Use of Ground based GNSS data in NWP at UK Met Office

Size: px

Start display at page:

Download "Use of Ground based GNSS data in NWP at UK Met Office"

Cody Goodwin
7 years ago
Views:

1 Use of Ground based GNSS data in NWP at UK Met Office Gemma Bennitt, E-GVAP workshop, 6th Nov 2008

2 Presentation Outline Introduction to the operational Met Office NWP models Pre-processing of ZTD observations Forward modelling What we assimilate and why Benefits Future developments Questions

3 Introduction

4 Global Model: 40km 4D-VAR Our main NWP models NAE Model: 12km 4D-VAR UK Model: 4km 3D-VAR

5 Pre-Processing

6 Processing incoming data Whitelist Stationlist Observation Processing System

7 Observation Processing System Forward Model Bias Correction Quality Control Checks Temporal Thinning

8 Observation Processing System Forward Model Bias Correction Quality Control Checks Temporal Thinning 28-day mean Obs-Model Site-Processing centre specific Fixed value of bias

9 Observation Processing System Forward Model Bias Correction Quality Control Checks Temporal Thinning Obs-Model ZTD 55mm Station heightmodel surface 300.0m

10 Observation Processing System Forward Model Bias Correction Quality Control Checks Temporal Thinning NAE Model: 1 Obs per station per hour UK Model: 1 Obs per station closest to analysis time

11 Forward Modelling

12 Forward Modelling

13 Forward Modelling ZTD 6 0 = 10 Ndz

14 Forward Modelling ZTD 6 0 = 10 Ndz N ap = + T be T 2

15 Forward Modelling ZTD 6 0 = 10 Ndz N ap = + T be T 2 = 6 i i i+ 1 i ZTD 10 N ( z z )

16 Forward Modelling Calculate delay for full layer p n+1 p n+1 θ,,q q n n n p n p n Zz a a (n+1) Z b (n) Zz (n) b (n) Zz a (n) a (n) Assume constant potential temperature and humidity within model layers Calculate delay for partial layer p 2 p 2 θ,,q q p 1 p 1 GPS station G PS station height Z a (2) z a (2) Z b (1) z b (1) Z a (1) z a (1) Z b (0) z b (0) Linearly interpolate pressure onto b layers Linearly interpolate pressure from top of model layer down to station height

17 Forward Modelling Calculate delay for full layer p n+1 p n+1,q θ n, q n n p n p n Z a (n+1) z a (n+1) Z b (n) z b (n) Z a (n) z a (n) Assume potential temperature and humidity are the same below model surface as at level 1 Calculate delay below model bottom p 2 p 2 θ,,q q p 1 p 1 Z a (2) z a (2) Z b (1) z b (1) Z a (1) z a (1) Z b (0) z b (0) GPS G station station height Linearly extrapolate pressure down to station height

18 What do we assimilate and why?

19 What do we assimilate and why? Assimilating into operational models since March 2007 Only assimilate data from GOP, GFZ and METO Monitoring we use this to help us decide what data to use

20 Monitoring

21 Monitoring

22 Monitoring

23 Data assimilated

24 Benefits

25 Benefits of ZTD Observations Foreca st Range (hours T+6 ) T+12 T+18 RMS fit to observat ions for 3.1% Surface 4.1% temperat ure 4.3% RMS fit to observati ons for 0.1% Surface -0.2% winds 0.1% Also small improvement in visibility, precipitation and cloud Overall weighted score showed 1.85% improvement T % 0.3%

26 ZTD Forecasts

27 Future Work

28 Possible future developments Global data may become available Use a site specific whitelist approach Automatic bias correction updates New forward model

29 New Forward Model Monitoring

30 Summary Operational assimilation into NAE and UK models since March 2007 Assimilate GOP, GFZ and METO Static bias correction scheme Basic forward model assumes constant refractivity within model layers Overall 1.85% improvement in forecast

31 Questions and answers Further questions

Cloud verification: a review of methodologies and recent developments

Cloud verification: a review of methodologies and recent developments Anna Ghelli ECMWF Slide 1 Thanks to: Maike Ahlgrimm Martin Kohler, Richard Forbes Slide 1 Outline Cloud properties Data availability