Regional Applications of the NCEP Coupled Forecast System (CFS)

(1)

DBCP Capacity Building workshop for the Western Indian Ocean, Cape town, 19-23 April 2010

WMO IOC

1

Regional Applications of the NCEP Coupled Forecast System (CFS)

Dr. Wassila M. Thiaw

Team Lead African and International Desks

Climate Prediction Center

National Centers for Environmental Predictions

National Oceanic and Atmospheric Administration

(2)

WMO IOC

2

Acknowledgments

• Vadlamani Kumar and Nick Novella, CPC International Team

• EMC Ocean Modeling Team, Dave Behringer, Lead

• CPC Ocean Monitoring Team, Yan Xue, Lead

• NCEP CFS Team, Suranjana Saha and Hua Lu Pan

(3)

WMO IOC

3

Climate Prediction Center Mission

• National temperature and precipitation National temperature and precipitation outlooks, but not monthly / seasonal outlooks, but not monthly / seasonal rankings

rankings

• Focus: weeks, months, seasons, years Focus: weeks, months, seasons, years (i.e.

(i.e. short term climate short term climate ) )

• Forecasts in collaboration with other NCEP Forecasts in collaboration with other NCEP Centers, NOAA line offices, other agencies Centers, NOAA line offices, other agencies and labs

and labs

• Integral to NWS Seamless Suite of Integral to NWS Seamless Suite of Products

Products

We deliver climate prediction, monitoring, and assessment products for timescales from weeks to years to the Nation and the global community for the protection of life and property and

the enhancement of the economy.

Temperature Outlook

(4)

WMO IOC

4

Training Coverage

Climate Prediction Center

International Desks

African Desk, International Projects

Week1 Precipitation Probability of exceedance

• Support capacity building in weather and climate in developing countries: Africa, central America, and Asia

• Provide access to NCEP global data and products to domestic and international agencies and to the public

Global TC monitoring New Website

Africa Hazards

(5)

WMO IOC

5

CPC International Desks

• Deliver climate information and services

• Transfer technology to National Meteorological and Hydrological Services and regional institutions

• Build network of well-trained international scientists

• Provide data to support NCEP’s mission

(6)

WMO IOC

6

Outline

• NCEP Climate Forecast System (CFS)

• CPC Prediction and Monitoring Tools

• CFS Reanalysis and Reforecast Project

• Climate Dynamics of southern and eastern Africa

(7)

WMO IOC

7

The NCEP CFS Components

CFS Operational Implementation : 24 Aug 2004

• T62/64-layer version of the 2003 operational NCEP atmospheric GFS (Global Forecast System) model

• GFDL MOM-3 (Modular Ocean Model, version 3)

– 40 levels

– 1 degree resolution, 1/3 degree on equator

• Reference: Saha et al. (2006), Journal of

Climate, 19, 3483-3517.

(8)

WMO IOC

8

Data Assimilation Systems

• Atmosphere (Global Data Assimilation System – GDAS)

• Ocean (Global Ocean Data Assimilation System – GODAS).

• Models incorporate observations in a

dynamically consistent fashion to produce

analyses of initial conditions, which provide

starting point for making predictions of future

conditions.

(9)

WMO IOC

9

Skill in SST Anomaly Predictions

Predictions archived since 1997. The CFS shows a substantial increase in skill

over the previous NCEP operational forecast model (CMP14) and comparable

skill to operational statistical forecast models (CCA, CA, CONS, and MARKOV).

(10)

WMO IOC

10

SST Biases for DJF and JAS

DJF Climatology JAS Climatology

Lead0 - Obs

Lead3 - Obs

Lead6 - Obs

(11)

WMO IOC

11

SST Biases for DJF and JAS

DJF JAS

Lead0 - Obs

Lead3 - Obs

Lead6 - Obs

(12)

WMO IOC

12

Skill for Raw and Bias-corrected Niño 3.4 SST Predictions

Bias correction

improves the skill for all prediction

methods (compare

bottom panels with

top panels).

