Risultati da 1 a 16 di 16

Discussione: Aggiornamento del modello ECMWF

  1. #1
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314

    Aggiornamento del modello ECMWF

    ECMWF forecast and analysis system updated

    An updated version of the ECMWF forecast and analysis system, cycle 38r1, was implemented on 19 June. The new cycle includes a collection of improvements to the forecast model and the data assimilation system, affecting the high-resolution forecasts and analyses, the ensemble of data assimilations (EDA) and the ensemble prediction system (EPS) including its monthly extension.
    Upper-air scores changes are positive and statistically significant when verified against the respective analyses throughout the free atmosphere. Differences are generally smaller and less statistically significant when verified against observations. There are consistent improvements in precipitation forecasts, both in the tropics and extratropics. For 2-metre temperature (T2m) there is a shift to slightly colder values by about 0.1 to 0.2 K. The T2m errors are reduced in the tropics and extra-tropics. There are statistically significant improvements also in the probabilistic skill of the ensemble, consistent with the improvements to the high-resolution forecast.

    Cycle 38r1 includes a major improvement of the characterization of the background errors that are used in the 4D-Var analysis. These were obtained from re-computing their climatological structure by gathering statistics from the latest available version of the Ensemble of Data Assimilations (EDA) and model cycle 37r3. The spatial structure and amplitude of the background errors are particularly important because they, to a large extend, determine the impact of the observations in the analysis.

    Meteorological contents of the new cycle

    Updated background-error covariance statistics (for Jb) based on the cycle 37r2 EDA: New background error covariance statistics have been produced with data from EDA (the ensemble of data assimilations) using cycle 37R2. These provide updated error statistics for 4D-Var that are more consistent with the current forecast model and observing systems.
    Improved statistical filtering of EDA sampled background errors. This formulation of the filter overcomes some of the limitations of the original version and allows the EDA error statistics "of-the-day" to have a larger flow-dependent impact on the assimilation weights.
    Assimilation of MHS channel 5 over land and the all-sky radiance product from SEVIRI on board Meteosat-9 (the OSCAT scatterometer data on board the Indian Oceansat-2 satellite will be monitored): MHS channel 5 radiance assimilation over land improves mid-lower level moisture analysis. SEVIRI all-sky radiances now includes clear and overcast observations. OSCAT scatterometer data is monitored to complement ASCAT (on MetOp) in the future.
    Modified convective downdraught entrainment: This improves lower-tropospheric temperature and surface wind speeds in the tropics.
    Changes to cloud ice fall speed, ice supersaturation, melting of ice to rain and freezing timescale of rain: This reduces upper-tropospheric temperature bias and improves liquid/frozen precipitation.
    Improved de-aliasing of the pressure gradient term, reducing numerical noise, allowing a reduction of the horizontal diffusion in the forecast: This leads to a reduction of wind noise near lower latitude orography; the reduction of diffusion produces more realistic energy spectra for small scale motion.
    Improved description of swell: Improved description of wind sea and swell by a slight retuning of the wind input source function in order to agree better with observed growth by waves. As a consequence the dissipation source function was slightly modified as well
    Changes specifically to the EPS

    Use of a new surface reanalysis to initialize the surface fields in the EPS hindcasts.This improves the consistency between forecasts and model climate, visible for example in the Extreme Forecast Index for 2m temperature in particular.
    Extension of the EPS hindcast years from 18 to 20, all based on ERA-interim initial data. This provides a more stable hindcast climate.
    EDA perturbations redefined using the EDA ensemble mean instead of the EDA control as the reference: Up to cycle 37r3, the EDA initial perturbations are defined as differences between perturbed EDA members and the unperturbed EDA control member. From 38r1, the EDA initial perturbations are defined as differences between perturbed EDA members and the mean of the EDA. If the distribution sampled by the EDA is symmetric, the latter method retains the shape of this distribution. In addition, the variance of the singular vector initial perturbations is increased in 38r1 in order to compensate for the reduction of spread that is due to the change of the EDA perturbations.

