Slide 1© ECMWF Sub-seasonal forecasting, Forecasting system Design Frédéric Vitart European Centre for Medium-Range Weather Forecasts

Slide 2© ECMWF Sub-seasonal forecasting

Slide 3© ECMWF These studies explored the predictability at a subseasonal time-scale (beyond deterministic predictable limit), recognized that the subseasonal prediction can be seen as an initial value problem with external forcings (boundary value problem). “Predictability In the Midst of Chaos” Shukla (1998), Palmer (1993) Pioneers in subseasonal predictions Dr. Kikuro Miyakoda Source: Princeton Univ. webpage Pioneering and challenging work of Miyakoda et al. (1983), Spar et al. (1976), Shukla (1981) opened the door for subseasonal predictions. Miyakoda et at. (1983) Simulation of a blooking event in January MWR Spar et al. (1976) Monthly mean forecast experiments with the GISS model. MWR Spar et al. (1978) An initial state perturbation experiment with the GISS model. MWR Shukla (1981) Predictability of time averages. Part I. Dynamical predictability of monthly means. JAS 3

Slide 4© ECMWF January 1977 Source: NOAA/NWS 4

Slide 5© ECMWF January 1977 Miyakoda et al T850 Forecast (Day10-30) 5

Slide 6© ECMWF First report to the international community Cubasch, Tibaldi, Molteni: Deterministic extended-range forecast experiments using the global ECMWF spectral model Molteni, Cubasch, Tibaldi: Experimental monthly forecasts at ECMWF using the lagged-average forecasting technique 4 case studies in winter 1983/84 9-member lagged-average forecasts I.C. from operational analysis at 6-hour interval T21 and T42 spectral model Fixed SST, persisted from I.C. (no cheating!) Correction for systematic error, based on day integrations in winters 1981/82 and 1982/83, started at 10-day intervals Comparison w.r.t. deterministic forecast from last I.C. and persistence

Slide 7© ECMWF From A. Kumar (NCEP/CPC) Sub-seasonal forecasts Sub-seasonal forecasting is still in its infancy. 10 years ago, only a couple of operational centres were producing sub-seasonal forecasts. Now most of the Global producing Centres are producing and issuing or experimenting sub-seasonal forecasts.

Slide 8© ECMWF Experimental Week3+4 outlook From A. Kumar (NCEP/CPC)

Slide 9© ECMWF Time- range Resol.Ens. SizeFreq.HcstsHcst lengthHcst FreqHcst Size ECMWFD 0-32T639/319L91512/weekOn the flyPast 20y2/weekly11 UKMOD 0-60N216L854dailyOn the fly /month3 NCEPD 0-44N126L6444/dailyFix /daily1 ECD x0.6L4021weeklyOn the flyPast 15yweekly4 CAWCRD 0-60T47L1733weeklyFix /month33 JMAD 0-34T159L6050weeklyFix /month5 KMAD 0-60N216L854dailyOn the fly /month3 CMAD 0-45T106L404dailyFix1992-nowdaily4 Met.FrD 0-60T127L3151monthlyFix monthly11 CNRD x0.56 L5440weeklyFix /month1 HMCRD x1.4 L2820weeklyFix weekly10 Since 1983, most producing centres have developed sub-seasonal forecasts

Slide 10© ECMWF Sub-seasonal Forecast Configuration Different strategy for sub-seasonal forecasting:  In some centres, sub-seasonal forecasts use the same forecasting system as the seasonal forecasting system (e.g. UKMO, NCEP). More frequent start date or larger ensemble size.  In other centres, it is an extension of medium-range weather forecast (e.g. ECMWF/EC)  In other centres it is a separate system which contains characteristics from both medium-range and seasonal forecasting (e.g. ECMWF before 2008/JMA)

Slide 11© ECMWF Sub-seasonal Forecast Configuration Very different configurations of the sub-seasonal forecasting systems (much more than for medium-range or seasonal forecasting). There is currently no consensus on the optimal configuration. Differences in configuration include:  Frequency of forecasts (daily/weekly/monthly)  Ensemble size: e.g. large ensembles run once a week (burst sampling) vs small ensembles run daily (lag ensemble approach)  Model resolution: currently from about 250 km to 50 km  Time range: between 32 and 60 days  Different model set-up: Ocean atmosphere coupling/active sea-ice

Slide 12© ECMWF ModelsTime-rangeFreq.Hcst lengthHcst FreqOcean coupling Active Sea Ice ECMWFD 0-462/weekPast 20y2/weeklyYES Planned UKMOD 0-60daily /monthYES NCEPD 0-444/daily /dailyYES ECD 0-35weeklyPast 15yweeklyNO BoMD 0-602/weekly /monthYES Planned JMAD 0-34weekly /monthNO KMAD 0-60daily /monthYES CMAD 0-45daily1992-nowdailyYES Met.FrD 0-60monthly monthlyYES ISA-CNRD 0-32weekly /monthYESNO HMCRD 0-63weekly weeklyNO Main contribution to YOPP: S2S database