(13)

WMO IOC

13

Precipitation Biases for DJF and JAS

DJF Climo

JAS Climo

Lead0 - Obs

Lead3 - Obs

Lead6 - Obs

(14)

WMO IOC

14

OBS DJF Clim. 1981-2004 CFS DJF Clim. 1981-2004

(15)

WMO IOC

15

CFS Wind Clim. 1981-2004 CDAS Wind Clim. 1981-2004

Atmospheric Circulation

(16)

WMO IOC

16

CFS_2-tier DJF Precipitation Climatology 1981-2004

OBS

CFS

(17)

WMO IOC

17

(18)

WMO IOC

18

Precipitation Interannual Variability

RFE Oct 1 2005 – May 31 2006 RFE Oct 1 2006 – May 31 2007

(19)

WMO IOC

19

Time – Latitude 5-day RFE 30-40E

2005-06 2006-07

(20)

WMO IOC

20

Time – Latitude 5-day CFS 30-40E

2005-06 2006-07

(21)

WMO IOC

21

3 Time – Latitude 5-day P Difference 30-40E

RFE 2006-07 minus 2005-06 RFE 2006-07 minus 2005-06

(22)

WMO IOC

22

Central Africa Seasonal Rainfall Predictability

(23)

WMO IOC

23

(24)

WMO IOC

24

Summary of CFS Version 2004

• CFS skill in predicting Niño 3.4 SST anomalies is greater than the previous NCEP operational 2-tier

model (CMP14), and comparable to NCEP operational statistical models.

• For southern Africa, precipitation biases appear to be related to errors in low level wind fields and

atmospheric convection. Nonetheless, there is some predictability in the system for the austral summer.

• Seasonal rainfall predictability can be significantly enhanced in the areas where the ENSO signal is

strong such as portions of central Africa and eastern

Africa by downscaling CFS forecasts.

(25)

WMO IOC

25

CFS SST Forecasts

March 2010 initial conditions

The CFS ensemble mean (heavy blue line) predicts neutral conditions to return by late spring-early summer followed by a

development of La Nina that is expected to

last through fall 2010.

(26)

WMO IOC

26

IRI Forecasts for Niño 3.4 (March 2010)

Most models suggest a gradual return to ENSO- neutral conditions by May-July 2010, which persist into the fall.

Some models suggest the

potential of continued El

Niño conditions or the

development of La Niña

conditions during the late

summer or fall.

(27)

WMO IOC

27

Prediction, Assessment, and Monitoring of the climate system at CPC

• Global and Regional Climate Analysis ^{(R1, R2,}

NARR, CFSR)

• Real-time Global Climate Predictions and Monitoring

• Forecast Tool Development/ Improvement

• Verifications and Expert Assessments

• Climate Diagnostics and Attribution

• Applied Research

• Outreach and Service

• International Activities

(28)

WMO IOC

28

Predictions and Monitoring of the Global Ocean

• Weekly and monthly briefings

– State of Global Oceans

– Variations in tropical eastern Pacific--El Niño & La Niña

– Global Climate Impacts of Sea Surface Temperatures

(29)

WMO IOC

29

Global SST Anomaly (

⁰

C) and Anomaly Tendency

- El Nino condition (NINO 3.4 > 0.5^oC) weakened slightly in the tropical Pacific;

- PDO was in positive phase (slide 19);

- SST was above-normal in the tropical Indian Ocean and tropical North and South Atlantic;

- Large SST anomalies in the subtropical South Pacific, and North Atlantic

- SST decreased slightly in the eastern tropical Pacific;

- SST tendency was large in the Southern Ocean

- SST increased in tropical North and South Atlantic.

-SST tendency was small in the Indian Ocean

Fig. G1. Sea surface temperature anomalies (top) and anomaly tendency (bottom). Data are derived from the NCEP OI SST analysis, and anomalies are departures from the 1971-2000 base period means.

