Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Fuzzy Verification toolbox: definitions and results Felix.

Slides:



Advertisements
Similar presentations
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Improving COSMO-LEPS forecasts of extreme events with.
Advertisements

Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Quantitative precipitation forecasts in the Alps – first.
Exploring the Use of Object- Oriented Verification at the Hydrometeorological Prediction Center Faye E. Barthold 1,2, Keith F. Brill 1, and David R. Novak.
The Consideration of Noise in the Direct NWP Model Output Susanne Theis Andreas Hense Ulrich Damrath Volker Renner.
Ensemble Post-Processing and it’s Potential Benefits for the Operational Forecaster Michael Erickson and Brian A. Colle School of Marine and Atmospheric.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Synthetic future weather time-series at the local scale.
WWOSC 2014 Assimilation of 3D radar reflectivity with an Ensemble Kalman Filter on a convection-permitting scale WWOSC 2014 Theresa Bick 1,2,* Silke Trömel.
1 GOES-R AWG Hydrology Algorithm Team: Rainfall Probability June 14, 2011 Presented By: Bob Kuligowski NOAA/NESDIS/STAR.
4th Int'l Verification Methods Workshop, Helsinki, 4-6 June Methods for verifying spatial forecasts Beth Ebert Centre for Australian Weather and.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss High-resolution data assimilation in COSMO: Status and.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Data mining in the joint D- PHASE and COPS archive Mathias.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Task 1 of PP Interpretation 1.1Further applications of.
Verification methods - towards a user oriented verification WG5.
© Crown copyright Met Office Preliminary results using the Fractions Skill Score: SP2005 and fake cases Marion Mittermaier and Nigel Roberts.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss MAP D-PHASE – a WWRP Forecast Demonstration Project Marco.
Dubrovnik - EWGLAM/SRNWP 8-11/10/ 2007 COSMO strategy for Verification Adriano Raspanti COSMO WG5 Coordinator – “Verification and Case studies” Head of.
Latest results in verification over Poland Katarzyna Starosta, Joanna Linkowska Institute of Meteorology and Water Management, Warsaw 9th COSMO General.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Priority project « Advanced interpretation and verification.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss WG4 Activities Priority project « Advanced interpretation.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Accounting for Change: Local wind forecasts from the high-
Page 1© Crown copyright Scale selective verification of precipitation forecasts Nigel Roberts and Humphrey Lean.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Local Probabilistic Weather Predictions for Switzerland.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss WG4 activities Pierre Eckert, MeteoSwiss, Geneva.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Quantitative precipitation forecast in the Alps Verification.
Priority project Advanced interpretation COSMO General Meeting, 18. September 2006 Pierre Eckert.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss WG 4 activities.
Deutscher Wetterdienst Fuzzy and standard verification for COSMO-EU and COSMO-DE Ulrich Damrath (with contributions by Ulrich Pflüger) COSMO GM Rome 2011.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Status of the COSMO-1 configuration at MeteoSwiss Guy.
Observed & Simulated Profiles of Cloud Occurrence by Atmospheric State A Comparison of Observed Profiles of Cloud Occurrence with Multiscale Modeling Framework.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss A more reliable COSMO-LEPS F. Fundel, A. Walser, M. A.
U. Damrath, COSMO GM, Athens 2007 Verification of numerical QPF in DWD using radar data - and some traditional verification results for surface weather.
