1 / 67

XIII COSMO General meeting Rome, 5-8 September 2011

COSMO-LEPS: present status and developments Andrea Montani, D. Cesari, C. Marsigli, T. Paccagnella ARPA-SIMC HydroMeteoClimate Regional Service of Emilia-Romagna, Bologna, Italy. XIII COSMO General meeting Rome, 5-8 September 2011. Outline. Recent upgrades of COSMO-LEPS:

jovan
Download Presentation

XIII COSMO General meeting Rome, 5-8 September 2011

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. COSMO-LEPS: present status and developments Andrea Montani,D. Cesari, C. Marsigli, T. PaccagnellaARPA-SIMCHydroMeteoClimate Regional Service of Emilia-Romagna, Bologna, Italy XIII COSMO General meeting Rome, 5-8 September 2011 A.Montani; The COSMO-LEPS system.

  2. Outline • Recent upgrades of COSMO-LEPS: • use of soil-moisture analysis fields from COSMO-EU; • Performance of the system: • time-series verification of COSMO-LEPS using SYNOP; • Implementation of 00UTC COSMO-LEPS: • present situation and future developments; • Implementation of SOCHMEL (“Sochi ensemble”). • Work for LAMEPS_BC project. • Modifications to the clustering/selection technique. • Summary and plans. A.Montani; The COSMO-LEPS system.

  3. COSMO-LEPS suite @ ECMWF: present status 16 Representative Members driving the 16 COSMO-model integrations (weighted according to the cluster populations) employing either Tiedtke or Kain-Fristch convection scheme (randomly choosen) + perturbations in turbulence scheme and in physical parameterisations 3 levels 500 700 850 hPa 4 variables Z U V Q d+3 d+4 d d+5 d+1 d+2 d-1 Cluster Analysis and RM identification Cluster Analysis and RM identification older EPS 00 2 time steps younger EPS 12 European area clustering period Complete Linkage • suite runs as a “time-critical application” managed by ARPA-SIMC; • Δx ~7 km; 40 ML; fc+132h; • COSM0 v4.12 since July 2010; • computer time (20 million BUs for 2011) provided by the COSMO partners which are ECMWF member states. COSMO-LEPS Integration Domain COSMO-LEPS clustering area A.Montani; The COSMO-LEPS system.

  4. Outline • Recent upgrades of COSMO-LEPS: • use of soil-moisture analysis fields from COSMO-EU; A.Montani; The COSMO-LEPS system.

  5. 9 November 2010: • the deterministic run (LMDET) takes the ICs from the soil fields provided by the proxy-run mixed with some ECMWF fields (soil-proxy); • COSMO-LEPS output (up to fc+60h; last 3 months) is saved on ML to enable further downscaling; • 11 April 2011: • COSMO-LEPS members use, as ICs, the soil fields provided by COSMO-EU (soil-merge); • perturbation values are tuned and new perturbations are introduced; • upgrade of INT2LM to version 1.18 (with grib_api). • 5 July 2011: • ECMWF EPS members are saved on ML over a large domain to enable reruns (last 3 months). • No upgrades of COSMO since July 2010 (using model version 4.12) Upgrades during the “COSMO year” A.Montani; The COSMO-LEPS system.

  6. Oper system (interp) x = 7 km fcst range = 132h initial conditions: interpolated from EPS members; perturbations: type of convection scheme; tur_len; pat_len; crsmin; rat_sea; rlam_heat. COSMO-LEPS with soil-moisture fields from COSMO-EU Test system (merge) x = 7 km fcst range = 48h initial conditions: interpolated from EPS members merged with surface and soil-layer fields from COSMO-EU (T_SO, W_SO, W_ICE, W_I, FRESHSNW, RHO_SNOW, W_SNOW, T_SNOW). perturbations: type of convection scheme; tur_len; pat_len; crsmin; rat_sea; rlam_heat; mu_rain; cloud_num. “Oper” and “Test” were run in parallel from 1/12/2010 to 15/3/2011 (> 100 cases). A.Montani; The COSMO-LEPS system.

