is too coarse for our needs It may be sufficient for drawing pu
39. my C and JAVA programs Retrieval of information from a single grid point but that s not quite a matter of the language Automatic computation of geopotential for model levels if extracting only a few of them Increased checks on syntax logics or too large datasets Try date 1959040 1 to 195904030 if this hasn t been fixed It should be possible to retrieve ERA40 at model levels and interpolate them to arbitrary pressure levels at one step Metview command language change GRIB headers save lookup table of contents for large GRIB files to filter data faster after first sweep save table for future reference I have not encountered limitations for what we used yet Decoding of GRIB variables by means of Fortran language cannot say at this stage since I don t know what MARS is currently capable of doing zonal averaging Extraction of vertical Profiles I just started using MARS and for the moment I can t think of any additional feature that could be useful for my work Sometimes easier access to longterm monthly mean time series may be useful It would be nice if in the same MARS script you could save the data in different formats Specially NetCDF is this possible already to access observations Wind fields requested at lower resolution than original are not conservative LowRes fields are interpolated at request grid instead of degrading resolution by e g a mass weighting Better control on how the resulti
40. of Total Respondents Very satisfied lb M 44 Satisfied 96 SSS 37 Neither satisfied nor dissatisfied ca 18 Dissatisfied e 2 Veey dissatisted o 0 Total Responses 262 0 20 40 60 80 EL DE e Your comments I do appreciate his cooperation Can t think of any special requests I ve made so that question is not relevant Service generally excellent The courtesy and the professionalism of the User Support Section are incomparable Overall excellent not often contacted ECMWE user support is easily the best such service that I use Patience is not relevant here The responses to the requests I have made have all been excellent except for the last one sent to software services ecmwf int 2 months ago which has still not been answered I do have sent one reminder Again in general I am very pleased with the received service What more can I add Remember it is rarely used I am really very satisfied with ECMWE service The good thing is that user support afterwards checks that everything is OK It seems to me that there is a distinction between the User Support and the Call Desk In general I expect more from the User Support than from the Call Desk My evaluation here is for the User Support As computing representative I have frequent contact and a fruitful relationship very high professional level I am very satisfied with new user support who is Paul DANDO The user support have provided me a gr
41. of Total Respondents gt 5 5 6 Total Responses 374 0 20 40 60 80 e OpenMP parallel Each Respondent could choose only ONE of the following options Response Total of Total Respondents gt 5 10 I 3 Total Responses 353 0 20 40 60 80 Computing Services ecgate In this section we would like to find out about your use of the computing services provided by ecgate How often do you use the server ecgate Each Respondent could choose only ONE of the following options Response Total of Total Respondents Daily 37 al 9 Several times a week 85 a 20 33 of 141 29 06 05 17 21 ECMWFE User Survey Once a week o 9 Occasionally 202 iii 48 Never Em eal 9 Other please specify 22 5 Total Responses 423 0 20 40 60 80 E e Other please specify 5 intensively during limited time periods 3 once a month 1 always 1 daily a year ago and occasionally in the recent year 1 daily but by intermittency 1 during some periods 1 each time a get connected to ECMWF 1 hardly ever now 1 intermittently 1 I used to use it several times a week and plan to do so again by the summer 1 not sure 1 occasionally but last use over a year ago 1 occasionally to daily 1 once a month or more if needs 1 once a year 1 project status dependent up to daily 1 regularly for some periods when I need to get the data then not at all for extended periods 1 sometimes dai
42. on the web Guest login via NMS site and personal login These three give different data access It is often confusing to me what each level gives Especially this is a problem if I refer someone to the site The limited versions of ERA4O available for the public are a bad idea ERA4O is better than the NCEP data but the latter still gets used because everyone has a no strings attached access to the full data temporal resolution of meteorological parameters increase to 3 hours of the analysis data some are provided as 3 hourly forecast and 6 hourly analysis They are too restrictive with permissions on the system Sometimes I can t do the simplest things Real time forecast products are very limited 128 of 141 29 06 05 17 21 ECMWF User Survey more disc space Continue the service We use ECMWF data for scientific work Compared to other data provider e g UK Met Office ECMWF sometimes appears as a black box Contact and communication with experts in charge is sometimes difficult Access via SecurID card is not very convenient I would rather use sth like an ssh key or passphrase Carry on like up to now To change the data policy to be open and free Super computing facilities for people abroad Web access web page layout navigation documentation keep an eye also on non national services release the ocean model as a MARS product The PrepIFS server is not always stable but problems are solved very quick M
43. our Service for dynamical downscaling producing wind climatology over Hungary Later on we will use the same dataset for the validation of climate models Research scientist with focus on high resolution and probabilistic NWP Computing instability indexes Prepare run and test the COSMO LM code for MeteoSwiss development purposes 14 of 141 29 06 05 17 21 ECMWF User Survey Use ECMWF access to retrieve reanalysis data ERA 15 ERA 40 for doing climate analysis studies in combination with our own station based data sets Forecaster meteoswiss data extraction and treatment ECMWF Weather Watch Synoptic Meteorology Severe weather events I m working at Marine Meteorological Division Subjects numerical prediction of waves oil spill modelling activities at JCOMM and IOC groups and other activities related to marine meteorology and oceanography Setup and run hydrodynamic models dealing with waves and currents on the ocean Research and development of ALADIN model Mesoscale modelling Research amp development of numerical algorithms for the UK Unified Model dynamics and physics Assistance to climate researchers in data retrieval conversion Making software tools Web publishing of research results At ECMWF development of a data assimilation method now completed In my organisation similar work but now using local computing resources I still keep in touch with the seasonal forecasting group to learn about each o
44. past cases I m analysing the representation of extra tropical transition of tropical cyclones in the EPS Id like to have a closer look on the improvement of the forecast due to application of targeted singular vectors and to utilising moist singular vectors further north and want to calculate Ensembles by myself HIRLAM system management both at ECMWF and KNMI at ECMWF getting data for at UNIVERSITY calculating trajectories Aim develop STE climatology including ozone Research on the Climate variability in the Mediterranean region Numerical ocean modelling supervision of students project administration 20 of 141 29 06 05 17 21 ECMWF User Survey Hydrological Studies especially validation of climate models and data Land surface processes modelling Land surface datasets I use temperature and precipitation data from the era40 reanalysis as input for my model which tries to predict the mass balance of glaciers My work consists in performing high resolution numerical simulations to describe the dynamics of local winds Currently we are running several experiments to study the skill of the ECMWF model in the intraseasonal time scale as well as testing new techniques to improve forecasting in such band statistical downscaling using data from the DEMETER models In my organisation I m a junior scientist in radar application develop air quality models use meteo data as input for our models Developing and maint
45. public library using web access and login It s not easy to access the web site from elsewhere because I would have to carry my SecurID card with me all the time while visiting other workplaces eg research institutes If possible I would like to access web site from home computer with same level of access What is the main web browser that you use to access the ECMWF web site Each Respondent could choose only ONE of the following options Response Total of Total Respondents Microsoft Internet Explorer version 5 or later s AR 2 Microsoft Internet Explorer other version r e Mozilla version 1 or later 126 Mozilla other version 2s on hs w Netscape version 5 or later Netscape other version Opera bi Don t know 2 85 of 141 29 06 05 17 21 ECMWF User Survey Other please specify 35 aly 9 Total Responses 397 0 20 40 60 80 e Other please specify 19 Firefox 10 Konqueror A Safari 1 both Mozilla 1 and Netscape 6 1 Mozilla 1 7 2 1 Mozilla 5 0 X11 U Linux 1686 en US rv 1 2 1 1 Gecko 20030225 build 2003022516 1 Mozilla and opera 1 Mozilla IE5 Opera What would make you want to visit the ECMWF web site more frequently e Please specify 13 nothing I am satisfied happy 7 If I had more time I would do it 4 better search facility 2 more data 2 I already visit it quite a lot more information on immediate future forecasts see answers
46. reason that I haven t tried WEBMARS yet I usually build scripts to process a long timeseries several years of data which would be rather impractical to do interactively Because the telnet system is more flexible and convenient for our purpose 64 of 141 29 06 05 17 21 ECMWF User Survey I always use MARS client on Metview or script language on ecgate Queries during ssh session or via ECtrans are perfect If webMARS is the same as ECtrans I m sorry I m lost with the plethora of terms and acronyms If it is not the same I just don t know about it I tried once to retrieve data via the website but I remember it seemed not that convenient for repeated use and I gave up Because I have bash scripts to extract ERA40 analysis that work very well and that convert after GRIB format to NetCDF format just haven t had a chance to try it Satisfied with old fashion MARS requests Mainly I only do one daily retrieval from Metview because there was no need to retrieve the data because I have not the need I prefer to use MARS services via ecgate never tried i will It s not relevant to my work I use the MARS archive via HPCD I need the data there So I don t need webMARS for my work We use shell script to retrieve data from MARS and to run a ECMWF loaded FORTRAN program to decode the resulting BUFR data in order to format the data in our own ASCII output I prefer ftp I am happy with MARS client less
47. s already possible or restricted by our local sysadmin Get rid of the GRIB data format too complex and rigid Make the use of HPF easier accessible for member states For example organise local information meetings in member countries 129 of 141 29 06 05 17 21 ECMWF User Survey In reference to my present activities the ECMWE is quite satisfactory improve products free availability satellite data improve the canteen s food give Indian taste a chance So far it is OK for my purposes Increase the model resolution and extend the range of the EPS data Disseminate more information on weekly and seasonal forecasts facilitate web site navigation I would like to see a full access to MARS resources through a web site data retrieval via rsync I have no suggestions for improvements I am not a heavy user Therefore I have had no problems Abandon Metview buy in grads c dat Simply because nobody knows uses Metview outside ECMWF Metview does not catch on even after many years of development Or better advertisement free usage an efficient platform to exchange experience and problems needs to be put in place n Better portability of ECMWF software incl IFS more testing on different platforms Documentation of errors allow users to access all data i e realtime products not only historical data I am satisfied with ECMWE services as they are Perhaps an improvement of some documentations MARS for example more
48. the following options Response Total of Total Respondents High 7 I 4 Low js EEs 9 Not relevant 171 ii _ _ 87 Total Responses 196 0 20 40 60 80 5 e Lack of optimisation tools Each Respondent could choose only ONE of the following options Response Total of Total Respondents High a l 2 Low a Pel 11 Not relevant 171 87 Total Responses 196 0 20 40 60 80 Please specify the importance of each reason or tick Not relevant e Other non technical reason constraint please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 26 frag 159 Low s E 5 Not relevant 37 80 Total Responses 171 0 20 40 60 80 e Please specify other non technical reasons constraints 10 not necessary for my current work 7 Cooperating states cannot use Centre s HPCF resources yet 5 I know nothing very little about HPCF 2 Lack of time for learning 2 We have to use other our own HPFC facility am not allocated resources to use your HPCF Many years gt 10 years ago when I did use your supercomputer for a high resolution climate run the biggest problem I had was transferring the large amount of output data back to the Met Office Will use in future for testing future MPI code 45 of 141 29 06 05 17 21 ECMWF User Survey I am afraid of MPI until now I am working on parallel vector m
49. to ECFS When I first used it it was quite hard to understand how to proceed in batch mode so I did a lot of things wrong mainly because my expertise is small We use ECFS also at the DWD Sometimes normally after a maintenance session ECFS is very slow One the size of the GRIB files while the MARS request is being processed I just stepped in once It is a pity that it cannot be accessed with several standard Unix commands a I can understand els etc in stead of Is etc b less clear is why ecp and not rcp wot c I miss mv rm r etc I add that Tony Stanford has always been very helpful to carry out the tasks that I could not do myself due to these limitations So it has not hampered my work but possibly Tony s I am not sure if ECFS is the file system where the data is stored when retrieving data from MARS It appears to be a robust and reliable system Fast Files access is sometimes slow How easy do you find ECFS to use Each Respondent could choose only ONE of the following options Response Total of Total Respondents 69 of 141 29 06 05 17 21 ECMWE User Survey En am 23 Easy jos m 42 Neither easy nor difficult 74 WE 32 pict 0 3 Very Difficult o P 0 Total Responses 231 0 20 40 60 80 e Your comments jela 3 A possibility for renaming files in ECFS would be helpful sometimes some more Unix like manipulation tools e g to move files f
50. to at any time I find login process tempremental at times Recently I am involved with forecasts all over the world beyond my trajectories I find it very convenient ECMWF provides meteorological maps It saves me a the trouble of inventing a good lay out There is so much information that it can be difficult to find the information you are looking for I find really helpful the very complete description of the model physics parameterisations and so on it is not very easy to navigate and the search function does not often yield the information I was looking for not easy to find what you are looking for 72 of 141 29 06 05 17 21 ECMWFE User Survey The information content is very useful But it is structured in a very cryptic manner and I find things difficult to find It is very clear I like it the handouts of the computer courses are not all available Confuse to find the information and tools As I can retrieve model fields that way it is very useful I wish I would have more time to use it Its the best source of documentation on the model and the data e g for finding documentation on libraries MAGICS Sometimes it s rather difficult to find what I m searching for The documentation helped me to migrate scripts in 2004 Happy Please rate your level of satisfaction with the following characteristics of the ECMWF web site e Ease of navigation Each Respondent could choose only ONE of the following
51. using the HPCF this work is performed by a Ph D student within our Special Project who is very satisfied Korn Shell is not my favourite By the time I used the system I was very satisfied How easy do you find the HPCF to use Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very Easy jo w 6 Easy s Em 31 Neither easy nor difficult Em aoe 58 T gt a Difficult g 4 Very Difficult l 1 Total Responses 164 0 20 40 60 80 e Your comments I don t really understand all the options for submitting jobs etc but I have some templates that work fine Recent experiments have been carried out by colleagues Therefore I do not have experience on the current IBM system the only thing which can be unhandy is the scripting needed to get input output on the TEMP The shell environment is not very user friendly tesh or zsh would be nice to have instead of ksh I was using HPCA a year or so back and it was easy to use I will soon be using its replacement at which time I will be able to form an opinion I must thank the chief system manager for his help Otherwise I should forget to run my jobs on the HPCF As an occasional user it is difficult to keep up with the changes in computer platform over the years I still can use HPCF and I don t now how to use and and what is this I don t like the korn shell A sftp connection to t
52. values of the fields for a MARS request The information about the possible values are very dispersed in the ECMWF Web page documentation and not all values always work together having to do a trial error exercise every time a change in the parameters is needed Better documentation explaining the possible values of each field and how are they compatible incompatible would be very helpful The Web page interface to automatically generate the MARS request is still to cryptic to understand properly Maybe it would be good to give some instructions about which programs can read retrieved MARS files GRIB etc and where those programs can be downloaded MARS works but the products are not easy to use for people lacking meteorological packages A convenient tool for opening GRIB is required Use is much easier now we have access to the online catalogue Mostly I find it easy now but had problems in the beginning Again the problem is with programming and scripting languages I haven t had the time to learn 52 of 141 29 06 05 17 21 ECMWF User Survey It is mostly self explanatory While MARS is easy to use the documentation on the batch syntax is not very thorough Also it can be difficult finding the data you require using the content lists on the web e g there are a vast array of expver without any explanation of what they are I am not expert in this area so am happy to be able to find my way around in the MARS user guide the usag
53. various locations It is not obvious how to get the forecasts for a none specified town The loops of the maps are not working As you download a page you don t access immediately to the latest model run but some informations are very difficult to find RMDCN web site Very useful Some things are difficult to find the structure of the pages is quite complicated can get lost in the many levels of pages Search engine is extremely useful for finding documentation I find news sheets very useful structure of docs and access to data is not very transparent Multiple and redundant data sources I would prefer a simple hierarchical document structure eg BADC site It is used for retrieval of actual weather forecast in your room I constantly check the site for the latest medium term forecasts for various places I like the EPS products Often very difficult to find specific information manuals etc The search does not seem to work very well Pages such as the MARS queue are difficult to find Not used easy to find info about ECMWF services The search engine is not satisfied IFS documents not very up to date Can be hard to find things have to go though many levels I would like to see more plotted fields available Too restricted to users outside MS Met Services It would be a disaster if it was not there Local knowledge very useful when trying to find things the information on the www is useful to have access
54. 0 80 Your e mail address Answers not shown Your country of employment e Select a country Each Respondent could choose only ONE of the following options Response Total of Total Respondents Austria 23 5 1 of 141 29 06 05 17 21 ECMWFE User Survey re a i Find SB 3 France 5 a ee 7 Meeland O 1 treana O 2 Serbia and Montenegro 3 PG Swiverand O a 2 C a E United Kingdom 522MM 12 C ne i Total Responses 450 0 20 40 60 80 Which of the following best describes your organisation e Select one of the following Each Respondent could choose only ONE of the following options Response Total of Total Respondents National Meteorological Service 223 M 48 Regional Meteorological Service 16 3 Other Meteorological Hydrological Organisation 10 l 2 2of 141 29 06 05 17 21 ECMWFE User Survey University ES TA 25 Research institute EE E 18 Other please specify jis 3 Total Responses 460 0 20 40 60 80 pE e Other please specify 5 EUMETSAT 1 air traffic control 1 Defence Meteo Wing 1 JRC 1 Mercator ocean 1 Military 1 international Organisation 1 operational oceanography project 1 RTH 1 Software house 1 Tidal forecasting centre Which of the following best describes your position Select one of the following Each Respondent could choose only ONE of the following options Response Total of Total Respon
55. 35 61 Neither satisfied nor dissatisfied oT a 18 Dissatisfied 3 1 Very dissatisfied e 1 Total Responses 371 0 20 40 60 80 PBE e Web site speed response Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 103 E 27 Satisfied 20 57 Neither satisfied nor dissatisfied s2 Pa 13 74 of 141 29 06 05 17 21 ECMWFE User Survey Dissatisfied 10 3 Very dissatisfied 1 0 Total Responses 386 0 20 40 60 80 e Web site reliability Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 122 32 Satisfied 220 eel 58 Neither satisfied nor dissatisfied Em 10 Dissatisfied 2 Me Very dissatisfied 0 0 Total Responses 381 0 20 40 60 80 e Overall look and feel Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied GE E 18 Satisfied 233 62 Neither satisfied nor dissatisfied es 17 Dissatisfied jio 3 Very dissatisfied oO 0 Total Responses 376 0 20 40 60 80 BBE e Your comments I find it very hard to know where to look for the info I want and always end up using the search facility or asking someone else for a link The menu navigation doesn t seem very intuitive Other than that the site is brilliant everything I ne
56. Always l 2 p Sometimes 2 gt MM Total Responses 167 0 20 40 60 80 e Solaris Sun Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 4 Sometimes 4 on 3 2 7 Never 102 24 53 Total Responses 192 0 20 40 60 80 Tru64 Compaq Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always jo E 6 Sometimes a Eal 9 Total Responses 161 0 20 40 60 80 e Mac OS Apple Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 9 6 Sometimes 9 Bz 6 Never 3 88 Total Responses 156 0 20 40 60 80 29 of 141 29 06 05 17 21 ECMWFE User Survey e Windows XP Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always s ame 18 Sometimes 9s SS 42 Never o La 41 Total Responses 222 0 20 40 60 80 gt fe e Windows 2000 NT Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always a 12 Sometimes 74 36 Never 108 52 Total Responses 206 0 20 40 60 80 e Windows 95 98 ME Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always T 4 Sometimes 2 E 16 Never
57. C rerun of forecasts with HIRLAM EC forecast data from MARS modelling of source receptor relationship of air tracer the models are driven with windfields retrieved from the ECMWF inverse modelling of greenhouse gases data assimilation model development at ECMWF preprocessing of ECMWF meteo data Using IFS to assimilate novel retrievals from EOS MLS User support and model code development at Reading University Using ECMWF ERA 40 archive to re run the UK Met Office Mesoscale Model for cases of interest Typically convective initiation and banded cold frontal rain Planning the Convective Storm Initiation Project using daily forecasts from the Met Office and ECMWF to decide when o operate the observational instruments User of Forecasts deterministic eps monthly seasonal comparison of the IFS model output with GME SSO Parameterisation I am responsible for monitoring of the data assimilation system and data impact studies Managing provision of Metview services on Met Office systems Personal use of Metview and ECMWF data for research studies mainly involving evaluation and improvement of Met Office NWP models Advising on and supporting other user applications to maintain develop and support the PREVIN Ensemble Forecasting System that provides products for use by Forecasters for our medium range forecast services 18 of 141 29 06 05 17 21 ECMWF User Survey I am currently working on development for the Met
58. CAR NCAR is a very good tool and it has things that ECMWF does not offer however I don t think a lot of resources should be invested in having those things Output from GCMs am not sure if this is offered from ECMWF I never compared but everybody believe ECMWE are better than UKMO UARS ECMWF has large relevant data sets but not easy to download and apply Free access analysis observations daily forecast Very favourably especially ease of archiving retrieving data at ECMWF the other centres are very good especially Irz ECMWF is excellent Access to data is easier and does not require user authentification etc there is nearly no documentation for aladin data available Bigger and faster But further and english speaking free analysis and forecast data in real time Very high acceptance offer high reliability Comparable ECMWF compares very well 136 of 141 29 06 05 17 21 ECMWF User Survey Very good hardly unexpected failures and no intelligent data migration which in practice only slows down your io ECMWF is much better than our in house databases though ECMWF lacks sufficient resolution for most of our purposes e g wave data in the Baltic which shouldn t even be provided with the ECMWF resolution if you ask me generally ECMWF is more difficult because it is designed for the expert user easier access to data retrieve I think that ECMWF is at least as good as other centres or even better ECMWF pro
59. DKRZ is offering much more disk space beside that ECMWF is offering the best service I ve not really done similar things at the two centres Some provide facilities to do on line statistical analyses Eg correlate your series with ENSO Impossible to get technical assistance at NOTUR Large storage memory models used by the French community 134 of 141 29 06 05 17 21 ECMWF User Survey Easier local access Better user support at ECMWF but they give easier data access Easier access to data There I had personal contact with system managers but I do have a very good feedback with the system manager at ECMWF ECMWF is more reliable Satellite data in Mcidas format Less downtime than DKRZ and SARA The processors of the IBM at ECMWF are faster than those of the SGI at SARA Tape system at SARA is not as well organised as ECMWF ecfs ECMWP s offer much better support Ease of availability convenience of operating system I could used ECMWF more frequently but generally the local centre fits my needs ECMWF is very professional and quick to respond in case of problems Good yes I every day compare GRIB data from AVN GFS and ECMWF some times AVN GFS is better some times ECMWF data is better for now I don t see rule but we work on this and make some software for verification for EtaMN model start with AVN GFS data with EtaMN numerical area model with same parameter start with ECMWF GRIB data ECMWF offers a lot o