    New model output parameters

    There are six new parameters produced by the model; They are:

    Atmospheric model parameters from the high-resolution forecast and the EPS at all post-processing time steps

    paramId shortName name GRIB edition
    228001 cin Convective inhibition 1
    260121 kx K index 2
    260123 totalx Total totals index 2
    Note that K index and Total totals index are surface parameters encoded in GRIB edition 2.

    Meteorological impact of the new cycle

    High-resolution forecast performance


    The new cycle has undergone extensive testing in research mode from June to December 2011, and pre-operationally from January 2012. Upper-air scores changes are positive and statistically significant when verified against the respective analyses throughout the free atmosphere. Differences are generally smaller and less statistically significant when verified against observations. The comparison of verification scores of the new cycle to scores of the current operational model for different parameters and domains can be seen in the cycle 38r1 score card. There is clear improvement in terms of most forecast variables well into the medium range. The deterioration seen in geopotential at upper levels (around 100 hPa) in the tropics is due to the reduction of numerical aliasing noise which affects the deep-layer mean pressure evolution.

    There are consistent improvements in precipitation forecasts, both in the tropics and extratropics. The reduction of error as measured by the SEEPS score is in the order of 1%. Improvements are consistent across seasons and significant at the 95% level for most lead times, especially in the tropics.

    For 2-metre temperature (T2m) there is a shift to slightly colder values by about 0.1 to 0.2 K. The T2m errors are reduced in the tropics and extra-tropics. In the tropics the improvement is significant at 95% for most lead times, in the extra-tropics it is significant up to D+4 at 00UTC and D+6 at 12 UTC. However, the improvement is not apparent in the scores for Europe.

    Assessment of a number of synoptic cases including some severe events has not shown any significant or systematic difference in performance between the new cycle and the operational one from the forecaster's perspective.


    EPS forecast performance

    The EPS performance has been evaluated over 64 cases (12 September to 16 December 2011 every other day and 9 April to 2 May 2012). The upper air verification against the respective analyses is summarized in the cycle 38r1 EPS scorecard. The main results are:

    Improvements for the control forecast and ensemble mean are statistically significant and consistent with the changes of skill of the high-resolution forecast.
    There are statistically significant improvements also in the probabilistic skill of the EPS (evaluated in terms of the continuous ranked probability score). These improvements are consistent with the impact on control forecast, ensemble mean and high-resolution forecast.
    Changes in ensemble spread are small.

    http://www.ecmwf.int/products/changes/ifs_cycle_38r1/
    http://www.ecmwf.int/publications/cm.../1340270006042
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  2. #2
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Il ciclo aggiornato 38-R2 è quasi pronto e dovrebbe diventare operativo probabilmente nei mesi da Aprile a Giugno (Q2 2013), in quanto non c'è ancora una data ufficiale. Questa nuova versione di ECMWF aumenta la risoluzione verticale da 91 a 137 livelli, doppiando GFS anche sulla risoluzione in altezza. Si aggiungono quindi 46 livelli agli esistenti.
    Meteorological contents of the new cycle:

    Increase of number of vertical levels from 91 to 137 in high-resolution forecast model (HRES), the ensemble of data assimilations (EDA), the main assimilation (4DVAR) and the Boundary-Conditions (BC) optional programme.
    Revised background error variances at 137 levels based on IFS cycle 38r1.
    Revised EDA calibration and filtering for 137 levels.

    http://www.ecmwf.int/products/changes/ifs_cycle_38r2/

    Vedremo
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  3. #3
    Cumulus Congestus L'avatar di bigshot
    Data Registrazione
    13-06-2012
    Località
    Marcon (VE)
    Messaggi
    93
    Citazione Originariamente Scritto da mr tornado Visualizza Messaggio
    Il ciclo aggiornato 38-R2 è quasi pronto e dovrebbe diventare operativo probabilmente nei mesi da Aprile a Giugno (Q2 2013), in quanto non c'è ancora una data ufficiale. Questa nuova versione di ECMWF aumenta la risoluzione verticale da 91 a 137 livelli, doppiando GFS anche sulla risoluzione in altezza. Si aggiungono quindi 46 livelli agli esistenti.
    Meteorological contents of the new cycle:

