UMAC data callpage 1 of 25North American Ensemble Forecast System - NAEFS EMC Operational Models...

UMAC data call page 1 of 25North American Ensemble Forecast System - NAEFS

EMC Operational Models

North American Ensemble Forecast System

Yuejian ZhuEnsemble Team Leader, Environmental Modeling CenterNOAA / NWS / NCEP

[email protected]


Statement

The North American Ensemble Forecast System (NAEFS) combines state

of the art weather forecast tools, called ensemble forecasts, developed at the US

National Weather Service (NWS) and the Meteorological Service of Canada

(MSC). When combined, these tools (a) provide weather forecast guidance for the

1-14 day period that is of higher quality than the currently available operational

guidance based on either of the two sets of tools separately; and (b) make a set of

forecasts that are seamless across the national boundaries over North America,

between Mexico and the US, and between the US and Canada. As a first step in

the development of the NAEFS system, the two ensemble generating centers, the

National Centers for Environmental Prediction (NCEP) of NWS and the Canadian

Meteorological Center (CMC) of MSC started exchanging their ensemble forecast

data on the operational basis in September 2004. First NAEFS probabilistic

products have been implemented at NCEP in February 2006. The enhanced

weather forecast products are generated based on the joint ensemble which has

been undergone a statistical post-processing to reduce their systematic errors.


North American Ensemble Forecast System

International project to produce operational multi-center ensemble products

Bias correction and combines global

ensemble forecasts from Canada & USA

Generates products for:Weather forecasters

Specialized usersEnd users

Operational outlet for THORPEX research using TIGGE archive

UMAC data call page 4 of 25North American Ensemble Forecast System - NAEFS4

NCEP CMC NAEFSModel GFS GEM NCEP+CMC

Initial uncertainty ETR ETKF ETR + ETKF

Model uncertainty/Stochasti

c

Yes (Stochastic Pert) Yes (multi-physicsand stochastic)

Yes

Tropical storm Relocation None

Daily frequency 00,06,12 and 18UTC 00 and 12UTC 00 and 12UTC

Resolution T254L42 (d0-d8)~55kmT190L42 (d8-16)~70km

About 50kmL72

1*1 degree

Control Yes Yes Yes (2)

Ensemble members 20 for each cycle 20 for each cycle 40 for each cycle

Forecast length 16 days (384 hours) 16 days (384 hours) 16 days

Post-process Bias correction (same bias for all members)

Bias correction for each member

Yes

Last implementation February 14th 2012 November 18th 2014

NAEFS Current StatusUpdated: November 18th 2014


NAEFS Milestones• Implementations

– First NAEFS implementation – bias correction Version 1.00 - May 30 2006– NAEFS follow up implementation – CONUS downscaling Version 2.00 - December 4 2007– Alaska implementation – Alaska downscaling Version 3.00 - December 7 2010– Implementation for CONUS/Alaska expansion Version 4.00 - April 8 2014 – Implementation of 2.5km NDGD for CONUS/Alaska Version 5.00 – September 2015

• Applications:– NCEP/GEFS and NAEFS – at NWS– CMC/GEFS and NAEFS – at MSC– FNMOC/GEFS – at NAVY– NCEP/SREF – at NWS

• Publications (or references):– Cui, B., Z. Toth, Y. Zhu, and D. Hou, D. Unger, and S. Beauregard, 2004: “ The Trade-off in Bias Correction between Using the Latest

Analysis/Modeling System with a Short, versus an Older System with a Long Archive” The First THORPEX International Science Symposium. December 6-10, 2004, Montréal, Canada, World Meteorological Organization, P281-284.