Page 1© Crown copyright 2005 Met Office Verification -status Clive Wilson, Presented by Mike Bush at EWGLAM Meeting October 8- 11, 2007.
Statistical Postprocessing of Surface Weather Parameters Susanne Theis Andreas Hense Ulrich Damrath Volker Renner.
Deutscher Wetterdienst Preliminary evaluation and verification of the pre-operational COSMO-DE Ensemble Prediction System Susanne Theis Christoph Gebhardt,
Nathalie Voisin 1, Florian Pappenberger 2, Dennis Lettenmaier 1, Roberto Buizza 2, and John Schaake 3 1 University of Washington 2 ECMWF 3 National Weather.
Page 1© Crown copyright 2004 The use of an intensity-scale technique for assessing operational mesoscale precipitation forecasts Marion Mittermaier and.
Verification of ensemble precipitation forecasts using the TIGGE dataset Laurence J. Wilson Environment Canada Anna Ghelli ECMWF GIFS-TIGGE Meeting, Feb.
WRF Verification Toolkit Workshop, Boulder, February 2007 Spatial verification of NWP model fields Beth Ebert BMRC, Australia.
NCAR, 15 April Fuzzy verification of fake cases Beth Ebert Center for Australian Weather and Climate Research Bureau of Meteorology.
VERIFICATION Highligths by WG5. 2 Outlook Some focus on Temperature with common plots and Conditional Verification Some Fuzzy verification Long trends.
Eidgenössisches Departement des Innern EDI Bundesamt für Meteorologie und Klimatologie MeteoSchweiz Weather type dependant fuzzy verification of precipitation.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Probabilities from COSMO-2 derived with the neighborhood.
DOWNSCALING GLOBAL MEDIUM RANGE METEOROLOGICAL PREDICTIONS FOR FLOOD PREDICTION Nathalie Voisin, Andy W. Wood, Dennis P. Lettenmaier University of Washington,
VERIFICATION OF A DOWNSCALING SEQUENCE APPLIED TO MEDIUM RANGE METEOROLOGICAL PREDICTIONS FOR GLOBAL FLOOD PREDICTION Nathalie Voisin, Andy W. Wood and.
Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Numerical Weather Prediction at MeteoSwiss 8th of October.
Eidgenössisches Departement des Innern EDI Bundesamt für Meteorologie und Klimatologie MeteoSchweiz Weather type dependant fuzzy verification of precipitation.
WG4 Oct 2006 – Sep 2007 plans COSMO General Meeting, 21 September 2006 Pierre Eckert.
Global vs mesoscale ATOVS assimilation at the Met Office Global Large obs error (4 K) NESDIS 1B radiances NOAA-15 & 16 HIRS and AMSU thinned to 154 km.
11 Short-Range QPF for Flash Flood Prediction and Small Basin Forecasts Prediction Forecasts David Kitzmiller, Yu Zhang, Wanru Wu, Shaorong Wu, Feng Ding.
Deutscher Wetterdienst Long-term trends of precipitation verification results for GME, COSMO-EU and COSMO-DE Ulrich Damrath.
New results in COSMO about fuzzy verification activities and preliminary results with VERSUS Conditional Verification 31th EWGLAM &16th SRNWP meeting,
A few examples of heavy precipitation forecast Ming Xue Director
Hydrologic Considerations in Global Precipitation Mission Planning
Intensity-scale verification technique
Fuzzy verification using the Fractions Skill Score
Systematic timing errors in km-scale NWP precipitation forecasts
Verifying Precipitation Events Using Composite Statistics
Spatial Verification Intercomparison Meeting, 20 February 2007, NCAR
Multi-scale validation of high resolution precipitation products
Verifying and interpreting ensemble products
Daniel Leuenberger1, Christian Keil2 and George Craig2
Nathalie Voisin, Andy W. Wood and Dennis P. Lettenmaier
A. Topographic radiation correction in COSMO: gridscale or subgridscale? B. COSMO-2: convection resolving or convection inhibiting model? Matteo Buzzi.
COSMO-DE-EPS Susanne Theis, Christoph Gebhardt, Michael Buchhold,
Quantitative verification of cloud fraction forecasts
Verification of COSMO-LEPS and coupling with a hydrologic model
Numerical Weather Prediction Center (NWPC), Beijing, China
Christoph Gebhardt, Zied Ben Bouallègue, Michael Buchhold
Some Verification Highlights and Issues in Precipitation Verification
Short Range Ensemble Prediction System Verification over Greece
Presentation transcript:

Federal Department of Home Affairs FDHA Federal Office of Meteorology and Climatology MeteoSwiss Fuzzy Verification toolbox: definitions and results Felix Ament MeteoSwiss, Switzerland

2 Fuzzy Verification toolbox Which rain forecast would you rather use? Mesoscale model (5 km) 21 Mar 2004 Sydney Global model (100 km) 21 Mar 2004 Sydney Motivation for new scores Observed 24h rain RMS=13.0 RMS=4.6

3 Fuzzy Verification toolbox Fine scale verification: Fuzzy Methods observation forecast xxx xx xxx x XX XX XX xX X X x x x Intensity Scale Evaluate box statistics (Choose a threshold to define event and non-event) define scales of interest consider statistics at these scales for verification General Recipe “… do not evaluate a point by point match!”  score depends on spatial scale and intensity

4 Fuzzy Verification toolbox A Fuzzy Verification Toolbox Ebert, E.E., 2007: Fuzzy verification of high resolution gridded forecasts: A review and proposed framework. Meteorol. Appls., submitted. Toolbox available at Fuzzy methodDecision model for useful forecast Upscaling (Zepeda-Arce et al. 2000; Weygandt et al. 2004)Resembles obs when averaged to coarser scales Anywhere in window (Damrath 2004), 50% coveragePredicts event over minimum fraction of region Fuzzy logic (Damrath 2004), Joint probability (Ebert 2002)More correct than incorrect Multi-event contingency table (Atger 2001)Predicts at least one event close to observed event Intensity-scale (Casati et al. 2004)Lower error than random arrangement of obs Fractions skill score (Roberts and Lean 2005)Similar frequency of forecast and observed events Practically perfect hindcast (Brooks et al. 1998)Resembles forecast based on perfect knowledge of observations Pragmatic (Theis et al. 2005)Can distinguish events and non-events CSRR (Germann and Zawadzki 2004)High probability of matching observed value Area-related RMSE (Rezacova et al. 2005)Similar intensity distribution as observed

5 Fuzzy Verification toolbox Applying fuzzy scores Fuzzy scores provide a wealth of information, but the results seems to be contrasting their interpretation is sometimes difficult contain too many numbers goodpoor

6 Fuzzy Verification toolbox Application versus testbed Know the scoresForecast error is unknown !? ?! Scores are unknownKnow the forecast error Application Testbed

7 Fuzzy Verification toolbox A Fuzzy Verification testbed Perturbation Generator Analyzer Fuzzy Verification Toolbox Virtual truth (Radar data, model data, synthetic field) Realizations of virtual erroneous model forecasts Realizations of verification results Assessment of sensitivity (mean) [reliability (STD)] Two ingredients: 1.Reference fields: Hourly radar derived rain fields, August 2005 flood event, 19 time stamps (Frei et al., 2005) 2.Perturbations:  next slide

8 Fuzzy Verification toolbox Perturbations PerturbationType of forecast errorAlgorithm PERFECTNo error – perfect forecast!- XSHIFTHorizontal translation Horizontal translation (10 grid points) BROWNIANNo small scale skill Random exchange of neighboring points (Brownian motion) LS_NOISEWrong large scale forcing Multiplication with a disturbance factor generated by large scale 2d Gaussian kernels. SMOOTH High horizontal diffusion (or coarse scale model) Moving window arithmetic average DRIZZLE Overestimation of low intensity precipitation Moving Window filter setting each point below average point to the mean value

9 Fuzzy Verification toolbox Perfect forecast All scores should equal ! But, in fact, 5 out of 12 do not!

10 Fuzzy Verification toolbox Effect of „Leaking“ Scores observation forecast Problem: Some methods assume no skill at scales below window size! p obs =0.5 p forecast =0.5 Assuming random ordering within window yesno yes0.25 no0.25 An example: Joint probability method Forecast OBS Not perfect!

11 Fuzzy Verification toolbox Expected response to perturbations XSHIFTBROWNIANLS_NOISESMOOTHDRIZZLE Sensitivity: expected (=0.0); not expected (=1.0) Contrast := mean( ) – mean( ) Summary in terms of contrast: low high intensity coarse fine spatial scale

12 Fuzzy Verification toolbox Summary real Up- scaling Any- where in Window 50% cover- age Fuzzy Logig Joint Prob. Multi event cont. tab. Intensity Scale Fraction Skill Score Prag- matic Appr. Practic. Perf. Hindcast CSSR Area related RMSE Leaking Scores XSHIFT BROWNIANSMOOTH LS_NOISEDRIZZLE Contrast STD good Leaking scores show an overall poor performance “Intensity scale” and “Practically Perfect Hindcast” perform in general well, but … Many score have problem to detect large scale noise (LS_NOISE); “Upscaling” and “50% coverage” are beneficial in this respect Leaking scores show an overall poor performance “Intensity scale” and “Practically Perfect Hindcast” perform in general well, but … Many score have problem to detect large scale noise (LS_NOISE); “Upscaling” and “50% coverage” are beneficial in this respect