60. ECMWE User Survey Report on ECMWEF User Survey This report summarises the responses to a questionnaire sent on 1 2 2005 to all external registered users Of the 1267 users invited to take part in the survey 419 33 1 completed the questionnaire i e went through all pages and pressed the FINISH button on the last page whereas another 100 users 7 9 provided some useful answers but did not press the FINISH button 676 users 52 6 did not respond to the invitation One reminder was sent The survey closed on 25 2 2005 To protect the anonymity of the users some answers are not shown or rephrased Some answers are shown as n where n is the counter of identical or similar answers Additions are marked by rectangular braces For technical reasons the following represents users at Meteo France only through their free text answers About you personal information In this section we would like to find out about you If you would like to receive a copy of your answers to this survey and to be informed of the results then you must supply your e mail address Your name Answers not shown Your login name user id Answers not shown Can we get in touch with you to follow up or clarify any of your comments in this questionnaire if yes please supply your e mail address Each Respondent could choose only ONE of the following options Response Total of Total Respondents cs Total Responses 425 0 20 40 6
61. EP 7 SARA Amsterdam 6 CINECA Bologna Italy 6 KNMI 6 NCEP NCAR 5 INM 5 NCAR 5 NOAA 5 NOAA CIRES Climate Diagnostics Center 5 NOTUR Trondheim Norway 4 DMI 3 CSC Finland 3 CSCS Manno Switzerland 3 MPG computing centre Garching 3 NSC in Link ping Sweden 2 CSAR 2 Earth Simulator Yokohama 131 of 141 29 06 05 17 21 ECMWF User Survey 2 HLRN Norddeutscher Verbund fiir Hoch und Ho chstleistungsrechnen 2 ICTP International Centre for Theoretical Physics Trieste Italy 2 IRI 2 LRZ Munich 2 NASA 2 NCEP reanalysis 2 NILU Norwegian Institute for Air Research 2 NOAA ARL 2 NTNU National HPCF in Trondheim 2 PCMDI Program for Climate Model Diagnosis and Intercomparison Livermore 2 Rechenzentrum Juelich 2 ZAMG 1 Aladin data 1 Australian Bureau of Meteorology 1 Casablanca 1 CASPUR Inter university consortium for the application of super computing for Universities and Research Rome 1 CCRT CEA Bruyeres le Chatel France 1 CEA Grenoble 1 CGAM University of Reading 1 Computing services at MIT MA U S 1 CSAR Manchester 1 Czech Meteorological Service 1 DLR Computing Centre Germany 1 ENEA Rome 1 GFDL NOAA http www gfdl noaa gov 1 GISS http www giss nasa gov 1 HLRS Stuttgart 1 IFREMER Cersat France 1 Im aladin vienna 1 MAP data center 1 Meteo computing and data centres France 1 Meteorological Japanese Agency 1 MeteoSwiss 1 Montreal 1 NASA PODAAC 1
62. FE staff have an open mind and are always willing to think with you and solve your questions if they can Their response to questions and their expertise Everything I am very satisfied with all that I used till now My problem is that I don t have too much experience to tell you more comment concerning the improvements ECMWF it is a good example of European cooperation Then It provides good forecast fields I used only MARS service which is quite good MARS Maintain quality of forecasts and analysis MARS performance Those that I use user support is very good Very good HPFC facility people are kind competent and ready to assist you forecast production ECMWF remains the Centre of excellence that we are proud and priveledged to be associated with over the years Above all the reliability and timeliness of the forecast products The services ECMWF provides and the support is great The data are easily and fast available ECMWF does a lot to make there products easily usable Operations is very good archiving data Availability of archived data ECMWF training courses education Computing facilities Training presentation of forecasts and analysis newsletters education Professionalism of staff ECMWF is doing a great job in providing data to the climate community I think the helpdesk is also doing a great job in helping users in an efficient way 122 of 141 29 06 05 17 21 ECMWF User Survey meteorologi
63. My work at ECMWF mainly concerns analysis of atmospheric data I am currently a Professor head of the Laboratory of Climatology and Atmospheric Environment University of Athens Atmospheric modelling development of atmospheric models Study of predictability of the atmosphere Postprocessing and verification of ECMWF and MM5 products Running of METU 3 local wave model I have been working at NWP section I am using Metview MARS and ecgate I am also system administrator at TSMS We calculate backtrajectories corresponding to aerosol samples collected at different parts of Turkey Development of a hydrological ensemble prediction system supervision of operational activities at the weather room connection withe the modelling section participation in international projects 13 of 141 29 06 05 17 21 ECMWF User Survey I extract the ERA40 reanalysis on Greenland Europe to force my regional climate model at the lateral boundaries at RMI R amp D software management Supporting ERA40 work at EUMETSAT running Re processing MPEF and MSA Retrieving Fest and sfc upper air obs for RMPEF and MSA I also use Metview ECaccess i work for the meteorological operations dept at EUMETSAT Subject of my research interpretation of the isotopic composition of Antarctic snow Approach Modelling the isotopic composition of water vapour and precipitation along back trajectories for the Antarctic region ECMWF I use ERA 40 data as
64. NCAR research data archives USA 1 NCAR web archive 1 NCDC 1 NCEP NOAA 1 NCEP NOAA data archives USA 1 Norwegian data retrieving system 1 NWS http weather noaa gov 1 regional centers 1 SMHI 1 SSEC Space Science and Engineering Centre 1 Strato Rechenzentrum Germany 1 Tunis 1 UNIDATA 1 various climate data access services 1 various data centres 1 various met services 1 World Data Center System A 1 WOUDC World Ozone and Ultraviolet Radiation Data Centre Canada How does ECMWFEF compare to them What do they offer that ECMWF doesn t tr gt 6 ECMWF is better 2 ECMWE performs much better that other facilities ECMWF appears more user friendly also turn around time is shorter I m simply more used to working on their computers 132 of 141 29 06 05 17 21 ECMWF User Survey Well More customer driven ECMWF is much better in terms of reliability service documentation information ECMWF compares very well NCEP have a nice service that allows you to see exactly what parameters GRIB that are available for every forecast length of their GFS model Maybe this already is available at ECMWF website I don t think I ve looked at every page here Otherwise it would be a nice feature easier access the UKMO a more beautiful web based service The DWD nothing Different Products mesoscale information NADIR serves a different community and we collect atmospheric monitoring data and cam
65. Office PREVIN system which displays ECMWF ensemble data as a variety of products At ECMWF I ve been learning about MARS retrieval and ECtrans to acquire data Research Scientist Work in initial conditions preparation ensemble prediction systems breeding thecniques Research about high impact weather in the Mediterranean and cyclogenesis events I m Head of NWP team at the Met Service working mainly with HIRLAM model at home and at ECMWF Using PrepIFS from time to time for research Developments in data assimilation upper air and surface with the HIRLAm model Visiting Scientist position for the NWP amp CM SAF My main task nowadays is to develop a wind stress product from ERS data At ECMWF I am currently developing an ERS ECMWE collocation software which includes wind data as well as a set of surface stress parameters Developing operational 3D and ecosystem models for the Baltic Sea collaboration with the Predictability section at ECMWF to develop improved Stochastic Physics representations also running a cloud resolving model as a member state user to produce data that aid development of stochastic physics my work is to issue regional weather forecasts for the general public and the decision makers in the field of air quality control civil defence agriculture I get the ECMWF model output fields through the National Weather Service and access the MARS archive for hindcast activity with an air quality dispersion mo
66. Other please specify 4 1 Total Responses 404 0 20 40 60 80 e Other please specify I don t think I use this I don t use WebMars only scripts on ecgate Not yet but may try it in the future Not yet rarely How satisfied are you overall with the service provided by WebMARS Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 56 a 20 Satisfied 130 iii 47 Neither satisfied nor dissatisfied s 30 Dissatisfied s i 3 Ven dissatisned o 0 Total Responses 277 0 20 40 60 80 Your comments DE 4 did not use it so far 2 can be slow I haven t taken the time to learn I hope I shall use it in my future work It would be very helpful and useful that the resolution option was not fixed as with ERAI5 data happens fields with low resolution great 61 of 141 29 06 05 17 21 ECMWF User Survey Too many problems with headings on maps choice of data e t c Would like the Custom option for area when plotting data I do not know how to use it Perhaps if I could use this service I could solve some of the problems indicated in the above questions but unfortunately my limited time did not allow to me to get the useful informations about this service It is not difficult to get lost in the process but it is better than similar products by other organisations see for example http iridl ldeo columbia edu in
67. RS 1 HPCF docs 1 IBM technical documentations 1 Ihave only tended to access documentation on MARS 1 I only read MARS documentation 1 lecture notes 1 MAGICS training course 1 making crontabs 1 MARS archive 1 MARS catalogue 1 MARS modeldatafinder 1 Memos papers Meteorological aspects 1 model developments 1 most of the documentation available 1 no preference most documentation is presented professionally and easy to follow and in very helpful 1 on available products and available tools 1 on line material 1 paper based 1 prepIFS 1 Printed documentation sent when I first got my account 1 products description 1 proffesional 1 reports documentation 1 scientific 1 SMS 1 technical manuals 1 Technical Manuals ECaccess etc 1 Technical Memos especially on DEMETER and seasonal forecasts 1 Tech Notes 1 The documentation received in March 2004 during the training presentations copies 1 the guides 92 of 141 29 06 05 17 21 ECMWF User Survey 1 The old printed MARS manual 1 The theory behind the ECMWF model 1 The User Guide I like the following documentation least 5 MARS 5 MARS user guide 3 IFS documentation 3 PrepIFS 2 BUFR GRIB encoding decoding software 2 Metview 2 on line based 1 Accessing GRIB and BUFR data 1 access to Tables of parameters on web site 1 administrative 1 BUFR manual 1 BUFR needs more examples 1 Computer documentation out of date but I must conf
68. a retrieval and decoding post processing Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low 119 29 High 209 Not relevant 71 52 19 l Total Responses 405 0 20 40 60 80 e Other applications e g trajectories Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low 80 P 22 High 41 Pe 11 29 06 05 17 21 ECMWFE User Survey Not relevant 242 i 67 Total Responses 363 0 20 40 60 80 e Developing and executing your own model Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low 67 La 18 High 98 Si 27 Co operating with ECMWF using prepIFS Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low 51 eer 16 High 41 es 12 Total Responses 354 0 20 40 60 80 e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low js i 7 High 36 eS 14 Not relevant 195 78 Total Responses 249 0 20 40 60 80 e Please specify other activities that describe your work at ECMWEF 2 Metvi Please give a short description of your work both at ECMWE and in your organisation collaboration with ECMWF scientists on basic problems of me
69. ability after login procedure very strong The operational services works very well I m very pleased with all the work ECMWF have done in getting the targeting ensemble prediction system into operation I m very pleased for all the information advice and explanation we have got from Dominique Lucas In addition to this I m pleased for the data available at ECMWF which makes it possible to for us to run so many different models and to contribute to better forecasts and also hindcasts in research purposes Data availability forecast model products In my opinion ECMWF experts do their best They all always ready to help out and are very friendly Graphical outputs generally good I m most pleased with the MARS system It s easy to use and useful User service An exceptional service provider for meteorological data and computer resources In general I am overall satisfied Nothing in particular stands out skill of forecasts 121 of 141 29 06 05 17 21 ECMWF User Survey Efficiency user support and call desk are very efficient You are doing a very good job Overall professionality I am very satisfied with the way the MARS system works and with the ECMWF staff support Availability of historical forecasts for LAM and hydrological model applications Ease and speed of data retrieval Professionalism data availability user support efficency Very good data fast retrieval with MARS good user support ECMW
70. ability in system Large number of documentation 120 of 141 29 06 05 17 21 ECMWF User Survey website education on website It is a useful site which I have used since 1978 intermittently The huge amount of parameters the high space resolution of fields Help from people from Metview and MARS systems and from User s Support High quality meteorological products professionalism All in all the different kind of products informations software support etc from ECMWF are great the professionalism and high quality products I get practically all general data I need further I got several times immediate help for solving my problems call desk and support The cooperation between all contributing countries it is a success combination of meteorological competence educational program and computer resources improvement of forecast products forecast products MARS data archive Computing infrastructure is excellent with good support We used ECMWF for a very small part of its potentiality i e just for retrieving data we are satisfied about this part Also we are satisfied about ECMWF Newsletters news I use the web services most and am quite pleased with them The website however takes some learning to use effectively This being said it is true that most good things take some learning I think it is very hard to design a website to give access as varied data as you have in an easy fashion quality of the data and avail
71. achines for which openMP works fine we had enough computer resources for our previous work until now and have no experience on the HPCF but we plan to test our programs on the supercomputer soon and maybe use it them In fact cluster computing is of more interest to me i e running the same program lots of times on different data The vast improvement of ecgate over ecgate means that I can run all my serial model calculations on ecgate Often retrieval of forecasts analyses from MARS is the rate limiting step in the turn around of my forecast scripts In Zurich we had good supercomputing facilities now in Mainz it could be a very good option to run some models calculations directly at ECMWF I guess that the support would be excellent as it is for any other questions I had sofar to ECMWF The assistance is really efficient However the main problem is accessibility through a remote telnet ftp that sometime is quite low In addition disk space is usually limited for several tasks Indeed with my facilities it is not possible to transfer simulations outputs to my institution via ftp due to the large amount of data and disk resources on HPCD are limited I m now trying to realize a routine to extract main data from my simulations and to transfer them at home In addition disk quota is also limited on ecgate and I must use it to transfer data from HPCD to my computers The specified reason can change soon getting the permis
72. ain reasons for contacting User Support Call Desk BE e General advice on the use of computing resources Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 38 EE 14 Often 66 25 Sometimes 9 33 Rarely as 17 Never jo 11 102 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 268 0 20 40 60 80 e General advice on the handling of retrieved data Each Respondent could choose only ONE of the following options Response Total of Total Respondents Often Ea ee 22 Sometimes 0 _ amp 30 Rarely 22 Never 3 Mm 20 Total Responses 262 0 20 40 60 80 Specific meteorological questions Each Respondent could choose only ONE of the following options Response Total of Total Respondents Often Ea ee 5 Sometimes 2 ZZ 18 Rarely 26 Never 115 C 49 Total Responses 237 0 20 40 60 80 e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always j6 MM 13 Often 2 16 Sometimes j9 Mm 15 Rarely io jm 1 Never 63 i 50 Total Responses 127 0 20 40 60 80 e Please specify other reasons for contacting User Support Call Desk f BEEE i BBE PPE a 7 Technical problems 3 access problems 103 of 141 29 06 05 17 21 ECMWF User Survey 2
73. aining the operational Met Office seasonal forecasting system Research into quantifying uncertainty in model predictions using seasonal to multi decadal climate model integrations Numerical simulations of balloons trajectories Predictability Probabilities Uncertainty Retrieval of forecast fields mainly ensemble output regional atmospheric and ocean modelling data analyses Comparing AIRS retrievals with ECMWF profiles Mesoscale modelling of air pollutant dispersion trajectory calculations with own model using MARS data regional climate model System Administrator providing support for internal users Security Representative Retrieving ERA4O data for glaciers area in order to couple climate change with glaciers respond I work most of the time as a forecaster at the Finnish Meteorological Institute small part of the work time is for staff education specifically in Synoptic Meteorology At ECMWF I use Metview especially the macro language BIRA IASB assimilates ENVISAT chemical observations with BASCOE and develop stratospheric model ECMWF provides temperature and wind fields used by these models I was involved in DEMETER as a mathematical modeller of malaria I was involved thus in taking the DEMETER and ERA 40 output and creating a model of malaria risk Meteorological analysis Climatological studies Ensembles Neuronal Network Operational forecasting UK Met Office Unified Model System Support and Developmen
74. also referring to the first days of forecast as I am an University professor I am not doing commercial forecasts Some time ago I could make a request on the web and see the details of the request in text format I then used the salient part of those to write request that I submitted by telnetting in through ecgate This was useful since it gave my a way to find the right streams and GRIB numbers for fields I wanted to pick up in higher resolution only special need for my work I HOPE More forecast and analysis products Easier access to long range forecasts for particular towns More and better documentation more free forecasting products Access to forecast products More limited area model products Newsletter More timely updates on what is happening wrt reanalysis availability of products Information better information about updates possibility to few old forecasts more forecast products More recent publication lists A regular email advertising e g manuscripts of ECMWF people accepted by journals that can be downloaded from the web site would be very useful additional fc and EPS products for member state users My own demands It seems already quite optimal ECMWF user forum FAQ related to the members questions probably many questions are repeated data information With more scientific documentation better clearer structure More relevant documents Easier access One day forecasts in Your room But this
75. an input for the trajectory model Furthermore I use ERA 40 for e g snowfall data Analysis of air sea flux fields mainly from reanalysis Ocean waves Large scale ocean atmosphere interaction I have been working in the validation and dissemination of the ERA 40 wave data I am the chair of the Atmospheric Composition Modelling group of KNMI Climate model development and validation Use models to understand climate changes Regional climate modelling Mainly use ECMWF to retrieve various boundary conditions Develop HIRLAM codes related to DMI s operational activities and the international HIRLAM project data assimilation experiment for evaluation parallel experiments Developing and improving numerical weather prediction model Climate modelling and in this context critical appraisal of observational datasets including but not restricted to reanalyses Managing daily dissemination of ECMWF forecasts and analyses getting MARS data for customers and for developing MOS at AUSTRIA computing several forecast products for Austrian Weather service weather charts forecasts for several locations in Austria MOS forecasts Research in regional climate modelling especially feedback mechanisms within the physics package and model validation I use GRIB data from global model from ECMWF for 1 draw like profi meteo pictures 2 Run area Eta meso model We are heavily using ERA40 data from ECMWF database ERA4O data is used at
76. archive of the last week s forecasts My interests are in the EPS products Full information on deterministic forecasts from t 6 and not only from t 72 and also much more fields is not only a will but a necessity in the era of information society better indexing The need to find info on HPC at ECMWE and on other services as the storage system there may be a discisson forum concerning ECMWF tools sofwares programming etc The information exchange between member states will be useful as well as ECMWF to request for backtrajectories checking verification of some of the MPEF products checking weather forecast for the next ten days Most recent forecast easily available e g pressure temperature precip maps and epsgrams with just one or two mouse clicks Use the IFS and ERA documents Check the manuals for use of the computer and MARS facilities Look at the forecasts Need for other data services 1 A interactive trajectory computation similar to http www arl noaa gov ready hysplit4 html 2 If you give money each time I visit www ecmwf int ECMWFE other products and documentation higher resolution of atmospheric model Perhaps I would like to see more detailed graphics higher resolution contours for basic fields such as PMSL 4hPa 1000 500 hPa thickness 2m temperature from the deterministic forecast Email notification of news from ECMWF with links to articles updates on the web 88 of 141 29 06 05 17 21
77. ary system administrator of our local supercomputer and the administrator of the ECaccess gateway Thus I am not directly using the data provided from ECMWF but I assist other users in trying to access the data of ECMWF operationals and managerials now I use observations reanalysis products and models to examine Antarctic climate change I have used reanalysis data for climatological studies including back trajectory analysis Operational data have been used for validation purposes Numerical weather Prediction comparison between observations and model outputs Statistics on atmospheric fields Air quality modelling at urban and regional scale Regional climate modelling for Ireland Offline chemistry transport modelling The model is driven by ECMWF meteorological input OD as well as ERA 40 I am involved in several EU projects The goal of one of my projects aims to make a reconstruction of the changes in tropospheric composition during the period 1960 2000 on the basis of ERA 40 I am also involved in a study for the IPCC AR4 which aims to make a projection of the air quality in the year 2030 In my organisation my work concerns the operational models for the storm surge forecast in the Adriatic Sea These models are forced by ECMWF fields At ECMWF I retrieve MARS data of particular storm surge events to realise hind cast simulations with the hydrodynamic models The aim is the improvement of the sea level forecast in Venice
78. ast for non european states Possibly more real time forecast information such as forecast maps and meteograms e mail announcements of new content When I get a new task at my work place 1 tutorials 2 computer courses material handouts Forum Easier access to more products data that can be downloaded 1 We wish that the EC maps on the web that we look on could be seen with all in one window and a scroll list instead of change map for every 12 hour step 2 Another wish A map with daily accumulated precipitation on the same area as the others If I could retrieve observations and soundings If there would be a better weather forecast visualisation provided there I have no idea I find almost everything I need concerning the use of ECMWF through the documentation I received in March 2004 special training Availability in real time of analysis and next two days forecast Availability via web for user to specify local meteorological forecast products like information about wind direction in the EPS Meteogram If the problem with logging onto EPS products part of the site could be sorted out Notification of latest additions such as model and archive news If less busy a better webMARS 1 List of New items 2 Quick documentation of current meteorological configuration of operational system and of computer system this is a general improvement not mainly to make me use it more often The need to do so More easy access t
79. ation in PDS GRIB is not easy to use Use of WGRIB is inefficient requiring considerable preparation A standard routine for accessing data values is required In addition when needing large files there was not enough room for temporary storage Thus bandwidth and disk space became considerations Which data are you mainly interested in Type of data Please specify the importance of the data type or tick Not relevant e Fields Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 333 93 Low La a 5 Not relevant lo 3 53 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 359 0 20 40 60 80 e Observations Each Respondent could choose only ONE of the following options Response Total of Total Respondents o High 124 ia 37 cae 3 m 40 Not relevant 7 23 e Total Responses 334 0 20 40 60 80 Type of archive dataset class Please specify the importance of the archive dataset or tick Not relevant e Operational data Each Respondent could choose only ONE of the following options peme E 276 80 EEL 12 Not relevant 27 E 8 Total Responses 344 0 20 40 60 80 e ERA 15 Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 72 i 26 Low 105 SS 38 BEK e ERA 40 Each Respondent could choose only ONE of the following opti
80. ations thanks to Improve data access It would be nice if there was a RPM package available to the NMS s with Metview dependencies Then we could simply set up a computer with the correct architecture e g Linux Suse 9 0 or something else specified by ECMWF and then install the package without the hassle of compiling it ourselves Issues relating to MARS access etc could be handled by having more than one RPM packages It would have been an advantage if I could use my SecurID card from other machines than the one at work Faster MARS server and faster responses to problems 137 of 141 29 06 05 17 21 ECMWF User Survey Providing more realtime forecasts products recent level is good enough Permanent disc space for model code run scripts etc Greater computer resources easier access automated from outside May be a minor detail but would help us To provide easier routines tools to retrieve or convert data in formats other than GRIB I would like to participate some courses Lobby to widen our connections to the Centre The ECMWF is very closed in the use of format files I think that GRIB and BUFR files need to be complemented with tools or posiblities to manages other formats like HDF binary etc that could be more easy to use by people Larger disk space upload and download of large files one straight access to data sources no multiple gateways If the ECMWF data will available accessible remotely e g in a