  7. BIAS of T2M Ensemble Mean —— oper (interp) —— test (merge) T2m forecasts are corrected with height • bias computed over 3 different domains for the period 1/12/2010  15/3/2011 (> 100 cases). fulldom (~1400 synop) mapdom (~410 synop) • Bias closer to zero for “test” ensemble, which uses the soil moisture fields from COSMO-EU. • Smaller amplitude in bias oscillations for “test”. • fulldom and mapdom: the improvement is systematic for all forecast ranges; the cold bias of “oper” is reduced. • mapdom < 100m: large reduction of bias for day-time verification; increase of the bias for night-time verification. mapdom < 100m (~50 synop) A.Montani; The COSMO-LEPS system.

  8. MAE of TD2M Ensemble Mean —— oper (interp) —— test (merge) • mae computed over 3 different domains for the period 1/12/2010  15/3/2011 (> 100 cases). mapdom (~410 synop) fulldom (~1400 synop) • Lower mae for “test” ensemble, which uses the soil moisture fields from COSMO-EU. • fulldom and mapdom: the mae reduction is small, but systematic and lasts for about 2 forecast days. • mapdom < 100m: the mae reduction is more evident. mapdom < 100m (~50 synop) A.Montani; The COSMO-LEPS system.

  9. Summary of results for soil-merge experiments • T2M and TD2M: • reduction of bias and mean-absolute error if the COSMO-LEPS members take the initial conditions from the soil moisture fields provided by COSMO-EU; • the improvement is confirmed over larger and smaller domains; • the reduction of errors lasts for more than 2 forecast days; • TOT_PREC: • the impact is neutral (not shown). + • the transfer of SMA files from DWD to ECMWF is solid and timely. Soil-merge was successfully implemented in the operational COSMO-LEPS suite on 11 April 2011 (… and no problems since then!) A.Montani; The COSMO-LEPS system.

  10. Outline • Performance of the system: • time-series verification of COSMO-LEPS using SYNOP; A.Montani; The COSMO-LEPS system.

  11. SYNOP on the GTS Time-series verification of COSMO-LEPS Main features: variable: 12h cumulated precip (18-06, 06-18 UTC); period : from Dec 2002 to Jul 2011; region: 43-50N, 2-18E (MAP D-PHASE area); method: nearest grid point; no-weighted fcst; obs: synop reports (about 470 stations/day); fcst ranges: 6-18h, 18-30h, …, 102-114h, 114-126h; thresholds: 1, 5, 10, 15,25, 50mm/12h; system: COSMO-LEPS; scores: ROC area, BSS, RPSS, Outliers, … both monthly and seasonal scores were computed A.Montani; The COSMO-LEPS system.

  12. Time series of ROC area • Area under the curve in the HIT rate vs FAR diagram; the higher, the better … • Valuable forecast systems have ROC area values > 0.6. • Improvement of skill detectable during 2010 and 2011 for all but the highest threshold. • In 2011, few events with heavy precipitation: possible lack of significance of the results for the 15mm threshold. • Limited loss of predictability with increasing forecast range. • fc 30-42h: ROC area is low for the 15mm threshold: WHY? • fc 78-90h: in the last months, ROC area values have shown little dependence on the threshold. A.Montani; The COSMO-LEPS system.

  13. Seasonal scores of ROC area: the last 4 springs • Fixed event (“12h precip > 5mm”): consider the performance of the system for increasing forecast ranges. • Fixed forecast range (fc 30-42h): consider the performance of the system for increasing thresholds. • Valuable forecast systems have ROC area values > 0.6. • Need to take into account the different statistics for each season (last MAM was the driest). • Best performance for the spring 2011 (with a good advantage), for all forecast ranges and thresholds. • Similar results for longer forecast ranges or higher thresholds (not shown). A.Montani; The COSMO-LEPS system.