– Zhu, Y., and B. Cui, 2006: “GFS bias correction” [Document is available online]– Zhu, Y., B. Cui, and Z. Toth, 2007: “December 2007 upgrade of the NCEP Global Ensemble Forecast System (NAEFS)” [Document is

available online]– Cui, B., Z. Toth, Y. Zhu and D. Hou, 2012: "Bias Correction For Global Ensemble Forecast" Weather and Forecasting, Vol. 27 396-410 – Cui, B., Y. Zhu , Z. Toth and D. Hou, 2015: "Development of Statistical Post-processor for NAEFS" Weather and Forecasting (In process)– Zhu, Y., and B. Cui, 2007: “December 2007 upgrade of the NCEP Global Ensemble Forecast System (NAEFS)” [Document is available

online]– Zhu, Y, and Y. Luo, 2014: “Precipitation Calibration Based on Frequency Matching Method (FMM)” . Weather and Forecasting (in press)– Glahn, B., 2013: “A Comparison of Two Methods of Bias Correcting MOS Temperature and Dewpoint Forecasts” MDL office note, 13-1– Guan, H, B. Cui and Y. Zhu, 2015: “Guan, H., B. Cui and Y. Zhu, 2014: "Improvement of Statistical Post-processing Using GEFS Reforecast

Information", (accepted, in press)

http://www.emc.ncep.noaa.gov/gmb/yzhu/imp/i200711/1-GFS_bc.pdf

http://www.emc.ncep.noaa.gov/gmb/yzhu/imp/i200711/1-GFS_bc.pdf


Bias correction: • Bias corrected NCEP/CMC GEFS and NCEP/GFS forecast (up to 180 hrs)• Combine bias corrected NCEP/GFS and NCEP/GEFS ensemble forecasts• Dual resolution ensemble approach for short lead time• NCEP/GFS has higher weights at short lead time

NAEFS products (global) and downstream applications• Combine NCEP/GEFS (20m) and CMC/GEFS (20m) • Produce Ensemble mean, spread, mode, 10% 50%(median) and 90% probability forecast at 1*1 degree

resolution• Climate anomaly (percentile) forecasts• Wave ensemble forecast system• Hydrological ensemble forecast system

Statistical downscaling • Use RTMA as reference - NDGD resolution (5km/6km), Extended CONUS and Alaska• Generate mean, mode, 10%, 50%(median) and 90% probability forecasts

NAEFS Statistical Post-Processing


NAEFS Bias Correction (Decaying average method)

)()()( 0,,, tatftb jijiji

2). Decaying Average (Kalman Filter method)

)()1()1()( ,,, tbwtBwtB jijiji

1). Bias Estimation:

3). Decaying Weight: w =0.02 in GEFS bias correction (~ past 50-60 days information)

4). Bias corrected forecast:

)()()( ,,, tBtftF jijiji Simple Accumulated Bias

Assumption: Forecast and analysis (or observation) is fully correlated


Variables pgrba_bc file Total 51

GHT 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000hPa 10

TMP 2m, 2mMax, 2mMin, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000hPa

13

UGRD 10m, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000hPa 11

VGRD 10m, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000hPa 11

VVEL 850hPa 1

PRES Surface, PRMSL 2

FLUX (top) ULWRF (toa - OLR) 1

Td and RH 2m 2

Notes CMC did not process bias correction for Td and RH

NAEFS bias corrected variables

Last upgrade: April 8th 2014 - (bias correction)


Continuous Ranked Probabilistic Skill Scores

5-day forecast

10-day forecast

NH 500-hPa Height30-day running mean

GEFSNAEFS


Statistical downscaling for NAEFS forecast

• Proxy for truth– RTMA at 5km resolution– Variables (surface pressure, 2-m temperature, and 10-meter

wind)• Downscaling vector

– Interpolate GDAS analysis to 5km resolution– Compare difference between interpolated GDAS and RTMA– Apply decaying weight to accumulate this difference –

downscaling vector• Downscaled forecast

– Interpolate bias corrected 1*1 degree NAEFS to 5km resolution – Add the downscaling vector to interpolated NAEFS forecast

• Application– Ensemble mean, mode, 10%, 50%(median) and 90% forecasts


11

Variables Domains Resolutions Total 10/10

Surface Pressure CONUS/Alaska 5km/6km 1/1

2-m temperature CONUS/Alaska 5km/6km 1/1

10-m U component CONUS/Alaska 5km/6km 1/1

10-m V component CONUS/Alaska 5km/6km 1/1

2-m maximum T CONUS/Alaska 5km/6km 1/1

2-m minimum T CONUS/Alaska 5km/6km 1/1

10-m wind speed CONUS/Alaska 5km/6km 1/1

10-m wind direction CONUS/Alaska 5km/6km 1/1

2-m dew-point T CONUS/Alaska 5km/6km 1/1

2-m relative humidity CONUS/Alaska 5km/6km 1/1

NAEFS downscaling parameters Last Upgrade: April 8 2014 (NDGD resolution)