81. c Sea I m working on a PhD in which the influence of land use changes on the Sahelian climate is investigated using a mesoscale climate model which is implemented with ECMWF data Numerical mesoscale modelling Daily retrieval of operational data to keep up to date an archive for our trajectory model TRAJKS both in ECFS as at my home institute Run trajectories on semi operational base automatic during campaigns ECMWF Run trajectories for research KNMI chemistry transport modelling Model Development and Evaluation of the KNMI Regional Climate Model RACMO2 operating the RCM for present day climate forcings from ECWMF ERA40 or a global climate model and future climate forcings from a GCM work on ECMWF facilities primarily consists of compiling boundary forcings for the RCM but in the recent past some development work on the RCM was also carried out on the ECMWF facilities and part of this work might be resumed in the near future Diagnostics of climate variability Modelling studies of atmospheric large scale circulation simulations with simplified models Short and long term predictability studies Study of the atmospheric variability both on tropical and extra tropical regions Satellite remote sensing of geophysical variables over the ocean ECMWF retrieving data ERA 40 oper analyses oper forecast for past cases situations Home downscaling ERA 40 analyses coupling a LAM with ECMWF forecast or analyses for
82. cal archive There is general information on the ECaccess is a great thing for user convenience supplied software on ecgate and HPCD missing special issues are with the compile environment and the MPI libraries there is insufficient information on path levels and no information on pending and solved bugs The support I get from my contact person is very good The MARS and ecgate systems I use for my data retrievals work nicely Efficient computer environment in HPC Modelling quality of products ECFS file storage system is superb Very easy to use and I have never had a problem Provides a large amount of useful forecast and observed data DEMETER website is very good and comprehensive and contains interesting results Reliability in the dissemination of GRIB streams My impression of the group in User support is extremely good very helpful and professional Availability for meteorological forcing from MARS is also very good Archived met analysis easily accessible Good support computing environment and archive MARS retrievement Computing facilities ECFS In my opinion is difficult to find other Institution with the same high level of service ECMWF does research and education user support training documentation and assistance if needed data service computing environment I think that we can learn from ECMWF how a big computer system can be build up and can be held running in a very professional manner The
83. ce is the most reliable I know When something has to run I sure that will run ECMWF has much professionalism P utilise peu le Centre europ en une seule application quotidienne automatique de simple de d sarchivage et transfert de champs vers une machine de M t o France a Toulouse Les services assur s par le Centre conviennent tr s bien au petit utilisateur que je suis efficiency The best the user seems to be at the Centre s activity an attempt is made to have updated documentation products are generally user friendly mars ecfs user support computer systems reliability and availability Web services MARS access to forecast products 127 of 141 29 06 05 17 21 ECMWF User Survey meteorological data archive Well so many things I like meteorological products Access to computing facilities and PrepIFS cdp contact with users of computer resources possibility to archive tons of model output What should ECMWF do differently How can ECMWF improve more space for visiting scientists MARS data retrieval procedures I find sometimes slightly enigmatic However this is probably also partly due to giving it less attention than it needs Make the web site easier to navigate and more intuitive about where to look for things there are a lot of useful resources and I feel like they re wasted because I can t find them To allow easier access to real time forecast products like EPS and monthly s
84. censing Agent ECOMET contact point Italian Meteorological Service Catalogue responsible Wind and wave retrieval from ECMWF Wind and wave retrieval from SAR data PhD student working on Arctic Boundary layers Work at ECMWF is retrieving boundary conditions for meso scale simulations chief in Short and medium range department I study dynamic of the stratosphere tropopause in polar region I retrieve the ECMWF data and calculate some secondary parameters e g PV 7 of 141 29 06 05 17 21 ECMWF User Survey In our organisation 4D VAR chemical data assimilation using IFS winds forecast and analysed dynamical fields Emphasis on stratosphere Forecaster also responsible for some trajectory models contribution to the HIRLAM model surface parameterisation Developing a limited area ensemble prediction system Testing targeted EPS Studying forcing singular vectors and forcing sensitivities Group leader and working with atmospheric chemistry transport model on photochemistry and aerosol Long term hydrodynamic simulations of European Regional Seas using ECMWF atmospheric products for forcing the ocean model climatic summary and elaboration mesoscale modelling using ECMWF analyses as forcing climatological studies using ERA40 calculations with an Lagrangian particle dispersion model model improvement regional climate change investigations Feature based diagnostics of analyses re analyses and EPS Calcu
85. cess to computing resources at Meteo France however I currently under use HPCF for the reason I working on topics that use more metview and other tools that do not need to run on HPCF Why do you find the HPCF service unsatisfactory I find the HPCF unsatisfactory because e g insufficient processing power insufficient disk space poor turnaround difficult interactive use difficult batch use bandwidth limitation unavailability of software etc No answers received Archive Services MARS In this section we would like to find out about your use of the archive services provided by MARS How often do you retrieve data from MARS Each Respondent could choose only ONE of the following options Response Total of Total Respondents Daily 2 el 13 Several times a week as E 11 Once a week 38 9 Occasionally 219 53 Never a ain 10 47 of 141 29 06 05 17 21 ECMWFE User Survey Other please specify 18 i 4 Total Responses 413 0 20 40 60 80 e Other please specify 3 once a month 1 2 month sometimes daily 1 or 2 times a year Intensively during campaign periods daily automatic retrievals for balloons trajectories every couple of months I have to update our windfields for trajectory calculations once a month or more if it needs during some periods occasionally to daily On and off intensive in periods hardly ever now sometimes more often than others several
86. cipate At the moment support is only given to met services User friendliness of web page and documentation Probably some or another user interface artist expert is needed Giving the link about other data centres which can provide with the data ECMWF does not Provides conservative wind fields at lower resolution than original I think everything I ever needed is correct Iam sure there is a lot of data available that we could use in our research but the amount of time it would take me to learn your different retrieval systems may not always justify the benefit sometimes it is hard to find on the webpage Making retrieving of observations and soundings possible User interactive web for producing forecasts very basic and straightforward To a beginner the documentation of the MARS content is not very easy to use improve webMARS Increasing the number of documentations information available putting new products on the web allowing more customisation documentation on website is very large and sometimes difficult to handle Metview macro language is sometimes complicated It s not clear how users area on ecgate are backuped Other centres offer a backup service like Time Navigator and the user can select a restoration What other computing data centres have you used 38 none 20 DKRZ Hamburg 15 UKMO UK 13 University Computer Centre s 11 DWD 11 IDRIS CNRS computing centre Paris 11 Meteo France 9 BADC 8 NC
87. ck in terms of speed improvements are always possible It is easy to surpass the 2GB limit in MARS for files In this case the job simply aborts I don t know Here again I m reflecting the views of users who routinely retrieve data from MARS It would be nice if the data could optionally be decoded from GRIB to ASCII in a Mars job MARS itself is fine The difficulty is more the varying contents of the database But that s of course in the operational part not preventable Rather slow especially if need much data satellite data is not very well represented in the archive There is only one thing which is not great I have to make large request retrieving all the forecast over a month for instance and it is sometime very difficult to get it to do the job to the end Lately we have retrieved data namely upper air data for a case study and there were a lot of missing data in the European Area Looking at the Internet we found those missing data in a database in the USA We were wondering why as these data are available they are not included in MARS though we have since the introduction of frames as boundaries for HIRLAM been unable to extract from MARS frames in the same grid as those disseminated This problem was finally solved last month 49 of 141 29 06 05 17 21 ECMWE User Survey Excellent archiving system Error trapping still isn t good enough batch interface Once I understood the scripting methods I could d
88. cy in meteorological and computational areas I am very satisfied mainly with the overall performance and excellence of ECMWF statutory duties MARS archive Large capacity data storage with fast retrieval The professionalism and willingness to help of those who are responsible for ECMWF services is greatly appreciated The ECMWF services I am most pleased with is the MARS service User support activity is essential forecasts retrieval of analysis data Reanalysis analyses and forecasts continuity of services overall reliability workshops seminars and courses the library Extended range forecasting products and ensemble data I don t use ECMWF services intensively but I like the ECMWF educational program I ve attended training courses and it s always quite easy to achieve any task you begin High Performance Performance Facilities MARS archive data archive efficiency reliability User support is very helpful to the somewhat cryptic access to MARS for beginners Provision of climate meteorological data my room EPSgrams MARS I think ECMWF is a model centre for computing services I am very pleased with ECMWF services It has a very friendly access I like the visits User support They are extremely useful ECMWF gets the job done Expert knowledge is made available to whoever needs it Staff at ECMWF are most collaborative 124 of 141 29 06 05 17 21 ECMWF User Survey The mighty web service that
89. d o 0 Total Responses 282 0 20 40 60 80 e Responsiveness efficiency Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 198 10 Satisfied s 27 Neither satisfied nor dissatisfied 7 l 2 Dissatisfied o 0 Very dissatisfied ia 0 107 of 141 29 06 05 17 21 i BE ECMWFE User Survey Total Responses 281 0 20 40 60 80 e Knowledge Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 190 ii 67 Satisfied 86 eS 30 Neither satisfied nor dissatisfied eo l 2 Dissatisfied o 0 Very dissatisfied o 0 Total Responses 282 0 20 40 60 80 e Understanding of problem Each Respondent could choose only ONE of the following options BE Response Total of Total Respondents Very satisfied 13 i 61 Satisfied 101 SS 36 Neither satisfied nor dissatisfied ia 3 Dissatisfied oO 0 Very dissatisfied oO 0 Total Responses 282 0 20 40 60 80 pE e Patience Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 188 Ei 68 Satisfied 7 SS 28 Neither satisfied nor dissatisfied mo i 4 Dissatisfied ia 0 Very dissatisfied oO 0 Total Responses 277 0 20 40 60 80 e Professionalism Each Respondent could choose only ONE of the following options
90. d to receive newsletters directly by email Where do you access the ECMWF web site from e Place of work Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 305 m 77 Often ss E 21 Sometimes s fi ca Rarely o o Never ho 0 Total Responses 396 0 20 40 60 80 e Home computer Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always s u Often a2 m Sometimes EE e 28 Rarely s2 Never 105 Total Responses 286 0 20 40 60 80 e While travelling Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always ho 0 Often ja K 6 Sometimes 2s Pa 11 so C Rarely Never 84 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 249 0 20 40 60 80 e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always ft 1 Often o 0 Sometimes h2 l 2 Rarely 2 l 2 E Never 118 96 Total Responses 123 0 20 40 60 80 E e Your comments or other locations from where you access the ECMWF Website Only use home PC when work one is broken Service appears to only work at place of work Despite several trials I never managed to log in at the website from home Research stays Occasionally from a
91. data in our own database operational and research meteorologist surface meteo data retrieval and conversion at the ECMWF site application of the data in the studies of the meteo marine conditions mainly in the Adriatic and Ionian Seas air sea fluxes I am studying stratospheric variability with a special focus on the QBO For that purpose I perform integrations with the middle atmosphere version of the ECHAMS5 GCM at my own institute as well as at ECMWE The integrations at ECMWE have been carried out in cooperation with other external users of ECMWF Development of new UK microscale model of the atmosphere Development of techniques to interface meso scale models UM others with hydrological models for flooding initiation using ECMWF data sets I used the system 3 years ago for turbulent airflow studies amp precipitation forecasting I hope top start again to reuse the system after a gap of about 3 years Courses in meteorology and climatology in the framework of the degree in physics Assistance in the elaboration of Degree Thesis Senior scientist in a University Group G MET Influence of wind in mesoscale marine circulation in Western Mediterranean Sea GIS and Remote Sensing At INM Met Service Medium range operational forecaster At ECMWF Use and interpretation of ECMWF products only All activities related with data policy ECMWF and other data products ECMWF SecurID administrator EUMETSAT MSG and HRI Li
92. del I m analysing retrieved aerosol optical depth based on level 1 data from SeaWifs Seastar NASA By including ECMWF data I get a better understanding of my work One of my responsibilities in our organisation is to try coordinate the products that we order via ECMWE dissemination system dynamical modelling and data analysis of the middle atmosphere orography and boundary layer research and parametrisation Extracting model dumps as a possible source for starting global model at UK Met Office I manage the day to day support and maintance for our Operational Suite on the NEC Supercomputer Tropical meteorology tropical cyclones cloud lines squall lines Developing scatterometer products reprocess scatterometer data on request High resolution simulations over the Alps for MAP case studies Test and implement improvements to Limited Area Models Investigate operational problems We are trying to assimilate EOS MLS observations Retrieval of Temperature and water vapour profiles to use as initial guess in our retrieval algorithm Analysis of stratospheric water vapour profiles ECMWF data is used to provide temperature and pressure information in the stratosphere This is necessary for our profile retrieval from microwave spectra that were recorded from the ground and from an aircraft Lately we have started to compare the ECMWF water vapour product in the stratosphere with our measurements The first results look very promi
93. dels WGNE comparison Weather control service at MeteoFrance using ARPEGE for seasonal forecasts and climate simulation How long have you been a registered ECMWE user Each Respondent could choose only ONE of the following options Response Total of Total Respondents Less than 6 months just registered 26 m 6 Between 6 months and 1 year 38 a 9 Between year and 3 years 134 aa 32 More than 3 years 226 iT 53 Total Responses 424 0 20 40 60 80 How often do you use ECMWF services DE Each Respondent could choose only ONE of the following options Response Total of Total Respondents Daily 73 ar 18 Several times a week o 21 Once a week s 12 Occasionally 183 43 Other please specify Em 71 24 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 430 0 20 40 60 80 e Other please specify 4 frequently during periods 3 once a month 2 intermittently 2 occasionally to daily 1 1 2 per month 1 1 2 per year 2 a couple of weeks with intensive work within the year 1 at times and in the near future very frequently but at present not more than once a week 1 daily but by intermittence 1 daily in periods 1 daily with cron occasionally for interactive work 1 every couple of months I make intensively use of the MARS archive to update our wind fields collection 1 hardly ever now 1 intensive several times a week but only for limited time periods
94. dents Permanent 329 E 72 Temporary long term gt 2 years ss Ea 18 Temporary short term lt 2 years 2 m Student jis fum Other please specify so l Total Responses 460 0 20 40 60 80 e Other please specify 5 PhD student 1 Change of position 1 post doc 1 semi retired Bil PP Which of the following best describes your level of education Each Respondent could choose only ONE of the following options Response Total of Total Respondents Secondary education 14 3 University education 16 iii 43 3 of 141 29 06 05 17 21 ECMWFE User Survey Postgraduate education 235 ii 52 Other education 11 l 2 Total Responses 456 0 20 40 60 80 Please specify your level of education e g title of degree etc Answers not shown About you technical information In this section we would like to find out some technical information about you and your use of ECMWF systems Looking at the following definitions please indicate which category best describes your work at ECMWF 4 of 141 Please specify the importance of each category or tick Not relevant if your work at ECMWF does not include the category e MARS data retrieval only Each Respondent could choose only ONE of the following options Response Total of Total Respondents Low 92 24 High 237 i 61 Not relevant DE as 16 Total Responses 390 0 20 40 60 80 PBE e MARS dat
95. dents High 44 aie 15 Low 69 24 r Seasonal forecast Each Respondent could choose only ONE of the following options Response Total of Total Respondents High s hz 17 Low 38 a 20 Not relevant 13 I 63 Total Responses 290 0 20 40 60 80 e Other specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 8 i 4 57 of 141 29 06 05 17 21 ECMWE User Survey Low 9 Pl 5 Not relevant 181 91 Total Responses 198 0 20 40 60 80 e Please specify other data types Daily forecast Ocean wave data At moment I am not involved in experiments which need forecast data knowledge In the future perhaps they could become interesting to me Analysis feedback data WAM wave data Sensitivity fields Demeter ERA 40 Relative vorticity coeff in sh of analysis and or 24h forecast depending on my experiments Unfortunately we do not have access to real time Deterministic forecasts Only used for checking Metview Glosea seasonal forecast trials Future members of seasonal multi model ensemble Depends on date what is available Looking for best data observations These are data that we produce plus ERA4O analyses observations I would like to have the LORENZ FILES 0 to 10 days forecasts of the ERA40 dataset I don t know if they are available Ocean model data Depending on what
96. description of all the possibility to use ECMWF facilities Easy access to a procedure to transform data set from GRIB in GRID by interpolation and degribing on local my computer Speed in connections but it is a bilateral matter ECMWF NMSs Provide HIRLAM countries with full fields for timely forecast initialisation It would be nice to know about the educational programs courses a bit more in advance Allocation of disk space could be more generous assimilation scheme strength of residual circulation vertical diffusion extension to higher levels up to 100 km I took me long time and several attempts to actually get an user account But I believe that this was not due to ECMWF but rather on the side of the meteorological service It would be great if it would be possible to retrieve larger amount of data from MARS archive or if I could retrieve NetCDF files on telnet ECMWF should provide more facilities for GRIB data decoding More availability of forecast data online I need to have more experience with ECMWF services to answer that question It s too early for me to say It would be able to perform an online tutorial Handling of observational data in buffer format is currently not so easy the download speed 130 of 141 29 06 05 17 21 ECMWF User Survey There could be regional or country training organised by ECMWF in which students or researchers from outside meteorological services could have a better chance to parti
97. deterministictensemble 2 If it is possible we would like to get access to model data on the reduced Gaussian grid for just the European area See web user interactive and automatic dissemination of data to user The ability to convert data to NetCDF and the use of NetCDF tools such as ncdump and nco would be very useful The manipulation of spectral coefficients in MARS would be useful too some small improvements of MARS webMARS mentioned above Send you and email telling you what it is going to be removed from scratch and then wait for a few hours to do it And changing system sessions to Saturdays or Sundays for example instead of Wednesday morning I think it depends much more on my own Met Service Never stop for sessions Perhaps I miss existing tools methods Aside creating new tools make sure people know what you offer now to continue Have you any more comments suggestions to make Are there any other issues which have not been raised in this questionnaire that you feel are significant 20 No comments or suggestions To achieve ways e g seminars in the member s countries directly by ECMWF personnel or in coordination with the countries centres to spread better the information about their products and their use Could you make the ensemble mean MSL pressure available on the web for all continents I always feel guilty contacting the help desk user support with what are probably trivial s
98. dex html only used as an information not to get data This interface is good for finding out in which datasets a particular parameter is located very useful to find out about the availability of data also useful to assemble the MARS retrieves NetCDF functionality for quick look at data in particular since CF 1 is supported VERY NICE Very useful to try things out before making scripts or to help making scripts It is an excellent tool to verify justify our simulations on the basis on well defined analyses because of the large amount of data I retrieve it is often impractical to use WEBMARS For small amount of data it is OK not simple to find what is in the archive unless you already know precisely where to look at did not know it existed but I will now find out about it It would be very useful to have more metadata explanations with some of the more obscure choices method It would be useful to have a few default data operations built in such as differencing to get fluxes from accumulated fields summing LSP and CP etc finding out which parameters are available for a give source forecast analysis ERA is difficult because they sometimes have different GRIB codes for no apparent reason It would be helpful if they would be grouped and that only for the final selection you find out which parameters are available exactly i e I simply want vertical windspeed and do not want to decide beforehand whether this should
99. directives for fields absence of post processing directives Probably there was no need to use it was enough satisfied with normal MARS via scripts It is faster for me to get fields observations via ECaccess or csoMARS tempremental login process Faster to use MARS than webMARS Moreover both MARS calls and collocation software is already set and running at ecgate I am new to the service ignorance Because I continue to run a batch job I do not know how to specify rotated grid I have a number of MARS scripts that fit well my needs However I will try WebMARS as asson as I have to retrieve a new dataset there are easy ways to make MARS scripts available so there is no need th use webMARS I should check it out though Because I m just registered and I don t yet have done the training course about MARS I am able to obtain data by other means Don t know that service yet and we are familiar by working through ecgate I do not need it our data is usually extracted daily or covers a complete year so scripting is much more useful Ihave not found the data I need 65 of 141 29 06 05 17 21 ECMWFE User Survey I am used to connect to the Centre and use MARS directly on ecgate Since usually I have to retrieve a lot of data by means of an automatic launching system it is more convenient to directly manage the operation For a new job I normally copy from an old one which is not very difficult to do use the MARS
100. ducts available on the web could be improved the possibility to customise should be given like in case of other centres The background is very different In 1990 I was in formation to make support to other users Since 1997 I m a user of ECMWF computer to get data for another service studying chemistry or precipitation As I said before ECMWF is not perfect but it offers the most complete an relaible service it s different ECMWF is a reference for me Other centres have also good qualities just geographical proximity inside the firewall of m t o France It is more easy to download file results to my personnal computer in order to perform diagnostics My simulation files are very big 500mo ECMWF is the best in particular as far as archiving is concerned What could ECMWE do that would significantly improve your work productivity more meteorological consulting services I need that the documentation would be more quickly searchable Change colour at present white on Pale of FORMOST o p on hp ux Speed up MARS Editing programs would be faster if I could use e g emacs instead of vi supply a MOS or Kalman filtering with the operational products for European weather stations If I was more knowledgeable This is probably my fault but I am one of probably very few university users and therefore not worth the effort see previous remarks more storage in SCRATCH make easy access to upper air data and surface observ
101. e meteorological fields ECMWF usually provides routine meteorological model services on a very reliable basis I appreciated the level of the training I had last year in March to learn how to use the computing facilities ECMWF documentations are really accurate Forecast and analyses The seminars and conferences held at the centre are of a high standard The tours and talks given to visiting courses from the Met Office College were excellent a shame we re based so far away now to Data access is easy and the computing resources easily cope with my data requests User Support and Call Desk always resolve any problems very quickly Ecaccess to get data from ecgate ecfs to local systems Daily production on ecgate triggered by sms control MARS Good response on ECFS MARS etc and lots of resources User support has enabled me a non expert to get what I need Top give frequent info about ECMWF system sessions state very professional services and help I m very pleased about being able to retrieve data from the MARS archive Also I find the work of call desk and user support very efficient and useful People working at ECMWF in specific areas are also very kind patient and helpful if I have problems I only change parameters in scripts written by someone else ECMWE Support help me when I meet difficulties to run jobs I use to work on several computing platforms all around the word and I want to say that the ECMWF computer servi