Slide 13© ECMWF Example. The new ECMWF Ensemble fc. system Coupling in single executable NEMO 1/1-0.3 d. lon/lat 42 levels H-TESSEL IFS 41r1 32/64km grid (T636/319) 91 levels 4-D variational d.a. 3-D v.d.a. (NEMOVAR) EDA pert. sing. vectors 5 ocean analyses CGCM 51 runs T639 to 10 d T319 to 46 d Initial conditions perturbations Ens. Forecast The ECMWF ensemble prediction system for the medium and sub-seasonal range

Slide 14© ECMWF Short and medium-range forecasts: instantaneous/daily values Seasonal forecasting: Main products are seasonal or monthly means. Sub-seasonal forecast: Beyond 2 weeks, there is little predictability in the day to day variability, but there is some skill in predicting weekly mean anomalies. Sub-seasonal forecast products

Slide 15© ECMWF Anomalies (temperature, precipitation..) - ECMWF sub-seasonal forecasts

Slide 16© ECMWF Probabilities (temperature, precipitation..) -

Slide 17© ECMWF Weather Regimes

Slide 18© ECMWF Tropical cyclone activity

Slide 19© ECMWF MJO Forecasts

Slide 20© ECMWF Model systematic errors grow during the model integrations and after 2 weeks can be as big as the signal we want to predict. Two options: 1. Make corrections during the model integrations (bias or flux correction) (popular in the climate simulations) 2. Make a-posteriori corrections. The coupled ocean-atmosphere model is run freely and the model systematic errors are estimated from a set of model re-forecasts (same technique as for seasonal forecasting). Implicit assumption of linearity. We implicitly assume that a shift in the model forecast relative to the model climate corresponds to the expected shift in a true forecast relative to the true climate, despite differences between model and true climate. Most of the time, assumption seems to work pretty well. But not always. Sub-seasonal forecasts and re-forecasts

Slide 21© ECMWF Biases (eg 2mT as shown here) are often comparable in magnitude to the anomalies which we seek to predict

Slide 22© ECMWF Time- range Resol.Ens. SizeFreq.HcstsHcst lengthHcst FreqHcst Size ECMWFD 0-32T639/319L91512/weekOn the flyPast 20y2/weekly11 UKMOD 0-60N216L854dailyOn the fly /month3 NCEPD 0-44N126L6444/dailyFix /daily1 ECD x0.6L4021weeklyOn the flyPast 15yweekly4 CAWCRD 0-60T47L1733weeklyFix /month33 JMAD 0-34T159L6050weeklyFix /month5 KMAD 0-60N216L854dailyOn the fly /month3 CMAD 0-45T106L404dailyFix1992-nowdaily4 Met.FrD 0-60T127L3151monthlyFix monthly11 CNRD x0.56 L5440weeklyFix /month1 HMCRD x1.4 L2820weeklyFix weekly10 Since 1983, most producing centres have developed sub-seasonal forecasts

Slide 23© ECMWF Sub-seasonal Re-forecasts Two strategies for re-forecasts in S2S database:  Fixed re-forecasts (e.g. NCEP/BoM/JMA) The model version used to produce the sub-seasonal forecasts is “frozen” for a number of years (e.g. CFS2). The re-forecasts have been produced once for all before the system became operational. Advantage: More user friendly. The user can compute skill and calibration once for all.  “on the fly” re-forecasts (e.g. ECMWF/UKMO/EC..) The model version changes frequently (at least once a year). Therefore re-forecasts have to produce regularly since the model version of the re-forecasts has to be the same as the real-time forecasts. Advantage: This methodology ensures the best model version has been used to produce the sub-seasonal forecasts.

Slide 24© ECMWF The ENS re-forecast suite to estimate the M-climate 20y 51 T 639 L91 51 T319 L … March … ….. Initial conditions: ERA Interim+ ORAS4 ocean Ics+ Soil reanalysis Perturbations: SVs+EDA(2015)+SPPT+SKEB

Slide 25© ECMWF Why not using a 5-week window? Week 0 Week -1 Week -2 Week +1 Week +2 Example: Climate of 06/06 day 26-32: 1-week climate – 5-week climate

Slide 26© ECMWF Re-forecast strategy Re-forecasts are used for model calibration and also for skill assessment.  A large reforecast database is needed for calibration to distinguish between random error and systematic errors and also to estimate flow dependent errors.  A large reforecast database is also needed for verification and for flow dependant skill assessment, like assessing the concurrent impact of ENSO and specific phases of the MJO on the forecast skill scores. Signal to noise ration is also improved in long reforecast datasets (Shi et al, 2014)  Large ensemble size is also important for skill assessment, since some probabilistic skill scores are impacted by the ensemble size. However  Large re-forecast datasets with large ensemble size are often not affordable. Not clear what is more important: ensemble size, number of years?  Long re-forecasts suffer from inconsistent quality in the initial conditions (pre-satellite period).