All downscaled products are generated from 1*1 degree bias corrected fcst. globally Products include ensemble mean, spread, 10%, 50%, 90% and mode


12hr 2m TemperatureForecast Mean Absolute Error w.r.t RTMA for CONUSAverage for September, 2007

GEFS raw forecast

NAEFS forecast

GEFS bias-corr. & down scaling fcst.


13

NCEP/GEFS raw forecast

NAEFS final products

4+ days gain from NAEFS

From Bias correction (NCEP, CMC)Dual-resolution (NCEP

only)Down-scaling (NCEP,

CMC)Combination of NCEP

and CMC


14

From Bias correction (NCEP, CMC)Dual-resolution (NCEP only)Down-scaling (NCEP, CMC)Combination of NCEP and CMC

NAEFS final products

NCEP/GEFS raw forecast

8+ days gain


Future Plan for NAEFS

• Short term plan– Exchange 0.5degree data every three hours for selected variables out to 7 days

between NCEP, CMC (and FNMOC).– Exchange wave ensemble data between NCEP, CMC (and FNMOC)– NAEFS (SPP) upgrade (v5.0) – Q4FY15

• Improving bias correction process (method)• Adding cloud cover variable for bias correction• The same algorithm for processing ECMWF ensemble for internal use only• Downscaled products from 5km to 2.5km, and extend CONUS domain to cover

most of Canada and Mexico• Downscaled products from 6km to 3km for Alaska

– NAEFS (SPP) upgrade (V6.0) – Q2FY16• Improving bias correction process (method)• Adding additional variables (wind gust and wave height) for bias correction and

downscaling• Adding ECMWF ensembles to NAEFS (for internal use only)• Adding new products (EFI and Anomaly Forecast) for extreme weather• Expand downscaling products for other domains (Hawaii, Guan and Puerto Rico)


Future Plan for NAEFS

• Long term plan– Collaborate with CMC and FNMOC (and MDL)– NAEFS welcomes other international center(s) to join (and/or

contribute)– Exchange extend forecasts (out to 30 days) between NCEP, CMC

(and FNMOC).– Exchange GEFS reforecast between NCEP, CMC (and FNMOC)– Extend downscaling products to cover whole North American– Improve NAEFS SPP – Develop new guidance for extreme weather events those include

weather and week 2, 3 & 4


Bias correction for

each ensemble member

+

High resolution

deterministic forecast

Mixed Multi- Model

Ensembles(MMME)

Probabilistic products at 1*1 (and/or) .5*.5

degree globally

Down-scaling (based on RTMA)

Probabilistic products at NSGD resolution(e.g. 2.5km – CONUS)

NCEP

Others

CMC

Reforecast

RBMPFor

blenderVaried

decaying weights

Auto-adjustment

of 2nd moment

Smartinitializatio

n

Future NAEFS Statistical Post-Processing


National Unified Operational Prediction Capability

• NUOPC (National Unified Operational Prediction Capability) is an agreement to coordinate activities between the Department of Commerce (National Oceanic and Atmospheric Administration) and the Department of Defense (Oceanographer and Navigator of the Navy and Air Force Directorate of Weather), in order to accelerate the transition of new technology, eliminate unnecessary duplication, and achieve a superior National global prediction capability.

• The NUOPC partners determined that the Nation’s global atmospheric modeling capability can be advanced more effectively and efficiently with their mutual cooperation to provide a common infrastructure to perform and support their individual missions.