102. e of referencing field sets should be described in more detail or with more examples An exact detailed definition of all parameters archived in MARS would be needed in order to prevent misunderstandings There are many parameters with names that don t make their meaning obvious Most requests are made through webmars except when using facilities not available at this level i e ACCURACY PACKING I don t often use MARS I modify scripts written by someone else five years ago The documentation for ERA40 experiment isn t obviously clear By example how find the surface parameter idem for the units used I think I use it at a very basic level That is not difficult Depending on the choosen date you need to know which are the variables available You must perform separate asks to get not compatible variables We developp a shell script to create the good Mars request s according to the date and the variables lists Why do you find the MARS service unsatisfactory e The MARS service is unsatisfactory because e g poor turnaround difficult organisation of data difficulties in handling the GRIB BUFR format of the retrieved data bandwidth limitation for transferring of data availability of service etc As I mentioned before main problems are availability and delays in the completion of the retrievals see previous Rather slow especially if need much data MARS can formula compute GRIBs but not change the meta inform
103. e system is changing so that I will be prepared to access it next time around without having to look for this information I am not well informed about all your services but mainly because I have never sought this information I get these through our Computing Representative but like to have it send to me by e mail directly to avoid delays I keep myself informed mainly by reading the ECMWF Newsletter I have answered no but really I need to put in more effort as I feel it is also my fault well may be I could find such info in the ECMWF newsletter the changes in the operational implementation could be directly communicated by email Notice of sessions can be a bit short I am in general very satisfied with the support received changes in GRIB codes could be announced more systematically since this often requires an update of libemos locally Maybe it s my own fault to be rather uninformed by not seeking the information It would be nice if the red orange green blue console window that the internal users have would be available for MS users at least with the information relevant for them It would also help if one could be put on a mailing list that informs about changes in the computing environment meteorologicaly relevant changes I usually get informed about the large changes but not about more subtle changes e g new Metview compiler versions Quite some times I encountered IFS bugs which were already known at ECMWF A go
104. easonal forecasts please work on easier access and more user friendly data format How about unformatted binary formats or DODS access for Grads users the web services I sometimes wish that it would be easier to get access to forecast data and that we do not have to pay our national met services for providing them for us I guess this topic is difficult to do anything about for you making more products available on the web for the general public Take the open web policy of the US as an example After all people pay high taxes in Europe I am probably a group of very few university users and therefore not representative but I would like to see a training session may be once a year for 1 2 day with a session on how to run jobs from outside Implementing a more friendly and flexible retrieval system a ECMWF trajectory amp dispersion model easier access in the end these cards and high security remain a pain For access to higher resolution data is important but I m not sure if the exist availability of climatic data and observations Since disk space is reducing in price all the time the ability to hold recently produced data on disk for faster access A DODS type data server would also be useful for some data Actually web is very efficient and easy to use I would like to receive more informations about how to use it Also the use of the electronic card is a little complicated I have three different kinds of login levels
105. eat help They are extremely responsible and efficient I d like to express my great appreciation and regards to them many positive experience with Dominique Lucas Carsten Maass does an excellent job In the last few years I was always very well supported I can t thing of a problem which couldn t be solved with the help of ECMWF assistance Many thanks to Norbert Kreitz and Carsten Maas the contact person at ECMWF 110 of 141 29 06 05 17 21 ECMWF User Survey Last question not relevant to me What about Not Applicable as rating above Until recently questions regarding the security of transfers from the point of view of the user were not highly regarded The focus on the user at ECMWF will always win the day Nobody hides behind a process everybody falls in line to get whatever problem solved My answer to the last item means It is my business to apply their help to my environment You got our best woman Petra from DWD I have had only one occasion when I have contacted the Support Desk so cannot comment on the quality of service other than I received a prompt reply and advice It s not easy for me to estimate that yet because my question was very basic Nothing to add Again I couldn t say since I only contacted the Call Desk once That time the answers were prompt and clear The contact person my colleague is very satisfied with the user support Neither satisfied nor dissatisfied I use a Macintosh
106. ecially define the problems of X windows My guess is that system does nor allow opening too many X windows in our institute The first in the morning gets the X window possibility However there may be some improvements very lately Add more informative help messages 38 of 141 29 06 05 17 21 ECMWF User Survey Provide flexibility in the SCRATCH disk memory space Possibility of increase for short time period On ecgate SCRATCHDIR should not be a subdirectory of SCRATCH as TMPDIR is not a subdirectory of TEMP on HPCF I m just use ecgate to retrieve data Better indexing Adding all the explanations about this on the ECMWF web site We are highly satisfied with current service no additional action is neccesary is OK your guide is fantastic only my English is bad but I will try and try to learn better to write less downtime sessions on Wednesday But the largest disadvantage is the slow connection to ecgate which is hardly a ECMWF issue vnc would be useful however I m not sure if this is a request for ECMWF or for my home IT organisation It is not possible to checkout the status of an data transfer initialised by ectrans within a LoadLeveler batch job so it s not possible to set up conditioned job steps which are using ectrans Provide a better environment allow the use of commands previously used add more scripting examples not only csh but also bash python aso I would appreciate being able to use
107. ed areas plotting X sections time series plots and analysis tools it would be useful to retrieve observations of a particular type of parameter as temperature or precipitation data The possibility to access at the meteorological observations of the WMO network in a simple way i e possibility to retrieve and automatically decode the data I am referring particularly to some categories of data like synop temp metar Messen when a questioned field is not available The MARS interface should feature the same possibilities as the dissemination interface syntax to make simple calculations e g total precip from stratiform and convective precip is quite clumsy not a major problem One that could describe the content of the surface measurements e g SST I would like to have a more extensive information guide I would like to have a graphical tool that could save in a ASCII file the command to be submitted to the MARS as example to have an web page in which you can the possibility to choose the parameters latitude longitude extension etc and the end you can obtain the ASCII file with all the syntax The ASCII file could be used then as a pattern to put on the MS jobs system or in ksh scripts I would like to have the possibility to obtain the fields in order format than GRIB or BUFR I would like to have tools to convert it to HDF McIDAS format binary ASCII etc correct the wrong GRIB headers of the results when operations w
108. ed both in MARS and in dissemination Detailed definitions and description of parameters archived in MARS I have not read a lot of documentation I ve read especially when I have to migrate scripts in 2004 history of model modifications type of data coming into analysis MARS the ERA 40 part is very poor concerning how much data are used on each interest area for the different run periode MARS documentation an up to date list of parameters and their description Security representatives meetings reports not up to date last one is from 2003 transfer to elevctronic format PDF of some old paper based tech memo for example 97 of 141 29 06 05 17 21 ECMWF User Survey User Services In this section we would like to find out about your use of the user services Call Desk User Support at ECMWF How frequently do you contact the ECMWF User Support Call Desk on average Each Respondent could choose only ONE of the following options Response Total of Total Respondents Weekly 2 l 3 Fortnightly s l 2 Monthly 2 E 7 Occasionally 248 a 62 CE ao w gt Never Other please specify I 3 Total Responses 398 0 20 40 60 80 Other please specify twice weekly 2 Not for a long time but then I have not used the system recently So far only when I had problems with certificates or with my SecurID card Once every time I need Occasionally I am working a lot
109. ed is in there somewhere and the documentation is excellent really high quality It can sometimes be a bit difficult to find what you want but usually you can In fact I have for a long time meant that your web page is one of the best information services available on the net Congrats For running computer jobs ocassionally it is very useful to have good information available The site response speed is high on weekends and holidays but not the rest o f the days To visualise one specific graphic is NOT possible to select several options at the same time for example base time step and arguments It is neccesary to select first one argument second another argument Once I have found the right path to the info I am after things go smoothly But it often takes time to find this path I have never found anything useful with the search facilities 75 of 141 29 06 05 17 21 ECMWF User Survey very strong The loops of the maps are not working As you download a page you don t access immediately to the latest model run I don t understand why there isn t a map with the 500 hPa geopotential and the temperature at 500hPa search functionality is not helpful Overall ECMWF has a somewhat anonymous appearance It is difficult to get direct contact to people at ECMWF to discuss questions problems etc We have a very poor Internet access at INM 2Mb for about 300 500 people I believe which is a shame considering I have
110. eed from ECMWF not much local observations ECMWF is among the best of those mentioned because of high availability fast data retrieval and very good user support Very well Much better computer support at ECMWF than at these centres Difficult to compare very good Better overall Headaches ECMWF is in some extend less friendly and of course less open than NASA archive Also the documentation is less didactic than that of NASA Faster but data storage can be a limitation BADC better navigation basic documents NASA better data access documentation ECMWF provides a service of operational basis and therefore available with high reliability The other sources are of research institutes There personal contacts and technical problems are sometimes to be discussed individually Large database of meteorological data MARS very effective user support No more a large vector machine at ECMWF So for my applications not ideal an alternative source for ERS scatterometer data in case of incorrect data in the MARS archive Matlab DKRZ is not occupied by forecast computations on a daily basis Databases organised so you can see all the data they have got as a series of overviews instead of having to search for specific fields although the data I was using was not as complicated They offer a very fast supercomputer Apparently this is also accessible at the ECMWF for me but I didn t know that till recently
111. efault disk space is quite small Even the 40 MB I got is small It s not clear if the limit can be brought up easily or not Same thing for SCRATCH I got an increase since I knew others got one How could ECMWFE help you to make more use of its ecgate service e Please provide your suggestions 37 of 141 29 06 05 17 21 ECMWF User Survey 12 no suggestions Improving the HPCD remote connectivity rcp considering the protections required it is ok as it is Recent lack of use is due to reasons at this end nothing to do with the service you provide can t think of any improvements necessary at the moment I d like to have access without special cards Could we just have a userid 1 passwd and 1 server with no intermediate steps Easier access to data without dynamic password change Somehow make the use of X and ftping easier However I have no idea how because the need for security is inconsistent with this Can t really suggest anything other than wish To use more this service does not depend on ECMWF it depends on me and my specific job it could be useful a more friendly interface The procedure for retrieving data is simple but the file transfer is sometimes difficult if it was possible to get via web it could be better and simpler also a rapid way to identify the data availability and codification could improve the use of this service We obtained pin generator trough our meteorological service e
112. ells for handling international GRIB in same manner although their differences puzzle thinned other parameter tables a s o Climatic studies and numerical weather simulations 17 of 141 29 06 05 17 21 ECMWF User Survey Use of MARS products and EMOS lib Data are used as initial and boundary condition for our application of a NWP model Statistical processing of analysed fields from ERA40 I am interested in precipitation forecasting nowcasting by application of statistical models and very short range forecast of severe convective storms by the NWP model LM COSMO Use of ERA40 and DEMETER simulations for prevision application on the African domain numerical modelling Meteorological Modelling on mesoscale and local scale Pollutant Diffusion Modelling on local scale Monitoring stratospheric ozone solar radiation and aerosols Use of meteorological data archived on MARS Implementation of aerosol code in ECMWF IFS system Ocean circulation studies ocean forecasting Computing Representative Research and development of statistical methods for adaptation and verification model development HIRLAM other NWPs verification At ECMWF I do MARS data retrieval At SMHI I apply my knowledge in multidimensional signal analysis non linear regression optimal control and optimisation to a diversity of problems in the fields of meteorology hydrology and oceanografy Verification of Model cloud forecasts using satellite data At E
113. ematical symbols often appear incomplete or do not appear at all In order to circumvent this problem I usually download the PS PDF file available Is it possible to improve the web documentation It might be a problem in my computer settings though the physical meaning of many forecast parameters It would be nice to have a GRIB User Guide and GRIB Reference manual like the one s for BUFR in IFS documentation the parts on diffusion and frictions parametrisations including used values of parametrisations Metview Metview Macro Language can do with an alphabetic index of functions more programming examples or tutorials Extraction of vertical Profiles documentation about SST retrieval and use of SST as boundary condition in ERA 40 modelling the skin temperature and connection between SST and skin temperature in ERA 40 Model updates When are updates introduced in the operational daily data stream and what are they about too little experience to have an opinion IBM MPI configuration parameters versus standard MPI ones 96 of 141 29 06 05 17 21 ECMWF User Survey on data retrieval via MARS see comment on different data storages Introduction to computing facilities More details about the current model and history of the main changes computer technical documentation when retrieving meteo data from the MARS archive I need a list of meteo parameters and their exact definition and units and a link where to find more informa
114. er support and Call desk have always been friendly and professionally helpful user support computer facilities interaction with personnel The MARS archive and the ERA 40 data Professional adequate swift and pleasant call desk service My compliments to ECMWF Everyone is really helpful and knowledgeable Training courses are brilliant and people are happy to spend extra time discussing other issues outside the classes Always had first class support from everyone I ve dealt with the staff are one of your great strengths Computer systems are very well designed and easy to use too and documentation is excellent The computing facilities incl service and support are done very well Training courses are said to be good I haven t attended them personally I m pleased with the services I use that is ecgate and the website I see no reasons for any changes The products ERA15 ERA40 are top of the line otherwise I would rather avoid having to go through the hassle of getting all the access info using MARS and having to ungrib the data the Computing facility the MARS retrieval being this the main service I use Speed of retrievals from MARS Web site is an invaluable source of information I m particularly impressed with the set up for accessing MARS through the Web MARS data retrieval Your data are probably the best in the world and your systems for disseminating data are very good Very high availability and st
115. er ten days to see trends Educating meteorologists in practical forecasting after they ended their time at the university before they start forecasting Operational calculation of Source Receptor Sensitivity Information for a global radionuclide measurement network for this purpose ECMWF analysis data are retrieved two times a day as soon as they become available Research scientist produce scientific results for publication in peer reviewed journals The tools of my research are atmospheric general circulation models My work at ECMWF involves testing such numerical models as well as running simulations for studying Earth climate with the focus on the middle atmosphere Specifically I use the echam models Through my PhD I m using a Regional Climate Model MAR to study climate in Europe and West Africa I m trying to run the model on the ECMWF supercomputers Research in the field of wave field reconstruction in connection with wind fields Wave climate analysis both offshore and near shore Meteorology Trainer Met Office College Main use of MARS data in retrieving ECMWF model synops for visualisation in Metview in order to create weather forecasting case studies I extract operational and re analysis data and archive it at Manchester University where it is made available to the UGAMP community part of the UK academic community Additionally some post processing is performed e g conversion to NetCDF I then make a sub set of th
116. ess this is only from limited attempts 1 Computer Security Policy 1 Computing aspects 1 computing service documentation it is a bit fuzzy chaotic 1 Descriptions of OLD formats such as BUFR CREX and GRIB indeed GRIB is not my favourite 1 ECFS 1 ECMWF Computer Bulletin B 1 0 5 1 Passwords amp SecurID cards possibly outdated but it did not work as cookbook for accessing ecgate when I tried it the very first time interaction with helpdesk was necessary to properly set up the access 1 ECMWF forecast analysis 1 ECMWF products dissemination 1 emos preprocessing observation usage 1 Encoding and decoding GRIB data 1 ERA 40 1 Error encoding or decoding 1 GRIB and BUFR libraries 1 GRIB codes description of parameters 1 GRIB conversion 1 hard to say All the documentation I saw is of high quality 1 IBM specificalities of MPI 1 Introduction to computing facilities 1 lack of documentation on observations in the stored MARS 1 lack of the WEB tutorials for the newcomers 1 library information 1 list of all updates to the model difficult to find on the web site 1 LoadLeveler for AIX 5L 1 MARS contents documentation 1 MARS documentation was very good 5 years ago paper version then there was a too short online version don t know how it is today 1 MARS User Guide for data retrieval revision 11 November 1995 Would it be possible to have a new version of this document 1 Metview Lacks examples 1 model ve
117. evelopment of the ICT infrastructure and applications at Met Eireann I keep users informed of developments at the Centre and note their requirements and take action to contact User Support where necessary At KNMI I am working on preparations for the ESA Atmospheric Dynamics Mission Aeolus More specifically performance simulation and ground air campaign preparations My work at ECMWF involves sensitivity computations and experimental setup with PrepIFS and Metview use for plotting results I work in the Atmospheric Composition Research group of KNMI Here I focus on stratospheric dynamics A major part of this consist of diagnostic work for which I use ECMWF data For this work I also sometimes use a trajectory model that is driven by ECMWF data Supporting users of ocean models and coupled ocean atmosphere models ENSO and seasonal forecasting research theoretical work data assimilation forecast validation data acquisition and postprocessing developing scientific software for processing of satellite data ERS scatterometer ADM Aelus lidar Research and teaching on Dynamic meteorology Use of MARS archive to provide data on cases te analyze a o with Dionysos http www dionysos uqam ca Furthermore to obtain forcing data for mesoscale and boundary layer models and to obtain climatological data Initialising and running a spectral barotropic model to study scale interaction and spectral error growth under the supervis
118. ew version which is much improved running Metview as MS user with direct MARS requests so no need to retrieve the data first ECaccess user guide with remarks for co operating state users Research and operational description of the model Documents on data available ECMWF forecast analysis about content of MARS archive BUFR GRIB Mainly there use in FORTRAN code Examples of how to read observation and fields in cfortran access to the documentation at many pages perhaps every page of the EC website documentation on observations in the stored MARS And documentation on GRIB and BUFR libraries But then I don t particularly like the user interfaces of these libraries either The docs on the use of the MARS archive could be improved Or alternatively a MARS script generator would be helpful none for my use it s satisfying In general I like documentation that is rather handy than comprehensive More detailed and better indexing contents lists documentation about computing services and make clearly known what is available for a user when he is a registered ECMWF user in some kind of easy to read quickstart manual MARS and examples MARS on line documentation is rather difficult to navigate strategic documents technical and scientific documentation I would like to have a MARS index of the directives at the moment I have no area which could be improved Computing documentation and products documentation updated techn
119. f products but a lot of things are not freely available ECMWF is an order of magnitude bigger i can t think of something the ECMWF doesn t offer Fast computers with a higher bandwidth connection to my workplace SARA local information meetings many colleagues that use same platform IDRIS good telephone assistance regular visits to users at LODYC It is impossible to compare because I m running very different jobs on to them ECMWF has more documentation and services in comparison of most of them furthermore the ECMWE have higher quality but almost all the others have free access to all the resources i e operational forecasts and analysis without restrictions NCEP data are more easy to access and retrieve because they do not need special permission I use the IDRIS center for high performance computing because I work on French projects As I don t use ECMWF resources for this purpose I can t give any relevant comparison The same in general However they allow retrieving data without any certificate User programming assistance ECMWFE is by far the best But obviously the scale of resources used in them can not be compared lt realtime data Linux clusters support on specific numerical models For me ECMWF offers mainly meteorological data Their own implemented models and knowledge about their own models The data from above centers are and ASCII or in GRIB but without spectral coefficients repre
120. fic questions with respect to the systems and trouble solving Most often problems in computer telecommunication problems Mostly questions why a job did not work Reporting problems detected either on systems or on applications un resolvable problems with my batch MARS jobs to receive support about parallel computing issues problem in connecting to ECMWE or in authorisation for securID card management often unauthorised general problems for Belgian users Specific computer problem Signalling problems with ECMWF software or hardware How to retrieve certain streams of MARS data 104 of 141 29 06 05 17 21 ECMWF User Survey Problem related with SecurID card Computer problems such as MARS requests not working webMARS not showing expected data etc At the beginning registration errors in the ECMWF tools installed at INM Some specific information concerning a particular subroutine clarify misinterpretation of the Metview manual or trying new things in Metview that are not based on 2 d visualisation of global fields chasing failures in semi operational processes that occurred after changes that were not announced or the announcement was overlooked Operational Issues email problems We have some problem with the forecasting fields in spectral coefficients in GRIB format Dissemination member state gateway installation problems problem with ssh connection and use of text editors nedit and vi Problems with ru
121. for a certain area during a certain time I m sure that data is available but have not managed to locate it Don t know who to ask I don t use MARS on the web I m very conservative and stick to batch jobs edited and submitted from a Unix prompt When I retrieve large amounts of data they get stored on scratch disks of ECMWF Before I manually move them to SMHI they occasionally grow beyond the permitted disk usage I find the temporal storage at ECMWF and manual moving of files from ECMWF to SMHI a bottle neck I suspect there might be ways to automatically put the data on my disks at SMHI but have not figured out how There seem to be different ways of moving data between ECMWF and SMHI I am sometimes puzzled on what to use the benefits of the different methods and the syntax and way to use the different methods more flexibility is needed in data extraction and plotting for limited geographical areas and cross sections very good instruction pages on the Internet Since I can extract ncdf format I m very satisfied There is a large variability in speed of retrievals Sometimes it is the MARS retrieval that is the rate limiting step in my forecast scripts It is down every Wednesday morning and seems to go down regularly other days as well It is also slow especially as I have low priority on my jobs occasionally get problems but less than a few years ago For some of my applications MARS access is still the bottlene