13 Fuzzy Verification toolbox Spatial detection versus filtering  x=25km   x=10km   x=5km  Horizontal translation (XSHIFT) with variable displacement  x “Intensity scale” method can detect spatial scale of perturbation All other methods like the “Fraction Skill score” just filter small scale errors

14 Fuzzy Verification toolbox Redundancy of scores Correlation (%) of resulting scores between all score for all thresholds, window sizes – averaged over all types of perturbation:  Groups of scores: UP, YN, MC, FB, PP FZ, JP FB, PP, (IS)

15 Fuzzy Verification toolbox August 2005 flood event Precipitation sum : Mean: 106.2mm Mean: 43.2mm Mean: 73.1mm Mean: 62.8mm (Hourly radar data calibrated using rain gauges (Frei et al., 2005))

16 Fuzzy Verification toolbox Fuzzy Verification of August 2005 flood Based on 3 hourly accumulations during August 2005 flood period ( ) bad good Intensity threshold (mm/3h) Scale (7km gridpoints) COSMO-7 COSMO-2

17 Fuzzy Verification toolbox Fuzzy Verification of August 2005 flood COSMO-7 better COSMO-2 better neutral Difference of Fuzzy Scores Intensity threshold (mm/3h) Scale (7km gridpoints)

18 Fuzzy Verification toolbox D-PHASE RADAR Operational phase (June until November 2007) is running 33 atmospheric models take part … … and store there output in a common format in one data archive Demonstration of Probabilistic Hydrological and Atmospheric Simulation of flood Events in the Alpine region Standard verification (see Poster) Let’s apply the fuzzy toolbox Models: COSMO -2, -7, -DE, -EU Period: August 2007 Lead times: most recent forecast starting at forecast hour +03. Observations: Swiss Radar data aggregated on each model grid To be verified: 3h accumulation of precip.

19 Fuzzy Verification toolbox D-PHASE: August 2007 Intensity Scale score (preliminary), 3h accumulation COSMO-7 COSMO-2 COSMO-DE COSMO-EU

20 Fuzzy Verification toolbox Conclusions Fuzzy Verification score are a promising framework for verification of high resolution precipitation forecasts. The testbed is a useful tool to evaluate the wealth of scores (not necessarily fuzzy ones): Not all scores indicate a perfect forecast by perfect scores (Leaking scores). The “intensity scale” method is able to detect the specific scale of an spatial error. MeteoSwiss goes for: Upscaling, Intensity scale, Fraction skill score ( and Pracitically perfect hindcast) methods. First long term application for D-PHASE has just started.

21 Fuzzy Verification toolbox Summary ideal Up- scaling Any- where in Window 50% cover- age Fuzzy Logig Joint Prob. Multi event cont. tab. Intensity Scale Fraction Skill Score Prag- matic Appr. Practic. Perf. Hindcast CSSR Area related RMSE Leaking Scores XSHIFT BROWNIANSMOOTH LS_NOISEDRIZZLE Contrast STD good

22 Fuzzy Verification toolbox D-PHASE: August 2007

23 Fuzzy Verification toolbox D-PHASE: August 2007

24 Fuzzy Verification toolbox D-PHASE: August 2007 – cosmoch7

25 Fuzzy Verification toolbox D-PHASE: August 2007 – Cosmoch2

26 Fuzzy Verification toolbox D-PHASE: August LME

27 Fuzzy Verification toolbox D-PHASE: August LMK

28 Fuzzy Verification toolbox August 2005 flood event Precipitation sum : Mean: 73.1mm Mean: 62.8mm Mean: 106.2mm Mean: 43.2mm 7

29 Fuzzy Verification toolbox August 2005 flood event Fuzzy Verification (hourly accumulations): COSMO-7 COSMO-2

30 Fuzzy Verification toolbox August 2005 flood event Fuzzy Verification COSMO-2 – COSMO-7: Suprisingly, small differences However, COSMO2 seems to be slightly better slightly better at