• The NUOPC Tri-Agency (NOAA, Navy, Air Force) agreed to work on a collaborative vision through coordinated research, transition and operations in order to develop and implement the next-generation National Operational Global Ensemble modeling system. This NUOPC plan consists of the following elements:


NCEP CMC FNMOCModel GFS GEM Global Spectrum

Initial uncertainty ETR EnKF (9) Banded ET

Model uncertainty Stochastic

Yes (STTP) Yes (multi-physics and Stochastic)

None

Tropical storm Relocation None None

Daily frequency 00,06,12 and 18UTC 00 and 12UTC 00 and 12UTC

Resolution T254L42 (d0-d8)~55kmT190L42 (d8-16)~70km

~ 50kmL72

T239L50 ~ 55km

Control Yes Yes Yes

Ensemble members 20 for each cycle 20 for each cycle 20 for each cycle

Forecast length 16 days (384 hours) 16 days (384 hours) 16 days (384 hours)

Post-process Bias correction for ensemble mean

Bias correction for each member

Bias correction for member mean

Last implementation February 14 2012 November 18 2014 May 21 2014

NUOPC Current StatusUpdated: May 21 2014


NUOPC (FNMOC) Grid Exchange VariablesUpdate: May 21 2014

21

Variables Pgrba file Total 80/73

GHT Surface, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000 hPa 11/(11)

TMP 2m, 2mMax, 2mMin, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000 hPa 13/(13)

RH 2m, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000 hPa 11/(11)

UGRD 10m, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000 hPa 11/(11)

VGRD 10m, 10, 50, 100, 200, 250, 500, 700, 850, 925, 1000 hPa 11/(11)

PRES Surface, PRMSL 2/(2)

PRCP APCP, CRAIN, CSNOW, CFRZR, CICEP 5/(4)

FLUX (surface) LHTFL, SHTFL, DSWRF, DLWRF, USWRF, ULWRF 6/(2)

FLUX (top) ULWRF (OLR) 1/(1)

PWAT Total precipitable water at atmospheric column 1/(1)

TCDC Total cloud cover at atmospheric column 1/(1)

CAPE Convective available potential energy, Convective Inhibition 2/(2)

SOIL/SNOW SOILW(0-10cm) , TMP(0-10cm down), WEASD(water equiv. of accum. Snow depth), SNOD(surface)

4/(1)

Other 850 hPa vertical velocity 1/(1)

Notes Original NAEFS grids currently being sent to NCEP by FNMOC, Require model change to add. (future plan)Not available

FNMOC=72


National Unified Operational Prediction Capability

• NAVGEN ensemble (FNMOC) data at NCEP– Twice per day (00UTC, 12UTC)– Every 6 hours, out to 16 days– 20+1 ensemble members for each lead time– 1*1 degree raw ensembles (72 variables) forecasts

• Post process – bias correction– NCEP runs bias correction for NAVGEN ensembles– The same algorithm as NCEP and CMC

• Public data access– NCEP ftp – real time– NCEP NOMADS – real time– NCDC NOMADS – real time?


Northern Hemisphere 500hPa height (against consensus analysis):

Latest 3-month winter scores: CRP scoreRMS error and ratio of RMS error / spreadAnomaly correlation

NH CRPS skill scores

NH anomaly correlation

500hPa Height

NH RMS errors


Ratio of RMS error over spread

Northern Hemisphere 500hPa height (against consensus analysis):

30-day running mean scores of day-5: CRP score RMS error and ratio of RMS error / spread Anomaly correlation

NH 500hPa anomaly correlation

5-day forecast

NH 500hPa CRP scores

Under-dispersion

Over-dispersion

NH 500hPa RMS errors


The results demonstrate:1. BMA could improve 3 ensemble’s mean, but

spread could be over if original spread is larger2. RBMP could keep similar BMA average future, but

2nd moment will be adjusted internally3. All important time average quantities are decaying

average (or recursive – save storage)

NUOPCIBC – simple combine three bias corrected ensembles

DCBMA – decaying based Bayesian Model Average

RBMP – Recursive Bayesian Model Process (built in decaying average and internal 2nd-moment adjustment)

Solid line – RMS errorDash line - Spread

Over-dispersion

Multi-modelPost-processing

UMAC data callpage 1 of 25North American Ensemble Forecast System - NAEFS EMC Operational Models...

Documents

Transcript of UMAC data callpage 1 of 25North American Ensemble Forecast System - NAEFS EMC Operational Models...