122. he Internet would make it easier to transfer data to computers outside ECMWF 42 of 141 29 06 05 17 21 ECMWFE User Survey more shells bash would make interactive use easier LoadLeveler is not very user friendly It is OK for users that can concentrate on IBM but most users will also have to know other submission methods It would be useful to describe the processes through more examples That s easy once you ve understood how to change the shell into Korn shell and that you have keyboard shortcuts for online command That may seem absurd but you can loose a lot of time without them If not I appreciate the rapidity of the machine the job submission and all the special commands designed to access the mass storage Highly powerful Quite easy thanks to PrepIFS and CDP If you are not using the Centre s HPCF resources at all or less than you would like to what are the key reasons Please specify the importance of each reason or tick Not relevant e You have easier access to supercomputing resources at your organisation Each Respondent could choose only ONE of the following options Response Total of Total Respondents High ss a 36 Low 34 eS 14 Not relevant 120 iii 50 Total Responses 242 0 20 40 60 80 HBE e You have easier access to supercomputing resources elsewhere Each Respondent could choose only ONE of the following options Response Total of Total Respondents High
123. hen only change time period and domain In the beginning when I had time there were some technical problems and later I had more administrative duties prefer running batch jobs because my old job implicates daily data retrieval But not now Not relevant for my work I use MARS through telnet and ftp my meteological service bought ERA15 and ERA40 on cassettes For my work I used ERA40 data from the mirror site at DKRZ Therefore I do not have the need to extract these data from MARS Also for my work I usually do not need operational analyses or other data which are available only on the MARS system because I use it through scripts no specific reason I used MARS directly As indicated above Working with shell language is more simply for me I handle the data with programs only We have no Unix happy with telnet service It wasn t necessary I did all through ecgate Obtaining data through our implementation of Metview is sufficient for my needs What is webMARS I have scripts for normal MARS and they are efficiently running scripts offer more flexibility Because for supply the NWP files to our SAFNWC s reference system we need regularly then the MS jobs approach is the better MARS is sufficient for my purposes Don t use MARS either I have been using MARS for quit some time now and maybe because I am maintaining the operational dissemination requests as well This is very much the same work so this might be the
124. ic tools that I developed when I was in the ECMWF satellite section retrieving analyses fields and radiosounding data for research purpose within the framework of the European COCO project ECMWF retrieving initial and driving data for mesoscale model MM5 I am not running it myself My scientific work develop simple e g 2 dimensional mainly idealised models e g of hurricanes research on cirrus clouds condensation trails and ice supersaturated regions Mesoscale cloud resolving modelling of ice supersaturated regions and cirrus clouds with the model EuLag Developing evaluating and analysing a regional atmospheric climate model for Antarctica I am retrieving ECMWF fields to force a transport model At ECMWF I run some routines by means of Ecaccess At my Institute I am head of the group for research for long range forecasting I applied statistical method and postprocessing of ECMWF and other centres models for obtaining the regional signal At ECMWF see above At DWD scheduling with ECMWF s SMS several suites with approx 4500 batch runs day in civil and military compute centres Backup planing between that two centres customer advisory data distribution mostly GRIB and BUFR less in ASCII code Graphic production with own written software GRIB BUFR obs temps trajectories HTML based survey of my 24h production called modelmarket Confrontation of several inter national forecast models Intelligent database sh
125. ical info about parallel computing Usage of the GRIB library how to code and decode data Instruction about HPCF usage MARS how to retrieve all available data streams with the appropriate parameters to set and choose from PrepIFS XCdp model code newsletter GRIB Format of files and how to create use GRIB files model modifications log documentation on de gribbing facilities at ECMWF MARS fields characteristics 95 of 141 29 06 05 17 21 ECMWF User Survey seasonal forecast hovmoeller diagram LoadLeveler documentation adapted for the existing HPCD available environment on ecgate Methods of computation of weekly and seasonal forecasts kind of how to or cookbook examples for typical beginners access to MARS scripting examples included and discussed in the manuals the documentation about ERA40 Metview Lacks examples Why do users not use it more not necessary to receive email fax and regular mail on the same subject Email and mail is enough SMS ODB Technical documentation of IFS Interpolation and Error encoding or decoding Skill of the models over the south eastern part of Europe Error messages from MARS software documentation easier navigation in online documentation of MARS information on content of the different data types ECFS Using prepIFS what do all the switches mean Or maybe I haven t yet found the right document IFS code and documentation In the web documentation math
126. ide 1 Status of ERA 40 in the past 1 training courses 1 Very wide use very useful calendar all kind of reports Order Software When you access the ECMWF web site do you Each Respondent could choose only ONE of the following options Response Total of Total Respondents Access the web site without explicitly logging in 82 Es ale Log in from your authorised Internet domain e g meteo fr 52 Eal 13 Use your own ECMWF user certificate to log in 102 26 Use your userid and SecurID 153 39 passcode Other please specify 7 l 2 Total Responses 396 0 20 40 60 80 e Other please specify HERE all of the above occasionally I do not explicitly login but the site does recognise me cookie probably All the above except using the certificate 78 of 141 29 06 05 17 21 ECMWF User Survey All of those use all last three methods e Your comments Depends what I m doing usually just use the Met Office domain but sometimes my certificate as well Sometimes I log in using my userid and passcode but quite often I am only consulting the documentation and then I just access the web site without specifically logging on Sometimes I use my ECMWE user certificate Depending on the purpose I use gt Access the web site without explicitly logging in gt Use your userid and SecurID passcode And using my own ECMWF user certificate certificate didn t work anymore anyhow i
127. iles on Model Levels loop over several requests more years described in MARS user guide didn t work for me and if NetCDF format would be available for retrieval on telnet it would be great see comment on different data storages Easy example scripts after installing the new IBM computer were missing HDF format instead of GRIB make more easy the data access ease of use 115 of 141 29 06 05 17 21 ECMWE User Survey Define which set of data is available for users and which is free shared can save some time from beginning of search User friendly output with ability for downloading data in text arrays for general programming use or a set of C routines for accessing gribbed data Clear information of the stored fields See earlier comments Would be nice to retrieve observations and soundings Detailed description of archived parameters needed idem solutions to extract long series of data for example 2 years of EPS precision on the data s avalaibility documentation catalogue post processing Detailed description in the documentation of archived parameters because often asked by users Services offered on the website Each Respondent could choose ANY of the following options Response Total of Total Respondents Services offered on the website 53 100 Total Responses 53 0 20 40 60 80 e Your comments 3 Search tool for information More meteorological variables available fo
128. ing options Response Total of Total Respondents Very satisfied s Tian 24 Satisfied 106 41 Neither satisfied nor dissatisfied a E 30 Dissatisfied o 0 Very dissatisfied o 0 Total Responses 227 0 20 40 60 80 e Your comments 5 Ido not use ECFS BBE 2 no comments slow sometimes Have not really had the need to store big files for long periods on the ECMWF The home directory is sufficient for most of my files For temporary storage of MARS retrievals I use the scratch disks 68 of 141 29 06 05 17 21 ECMWF User Survey Sometimes there can be long delays if multiple files are being copied from ECFS to ecgate SCRATCH Comparably quick and files are always OK Wildcard expansion in ecfs commands would sometimes be useful too few commands are available It would be nice if documentation on the gt ECMWF tools such as MARS and ECFS would be very easy availably via the Web for example via a link on www ecmwf int research and also via links on many places on the ECMWF site Slow Sometimes ECFS is not available relative long reload times missing support of large files It s overall performance is good Easy to use but not fast It would be nice if files and directories in ecfs could be renamed moved Very slow can t answer correctly Could be useful to have possibility to ftp files from ECFS simple fast efficient would like to have it here Associated Member State no access
129. ion of Prof Dr Martin Ehrendorfer Therefore MARS data is needed Computation is performed locally Linux Cluster and at the ECMWF s ECGATE server Further development of our NWP model Facilities at ECMWF mainly used to run model test versions for later verification Assimilation of Satellite Radiances definition of background error covariances Development of IASI and ATOVS instrument processors in the frame of the EUMETSAT Polar System s core ground segment development 11 of 141 29 06 05 17 21 ECMWF User Survey head of regional forecast centre using EPS and similar products Developing a malaria model which takes ERA 40 and DEMETER data as input In my own organisation I m mostly doing simulations using a mesoscale meteorological model combined with an extensive landsurface hydrology model to investigate feedbacks between the surface and the atmosphere I use ECMWF data MARS to initialise the model and observations BUFR to validate the model calculation of singular vectors and additional quantities in the Eady model Model development optimisation parallelisation model coupling model porting At the Belgian Institute of Space Aeronomy BIRA IASB my work consists partly of data retrievals both for modelers and other scientists analysts validation people etc This data consists mainly of satellite and ground based atmospheric measurement data but I retrieve also auxiliary data like ECMWF data wind fields e
130. is main archive available to scientists and students in the Department of Meteorology University of Reading Assimilation of satellite measurements in a CTM Processing MARS data for input to the CTM Development of HIRLAM and ALADIN retrieval of data for metocean condition consultancy Project Leader Science manager therefore not much active user of the facilities Relating surface atmospheric and satellite observations of desert dust clouds and Maintenance ofreprocessing software maintenance of software receiving GRIB and storing in GRIB db Maintenance og software handling written forecasts and sending to customers General support research activity on extreme flood events over the Mediterranean use of backtrajectory methods to identify water vapour source for Mediterranean floods 22 of 141 29 06 05 17 21 ECMWF User Survey bench forecaster for aviation development of ECMWF derived products for operational use in our service Verifying and evaluating models used by our operational forecasters identifying typical model errors preparing studies including case studies Developing and testing new methods to improve operational weather forecasting Retrieving and processing archived model forecast and analysis ECMWF data to complete tasks both governmental commercial sector data retrieval at ECMWF informatic assistance and development at METEO FRANCE Head of the team in charge of developing physical processes in o
131. ist I check that the Met Office stratospheric model output is consistent with the ECMWF analyses Simulation of meteorological episodes with numerical atmospheric limited area models and ocean models initialised by ECMWF data Generation of 3D trajectories using ECMWF analyses to handle DISS stream to develop programs and scripts using MARS data on ECGATE and our computers 8 of 141 29 06 05 17 21 ECMWF User Survey Investigating the output from the physics of the ECMWF model in the context of stochastic modelling of unresolved processes Operational weather forecast ozone and UV forecast I m a forecaster I use only your web s products Weather forecaster and technician in air quality modelling Meteorologist working with a LAM which I contribute to maintain and develop Working on NWP models and operational procedures at National HMS Computing representative lt Tim leader of Numerical Weather Prediction Group R amp D Dept NWP Div Met Contact Point amp Computing Representative develop tools for ECMWF related applications Running a global Chemistry Transport Model CTM on the ECMWF computer and retrieving meteorological data from the MARS system mainly 6h forecast Regional climate modelling HIRLAM development Tropical Meteorology Air sea interaction Monsoon and Intraseasonal Variability Seasonal forecasting ECMWE ERA4O data retrieval here ERA40 data dissemination I am among other items the prim
132. ithin MARS are performed retrieving profiles at a given lat lon would be very useful to me This is not yet possible as far as I know Despite the online catalogue and other tools I find the composition of queries for MARS not easy Furthermore the type of data archived differs considerably between operational forecast analysis ERA etc This makes it difficult to use those different datasets for a given application i e Dionysos A tool to generate the scripts automatically Better explanation of how to order a job to optimise retrieval time More detailed documentation of the syntax and how fields are archived vertical interpolation on model level predefined for the older 31 50 and 60 levels when 90 are available improved file manipulation routines An easy to use tool callable from a shell script which can give a short answer about the availability of requested fields without parsing the output from the MARS list command MARS allow computation but this could be improved by allowing for instance to compute formula like int U sin theta dS dp where theta is the colatitude U the zonal wind and dp the pressure increment I need vertical profiles of pressure temperature and humidity above certain sites latitude and longitude is given With MARS I can only download a grid at least four points around the 59 of 141 29 06 05 17 21 ECMWF User Survey site and not the profile at the site directly API to
133. ity Better documentation Maybe more examples An update of the MARS User Guide for data retrieval will be very useful In my opinion it is preferable to increase the description of the products and their changes in the MARS guide frequent updating information quick documentation I retrieving derived data in GRIB format the GRIB header is copied from the first field encountered It would be helpful to be able to handle the header May be this is already possible but I was not able to avoid to get the same variable identification GRIB header in retrieving two derived fields having the same first field as operand Vertical interpolation algorithm for the model level pressure level transformation for arbitrary pressure levels should be implemented in the retrieval process Information about the reanalysed meteorological fields resolution What is the resolution of ERA40 default value and so on There are possibilities to choose resolution but is it interpolated values or not Documentation as earlier stated The functionality under compute could be more similar to what may be done in Metview I bet there is a lot of overlap More examples may be provided in the manual access to realtime products to users like NCEP is offering to the public an exhaustive description of the requests we can make To increase retrieving speed if possible more postprocessing zonal averaging More descriptive error messages Extraction of Vertical Prof
134. kind of observations I can retrieve that might be interesting Data from other centres The interest for seasonal monthly forecast will soon grow sensitive areas and singular vectors Which additional features in the MARS language other than the ones already available would be useful to you e Please specify 9 no suggestions don t know Have the possibility to change the scanning mode flag ISEC2 11 in a field to get the field value in a grid point given lat lon coordinate or array index to be able to calculate covariance between a point and a field output a field As indicated above 1 Rotation of fields 2 The possibility to select any domain and grid spacing 3 Easier retrieval of precipitation fields Why do you have to choose forecasts and then make 58 of 141 29 06 05 17 21 ECMWF User Survey differences Could it not be possible to store the best estimate of precipitation for a certain period of time i e every 6 hour which is typically the interval I d retrieve winds and temperatures etc I have not learned the on line programming language and instead retrieve a bunch of forecasts and then invoke a home written postprocessing program to construct the right precipitation by subtracting two fields for a certain period and finally append the results to the corresponding wind and temperature GRIB file a friendly interface To be available more levels with ensemble forecast Plotting for limit
135. l of Total Respondents General purpose server ecgate 55 100 Total Responses 55 0 20 40 60 80 e Your comments 2 X connections has sometimes problems not available or drops out I would expect it to be faster at least at evenings nights An alternative server would be good Clear but brief information on how to log in move files to member state computer Better interface though it might be there today I just didn t figure out how to get it I would like improvements to the grib2netcdf transfer routines when a run of several weeks is launched you have to restart this run yourself That was quiet time consuming good but there is always improvement ecput ecget data transfer on a common disc with HPCD transfer from to mpi gt ecgate gt HPCD almost double work easier transfer of data over Internet to external servers Need some written basic document to read and understand to use more I am extremely happy with the VNC service at ecgate It is fast and reliable but there is little documentation I think it should be advertised more strongly How can I get larger than 1024x768 pixel windows Disk space on scratch is sometimes a limitation linking libraries for Fortran and C update software environment to the one on HPCD Longer time of inactivity allowed before the connection is automatically closed More helpful documentation amp messages when logging in SCRATCHDIR should not be a subdi
136. l Organise mailing lists directly related to specific topics as MARS usage Allow subscribers to search history data to match their interests Learn from other people s mistakes Not at the moment Not too much Do not introduce very often substantial changes in the computing facilities that imply to the users a lot of training and migration tasks of their own developed tools at ECMWF availability of ready to plot products reports and communications about operational aspects of the models larger scratch for maybe shorter time If I would be able to retrieve NetCDF files not only on the web and if MARS archive on telnet could work for several requests more years in one request and maybe the speed for retrieving smaller requests Availability of standard script tool for diagnostic and predictability online I need to have more experience with ECMWF services to answer that question It s too early for me to say ftp links for downloading data Nothing I can think of an easier access to the data I would Like to participate to a stage of several weeks in ECMWF It would be very helpful to get conservative wind fields at requested resolution lower than original in HDF format 139 of 141 29 06 05 17 21 ECMWF User Survey Free access to daily observations analysis and forecastings Improve the web environment Conduct a technical seminar for beginners in the Netherlands 1 Earlier deliverance of grib files from ECMWF 00
137. l gateway 1 ftp 1 ftp push by ECMWF 1 ftp service from pc to ECMWF for retrieving the results of tasks when not within the organisation network 1 ftp thro Met Office dedicated link or ftp from other locations e g home 1 ftp to download data from ECMWF account to my own workstation in my organisation 1 ftp to scratch 1 Haven t done it for a while so can t quite remember Not used ECaccess think I used the predecessor ecbatch May use Metview running on our system in future but not tried it yet 1 I haven t connected for a while and apparently the server names have changed 1 I have telnetted in but recently I mostly use a local MARS client in combination with web access 1 javaws prepIFS jnlp 1 On site 1 Our National Met Institute has a dedicated provision from ECMWF 1 pftp ecaccess ecmwf int 1 Public MARS server web 1 secure connection via the web pages 1 using the web interface to MARS to query and create MARS commands then using the generated commands through ECtools for file retrieval and transfer 1 using vncviewer 1 web access to local gateway http ecgate smhi se 9080 1 web browser interface to MARS seasonal forecasts data server 1 web I am looking daily at the weather maps epsgrams ensemble forecasts 1 we receive daily data for our LAM by means of the national state member By what other means do you access ECMWF services e Select all that apply Each Respondent could choose ANY of
138. lating trajectory forecasts and analyses relevant to atmospheric chemistry field campaigns as part of the special project routine back trajectories I visited ECMWF only for retrieving data before it was possible to do via ftp telnet or ssh In our organisation we use numerical models atmospheric limited area models like RAMS ocean models like DieCAST and POM SVAT schemes like LSPM SiB and BATS and we use ECMWF data analyses synop temp for model initialisation regional climate modelling I make use of ECMWF data most often ERA40 both for climate model BCs and also for climate studies Working mainly with observations decoding encoding routing visualisation documentation ECMWF Data retrieval for special project SPDEWVUT Local Development of small scale flow and dispersal models ECMWF data retrieval of several parameters purpose use of atmospheric mass distribution in 3d and input for ocean circulation model in the context of satellite gravity field missions GRACE GOCE Computation of time variable gravity effects Making forecasts with data from ECMWF models Used PrepIFS to run ensemble prediction tests as a preliminary study for the TEPS Participated in the work making the setup of TEPS in cooperation with ECMWF and my group at the Research and Development Unit at met no Retrieve era40 data from MARS to get forcing data to the regional climate model HIRHAM project ENSEMBLES Stratospheric scient
139. libraries that are implemented in my local linux system I didn t know about the existence of this service but l 11 get information about it I think I should be using more ECMWF resources because I am used with scripts Archive Services Data Server In this section we would like to find out about your use of the Data Server Have you used the ECMWF Data Server http data ecmwf int data Each Respondent could choose only ONE of the following options Response Total of Total Respondents How satisfied are you overall with the service provided by the ECMWE Data Server Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied jis el 9 Satisfied s Neither satisfied nor dissatisfied 9 ae 51 Dissatisfied 3 l 2 Very dissatisfied o 0 Total Responses 155 0 20 40 60 80 e Your comments 6 I haven t used it yet 2 no comments It is too slow to download the fields in a reasonable time Occasional NetCDF problems seem to be solved by now It is not always clear which fields you get e g is he precip in the verification section of the Demeter data 6 30 or 12 36hr low resolution 66 of 141 29 06 05 17 21 ECMWE User Survey NetCDF conversion needs improvement e g problems with land sea mask and or missing values Try to retrieve ERA 40 SWH in NetCDF Difficult to retrieve large chunks of data limit If you use
140. llowing options Response Total of Total Respondents Always jis E 5 Often a Eal 12 Sometimes EE 24 Rarely o EAR 2 81 of 141 29 06 05 17 21 ECMWFE User Survey Never 16 M 37 Total Responses 314 0 20 40 60 80 e ECMWF Newsletter Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 2 Often Em Sometimes s Ln 28 Rarely oe Never 109 Total Responses 314 0 20 40 60 80 e Search tool Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 37 11 Often 106 m 33 Sometimes o9 m 31 Never MM 11 Total Responses 322 0 20 40 60 80 e Sitemap Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always e a 5 Often o 20 Sometimes Ea 32 Rarely Em aes 19 Never 74 Za 24 Total Responses 306 0 20 40 60 80 e Colleague Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 10 3 82 of 141 29 06 05 17 21 ECMWFE User Survey Often oe 21 Sometimes 124 38 Rarely cm 20 Never s6 17 Total Responses 323 0 20 40 60 80 EDGE e Other Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always 2 poo
141. ly 1 sometimes daily sometimes occasionally depending on the work I am doing at the moment 1 varies widely see above Overall how satisfied are you with the service provided by ecgate Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 126 Satisfied 224 Neither satisfied nor dissatisfied 34 Peal 9 Dissatisfied a 1 Very dissatisfied 2 1 Total Responses 390 0 20 40 60 80 e Your comments Is it possible to extend the timeout limit for a telnet connection MSaccess Some times the xterm connection does not remain long enough for a proper work now it is simpler than it was sometimes it is too busy and there is no alternative 34 of 141 29 06 05 17 21 ECMWF User Survey Iam very happy but find the loadlever system currently a problem I must study the supplied information more No real problems apart from downtime I now run my trajectory model on ecgate since it replaced ecgate1 because it has a very fast turn around time faster than when I used to run it on vpp700 I no longer require computing units There were problems with LoadLeveler when it was first introduced which caused my forecast jobs to fail several times during the flight campaign in July 2004 However from August onwards this problem was fixed unfortunately the campaign finished on 4th August we have a pin generator problem and we can not login the serve
142. makes it easy to use other services Impressed how you meet your project schedules Like you inform with email with upcoming downtime or updates General professionalism and quality of work Feel there is solid team of excelling employees working hard and always doing better medium range NWP MARS retrieval distributing information to the users user support ECMWF is a mine of knowledge observations and high quality meteorological fields Overall I am very satisfied of ECMWF services Easy connect to ECMWF members it always works and when I have a question the answer is quick and clear Multiple data set both observational and forecasting fields You already make a good job but that s endanger national meteorol services my job too User Support I am satisfied with services that ECMWF provides High quality services products hardware and interfaces Quick response and very willing to help professionality in all fields computing meteorological science data handling data use Ensemble forecasts ERA 40 The MARS archive The way to contact the computing facilities is quite easy and I feel quite free to even try out things based on scripts that are already given in the documentation The education material is very suitable Personally the response to my my problems and queries Also the provision of an easy to use model and forecast system For my purpose obviously the forecast products MARS Documentation ECFS Web