Slide 27© ECMWF VERIFICATION

Slide 28© ECMWF ECMWF Extended-range forecasts 28

Slide 29© ECMWF Precip anomalies : 26 July 2010 – 01 August 2010

Slide 30© ECMWF ECMWF Monthly Forecast Skill scores ROC area – Probability of 2mtm in upper tercile

Slide 31© ECMWF Skill of the ECMWF Monthly Forecasting System 2-meter temperature in upper tercile - Day ROC scoreReliability diagram Persistence of day 5-11 Monthly forecast day Day Day Persistence of day 5-18 Monthly forecast day 19-32

Slide 32© ECMWF Impact of MJO on forecast reliability T_850 > upper tercile, fc. day Blue line: no MJO in IC Red line: MJO in IC Skill can be flow dependant – Windows of opportunity

Slide 33© ECMWF Linkage with SNAP 33 From Om Tripathi Impact of SSWs on forecast skill scores

Slide 34© ECMWF Model development

Slide 35© ECMWF Resolutions of One-month EPS at JMA Grid resolution Wave number Num. of vert. lev. Model top Ensemble size hPa km Year GSM1304 GSM0801C GSM0103 GSM0603C GSM9603 * Indicates changes with resolution/ensemble size upgrades, only x3 horizontal resolution, x1.5 vertical levels, x5 ensemble size 35

Slide 36© ECMWF Evolution of the ECMWF sub-seasonal ensemble forecasts Frequency Every 2 weeks Once a weekTwice a week Horizontal resolution T159 day 0-32T319 day 0-10 T255 day T639 day 0-10 T319 day T639 day 0-10 T319 day Vertical resolution 40 levels Top at 10 hPa 62 levels Top at 5 hPa 91 levels Top at 1 Pa Ocean/ atmosphere coupling Every hour from day 0Every 3 hours from day 10Every 3h from day 0 Re-forecast period Past 12 yearsPast 18 years Past 20 years Re-forecast size 5 members, once a week11 members, twice a week Initial conditions ERA 40ERA Interim Mar2002 Oct2004 Feb2006 Mar2008 Jan2010 Nov2011 Nov2013 May

Slide 37© ECMWF A success story: forecasting the Madden-Julian Oscillation Wheeler – Hendon (2004) MJO metric based on composite EOFs

Slide 38© ECMWF MJO skill scores

Slide 39© ECMWF MJO teleconnections in October-March 500 hPa height, MJO phase days

Slide 40© ECMWF Skill scores are improving!

Slide 41© ECMWF October 29, 2014 Grid mesh/resolution and sp. harmonic truncation in spectral models Linear grid: spectral truncation N-1, 2N grid points at the equator Quadratic grid: spectral truncation N-1, 3N grid points at the equator Cubic grid: spectral truncation N-1, 4N grid points at the equator “Reduced” grid: No. of points in longitude decreases in steps 41 Octahedral grid: No. of points in longitude decreases continuously

Slide 42© ECMWF Oper 41r1 T+0 Example: 2m temperature over the Alps valid 01 June z Oper 41r1 T+48 TCo r2 T+0 TCo r2 T+48

Slide 43© ECMWF October 29, 2014 Impact of resolution upgrade on sub-seasonal scores 43

Slide 44© ECMWF Impact of resolution on track probability- Tropical cyclone PAM, 9-15/03/ 2015 Oper TL639/31 9 High Tco639/31 9 Tco639 Day Day Observed track

Slide 45© ECMWF

Slide 46© ECMWF Sea Surface Temperatures U50 T850 RPSS over NH Obs SSTs Coupled 80 case, starting on 1 st Feb/May/Aug/Nov WEEK1 WEEK2 WEEK3 WEEK4 MJO Bivariate Correlation Coupled Obs SSTs Pers SSTs

Slide 47© ECMWF Correlations for week 4 Northern Hemisphere WinterSummer Current system With sea- ice model (LIM2)

Slide 48© ECMWF Active sea ice model: Z500 Forecast Skill (weeks 1-4) 80 cases – The vertical bars represent the 95% level of confidence SEA ICE Control

Slide 49© ECMWF New Higher-resolution Ocean Reanalysis

Slide 50© ECMWF 1/4 vs 1 degree – Z500 skill scores -NH New higher-resolution ocean model

Slide 51© ECMWF Conclusions Sub-seasonal forecasting is still in its infancy. There is no consensus on the optimal forecasting system. S2S database will help compare the various forecasting systems. S2S forecasts need calibration. Flow dependant calibration however would need more re-forecasts than currently produced. Sub-seasonal forecasts have improved over the past 10 years, but skill at week4 is still marginally better than climatology. Model are getting more complex, with higher resolution and more components of the earth system.