Benefit of high resolution data assimilation and observing systems in the Met Office UK NWP model

1. Benefit of high resolution data assimilation and observing systems in theMet Office UK NWP model G.T. Dow and B. Macpherson 13th EMS Annual Meeting & 11th European Conference on Applications of Meteorology (ECAM) 09 – 13 September 2013, Reading, United Kingdom

2. Outline • Purpose • UK4 Model and Trial Configurations • Results • General • Case Studies • Observation Network Denials • Summary

3. Purpose Local hi-res DA - why bother ??! Want to demonstrate added benefit relative to cheaper initialisation options eg downscaling

4. Outline Purpose UK4 Model and Trial Configurations Results General Case Studies Observation Network Denials Summary

5. Resolution:4km, 70 levels LBCs:from25kmGlobal model 3DVAR (with FGAT) + IAU for all observationsexcept Latent Heat Nudging for radar-derived surface rain rate Uniform 4km analysis grid 8 three-hourly cycles per day UK4 Vital Statistics

6. UK4 Trials to measure and apportion marginal benefit Total Benefit of high-resolution DA A vs C • Full local DA (CONTROL) • Partial local DA - omitting UK4 onlyObs types • Downscaler - run from reconfigured 25km → 4km Global Analysis Benefit from observations only used in the high-resolution DA system A vs B Benefit from higher-resolution alone B vs C • MOPS cloud fraction profiles • radar-derived surface rain rate • visibility from SYNOPs • T2m & RH2m from Highways Agency roadside sensors • Doppler radial winds These observations were denied to the ‘Partial Local DA’ trial

7. Trial Periods Forecasts to T+24 at 00Z, 06Z, 12Z & 18Z Period Dates No. of Forecasts Summer Jul 1st→ Aug 10th 4x40=160 Autumn Nov 1st→ Dec 14th 4x44=176 Winter Jan 3rd→ Feb 10th 4x38=152 Spring Mar 10th→ Mar 31st 4x21=84 Period picked at Random Period picked due to specific (SCu) event

8. UK Index Metric • Weighted Basket of Indices • 6 elements • Combo of ETS & RMS scores • UK4 Impact trials verified to T+24 at 00, 06, 12, 18 UTC

9. Outline Purpose UK4 Model and Trial Configurations Results General Case Studies Observation Network Denials Summary

10. Verification by Period UK Index Benefit Period Days of full local DAsystem of extraobs types used in the UK4 model of higher resolution alone Summer 40 +2.54% -0.96% (2.54--0.96) = +3.50% Autumn 44 +1.17% +0.28% (1.17-0.28) = +0.89% Winter 38 +0.78% -0.18% (0.78--0.18) = +0.96% Spring 21 -4.83% -4.47% (-4.83 -- 4.47) = -0.36% Cf. typical annual UK Index progression of 2% per annum

11. = Case Study Verification by Element Colors indicate best setup for element/period Full DA(9) Partial DA(11)DownScaler(4) Period Vis Precip Cloud Amount Cloud Base Height Temp Wind Overall Jul 2011 Nov 2011 Jan 2012 Note: Some boxes are more significant than others Mar 2012

12. False Detection P( ) 0.5mm/6hr 1.0mm/6hr 4.0mm/6hr 8.0mm/6hr Summer Precipitation 6-hour accumulations July 2011 Threshold E T S Frequency Bias P(Detection) Full DA Partial DA DownScaler Full DA – LHN

13. T+7 Partial DA Full DA DownScaler Stratocumulus PeriodMar 10th -15th 2012 • Blocked episode • Cloud not breaking soon enough • Significant T2m errors • Suspect analysed cloud depth too large

14. Similar Bias at T+0 Large Bias in Control at T+12 Cloud Cover March 2012 10-15th (SCu) 16-31st (post-SCu) Bias Bias RMS RMS Partial DA ~ Control – MOPS Cloud Full DA Partial DA DownScaler Full DA – MOPS Cloud

15. 10-15th (SCu) Bias T+1 SCu 10th-15th RMS T2m impact from SCu errors Full DA Partial DA DownScaler Full DA – MOPS Cloud Timeseries 10th-31st T+12 Forecast Range →

16. Outline Purpose UK4 Model and Trial Configurations Results General Case Studies Observation Network Denials Summary

17. UK4 Observation Network denial experiments(Autumn period)

18. Summary • Consistent Benefit for all elements from full higher-resolutionanalysis (except perhaps for wind) relative to downscaled analysis • Mixed performance from the extra observations • Sometimes detrimental to the UK Index scores • Consistent summer precipitation benefit up to T+6 from Radar RainRate (LHN)and for some thresholds to T+12 • 3D MOPS cloud analysis (subsequently replaced) shows overall benefit for cloud cover, but not so good for blocked SCu • Visibility (not presented here) – higher thresholds benefit from vis assimilation, lower thresholds sensitive to RH bias • Beware – largest DA (cloud) impacts can be later than T+1 ! • Greatest contribution to UK Index from surface observation network

19. Questions ? With thanks to: Bruce Macpherson, Mark Weeks, Dale Barker, Jorge Bornemann, Richard Renshaw And others... Thank you for listening

20. Extra slides...

21. UK4 Downscaler Configuration & Analysis As UK4 except • Runs 4x daily • 2stage 4D-Var assimilation at 120km/60km resolutions, resulting 25km Global Analysis then reconfigured to 4km/UK4 70 levels • Prognostic ‘murk aerosol’ value for visibility (constant value in operational) • Global Obs Cut Off ~ hh+160 mins

22. 5th Dec 2011 T+1 Full DA Partial DA Downscaler Precipitation

23. UK4 Observation Network denial experiments (November 2011 period)

24. UK4 Observation Network denial experiments

25. T+7 Partial DA Full DA Full DA Partial DA DownScaler DownScaler Arctic Ice Syndrome ….. the thickness also matters Stratocumulus PeriodMar 10 -15 2012 • Cloud not breaking soon enough • Significant T2m errors • Analysed cloud depth too large • Suspect Cloud Top Height too high T+1

26. 16-31st 10-15th (SCu) Cloud Cover March 2012

27. T+1 10-14th T+12 16-31st T2m impact from SCu errors

28. Awkward Period (see later) Verification by Element

29. Roadside sensor network OpenRoad – full network SYNOP

30. Roadside sensor network impact Mean T2m error control test RMS T2m error (2nd half of Dec 2010)

31. Further Work • Do additional trial configurations to decipher signals from individual obs types • Assess benefit from staggered Data Times • Look at sensitivity of DA signal to synoptic conditions • Adjoint-based impact studies in the UK Model (see Richard Marriott’s talk tomorrow)