143. me Once stream expid method type and class are known stress usually recedes Compute on field sets can be very tricky Would bringing the compute engine from Metview into MARS not be an option Sometimes it is not easy to find the relevant information in the documentation But it is always possible to obtain an efficient help from the ECMWF support team help desk Sometimes hard to find info on possible directives e g what ident number should I use to retrieve data from the NOAA 17 satellite I have little knowledge of how the system is constructed and behaves Presently I hack existing scripts far easier to use than met office archiving system Navigating through the calatogue without knowledge of some activities at ECMWF can be difficult Error messages are very cryptic Default values are not always consistent with the assigned parameters User manual reports some errors The list of parameters as documented under services archive d parameters proved to be very useful to me once you have developeved a modus operandi it is fine but it takes a while to find that modus sometimes when MARS is not retrieving what you asked for it requires an effort to find out why it isn t I use MARS for one single application Once that is configured use is easy I cannot say anything about other more variable applications giving examples to retrieve data The MARS syntax is difficult for me It is difficult to find out the
144. member of the NWP group at MeteoSwiss climatological studies of ERA40 data and analysis of monthly and seasonal forecast data Decoding of the GRIB data postprocessing and plotting is not done at ECMWF but CSCS libemos is used for some of these tasks I work to develop the global model and diagnose global model problems at my NMS A large source of validation data is from the ECMWE MARS archive 9 of 141 29 06 05 17 21 ECMWF User Survey Regional climate modelling at ECMWF ERA 40 retrieval to get boundary initial condition for climate model runs and climatology for validation Clouds retrievals parameters coverage phase cloud top height vertical structure classification etc using IR MW satellite data Investigate the reliability of flood predictions obtained by hydrological models coupled to a limited area model initialised using ECMWF predictions climate model development and application teaching students in using models doing model experiments Ozone climate interaction research using ERA40 and other meteorological data sets ERA40 data e g temperature data is also used as a priori information for improving satellite trace gas data retrievals Numerical modelling chemical transport model and data analyses in order to interpret atmospheric trace gas measurements mainly satellite data Diagnostic studies with ECMWF reanalyses and operational ana fc data Transfer of GRIB files extracted from MARS to home institute
145. n login problems Better coverage for Africa knowledge of the latest changes Overall your services are very good Graphics from deterministic model are a bit slow make the website easier accessible for other browsers than Internet explorer More detailed explanation on Weekly and seasonal forecast products A GRIB exe for Windows with the concrete documentation related to input parameters for interpolation and decoding easier access to documentation Additional forecast info as mentioned earlier Clearer catalogue browser possibility to retrieve more data from MARS archive on the web whole year for example like it is possible from Data server ease of use 1 tutorials 2 courses handouts Web site could be improved in many areas download of files the few times I tried I failed On line availability of analysis and next two days forecast As mentioned before more forecast products based on user interactivity for local and regional areas Web MARS the possibility to customise plotting Information documentation available on the website Each Respondent could choose ANY of the following options Response Total of Total Respondents 117 of 141 29 06 05 17 21 ECMWE User Survey Total Responses 72 0 20 40 60 80 Information documentation available 79 100 on the website e Your comments More scientific references to various parts of the forecast and analysis system See
146. nd why Strange interface inconvenient commands for example pressing up arrow I don t get previous commands neither does tabbing fill filenames I guess this could be improved by SSH No problems as long as the right ssh version is used 36 of 141 29 06 05 17 21 ECMWF User Survey It is particularly useful to be able to sftp multiple files out of the ECMWE as part of a script Please keep this functionality as it enables me to automatically transfer my derived forecast products mainly trajectories to the BADC Internet connection is very slow particular X forwarding could be a problem on our side MeteoSwiss In my capacity as Computing Rep I do not use ecgate very much For the little work that I do mainly MARS data retrievals via Shell scripting the system is quite easy to work with Recently have done no more than login and do some listings Have not used it since last April some of the messages are not very helpful Problems sometime occur when handling large numbers of files UNIX knowledge required I have some problems finding the time to adapt my older applications to keep running in the over the years changing environment I guess computer novices would be better off using the web server For fast large and flexible transfers I use ECtools and my own scripts Some information for data retrieval like satellites are not easy to handle At present we are not de gribbing data using ECMWF facilities and
147. ng researches the pollution transport across the English Channel in collaboration with local authorities and other Universities in Southern England and France The research is based on pollution transport and source identifying simulations based on a a Lagrangian Particle Dispersion LPD model b an Eulerian Receptor Based model developer section of activity DA in particular use of satellite data Head of Networking at the Italian Air Force Met Service Operational ocean modelling development and assimilation Participation to international research campaigns related to the transport of pollutants Analysis of aircraft data Use of met analysis for description of meteorological context related to research campaigns PV field trajectories preparation of met analysis for FLEXPART trajectory code and Meteo France Meso NH model Model development and testing at DWD System support team HPC head at ECMWF Comp Rep 12 of 141 29 06 05 17 21 ECMWF User Survey My work is involved in climate variability studies At my organisation I work on the Computer Division as a technical consultant for new projects and big systems High performance computer and Data handling At ECMWF I mainly work giving first step support to INM users of ECMWF systems at ECMWF retrieving MARS data and preparing input data for trajectory and dispersion models and mesoscale models MM5 WRF in the organisation emergency response m
148. ng GRIB records are identified If I use the MARS language to compute the difference between two fields the GRIB identification for the resulting field will be misleading The experimental NetCDF output option should be made official and incorporated into the MARS language With NetCDF better data description identification will be possible This should be utilised by providing possibilities in the MARS language to set attributes on the data fields describing their origin The ability to do simple computations on the data would be useful e g computing stream function and velocity potential and performing spatial smoothing are all very easy to do from the spectral coefficients Adding fields e g CP LSP would be useful too POINT like grid but to extract a times series for one or more lat lon position A soft to create mars requests entries should be begin date end_date list of variables names not the number values Analysis or Forecast Dataset geographical area output format storage_directory Archive Services WebMARS In this section we would like to find out about your use of the service called WebMARS How often do you use webMARS Each Respondent could choose only ONE of the following options 60 of 141 29 06 05 17 21 ECMWFE User Survey Response Total of Total Respondents Daily a 1 Several times a week E 5 Once a week 2 Fal 5 Occasionally 204 50 Never 153 La 38 hal pE
149. ng large computer codes is an important problem in our group I will like to learn how to use centre s HPCF Until now the work developed did not require the use of HPCF 46 of 141 29 06 05 17 21 ECMWE User Survey HPCF at ECMWF is the best I can get there s no alternative I have only used the Single Column Version of the IFS which runs quickly and safely on ecgate Hence for that there was no need for me to use HPCF Programming for HPCF resources is not part of my job at the moment I mainly use a non parallel version of TMS for inverse modelling data assimilation which can be easily run at our local SUN cluster for full chemistry simulations parallel TM5 version however ECMWF is very attractive for us mainly done by several colleagues most tasks are easier to perform on our local cluster disk space and access for large is the bottleneck for ECMWE use I m not using it right now but I will hopefully Plan to use HPCF in near future Other work commitments as writing reviewing editing education Mainly I have not enough time to perform and finish the optimisation of my code on your platform I want to do that carefully and understand what I m doing But finding the good compilation options is very hard and the model code is huge I don t really know how to put up with all the subroutines mainly graphical applications Jobs run by Ilsubmit on ecgate access to results That s true that I ve also ac
150. nning programs problems with X11 connections something not working specific advice on the use of computing resources Specific advice on problems encountered My colleague is a contact person so I ask him if I have problems Specific problems encountered while using the computing resources If our member state SMS jobs are not started by the system if the jobs crash ECFS failures compiling problems format and retrieval of wave data problem with disk space User support to help me to solve users problems or questions to give me advices to answer specific users requests Call desk to help me with SECID cards management to inform them when there is a problem and ask for information about it not relevant compilation errors What is your preferred means of contacting the ECMWF User Support Call Desk Each Respondent could choose only ONE of the following options Response Total of Total Respondents E mail 252 P 84 Telephone 49 Fa 16 Other please specify 0 0 105 of 141 29 06 05 17 21 ECMWE User Survey Total Responses 301 0 20 40 60 80 e Other please specify both email and tel When needed how helpful have User Support s and Call Desk s services been to you Have you any suggestions for improvements Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very helpful 192 O 65 Helpful 8 z 33 Neither helpful nor unhelpful Is
151. nt 274 i i 70 Total Responses 390 0 20 40 60 80 BBE Do you think you are adequately informed about the following topics e Computer system sessions Each Respondent could choose only ONE of the following 99 of 141 29 06 05 17 21 ECMWFE User Survey options Response Total of Total Respondents e o 11 Not relevant 47 ioe 13 Total Responses 368 0 20 40 60 80 e Changes in the computing environment software hardware Each Respondent could choose only ONE of the following options Response Total of Total Respondents o No e Ee 20 Not relevant w E Ca Total Responses 371 0 20 40 60 80 e Meteorologicaly related changes Each Respondent could choose only ONE of the following options Response Total of Total Respondents Yes 214 58 15 a No 100 a7 27 Not relevant 57 Total Responses 371 0 20 40 60 80 PBE e ECMWF educational programme Each Respondent could choose only ONE of the following options Response Total of Total Respondents Yes 232 63 15 No 84 23 Not relevant 55 as Total Responses 371 0 20 40 60 80 e Your comments 2 It s too early for me to say It could be useful a mailing list for each topic You never seem to have time to take in the information that you know is available 100 of 141 29 06 05 17 21 ECMWF User Survey I would appreciate receiving email notices updating how th
152. ntre s High Performance Computing Facility HPCF resources Each Respondent could choose only ONE of the following options Response Total of Total Respondents Daily is 4 Several times a week i i 4 Once a week a 1 Occasionally o OAS 17 Never 280 M 68 Other please specify 2 E 5 Total Responses 409 0 20 40 60 80 e Other please specify 2 I don t know daily a year ago occasionally in the recent year daily in periods PE 40 of 141 29 06 05 17 21 ECMWE User Survey during some periods hardly ever now intensive several times a week but for limited time periods intensively in junks of weeks throughout the year in the future and in the past near daily but at present occasionally monthly never so far but having plans to do so not used to my direct knowledge not used it for months only once Recent experiments have been carried out by colleagues Therefore I do not have experience on the current IBM system see above comments I have not used it for 3 years but will use it several times a week after summer seldom sometimes daily Sometimes occasionally sometimes daily when running HIRLAM experiments sometimes occasionally until now occasionally from now several times a week want to use it in future when visiting ECMWF Overall how satisfied are you with the service provided by the HPCF Each Respondent could choose only ONE of the following options
153. o O 1 C Sometimes j6 Em OOO 12 Rarely s m O 6 Never 102 C 13 Total Responses 139 0 20 40 60 80 e Please specify other ways of locating information or give your comments BBE 8 Just browsing 2 Looking at the home page Italian user support ECMWF friends and user support As you can gather from above there are no good means to hear about potential new information on the web site read an on site announcement new features of the website are not announced enough New information New to me is what I was looking for explicitly or came across by accident I am not very aware of what is new on the site still some information are not at site map e g news sheets phone to the help desk of ECMWF I question of logics function of my needs looking for information I mean that it could reside on ECMWE site try hard By bumping to it by hazard by visiting the webpage every so often or read somewhere else about s th that took place in ECMWF By references in the paper documentation I got when my user account was established No time to visit your website to browse what is new 83 of 141 29 06 05 17 21 ECMWFE User Survey Using logic to assume where to find it some call this intuition By accidentally stumbling over it Reading the web pages By chance no way to see from the site Sometimes I am looking at the latest release of EMOS related software published I would have preferre
154. o everything I needed to do Manipulation of GRIB data headers should somehow be enabled Ideally this should be automatic I wrote it before your own parameter tables shortcuts are strange We now often use CSOMARS Aubert Guillaume Again I look after scripts developed by my predecessor different data retrievals and storages are sometimes difficult to understand poor documentation see later It is relatively fast but not easy to use The system is hard to use for people outside the meteorological world The Data finder web application is nice to use for finding out which data are available But it gives sometimes incorrect answers more data are in fact available than those displayed by the Data finder Another issue For large data retrievals from the MARS system there is a risk for constructing inefficient retrieval jobs which lead to mounting the same tapes several times A remedy could be to give better information on how data are organised on the tapes or even better to provide software to analyse a sequence of MARS requests which would discover repeated tape mounts and advise the users on how to improve their MARS scripts You get the data out on the web very fast MARS is too slow to let a running model read meteo directly We now store data intermediately on ecfs but I guess that this is not what ecfs was meant for problem if you need to extract long series of data I got some problems with the smart features
155. o forecast Maybe if there was more customising possibilities at Web MARS data plotting Studies for Global Atmosphere Watch Network project account frsctraj forecasts and verification products 90 of 141 29 06 05 17 21 ECMWE User Survey Easy facility of navigation Animated forecast maps Pseudo satellite maps based on forecasts animated as well there s no limitation profession wise I don t see from now what could be the key for more frequent visits Search for documentation that helps me in my work as how to use fortran libraries ecfs compilation tools etc If i receive an email for news on this web site This does not dependent on the ECMWF web site that I find very good Discovering it can be an advantage for me to use it the need for a specific information to find there Documentation Here we would like to find out about your satisfaction with the documentation ECMWF provides How satisfied are you with the documentation available both on line and paper based Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied co EEJ 17 Satisfied 237 i 60 Neither satisfied nor dissatisfied Gal 20 Dissatisfied a 2 Very dissatisfied El 0 Total Responses 392 0 20 40 60 80 Which documentation do you like best and least BPE I like the following documentation best 17 MARS 17 Metview 11 MAGICS 9 Newsletter 8
156. o it is not easy to find or it is found somewhere you won t expect it It would be useful greater instructions and examples for young users when retrieving meteo data from the MARS archive I need a list of meteo parameters and their exact definition and units and a link where to find more information on the meteo parameter It is sometimes difficult to find the information I am looking for This I think is mainly due to too short descriptions on what a link leads to Also it can be difficult to find out where you are in the information hierarchy For example if you click on Order Data under Products you will enter a page identified by Home gt Products gt Data Services which is almost but not quite the same you would expect Why not be consistent and use exactly the same words in the menu and in the identification sequence for each page Sometimes I encounter messages like An object was lost Try for example from the main menu Seasonal gt Seasonal Forecast Products gt Forecast However getting metadata is not always easy such as information on units data source reliability The appropriate documentation is sometimes difficult to find My only problem was that I failed to download reanalysis products ERA40 A catalogue of available observational data would be useful generally improving I ve found easily information in 2004 to migrate scripts I don t really need to use the web site My feeling is that I
157. od start would be sending the IFS memoranda listing the code changes I am NEVER informed by the National weather service but fortunately from user support As Computing Rep I receive and pass on information relating to each of the areas mentioned above User Support is excellent and Call Desk too E Mail notice about computer system sessions might be helpful very good user support Info about courses is sometimes transmitted rather late by the Computing Representative Is there an obvious place on your website where you document past and future model changes and where you alert when a model change happened I usually find out via the ECMWF Newsletter but that s not a very timely way and not very detailed The information would be distributed by e mail Our local representative forwards most information to us use more email to circulate info If I wanted to be informed it probably would not be a problem The change to handling GRIB tables in MARS during last week of January went unnoticed until operational retrievals of GLOSEA data started failing Who I contact depends on the problem If it is an operational problem e g disseminated data has not arrived then I contact Call desk If it s a general question or something which concerns a local user then I contact User Support The educational programme comes quite late for the application I have to plan it early e g in the year before it takes place You never know when
158. odelling oper dispersion modelling occas mesoscale modelling occas I am working on our emergency response modelling system TAMOS so sometimes I have to retrieve meteorological data from the MARS archive Retrieve MARS data calculating 3D wind fields and based on that back trajectories Wind fields span several years Trajectories are then summarised with special statistical tools I compute the interaction between the Earth and the atmosphere in order to better model the Earth rotation Unix Linux Administrator in my organisation Liaison with ECaccess team at ECMWF I download vertical profiles of pressure temperature and humidity above certain sites GPS antennas to derive so called mapping functions which are used in geodesy Mapping functions are used to map down path delays in zenith direction onto arbitrary elevation angles Downscaling models Ensemble based predictions Statistical Downscaling of NWP from Short to Seasonal forecast Analysis of EPS particularly seasonal R amp D DEMETER and ERA 40 plan to look at operation EPS archives seasonal monthly and medium range We have access in the ECMWF MARS data base We download data for case studies analyses of high impact weather in the Mediterranean We also perform operational weather forecasting at regional scale based on the GFS initial and boundary conditions because unfortunately ECMWF has a very restricted policy on the dissemination of real time data
159. of MARS that is modifying or completing the directives provided For old Centre s users it is difficult to know the new things options in MARS Since the paper documentation disappeared I am not comfortable with the use of web documentation at the beginning it was minimalistic and I have no view how it has changed since Perhaps things are described in newsletters but I don t go and see them I preferred the old system giving newsletters by email I m very staisfied by MARS at ECMWF the local interface via metview ran locally in my organisation is less satisfactory How easy do you find MARS to use Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very Easy o 16 Easy 190 50 Neither easy nor difficult 102 2 Difficult 26 7 Very Difficult fe 0 50 of 141 29 06 05 17 21 ECMWF User Survey I r 0 20 40 60 80 i Total Responses 379 e Your comments From memory I think the one thing I used to struggle with is getting a complete definition of each field horizontal grid vertical levels units spot value or time averaging interval etc Apart from that no problems once I got used to the system Easier access without dynamic password change It could be much more friendly such as the American facilities implemented by CDC and ARL A solid and good system in my case it is hidden behind scripts p
160. on from scientists but also general public reminds us the Middle Ages the path to find specific information is not always straightforward I have sometimes difficulties to come again to information which I found previously For some security reason I cannot access all ECMWF pages with the browser that I normally use safari It bothers me to have to use another browser every time I visit the ECMWF web site Too cryptic in terms of data availability to which user who sees what Search function to be improved sometimes difficult to find an information already found in another search on the site response speed probably slow because of local constraints Access via broadband to a home PC is satisfactory The web site has been a labyrinth As the amount of information provided by the website is huge it is not always easy to find the right place Anyway I have no helpful suggestion to make in order to improve this Overall a very good job was done I ve a old SGI INDY with 9600Baud so it might possibly the reason why your web sites are as slow 76 of 141 29 06 05 17 21 ECMWE User Survey it is too difficult to find the information needed Sometimes we feel rounding about without getting to the right page As this is not my preferred method of connection I cannot really comment The search facility never gives me reasonable results probably I don t use it in the right way overall it is OK sometimes looking for specific inf
161. on you website on this subject Have example scripts but more online help would be useful By using my scripts issuing commands to the MARS archive system for retrieving data is proven to be quite easy In the MARS user guide there should be a list or table of all options in request with short explanation to easily check the correct parameters Some of the file manipulation routines are too inflexible e g the retrieval of monthly maximum temperature data from DEMETER required the creation of 1000s of temporary files because I was unable to use looping facilities Maybe the observations could be retrieved by way of the Internet one day Already years ago I expressed the wish for an index of the MARS directives In the user guide I have to guess the category to find an explanation 51 of 141 29 06 05 17 21 ECMWF User Survey Need a centralised documentation of how to retrieve various data streams I could sometimes find that information only in the dissemination area The documentation of the parameters is somehow leaky WebMARS has solved some issues Easy on the web difficult with syntax in batch Sometimes it was hard to find out exactly what was wrong with my script MARS is very difficult to use for a beginner and even with experience requires a lot of attention to understand what one can get and how Many parameters are badly documented Easy to use via Metview Getting started on a new data stream seems difficult each ti
162. one Sometimes it seems the catalogue is not 100 up to date Graphical tools are not flexible I use it to check for availability of data in general as well as specifically and for the MARS request It seems like one can retrieve data in many ways DataFinder or archive etc And sometimes they give different results depending on which path one choose This can be very confusing and I m not sure if I m doing wrong or if it should be like this Not much help for non operational data brand new trials Have to go and ask around for stream method type etc Sorry I vent enough time surfing your web Easy to access and simple GUI I find login at the web interface quite tempremental so it is not my preferred method of connection I m looking forward to be able to use webMARS to get information about operational observations Would be nice if several fields from different level types e g 500 hPa geopotential and surface precipitation could be easily plotted together Maybe it is possible but I haven t managed poor documentation Would be useful if it could deal with archived ocean data Some ongoing problems like retrieving multiple datasets but only one of them is displayed difficulty on selecting the working area Difficult to use not intuitive It is very cryptic and the level of user friendliness is almost as the MARS scripting so in the end is better to use a regular MARS script request Even so it is of some