  14. Outliers: time series + seasonal scores • How many times the analysis is out of the forecast interval spanned by the ensemble members. • … the lower the better … • Performance of the system assessed as time series and for the last 4 winters. • Evident seasonal cycle (more outliers in winter). • Overall reduction of outliers in the years up to 2007; then, again in 2009 and 2010, but not in 2011. • Need to take into account the different statistics for each season. • In the short range, best results for last winter. • For longer ranges, the performance of the system is “stable”. • Outliers before 10% from day 2 onwards. A.Montani; The COSMO-LEPS system.

  15. Ranked Probability Skill Score: time series + seasonal scores • A sort of BSS “cumulated” over all thresholds. RPSS is written as 1-RPS/RPSref. Sample climate is the reference system. RPS is the extension of the Brier Score to the multi-event situation. • Useful forecast systems for RPSS > 0. • Performance of the system assessed as time series and for the last 4 springs (MAM). • the increase of the COSMO-LEPS skill is detectable for 3 out of 4 forecast ranges along the years; • RPSS positive for all forecast ranges since May 2008; encouraging trend in the last year. • Good results for MAM 2011 (12h-cycle of the score). A.Montani; The COSMO-LEPS system.

  16. Outline • Implementation of 00UTC COSMO-LEPS; • present situation and future developments; A.Montani; The COSMO-LEPS system.

  17. COSMO-LEPS @ 00UTC • Background: • Due to the increasing use and interest in COSMO-LEPS products, a new ECMWF time-critical application was prepared and sent to ECMWF as for the implementation of COSMO-LEPS also at 00UTC (cleps00). • Main features: • same configuration as present cleps12 (modify cluster analysis??); • post-processing will be done using Fieldextra; • cleps00 will run in the late morning, interacting with ECMWF system sessions  product delivery might be occasionally delayed a few hours. • Present situation: • depending on ECMWF comments/suggestions, cleps00 could be implemented by the end of 2011 (test products should be available earlier!). A.Montani; The COSMO-LEPS system.

  18. Outline Sochi (RU) Rome (I) • Implementation of SOCHMEL (“Sochi ensemble”); A.Montani; The COSMO-LEPS system.

  19. SOCHMEL (SOCHi-targeted Mesoscale EnsembLe system) • Next Olympics and Paralympics Winter Games will take place in Sochi, Russia (7-23 February and 7-16 March 2014). • For this event, WMO launched a WWRP RDP/FDP initiative, named FROST-2014 (Forecasting and Research: the Olympic Sochi Testbed)???????? • COSMO will perform a number of activities related to Sochi-2014 WWRP RDP/FDP: • deterministic forecasting, • probabilistic forecasting: • b1) FDP initiatives (“clone” of COSMO-LEPS over the Sochi area), • b2) RDP initiatives(development of a COSMO LAM-EPS system for the Sochi area). • post-processing and product generation, • verification. A.Montani; The COSMO-LEPS system.

  20. FDP (Forecast Demonstration Project) initiatives SOCHMEL7: limited-area ensemble system based on COSMO (“relocation” of COSMO-LEPS) Horizontal resoluzion: 7 km. Vertical resolution: 40 model levels. Forecast range: 72 hours. Starting time: 00UTC, 12UTC. Ensemble size: 10 members. Boundary conditions: from selected ECMWF EPS members. Initial conditions: interpolated from ECMWF EPS members. Timetable Pre-trial: 2011/12 winter Trial: 2012/13 winter Olympics: 2014 winter. A.Montani; The COSMO-LEPS system.

  21. SOCHMEL suite @ ECMWF: present status 10 Representative Members driving the 10 COSMO-model integrations (weighted according to the cluster populations) employing either Tiedtke or Kain-Fristch convection scheme (randomly choosen) + perturbations in turbulence scheme and in physical parameterisations As for the future, apply for an ECMWF time-critical application 3 levels 500 700 850 hPa 4 variables Z U V Q d d+1 d+2 d+3 Cluster Analysis and RM identification Cluster Analysis and RM identification ECMWF EPS 2 time steps clustering interval Black-Sea area Complete Linkage SOCHMEL clustering area • Δx ~ 7 km; 40 ML; fc+72h; • initial time: 00UTC, 12UTC; • for the moment, computer time (10 million BU for 2012) provided by an ECMWF Special Project; • suite managed by ARPA-SIMC; • contributions from ECMWF member states could be needed in the future. SOCHMEL Integration Domain A.Montani; The COSMO-LEPS system.