    Increase of number of vertical levels from 91 to 137 in high-resolution forecast model (HRES), the ensemble of data assimilations (EDA), the main assimilation (4DVAR) and the Boundary-Conditions (BC) optional programme.
    Revised background error variances at 137 levels based on IFS cycle 38r1.
    Revised EDA calibration and filtering for 137 levels.

    http://www.ecmwf.int/products/changes/ifs_cycle_38r2/

    Vedremo
    Beh, ma finché il NOAA non si dota di un supercomputer di potenza simile a quello del consorzio ECMWF la vedo dura un recupero.

  4. #4
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Citazione Originariamente Scritto da bigshot Visualizza Messaggio
    Beh, ma finché il NOAA non si dota di un supercomputer di potenza simile a quello del consorzio ECMWF la vedo dura un recupero.
    Questo non è esatto

    http://www.ncep.noaa.gov/newsletter/...rintable.shtml
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  5. #5
    Cumulus Congestus L'avatar di bigshot
    Data Registrazione
    13-06-2012
    Località
    Marcon (VE)
    Messaggi
    93
    Citazione Originariamente Scritto da mr tornado Visualizza Messaggio
    http://www.top500.org/list/2012/11/?page=1

    200 teraflops contro oltre 600 è un confronto un po' impietoso...

  6. #6
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    c'è crisi
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  7. #7
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Scusate, uso questo elemento per porre alla vostra conoscenza importanti e recenti considerazioni riguardo i modelli globali.
    Allora, il buon Yang, la mente di GFS, ha pubblicato una revisione delle performances di GFS nel 2012. Vengono sfatati diversi luoghi comuni... Leggere leggere leggere!!!

    http://www.emc.ncep.noaa.gov/gmb/wx2...eview.2012.pdf

    Un'altra recente pubblicazione riguarda l'analisi della differenza tra i run 06 e 18z dagli 00z e 12z. Vengono sfatati (ma talora confermati anche) molti luoghi comuni... Leggere leggere leggere!!!!

    http://www.emc.ncep.noaa.gov/gmb/wx2...ycle_fyang.pdf
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  8. #8
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Nel 2015 due importanti aggiornamenti per EC:
    1. IFS cycle 40r3
    2. IFS T2047

    Con i nuovi supercomputer Titan della Cray stanno preparando il globale a 10 km di risoluzione, adesso gira a 16 km attuali, scendendo anche sotto la stessa risoluzione attuale dei due WRF (modelli LOCALI!) del LAMMA che operano a 12 km.

    La tabella di marcia di EC

    Clicca l'immagine per ingrandirla. 

Nome: Immagine.png 
Visualizzazioni: 12 
Dimensione: 92.1 KB 
ID: 17396
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  9. #9
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    La nuova versione del modello ECMWF dovrebbe entrare operativa il 14 di aprile a T2047 (10 km di risoluzione) contro T1534 (15 km) di GFS.

    http://www.ecmwf.int/en/forecasts/do...del/cycle-41r1

    E' lunghetta ma se volete leggere ci sono un sacco di novità.
    Vedremo i miglioramenti.
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  10. #10
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Forse avevo capito male io ma la risoluzione sembrerebbe rimanere a 16 km.
    Cossa me gero bevuo?
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  11. #11
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    ECMWF è pronta per uno dei più grandi passi nella previsione numerica.
    Tra l'altro ha rifilato una sonora sconfitta a GFS nella previsione dell'uragano Joaquin. Leggete il blog su twitter di Ryan Maue di WeatherBell

    https://twitter.com/ryanmaue

    Immagine.png
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  12. #12
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    Colpo grosso a Reading