163. one with the resolution eg 50 km I am the Head of the Met Service s satellite Unit I also develop SAFNWC s PGEO7 and PGE08 algorithms in these development I use RTTOV neural networks etc On the ECMWF I only use the MARS in order to supply GRIB files to SAFNWC reference system and in order to download analisys for the development of SAFNWC s PGEO7 and PGEO8 3D Var development at DWD 10 of 141 29 06 05 17 21 ECMWF User Survey Satellite meteorology at regional scale Mediterranean Sea Use of ECMWF surface analysis data My work is to develop and test Limited Area Models both for Atmosphere and Waves in the Mediterranean area and in particular in Southern Italy Research activity in order to produce down scaled seasonal forecasts for Northern Italy Climate research in general with particular attention to Europe Assimilation of remote sensing data into chemistry transport models In my group we use ECMWE ERA 40 or operational analyses to drive the atmospheric circulation model KASIMA see http www imk fzk de imk2 kasima Coupling of atmospheric chemistry and aerosol models with global circulation models implement develop and monitor a limited area ensemble prediction system run in an experimental operational basis on ECMWF machines I was Head of the Marine Unit until my recent promotion Regional Climate Modelling I am Head of IT Division Met Eireann and responsible for the support maintenance and d
164. ons Response Total of Total Respondents High 223 61 Low ol Pay 18 Not relevant oe E jis 54 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 333 0 20 40 60 80 e DEMETER Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 2 Lz 11 Low 2 Eg 16 Not relevant 190 i 73 Total Responses 261 0 20 40 60 80 BE e Research experiments prepIFS Each Respondent could choose only ONE of the following options Response Total of Total Respondents High s E 10 Low Em aed 21 Not relevant 184 69 Total Responses 267 0 20 40 60 80 e Member State projects Each Respondent could choose only ONE of the following options Response Total of Total Respondents High u e 16 Low s EE 19 Not relevant 176 ii 65 Total Responses 271 0 20 40 60 80 HBE e Other please specify below Each Respondent could choose only ONE of the following options Response Total of Total Respondents High 0 E 5 Low Pp l 3 Not relevant 184 92 55 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 199 0 20 40 60 80 e Please specify other archives datasets member state models If I have a problem I usually get a copy of the user s data and I don t need to question how they obtained it sorry I don t know about datasets Ocean waves fcts
165. ore flexibility in plotting capability of MARS retrievals through web such is control over wind speed vector length contouring colour codes etc Iam satisfied providing better libraries to read this data BUFR GRIB Maybe you could consider providing them with an open source like license to enable others to improve on them Openness information about current and passed model parameterisations More training courses Website could have more support information and be better organised Make clear which documents are current and which are old versions Give more information about what services a new registered user can use even more training courses about the operational model improvement hardly possible keep uptodate as all the time in the last years Some facilities not that user friendly Eg need for an easy to use GRIB decoder like NCEP s werib The ECMWF system for the maintenance of dissemination requirements should be improved Improve the policy on data dissemination and mostly real time Data It is unacceptable that European can have ftp access to real time data through USA instead of through Europe It make ECMWF look oldfashionned Become more user friendly Already answered We will be satisfied if the current service level is maintained Any improvements that can make our lives simpler are welcomed Perhaps make running experiments by scientists at met services easier MOS for chosen locations VNC maybe it
166. our analysis tool VERA using of ECMWF model results in lectures and exercises Data Coordination Project Leader of Data Warehouse Project Radiosonde Climatology Research in Mesoscale Alpine Programme ECMWF Regional climate modelling using ERA 15 ERA 40 data My institution Regional climate modelling 6 of 141 29 06 05 17 21 ECMWF User Survey Tropical prediction in global model Operational Archive Supervisor Working as member of Operational Application Core Suite Team Work with the visualisation tools used within the Met Office to derive information from the EPS forecasts In the last year I have not needed to use the ECMWE system for my work In the past I have mainly used the system to retrieve data from MARS The data I have retrieved has mainly been from ERA I now work in the convection group at the Met Office About 18 months ago I was working in the climate model validation group I am a forecaster in a regional meteorological service and I am greatly occupied in meteorological data acquisition storage and postprocessing Responsible for operational download of met data Responsible for database activities going on at the institute Work general assimilation of satellite measurements of greenhouse gases in order to estimate sources and sinks Work at ECMWF running the chemistry transport model which is driven by ECMWF meteorology winds etc I am a system developer and using MARS archive when there are missing
167. pages when new infos come or what is planed next More forecast products More information on Weekly and seasonal forecast products especifacally in my country and with my problems 118 of 141 29 06 05 17 21 ECMWF User Survey Metview and MARS Compute could be more aligned More examples would be helpful In particular in multi model ensemble data handling SMS ODB IFS technical docs Porting hints for ECMWF exported software Some of the information should be updated Sometimes it is quite hard to find the doc I m looking for But if I manage to find the correct documentation it is very useful then prepIFS perhaps linked to each window of prepIFS IFS model documentation It would be nice to improve the mathematical visualisation in the web documentation values of parametrisations Metview online manuals glossary of functions in macro language It would be able to perform an online tutorial when retrieving meteo data from the MARS archive I need a list of meteo parameters and their exact definition and units and a link where to find more information on the meteo parameter simpler documentation understandable by physicists not NWP experts easier to navigate web page See earlier comments Too often you need to log in with your user identification for information and documentation which can be tedious Especially when working for different computers or locations clarify the way to search the relevant documenta
168. paign data Some of these data we also provide in NRT to you ozone sonde data in CREX format In addition we provide ECMWF data for research campaigns and programs the EMEP program ECMWF has a faster computer and easy fast storage on ECFS They offer a easier access to observation data Most their own data NCEP makes almost everything available on the web I learned meteorology thanks to that If you want to attract young european people to the field of meteorology you should give them access to almost all your products so they can use and learn themselves The other centres often have easier transport of data in an out and a more standard type of Unix interface Clear and fast options for graphical presentation of data The use of these other centre s data is very low compared to ECMWF data Online plots of necessary data more professional a friendly interface we could use DKRZ computers with a german research project no need to write something special parallel vector machine which could be used with openMP no need for MPI ECMWF compares very well I use Manchester sometimes for the ERA40 data held on disk They allow the use of the web for retrieving data and they just use a password to connect to the data centre so the connection is easier For data services in general the level of support is similar The data obtained from US sources is simpler in that there are no controls on user access level So everyone
169. perational NWP models at Meteo France i am the Metview correpondent At ECMWF checking managing the ECMWF dissemination to Toulouse Less often data retrieval and some interaction with MetAppS about libemos At home Head of Operations division within Information systems management Nothing at ECMWF System administrator UNIX Linux in National Meteorological Service M t o France Climat Research GCM development Seasonal forecasting evaluation and development of products based on EPS and monthly forecasts Working on the surface forcing field workpackage 4 of the MERSEA IP Assessing the presence of spin up in forecast output meteorological variabiles to be used in the calculation of surface turbulent fluxes for forcing an ocean model The improvement and evaluation of the bulk parameterisation used to calculate surface turbulent fluxes in an ocean model I work in a research center M t o France about the tools for avalanche forecast From ECMWF we use principally the MARS archive and also the ERA40 experiment I use fields of the CEP analysis in order to compute scores for experimental runs I work in the forecast laboratory team We work about the tools of the forecasters the concepts that they can use to make their forecasts retrieve NWP EC output for comparison with other models Nous d sarchivons les donn es de vent dd ff stress aux heures synoptiques au ECMWF Nous utilisons ensuite ces donn e
170. potentially interested in using them more Actually for a project I ll run soon climate simulations on ECMWF supercomputer I m sure I will appreciate how much the service improved in recent years 141 of 141 29 06 05 17 21
171. r for a long time I normally connect using ssh2 which works 75 percent of the time The rest of the time I have to manually set up the xproxy server and connect to ecgate with telnet However the problem probably originates in Reading as much as it does at the ECMWF Mainly used to loggin at hpc The performance has improved considerably during the last years the connection is sometimes quite slow My WAM wave model retrieval programme was written in house by somebody else but help to modify it over the past number of years was always given by ECMWF staff It s setup is not identical to HPCD all software should have the same version number despite this I m satisfied Server ecgate is reliable and maintenance interventions and other system interruptions are announced properly Rather slow especially afternoons problems with X environment sometimes does not allow X windows opening weekly maintenance sessions sometimes interrupt the work Recently have done no more than login and do some listings This server provides a simple system to access the ECMWF facilities Only once and for a duration of several hours I was not able to connect to ECMWF through RMDCN in the evening Very stable Data transfer by using ectrans creates files as large as the files to be transfered within the hidden directory ecaccess_DO_NOT_REMOVE which are not removed automatically This provokes to overshot the quotas When submitting files a
172. r visualisation in Products gt Forecasts without the need of higher credentials for registered ECMWE users Specific information is sometimes hard to find through the search key Some type of advanced search to narrow the search criteria would be helpful to eliminate large number of hits I would like the same data policy applied by US weather centre a lot of free product and any commercial limitation more products on a regional scale able to zoom in on the European maps or those of other continents Maps for low high middle clouds separately now it is on one map and sometimes hard to read More forecast parameters on the epsgrams wind direction low middle high clouds type of precip in winter availability of others models products for climatic study it s difficult to retrieve data for single stations It would be nice to be able to be able to choose btw either data postscript or png once the request is over very strong 116 of 141 29 06 05 17 21 ECMWF User Survey to much limitations with my account Better access to ensemble products or clearer access methods More products and interactivity Very nice but should be improved observations in WebMARS good but there is always improvement I don t know why but I rarely was successful in finding sth with the search utility Perhaps you should have your pages crawled by Google news improved sitemap xcdp on web The ability to use the Safari browser to log i
173. rectory of SCRATCH more disk space in home currently 10MB more disk space in scratch currently 1 5GB Better access to ecgate SFTP and more control over VNC sessions 1024x768 is too much ectrans from ECMWF to our workstation is too slow higher quota limit Iam very satisfied The only things which bothered me sometimes are the frequent maintenance downtimes Wednesday during working hours 112 of 141 29 06 05 17 21 ECMWE User Survey Response time is sometimes too long speed ectrans should remove any temporary files it has been creating LoadLeveler should be able to recognise a successful termination of ectrans support for bash login shell More gates Faster interactive access Faster data transfer easier file transfer higher quota on localscratch SSH Logion using Private Keys More stability in the connections in particular X Get rid of the security card tesh shell scratch space and connection to the outside for data transfer I guess disk space as always sometimes slow response during telnet sessions occasional time out also occasional ftp time out No MARS or ECFS sessions between 8am and 6pm I know it s impossible better time response in interactive work For the ECMWE course possibility to have some session not always in your center but by exemple in some places near to end user like Toulouse Rome Telnet connection to ecgate to improve High Performance Computing Facility
174. rom one directory to another or one user to another would be nice If the tabbing feature of tcsh would work for files on ecfs that would be nice When creating systems which are use on many platforms would be easier if the ecfs commands would look normal Unix command cp rm then there would be no need of special treatment of ECFS same as above idem One problem is that ECFS related functions are not accessible under bash Problem with path Unix like file structure and command language Only a hand full of commands to learn There used to be a list of archived files It was very useful For us should be more easily if the fields in GRIB code to be in GRID form DWD is using similar software see comment above It could be easier to handle directories If this is the case see above the name of the scratch directory is different fora MARS request and ftp download This is confusing Also the scratch space is limited and I find it difficult or impossible to retrieve big files The set of special commands is great An option to add data to a file on ecfs would be nice for example to add a day file to a monthly tar file I don t often use ECFS I use ECFS at a basic level That s easy Why are you not satisfied with the ECFS service e Tam unsatisfied with the ECFS service because e g poor responsiveness difficult access to the data availability of service bandwidth limitation for data transfer etc
175. rovided by KNMI so I don t really know very helpful the Internet pages for searching the parameters check for availability and constructing the retrieval It is not always clear in which resolution the data are stored The sample retrieve commands generated by the archive section of the ECMWF website are very useful Occasionally I have problems with making MARS accept my date string Not always clear how to choose the best resolution Sometimes it is difficult for me to find which data are available and which are the data codes more easy now with the informations on the web but usually there is a problem when I try to get a new period of data as often something has changed the huge capabilites of MARS are great but make it difficult to use MARS retrieves resulting in errors are difficult to catch in automated retrieves Good support from User Support Although at first glance it looks simple it is in fact rather tricky which I understand due to huge variety of data it handles Sometimes it is not so clear to me what exactly is in which stream I sometimes have problems with land sea mask definitions which are not scale independent Occasional NetCDF problems especially finding out what data is stored in MARS is a problem since only model parameters are in the data finder on your website and not the observations For example I tried once to find out which buoy data are available in the MARS and could find nothing at all
176. rposes Very little use 3 to 4 times with very simple type of data request TOST data The ECMWF Data server is very fine I will use it in the future Archive Services ECFS In this section we would like to find out about your use of the ECMWF File Storage System ECFS How often do you use ECMWE s File Storage service ECFS Each Respondent could choose only ONE of the following options a eee Ld Daily 23 m Several times a week 2 m Once a week 6 u Occasionally p2 E 31 Never j192 49 Other please specify Em i 4 Total Responses 392 0 20 40 60 80 e Other please specify PEPEPEPE 67 of 141 29 06 05 17 21 ECMWFE User Survey when visiting ECMWF once a month Not for a long time I used to use it and hope to use it occasionally in the future after summer sometimes daily sometimes occasionally Intensively during campaign periods I m not sure to have ever use it possibly in the future daily in periods not used to my knowledge but this does not mean it is not being used by scripts that I now maintain used in the past gt 2years ago but not recently In the future and in the past daily At present not very often Once so far hardly ever now almost every time I access ecgate or HPCD monthly I used it only during my 3 month visit to ECMWF How satisfied are you overall with the ECMWF File Storage Each Respondent could choose only ONE of the follow
177. rsion and parameterisation description 1 no complete overview 1 paper because it can be out of date obviously 1 PrepIFS documentation 1 SEASONAL and MONTHLY 1 SMS 1 SMS CDP 1 SMS training course 1 Struggle with the model documentation hard to find what I want and lots in there which seems unclear 1 System documentation 1 technical descriptions of batch procedures completely geared towards the c Unix audience nothing for a windows slave like myself 1 the list of meteo parameters which I can retrieve from MARS 1 Use of the batch queue LoadLeveler 1 user guide to ECMWF forecast products 93 of 141 29 06 05 17 21 ECMWF User Survey 1 web documents 1 Weekly and seasonal forecast 1 workshop and seminar reports What areas of documentation would you like to see improved I would like to see the following documentation areas improved 8 MARS 2 IFS Documentation 2 documentation on how use the ECACCESS software emos preprocessing More documents on the accuracy of ERA 40 in various parts of the globe Model documentation maybe a quick reference bit for basics like grid structure field definitions etc details of the operational forecasts I would like to see a complete guide to emoslib in one document that is possible to download and to print on paper data format Metview documentation arranged in alphabetical order by icon including thumbnail icon pictures info on seasonal and monthly forecast produc
178. ry important to have their visit to our working home environment This way they can experience the actual problems we have at home 106 of 141 29 06 05 17 21 ECMWFE User Survey my problem it s a very specific one and machine related the user support it s all the same very useful They respond very quickly I have got perfect help from you Especially Umberto M and John G Always great to speak with former colleagues it is good to experience prompt reply and action Usually they help me out of trouble within short Or they make clear I demand the impossible e g missing data Find an equally good successor to John Greenaway I have had immediate help in the past but at the moment I am waiting for a reply to my e mail for weeks and I have resent that e mail after a week Nop perfect like this Great team very patient with user from a non computing background generally very fast response from both user support and call desk I ve always found their help very satisfying and as quick as possible No improvement needed very good service and efficient organization especially connection wise Please rate your satisfaction with the service you receive from User Support Call Desk e Courtesy Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 206 i 73 Satisfied oT 24 Neither satisfied nor dissatisfied lo 3 Dissatisfied oO 0 Very disse
179. s jobs from https ecbatch2 ecmwf int and looking at the Monitoring page the Stderr file and Stdout file are created by are empty and cannot explain why the job eventually failed Very poor environment compared to Linux GNU Very primitive vi No interaction of ssh agent with batch gt must split retrieval in steps with need of big buffers on EC system Very limited quota for user disk space Es muy lento porque hay mucha gente conectada a ecgate Deberia haber varios portales ecgatel 2 3 como antes Changes to ecgate need to be more widely advertised hardly use it can t tell I use midnight commander for data transfer Reliable and fast enough after the introduction of Ecaccess and ectrans performance has stabilised and sufficiently fast 35 of 141 29 06 05 17 21 ECMWE User Survey I didn t do much work yet on ecgate just retrieved data It is a nuisance though that the connection is lost after some time in particular X connections using usr local share ecaccess ecauth ecbin AIX ecauth apparently are unstable Mostly use regular scripts which run daily to retrieve data It is good in speed but not easy to use The security card is a hassle although maybe necessary because of the privacy policy Fast and efficient answer to all my little problems As a ignorant user I am surprised and pleased to be able to get what I need The space quota limit is a bit low slow answers in the afternoon for interactive work
180. s pour obtenir des donn es de vent co localis es avec les positions de bou es d rivantes afin de calculer les courants de surface dans le cadre du programme CORIOLIS Data base management Studies using ERA40 Monitoring of all observations assimilated in our operational NWP model at the Centre retrieval of long series of MARS data followed sometimes by some post processing fortran transfer of data to and from Toulouse special programs e g DEMETER at M t o France user support of a research team Work at ECMWF Porting ARPEGE ALADIN AROME and 4DVAR on IBM platform HPCD Work at Meteo France Project leader for the replacement of the VPP5000 23 of 141 29 06 05 17 21 ECMWFE User Survey Running ARPEGE Climat model and post processing at ECMWF Meteo france Computing representative at Meteo France system administration and users support on HPC Studies from meteorological data work on observation targeting issues use of sensitive area grib files and appropriate tools developed either at ECMWF or Meteo France Developing software Data analysis Reasearch activity Running global climate coupled models non urgent simulations for now I have not used much the connection Only for data retrieval but I could have to run our model in the future statistical adaptation daily extraction of forecasted rainfall datas and transmission towards MeteoFrance in order to compare them with others ouputs mo
181. sentation 135 of 141 29 06 05 17 21 ECMWF User Survey Keep searchable mailing lists FAQs They offer a faster data connection My experience is to small to really tell about it Ease of use and much more data As good as any other Very well ECMWF gives a much better service generally speaking not in computing resources but in archiving data retrieval and user support security issues make remote access to Met Office computer system much harder a comparison doesn t have any sense to me Access to BADC is much simpler and it is very easy to get an account However the data at ECMWF is much more interesting for my studies services are comparable except for BADC where I had some bad experience with scientists not being informed early enough about the closing down of a system But prob this was bad luck What other centers have offered is some help with porting the code to their machine i e BADC has been very helpful This is of course a great service and I am not aware that this is done for university scientist that want to run their codes at ECMWF ECMWF compares well with the other computer centres I have used I find they all offer very similar facilities better service at ECMWF very well they not you should learn in particular user support ftp links for downloading data The models I use are designed and tailored for use on the computer hardware offered at NOTUR Both centres are quite good MySCD portal from U
182. ser support is very good Good hardware although a bit changeable in the past year Good accessibility from a variety of sites but connections should live longer It is little time that I use ECMWF dataset for specific comments Ease of access MARS access of re analysis data I have been registered in the ECMWF community for a short time So at moment I have no suggestion on ECMWE services Best regards I am very impressed with the quality of the archiving system Availability of resources Level of help and support available the free access to the data Web services MARS Forecasting Research into NWP Meetings eg seminars User meetings Information content is very good The MARS service is correct and I can easily use it The user support is very good It is very good that ECMWF decided to share their data with the rest of the world Congratulations for that I very satisfied especially with the support offered by the specialist Providing quality atmospheric dynamical fields MARS archive 126 of 141 29 06 05 17 21 ECMWF User Survey Archiving retrieval of data Display of ensemble forecasts Computing Facilities Call desk User support When I have had problems and contacted ECMWF I received efficient and useful assistance and was very pleased with the service im very pleased with everything around data retrieval especially operational and the available information that is always helpfully Easy way to ret reiv
183. sing 19 of 141 29 06 05 17 21 ECMWF User Survey Lidar Instrument for sounding the Middle Atmosphere Integration of lower atmosphere Model data from ECMWF into out GCM for the Middle Atmosphere Case analysis of tropical cyclones idealised numerical simulation The retrieved ERA40 data is used as one of the sources to produce new meteorological analysis downscaled to 11 x 11 km over Europe ERA40 data is used as first guess fields and modified by observations selected from the observation archive at SMHI to achieve fine scale re analyses every 6h for the years 1990 2002 for temperature precipitation and wind The MESAN meteorological analysis system developed at SMHI is used for the re analyses MESAN is based on the optimal interpolation OI technique The obtained results will be evaluated by a cross validation If the results are found to be of good quality the obtained meteorological analyses will later on be used as a basis for regional scale air pollution modelling and assessments for Scandinavia and Europe on contract from the Swedish Environmental Protection Agency The obtained meteorological analysis will thus not be used for commercial purposes but for official duty purposes I work on climate variability and predictability on seasonal to decadal timescales The development and application of coupled GCM models is essential to this area of research Retrieval of ERA 40 data for analysis of atmospheric conditions over Adriati
184. sion to use the ECMWF HPCF via the national weather service is VERY difficult I have heard from others in here that it is very difficult to use and not worth the effort as we could never use it in an operational way which would be my interest Main reason is that I no longer do any development work personally My role involves enabling others to implement projects through provision of resources and training as required Data transfer from remote host to local work station for postprocessing reasons and visualisation Matlab is slowly Sometimes MARS access is so slow that my job is not completed in the usual time and I have to repeat it No appropriate environment to run weather models of our national centre for trial runs Until recently I didn t even know that I had access to the HPCF resources I might start using this facility more now that I know that I can use it Sometimes IBM is a bit different to other machines and makes porting more difficult Work to submit is still in process of preparation Better porting of the UM and UMUI to make easier to run it at the ECMWF s computer Lack of opportunity or priority on the home work In my work supercomputing resources are not necessary No priority on work with model migration The HPCF possibility is always at the back of my head and it is only a question of opportunity and a suitable project to make use of it Difficulty of code porting Lack of experience in parallelisi