  22. Outline • Work for LAMEPS_BC project; A.Montani; The COSMO-LEPS system.

  23. Aim: generation of EPS boundary conditions (at 6 and 18 UTC) for use of LAMEPS community Preliminary issue: to make progress in terms of increasing the efficiency of archiving and retrieving EPS data. The solution: to archive the data on the native reduced Gaussian grid of the ECMWF IFS, but only in a region of interest over the Northern Atlantic and Europe (a bitmap mask is applied so that only data in the masked region are coded in GRIB). The next step: to check if the regionally archived EPS data can be used by the LAMEPS groups. LAMEPS_BC project: background A.Montani; The COSMO-LEPS system.

  24. For one case study, ECMWF fields were archived both globally (as generated by the operational EPS) and regionally (data are on a reduced gaussian grid N320 for all fields). • LAMEPS groups are running their LAMEPS systems taking the boundary conditions from both experiments and comparing the forecasts. • As for COSMO-LEPS, cleps_g takes the globally-archived boundaries; cleps_r takes the regionally-archived boundaries: • upper-level fields: no differences are found for between cleps_g and cleps_r; • surface fields (2-metre temperature and precipitation): a few differences are found between cleps_g and cleps_r; differences are confined to very few grid points, but are not negligible. Further investigation is in progress. LAMEPS_BC project: present situation For problems relative to ECMWF data, contact M. Leutbecher. A.Montani; The COSMO-LEPS system.

  25. Outline • Modifications to the clustering/selection technique. A.Montani; The COSMO-LEPS system.

  26. Test modifications of clustering methodologies • always select control runs by ECMWF EPS: • a new 16-member ensemble (ctrl-LEPS) was created by taking the first 14 members from COSMO-LEPS and adding 2 more members nested on 00UTC and 12UTC the EPS control members; • ctrl-LEPS was compared to COSMO-LEPS for 3 months; • as for precipitation forecasts, the two systems have identical performance; • further investigation is needed to assess spread/skill relationship of both systems. • consider shorter forecast ranges for clustering intervals (e.g.: 48-72h, 72-96h); • select the Representative Members from only one EPS: • could be strategic in view of having COSMO-LEPS at both 00 and 12UTC. A.Montani; The COSMO-LEPS system.

  27. Outline • Summary and plans. A.Montani; The COSMO-LEPS system.

  28. Main results Operational implementation of the soil-merge • This should guarantee lower bias and mae for both T2M and TD2M in short range COSMO-LEPS forecasts. Time-series verification • ECMWF EPS changed substantially in the last year (introduction of EDA-based perturbations) and it is hard to disentangle improvements related to COSMO-LEPS upgrades from those due to better boundaries; nevertheless: • high values of BSS and ROC area for the probabilistic prediction of 12-h precipitation, • lower outliers. A.Montani; The COSMO-LEPS system.

  29. Study in more detail the impact of the “new” ECMWF EPS on COSMO-LEPS. Develop new products for users (don’t keep them waiting for too long!), using Fieldextra as much as possible. Test modifications to the clustering methodology. Increase vertical resolution (40  50 ML)? COSMO-LEPS@ 00UTC: implement it, finish it, put it operational; then, think about generation of products from lagged ensembles. SOCHMEL: have a first prototype (inclusive of post-processing and dissemination) by December  for the future, need of computer time from COSMO countries! LAMEPS_BC project  consider possible modifications to COSMO-LEPS suite. COSMO-LEPS for TIGGE-LAM (GRIB2 encoding). Future plans • Support calibration and verification. • Carry on collaboration within research project (e.g. EFAS, SAFEWIND, IMPRINTS, NURC, …). A.Montani; The COSMO-LEPS system.