    ECMWF plans to upgrade the horizontal resolution of its analyses and forecasts.The upgrade will have a horizontal resolution that translates to about 9 km for HRES and the data assimilation (the outer loop of the 4D-Var) and to about 18 km for the ENS up to day 15. The resolution of the ENS extended (day 16 up to day 46) will be about 36 km.
    A new cycle of the IFS will be introduced to implement the horizontal resolution upgrade. This cycle is labelled 41r2, and includes a number of enhancements to the model and data assimilation listed herein. The detailed specification of the resolution upgrades included in IFS cycle 41r2 are:
    Introduction of a new form of the reduced Gaussian grid, the octahedral grid, for HRES, ENS and ENS Extended;
    Horizontal resolution of the HRES increased from TL1279 / N640 to TCO1279 / O1280, where subscript C stands for cubic and O for octahedral;
    Horizontal resolution of the ENS increased from TL639 / N320 to TCO639 / O640 for ENS (Days 0 - 15) and from TL319 / N160 to TCO319 / O320 for ENS Extended (Days 16 - 46);
    For the medium-range ENS there will no longer be a decrease of resolution at day 10: the ENS Days 11 - 15 will be run at the same TCO639 / O640 resolution as ENS Days 0 - 10;
    Increase of the HRES-WAM resolution from 0.25 to 0.125 degrees and the ENS-WAM from 0.5 to 0.25 degrees;
    Horizontal resolution of the EDA outer loop is increased from TL399 to TCO639 with its two inner loops increased from TL159 / TL159 to TL191 / TL191, respectively;
    Horizontal resolution of the three 4DVar inner loops is increased from TL255 / TL255 / TL255 to TL255 / TL319 / TL399, respectively.

    atitude-longitude resolutions for atmospheric products provided in dissemination
    HRES
    HRES products will be available at the base grid resolution of 0.1° and multiples thereof.
    For backwards compatibility, HRES fields will also be available at the compatibility grid resolution of 0.125°.
    All product grids that can be generated from the base grid (0.1°) will be generated from that grid.
    Those grids that cannot be generated from the base grid will be generated with the compatibility grid (0.125°).
    For example, grids of 0.5° and 1.0° will be generated from the base grid; grids of 0.25° and 0.75° will be generated from the compatibility grid.

    Per assurdo il globale ha risoluzione migliore di alcuni LAM (tipo il LAMMA 12 km su medesimi dati che viaggia a 12 km).
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  13. #13
    Responsabile Scientifico L'avatar di Gigero
    Data Registrazione
    08-05-2012
    Località
    Cavazzale (VI)
    Messaggi
    4660
    Da quando questo upgrade?
    - Marco Rabito -

    Tecnico Meteorologo Certificato WMO 1083

    AMMINISTRATORE E PREVISORE
    https://www.facebook.com/pages/Seren...51079521608104

  14. #14
    Moderatore L'avatar di mr tornado
    Data Registrazione
    10-05-2012
    Località
    Padova (PD)
    Messaggi
    6314
    metà marzo
    Francesco De Bon

    "Nulla è difficile per chi ama" (Cicerone)
    Dati meteo disponibili su www.meteosangiorgio.altervista.org

  15. #15
    Responsabile Scientifico L'avatar di Gigero
    Data Registrazione
    08-05-2012
    Località
    Cavazzale (VI)
    Messaggi
    4660
    Sempre top.
    - Marco Rabito -

    Tecnico Meteorologo Certificato WMO 1083

    AMMINISTRATORE E PREVISORE
    https://www.facebook.com/pages/Seren...51079521608104

  16. #16
    Team Member L'avatar di Ale
    Data Registrazione
    10-05-2012
    Località
    Brugine (PD)
    Messaggi
    4515
    Bisogna sperare che l'uscita del regno unito dall'UE non porti problemi a ECMWF http://www.ecmwf.int/en/about/media-...oting-leave-eu
    Nell'avatar supercella del 17/7/09
    Le mie foto
    http://www.flickr.com/photos/87282230@N08/sets

Tag per Questa Discussione

Segnalibri

Permessi di Scrittura

  • Tu non puoi inviare nuove discussioni
  • Tu non puoi inviare risposte
  • Tu non puoi inviare allegati
  • Tu non puoi modificare i tuoi messaggi
  •