185. st change of userID I in fall 04 i was not able to re login It has always worked without a problem Not always is easy to update our certificate browser dependency see comment above SecurID waiting times are an acceptable evil see above there are browsers setups where it can be awkward to log on but this is not a problem Not easy to understand or get an overview of possibilities and why when log in is necessary When login fails it just cycles back to login page no indication why the login process failed Enough for my needs although may well not be the optimal A bit complicated I need my SecurID card even for the smallest informations Again I had to be told by a colleague that I could access the Web page through an ECMWF user certificate I could not readily see it on the Web pages Since I do not have to go through a specific login procedure Occasionally I needed to use the SecurID and have problems with it on several occasions We use SMHI connection and when you open a browser there are three questions to answer before you are authorised Why three instead of one See comment I have had problems accessing but always got good help telephone from ECMWE helpdesk problems with certificate using more than one desktop The use of passcode is convenient for me How do you locate new information on the ECMWF web site e By receiving the web address in an email from ECMWE Each Respondent could choose only ONE of the fo
186. t Seasonal model ensemble work at ECMWF Riceviamo ogni giorno attraverso Areonautica Militare i dati di analisi e forecast della corsa delle ore 12 fino a 144 ore e quelli su griglia ruotata della corsa delle ore 00 fino a 72 ore Sulla base di questi elaboriamo due modelli meteorologici commissionatici dal Ministero per le politiche agricole e forestali UCEA Atmospheric Remote Sensing with GNSS radio occultation ECMWF fields for initialisation and or validation 21 of 141 29 06 05 17 21 ECMWF User Survey I work at the INM HIRLAM Data Assimilation Group My tasks includes increasing the set of assimilated data in our model with new observations such as satellite observation especially since we have started to use 3DVAR work in remote sensing department Giving support to climate researchers regarding computer issues Programming web applications database systems At ECMWF retrieving data from the MARS archive mostly ERA 40 retrieving data ERA40 OD as boundary and initial conditions for regional model simulations MM5 We are using ECMWF data to assist in modelling various aspects of bird movements and behaviour in relation to weather data retrieval for oceanographic experiments performed with numerical oceanic models I m working on dispersion modelling and use MARS data as input for the model FLEXPART I m a local operational meteorologist The main purpose of my work is to use the control forecast aft
187. tc The remainder of my work includes some scientific analysis and informatics support programming problem solving website designs etc I also assist my colleagues at BIRA IASB when they try to access and use the retrieved data I work at the core group of the HIRLAM consortium and in this position I do a lot developing and testing of HIRLAM system All the long tests are done at ECMWF retrieving and visualising meteorological fields for operational forecasters retrieving of archived data for case study work for educational purposes within the organisation Retrieving the meteorological fields for synoptic research Developing scripts to run radiative transfer models and postprocessing routines on HPCA HPCD running the scripts storing results on ECFS for further processing Seasonal Forecasting Our model the GloSea model is run at ECMWF in forecast and hindcast mode My work includes assessing and calibrating forecasts which requires occasional retrievals from MARS and some processing Research on model error to be used in Ensemble Prediction Systems I work on the NWP data processing and archive section of a national meteorological national service which has ECMWE packages has basis of their operational systems ECMWF systems are used for comparison purposes Besides ECMWF computer resources are a complement of home resources The Trans manche pollution transport project Lagrangian particle dispersion and trajectory modelli
188. teorology and numerical modelling numero Braionpvar prad setyith applications to models of atmosphere and oceans turbulence stratified flows Check the weather forecasts for private usage f Pevepupnentiang amplgencetetiant gh SMR aaplications atmospheric prognostic model mAmileti onyen salves qlaernthredarpatprocessing 5 of 141 29 06 05 17 21 ECMWF User Survey Study and implementation of ocean wave modelling in different basins Comparison and validation of model results with measurements from buoys and satellite comparison of ECMWF data with model and recorded data run models to analyse specific cases use ECMWF data analysis reanalysis to hindcast wave conditions of the past Modelling of atmospheric pollutants using off line atmospheric dispersion chemistry deposition models management of forecasted and observed data in n historical observatory educational and research forecasting I am a head of the Weather Analysis and Forecasting Department I also participate in operative and research work I have attended two courses at ECMWE 1986 and 2000 Development and use of LAM in orographically complex terrain with implementation to the modelling of dispersion and deposition of pollutants ECMWF ECMWF fields as the boundary conditions for the local model archived data for building scenarios Extracting met data and processing it for input into atmospheric dispersion model ERA data for earlier dates than
189. the following options 27 of 141 29 06 05 17 21 ECMWFE User Survey Response Total of Total Respondents Web access only 3 I T_T 85 Visit to ECMWF 110 30 Other please specify tm a 5 Total Responses 369 0 20 40 60 80 BBE e Other please specify 2 e mail 2 ecgate 2 ftp 2 java webstart 2 telephone 1 Computer Courses 1 csoMARS 1 data download for testing 1 Dedicated linkup 1 dissemination system 1 Internet 1 ongoing collaboration with graphics team re Metview issues 1 professionist 1 web access What desktop systems do you use to connect to ECMWF e Linux Each Respondent could choose only ONE of the following options Response Total of Total Respondents 6 Always 189 ae 56 Sometimes n2 F 33 Never a7 11 Total Responses 338 0 20 40 60 80 e ATX IBM Each Respondent could choose only ONE of the following options Response Total of Total Respondents Always jo E 6 Sometimes ta 9 Total Responses 161 0 20 40 60 80 ae e HP UX HP Each Respondent could choose only ONE of the following options 28 of 141 29 06 05 17 21 ECMWFE User Survey Response Total of Total Respondents Always 2 Sometimes 2 on 0 12 1 Never Ea 12 76 Total Responses 172 0 20 40 60 80 e IRIX SGI Each Respondent could choose only ONE of the following options Response Total of Total Respondents
190. the new program is going to be online 101 of 141 29 06 05 17 21 ECMWFE User Survey I tend to pick this up via the newsletter it would be nice to get a personal electronic copy It is a problem with meteorological contact point at my Met Service not an ECMWF issue This may be partly my fault more scientific detail of changes needed The schedule for the educational programme is published quite late After publication you don t have much time to apply for a course There have been changes in the computing environment that were not properly announced Because we are all humans I accept that without serious complaining did not receive any newsletter up to now COS_INFO not always reliable up to date my non MS organisation is currently not adequately informed about relevant changes Computer sessions not even easily seen on the web even 5 mins before it happens or maybe hidden somewhere but should be seen easily Some minor changes were not announced however they had implications on some applications eg MAGICS date axis bug How satisfied are you with the overall quality of the advice you receive from User Support Call Desk Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied 166 P 55 Satisfied 117 39 Neither satisfied nor dissatisfied ta E 6 Dissatisfied o 0 Ver dissatished i 0 Total Responses 301 0 20 40 60 80 What are your m
191. ther s experiences I am working at IPSL a federation of six French laboratories located in Paris I am responsible in the IPSL organisation of the Atmospheric chemistry database called Ether I am in charge of retrieving ECMWE operational data for the IPSL scientific community study and numerical simulations of specific meteorological events ECMWE analyses provide initial and boundary conditions At ECMWF participation in committees and working groups At ARPA SIM Director I am the Head of the Weather Forecasting Room of ARPA SIM atmospheric and pollution modelling air sea interaction physical processes in the Planetary Boundary Layer Mesoscale numerical modelling and assimilation I work in the coastal protection field mostly dealing with atmospheric and marine numerical models on the regional and smaller scale Research on deep atmospheric convection and related severe weather phenomena like thunderstorms supercells mesoscale convective complex and systems tornadoes Numerical simulations over mesoscale domains with complex orography by means of the Weather Research and Forecast model Downscaling of global and limited area numerical forecasts for mesoscale and microscale subgrid forecasts Development of tools for forecasts verification both numerical and subjective ones Head of model and forecast verification section I am manly involved in developing new rainfall estimation algorithms from satellite measuremen
192. times per day for a few days a few times per year Daily when I work s on Mars archive Overall how satisfied are you with the service provided by MARS Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied i2 Satisfied 2u Neither satisfied nor dissatisfied 38 Ac 10 Dissatisfied a 1 1 Very dissatisfied 0 Total Responses 381 0 20 40 60 80 e Your comments 3 I don t use MARS 3 slow sometimes There are some problems in the interpolation to rotated lat lon 48 of 141 29 06 05 17 21 ECMWF User Survey I have been missing rotation possibilities for winds for several years but I understand this have been fixed I think it must be an error in the system that the domain can t be selected totally freely the Odeg long and Odeg lat must always be included why Maybe you d like to extract a domain with say 1 deg resolution but centred on half degrees I lack a printed manual I still consult the revision 11 sept 1995 document much more frequently than the web version As a not to initiated user I sometimes wonder what the difference is between certain data e g Forecast analysis initiated analysis first guess etc Again as a not so initiated user I would like to have some printed reference to the accuracy of the retrieved data which I use for other applications e g how good is the retrieved precipitation
193. tion documentation on how to download data I would like to have a place for instruction handbook s interpretation of data and models used Documentation of MARS content Current configuration of everything Better documentation of MARS and dissemination documentations are a bit hidden in the web menus docs link or menu should be on home page Increasing the number of available doc info and making it easier to find it if possible make the documentaion or information easier to find sometimes hidden you have to know the exact URL to reach it news should be send by email like the system sessions User Services User Support Call Desk Each Respondent could choose ANY of the following options Response Total of Total Respondents User Services User Support Call g 100 Desk 119 of 141 29 06 05 17 21 ECMWF User Survey I Total Responses 0 20 40 60 80 e Your comments Often user support personnel is not available because of alternative commitments the service is OK good but there is always improvement Perfect as it is Summary comments Finally we would like to know your comments about ECMWE as a whole What does ECMWE do well What aspects of ECMWFEF services are you most pleased with ECMWF truly assists a scientist with computational needs In general I m satisfied by the services ECMWF is offering to make my work smooth quick and productive Us
194. tion on the meteo parameter CDP SMS decoding encoding Not always easy to find just what you want without taking some time though I can t see a way of easily improving that Some SIMPLE introduction with SIMPLE words to the output of models for use of ECMWF data By SIMPLE I mean understandable by a Physics graduate or master not by a NWP ECMWE expert more easy to find a partically doc documentation about ERA4O can be more often updated if there is a possibility for those who work with that data to put their comments or reference to the paper programming Fortran 90 95 ECMWF s GRIB codification analysis observations I would like to see metadata documentation improved sometimes its just hard to find information How to move data I have produced at ECMWF to my home computers How to optimise a code help on the signification of the different compiling options for a really ignorant user help on how to debug a code with the programs you propose The MARS content documentation on the web could be improved At the moment there are a bewildering array of expver stream and class but few clues to what they mean Also what is the difference between the types 4D VAR analysis and analysis System documentation validation documentation up to date and compact wave and atmo As said on last page a quick clickable configuration of all components I find it very difficult to find the definition of products that can be request
195. ts available to guest users to M S users only to users with real time access Use my userid and passnumber when no other way to access How satisfied are you with the way you have to log in to the ECMWF web site Each Respondent could choose only ONE of the following options Response Total of Total Respondents Very satisfied aa 19 Satisfied 214 56 Neither satisfied nor dissatisfied so 21 Dissatisfied e 4 Very dissatisfied ye 0 Total Responses 381 0 20 40 60 80 e Your comments BBE It would help to have something on the site which makes me aware of what I can t see because I haven t logged in enough sometimes I don t realise that I m not seeing everything because I ve forgotten to log in fully I find it rather cumbersome sometimes to eventually get to the page I want to see It s OK when you know how User certificates were not satisfactory The Secure ID method is much better I do not know exactly how to access 80 of 141 29 06 05 17 21 ECMWF User Survey sometimes strange behaviour asking several times for password To complicate with the certificate I would appreciate having access with sth like a ssh key Using the SecurID card is somewhat inconvenient identification via IP address would be much easier It is often unclear what your status is logged in from domain via certificate not logged in I cannot log in from the default MacOS browser Safari Since the la
196. ts load leveller and running jobs etc exercises about ECMWF product applications Case studies Weekly and monthly forecasting forecast applications more examples regarding data and particularly observations retrieval Availability of pdf files in order to discharge on my computer print the necessary pages and read it at home I have not studied the documentation in detail The only one I have really read was the METVIEWS Export version installation notes It was a very difficult program to compile I wish you could make an rpm with it I have only looked at BUFR documentation Documentation at web site OK Source code documentation and instructions how to use the BUFR library could be improved documentation of data reading routines Most documentation needs an overhaul PrepIFS needs much better documentation and there needs to be a level of documentation that bridges the gap between the theory and the source code behaviour skill and defects of model some of the technical documentation are not always up to date ecgate HPCD some model specific documentation how many soil types does IFS have is not always easy to find it is difficult to find it once you have found it it is very good product description 94 of 141 29 06 05 17 21 ECMWF User Survey As said above it would be fine to be informed about papers from ECMWE people recently accepted by reviewed journals Metview Icon Reference but I just find a n
197. ts and NWP model data I am also involved in the hydrological model rain assimilation and validation section I m working as weather forecaster and developer in particular in the field of land surface model Senior Forecaster 15 of 141 29 06 05 17 21 ECMWF User Survey I work as research engineer on several research projects I work at ECMWF in the framework of a European data assimilation project I also use ERA40 data Development processing and dispatching of the ENACT diagnostics Chemical Transport modelling of stratosphere using ERA40 PhD studies in climate change monitoring using CHAMP radio occultation measurements I m using the MARS archive at ECMWF to download reference data At University of Graz Numerical regional climate modelling with mesoscale and microscale models At ECMWF Long term simulations of current and future climate using mesoscale models MM5 driven by ERA40 and ECHAMS datasets at the HPCF Mesoscale modelling for downscaling of meteorological fields use ECMWF data as boundary conditions for mesoscale models Aerosols effects on the atmosphere deduced from ECMWF reanalysis increments Saharan dust modelling and prediction analysis of atmospheric motions based on Doppler wind profiler and ECMWF fields Air mass trajectories and aerosol pollution detection based on in situ measurements and trajectories from ECMWF fields Research activity about Regional climate model applied other West Africa
198. uage do you regularly use in your work at ECMWF 31 of 141 29 06 05 17 21 ECMWFE User Survey and or in your organisation e Please select all that apply Each Respondent could choose ANY of the following options Response Total of Total Respondents Fortran 77 265 63 Fortran 90 95 307 73 26 C a ee 10 Java 6 None ja ee 6 Other please specify se mm 13 Total Responses 419 0 20 40 60 80 e Other please specify MATLAB IDL R Statistics language Delphi Grads Perl S PV WAVE MATLAB Visual Basic Metview programming language object Pascal OMP and MPI PHP Python Smalltalk Ultrix visual basic l 5 5 3 8 E How would you rate your knowledge experience of a batch environment like NQS NQE or LoadLeveler Each Respondent could choose only ONE of the following options Response Total of Total Respondents No experience at all 126 30 Very limited experience 105 25 Limited experience 104 25 Experienced 76 18 Very experienced lo 1 32 of 141 29 06 05 17 21 ECMWFE User Survey Total Responses 417 0 20 40 60 80 For how many programs have you modified the source code to exploit vector or parallel processing e Vector Each Respondent could choose only ONE of the following options Total Responses 361 0 20 40 60 80 e MPI parallel Each Respondent could choose only ONE of the following options Response Total
199. usefulness when I do not know exactly how to fill in a regular MARS script request Difficult to find clear information See comment made earlier Fast updating of latest model output and lot of data I only use it for finding out what s in the archive then I use scripts from ecgate See my previous comments on documentation it s handy to have webMARS write out the extraction MARS script and cut and paste into ECaccess extracting wave data works very poorly i haven t found a way to use the GRID command to limit the size of extracted fields It would be useful to specify your own contours for each parameter and to be able to overlay more parameters to visualise more parameters at the same map Easy to use but all features are not available ACCURACY PACKING I don t use WebMars only scripts on ecgate slow answer Not relevant I don t know this web I m going to look at it Please give the reasons why you haven t used the webMARS service yet 63 of 141 29 06 05 17 21 ECMWF User Survey 44 So far I had no need 26 I did not know about this service 6 don t know how to use it 3 did not yet have time to try out 2 lack of information I access the MARS archive using my own programs via ECGATE because we retrieve the data using the standard MARS commands 1 Haven t taken the time to learn 2 I m also reasonably happy with submitting scripts which I know work I save these scripts for later usage and t
200. ve system MARS 72 100 Total Responses 72 0 20 40 60 80 e Your comments Speed to handle requests including web MARS Printed complete manual Faster more older data in fast storage Numerical fields should be available also at 600 hPa Could be sped up Time series retrieval plotting analysis tools if you could offer data from a ECMWF trajectory amp dispersion model that would be a great benefit it should be more easy to extract data and observations also from temp observations good but there is always improvement Making MARS and the dissemination system behave identically to identical requests often the system is slow More easy to use Clearer MARS content documentation archive on MARS fields not produced by ECMWF model and be able to plot them using MAGICS Metview More detailed information about how to retrieve specific products 114 of 141 29 06 05 17 21 ECMWF User Survey a clearer table which variables can be downloaded and how to input them in the MARS scripting language Most of the times it is trial and error at least in my case more obvious information on changes in content Slow in getting data Would like to have Meteosat 8 data available on MARS More flexible file manipulation software avoid the need to create multiple temporary files A MARS directive index is missing I have some data request that need more than 6 hours in the LoadLeveler quick access to data Usabil
201. ven though we contacted to get a new one we did not hear about it for a long time The only hitch is that grib2netcdf tool does not work properly for certain advanced variables such as vertically integrated northward heat transport More work space to retrieve MARS data hpc disc mounted to do postprocessing on ecgate I can not understand that it is not possible to connect to ECMWF by ssh It is easy well known and secure provide more options for the vnc service choice of window manager selection of window geometry direct transfer of data over Internet to other servers via ncftp or scp instead of gate ftp no maintenance at Wednesday morning Provide the basic document for the bigginers Use is demand driven If and when I need it it is there I am very happy with the service Nothing comes to my mind at the moment I would like to have McIDAS on ecgate Not applicable for my work up to now because I did mainly retrieve of MARS data We are just happy easier use of X connection to machines outside the ECMWF It would be great if I didn t have to write the programmes for retrieval of MARS Data by another means eg web or windows method Personally in recent years I only use ecgate to update SecurID card user access provide better online documentation on your website I cannot always find what I need Have the setup identical to HPCD this means all software has the same version no running on both systems Esp
202. we need to get confidence with it It is difficult at the beginning until you get used to it or someone explains you how to use it AIX seems like many other Unixs Pd wish I could use tesh You are using other GRIB parameter tables shortcuts than international usual that is confusing me MARS computed fields are wrong labelled in meta information I have to write a own program to correct PDS because mostly I m using that fields in a temporary database There a wrong labelled field is nonsense problem of space disk max of 2 Gb problem of format with GRIB for DEMETER data the format cannot be read locally on our computers difficult seen from user s point of view Some problems initially but ECMWF user support very quickly got me going A little puzzling at times not had an opportunity to play with the commands I look after scripts written by my predecessor in which this command appears Easy once I ve experimented to obtain retrievals of the exact data I require Key set up del backspace etc and history commands are not set up correctly a bit annoying but never bothered to investigate it Documentation on ecmwf int is not up to date at all I don t think any of the access methods actually work as described there but one just has to know that it is e g ssh telnet etc disabled years ago I don t often use ecgate and I modify only parameters in scripts written by someone else five years ago The 20 MB d
203. web site down or products not available refreshed 2 Problems related to my account Specific advice on the submission of user application Contacted them a few times when things weren t working or to update contact details Usually get in touch with Vesa Karhila or his team for debugging building advice for Metview Technical programming questions e g MAGICS Problems receiving FORMOST product Certificates no longer expiring SecurID cards problems using the certificate to log in Failure of a forecast script trying to track down reasons and to re submit I have only once used the call desk regarding login problems I was surprised at how quickly the problem was resolved 10min Can t complain about that Specific principal questions regarding BUFR data arriving problems I usually email user support because something isn t working Help with SMS Depending on problem of our local registered users Also questions during local installation or customisation of ECMWF software are asked request of documentation papers reports suggestions for collaborations special runs specific questions about data Urgent problems User questions from NL Often it concerns technical problems in using ECMWF tools like PrepIFS PrepIFS problems removing experiment from HPCD xcdp Software bugs and problems about the contents of retrieved data questions on provided GRIB BUFR libraries and questions on data available in the MARS speci
204. ystem questions No solution It has always been a real pleasure to work with ECMWF Many thanks In general I am quite satisfied with the services and professionalism at ECMWF No further comments and suggestions Just thanks for your work I found very interesting visit your experts to Croatia at the user meetings held in mid June it looks as if everything has already been decided and little attention is paid to what users say at the meeting itself IT Division is usually the recipient of user comments complaints about the levels of service we provide in house However we do not receive user complaints since the HPCF at ECMWF is compatible with our IBM RS 6000 SP server and users can run parallel jobs using MPI without porting difficulties The Special Project team for C4I are based at HQ Met Eireann and so far have relied upon the HPCF facilities to run their compute expensive climate jobs 140 of 141 29 06 05 17 21 ECMWF User Survey SSH access without secure ID card would be good for some applications idem Ah looking at ECMWF site map I see that you have already improved since I last looked at that page Good Good idea to issue this questionnaire but it is quite long to fill in There could be a little more billing units I think I have received excellent and professional services with good understanding in details of many many things It is hard to understand the data policy in Europe in comparison to the USA

Report on ECMWF User Survey

Contents

Download Pdf Manuals

Related Search

Related Contents