  30. Thank you ! European Conference on Applications of Meteorology / EMS annual meeting 12 – 16 September 2011, Berlin (Germany) Session NWP4: Probabilistic and ensemble weather forecasting applications at short and very short ranges (including nowcasting) http://meetings.copernicus.org/ems2011 A.Montani; The COSMO-LEPS system.

  31. FROST-2014 vs SOCHMEL • Introduction to FROST-2014: • What is it? • What has to do with COSMO? • COSMO ensemble activities within FROST-2014: • introduction to SOCHMEL (the SOCHi-targeted Mesoscale EnsembLe system) • methodology; • phases of development; • planned activity. • Final remarks. A.Montani; The COSMO-LEPS system.

  32. Important ingredients • Develop experience with probabilities. • Feedback on the top-priority products. • Snow analysis. • Soil-field initialisation. • Observations to assess the quality of the system. • Computer time. • Timeliness in product delivery. • … More details in the plenary talk A.Montani; The COSMO-LEPS system.

  33. Test data for LAMEPS Boundary Conditions A.Montani; The COSMO-LEPS system.

  34. Proposed region for archiving LAMEPS BC data A.Montani; The COSMO-LEPS system.

  35. Outline • Introduction: • migration to the 7-km system. COSMO-LEPS 10 km (old) COSMO-LEPS 7 km (new) A.Montani; The COSMO-LEPS system.

  36. COSMO-LEPS (developed at ARPA-SIM) What is it? It is a Limited-area Ensemble Prediction System (LEPS), based on COSMO-model and implemented within COSMO (COnsortium for Small-scale MOdelling, which includes Germany, Greece, Italy, Poland, Romania, Switzerland). Why? It was developed to combine the advantages of global-model ensembles with the high-resolution details gained by the LAMs, so as to identify the possible occurrence of severe and localised weather events (heavy rainfall, strong winds, temperature anomalies, snowfall, …) generation of COSMO-LEPS to improve the Late-Short (48hr) to Early-Medium (132hr) range forecast of severe weather events. A.Montani; The COSMO-LEPS system.

  37. Operational set-up Additional products: • 1 deterministic run (ICs and 3-hourly BCs from the high-resolution deterministic ECMWF forecast) to “join” deterministic and probabilistic approaches: start at 12UTC; t = 132h; • 1 hindcast (or proxy) run (ICs and 3-hourly BCs from ECMWF analyses) to “downscale” ECMWF information: start at 00UTC; t = 36h. Core products: 16 perturbed COSMO-model runs (ICs and 3-hourly BCs from 16 EPS members) to generate, “via weights”, probabilistic output: start at 12UTC; t = 132h; A.Montani; The COSMO-LEPS system.

  38. Score dependence on the domain size (1) • Verification of COSMO-LEPS against synop reports over the MAP D-PHASE area (~ 470 stations; MAPDOM) and the full domain (~ 1500 stations; fulldom):  different statistics of the verification samples;  up to now, performance of the system over the 2 domains assessed only for 6 months (March-August 2007).  difficult to draw general conclusions A.Montani; The COSMO-LEPS system.

  39. OUTL RPSS ROC Score dependence on the domain size (2) • RPSS score… the higher the better… (and positive). • ROC area… the higher the better… (and above 0.6). • Outliers percentage … the lower the better. • Smoother transitions from month to month in “fulldom” scores. • Slightly better performance of COSMO-LEPS over the MAPDOM, but the signal varies from month to month. • Higher predictability with orographic forcing?  Need to check individual regions and/or to stratify for type of stations. A.Montani; The COSMO-LEPS system.

  40. Bias and rmse of T2M Ensemble Mean • Consider bias (the closer to zero, the better) and rmse (the lower the better). • Bias closer to zero (0.5 °C of decrease) and lower rmse for the 7-km suite. • The improvement is not “massive”, but detectable for all forecast ranges, especially for day-time verification. • The signal is stable (similar scores for 1-month or 3-month verification). • Need to correct T2M forecasts with height to assess the impact more clearly. A.Montani; The COSMO-LEPS system.

  41. Overestimation of Td2m and soil moisture (1) • Verification period: MAM07 and MAM08. • Obs: synop reports (about 470 stations x day). • Region: 43-50N, 2-18E (MAP D-PHASE area). • Larger bias and larger rmse in MAM08 rather than in MAM07 for COSMO-LEPS deterministic run (in 2007, no multi-layer soil model). A.Montani; The COSMO-LEPS system.

  42. Semi-diurnal cycle in COSMO-LEPS scores • BSS score … the higher the better … • Performance of the system assessed for 5 different Summers (JJA). BSS • Evident 12-hour cycle in BSS scores (the same holds for RPSS, while less evident for ROC area scores). • Better performance of the system for “night-time” precipitation, that is for rainfall predicted between 18Z and 6Z (ranges 30-42h, 54-66h, …). • The amplitude of the cycle is somewhat reduced throughout the years and with increasing forecast range. • The bad performance in Summer 2006 is confirmed. A.Montani; The COSMO-LEPS system.

  43. Main results • COSMO-LEPS system runs on a daily basis since November 2002 (6 “failures” in almost 6 years of activity) and it has become a “member-state time-critical application” at ECMWF ( ECMWF operators involved in the suite monitoring). • COSMO-LEPS products used in EC Projects (e.g. PREVIEW), scientific experiments (e.g. COPS, MAP D-PHASE, EFAS) and met-ops rooms across COSMO community. Time series scores cannot easily disentangle improvements related to COSMO-LEPS itself from those due to better boundaries by ECMWF EPS. • Nevertheless, positive trends can be identified: • increase in ROC area scores and reduction in outliers percentages; • positive impact of increasing the population from 5 to 10 members (June 2004); • some deficiency in the skill of the system were identified after the system upgrades occurred on Feb 2006 (from 10 to 16 members; from 32 to 40 ML + EPS upgrade!!!), but scores are encouraging throughout 2007 and, to a certain extent, 2008. • 2 more features: • marked semi-diurnal cycle in COSMO-LEPS scores (better skill for “night-time” forecasts); • better scores over the Alpine area rather than over the full domain (to be confirmed). A.Montani; The COSMO-LEPS system.

  44. COSMO-LEPS 16-MEMBER EPS 51-MEMBER EPS MAM06 Average values (boxes 0.5 x 0.5) tp > 1mm/24h tp > 5mm/24h • As regards AVERAGE precipitation above these two threshols, the 3 systems have similar performance. A.Montani; The COSMO-LEPS system.

  45. ENSEMBLE SIZE REDUCTIONIMPACT EVALUATED ON CASE STUDIES (1) A.Montani; The COSMO-LEPS system.

  46. ENSEMBLE SIZE REDUCTIONIMPACT EVALUATED ON CASE STUDIES (2) Observed precipitation between 15-11-2002 12UTC and 16-11-2002 12 UTC Piedmont case A.Montani; The COSMO-LEPS system.

  47. 2002111212 Piedmont 5 RMs 10 RMs All 51 A.Montani; The COSMO-LEPS system. 20 mm 150 mm

  48. COSMO-LEPSReal time products A.Montani; The COSMO-LEPS system.

  49. Why Limited Area Ensemble Prediction? • Global Ensemble Prediction Systems • have become extremely important tools to tackle the problem of predictions beyond day 2 • are usually run at a coarser resolution with respect to deterministic global predictions → skill inforecasting intense and localised events is currently still limited. A.Montani; The COSMO-LEPS system.

  50. Why Limited Area Ensemble Prediction? (2) As regards high resolution deterministic forecast in the short range, where limited-area models play the major role, a “satisfactory” QPF is still one of the major challenges. The same can be said for other local parameters. This is due, among other reasons, to the inherently low degree of predictability typical of severe and localised events. Probabilistic/Ensemble approach is so required also for the short range at higher resolution A.Montani; The COSMO-LEPS system.

More Related