(30)

WMO IOC

30

Evolution of Indian Ocean SST Indices

Fig. I1a. Indian Ocean Dipole region indices, calculated as the area-averaged monthly mean sea surface

temperature anomalies (^OC) for the SETIO [90ºE-110ºE, 10ºS-0] and WTIO [50ºE-70ºE, 10ºS-10ºN] regions, and Dipole Mode Index, defined as differences between WTIO and SETIO. Data are derived from the NCEP OI SST analysis, and anomalies are departures from the 1971-2000 base period means.

- Positive eastern (SETIO) pole SSTA persisted in Mar 2010.

- Both eastern (SETIO) and western (WTIO) pole SST have been persistently above-normal since April 09.

- DMI became above-normal in Mar 2010.

(31)

WMO IOC

31

- SSTA exceeding +1C presented in the tropical Indian Ocean.

- Net surface heat flux anomalies contributed to the positive (negative) SSTA tendency in the North (South) Indian Ocean.

- Convection was suppressed over the Maritime Continent, and intensified in southeastern Indian Ocean

- Consistent with the suppressed

convection was low-level divergence (up- level convergence) wind anomalies in the Maritime Continent

Tropical Indian: SST Anom., SST Anom.

Tend., OLR, Sfc Rad, Sfc Flx, 925-mb &

200-mb Wind Anom.

Fig. I2. Sea surface temperature (SST) anomalies (top-left), anomaly tendency (top-right), Outgoing Long-wave Radiation (OLR) anomalies (middle-left), sum of net surface short- and long-wave radiation, latent and sensible heat flux anomalies (middle-right), 925-mb wind anomaly vector and its amplitude (bottom-left), 200-mb wind anomaly vector and its amplitude (bottom-right). SST are derived from the NCEP OI SST analysis, OLR from the NOAA 18 AVHRR IR window channel measurements by NESDIS, winds and surface radiation and heat fluxes from the NCEP CDAS. Anomalies are departures from the 1979-1995 base period means except SST anomalies are computed with respect to the 1971-2000 base period means.

(32)

WMO IOC

32

CFS DMI SST Predictions from Different Initial Months

- Latest forecasts called for weak positive Indian Ocean Dipole (IOD) in fall 2010.

DMI = WTIO- SETIO SETIO = SST anomaly in [90ôE-110ôE, 10ôS-0]

WTIO = SST anomaly in [50ôE-70ôE, 10ôS-10ôN]

Fig. M2. CFS Dipole Model Index (DMI) SST predictions from the latest 9 initial months. Displayed are 40 forecast members (brown) made four times per day initialized from the last 10 days of the initial month (labelled as

IC=MonthYear) as well as ensemble mean (blue) and observations (black). The hindcast climatology for 1981-2006 was removed, and replaced by corresponding observation climatology for the same period. Anomalies were

computed with respect to the 1971-2000 base period means.

(33)

WMO IOC

33

(34)

WMO IOC

34

IESA – Integrated Earth System Analysis

(35)

WMO IOC

35

(36)

WMO IOC

36

(37)

WMO IOC

37

(38)

WMO IOC

38

Climate Forecast System V1 vs. V2

CFSv1 CFSv2

Atmosphere GFS 2003 (T62/L64) GFS2009 (T126/L64)

Land ^{OSU 2-L} ^{NOAH 4-L}

Ocean ^MOM3 ^MOM4

Sea ice Climatology Predicted

CO2 Fixed at 1988 level Evolving with time

Initial conditions ^R2/GODAS ^CFSR

Hindcasts ^15/month ^CFSrR

(39)

WMO IOC

39

New Global Ocean Data Assimilation

System

Ocean Model MOMv4 (GFDL)

1/2

^o

x 1/2

^o

z-coord, 40 levels

Data Assimilation 3DVariational Cost Function:

J = ½ (x-x

_b

)

^T

B

^-1

(x-x

_b

) + ½ (y-H(x))

^T

R

^-1

(y-H(x))

Model Data

First

Guess Analysis

Forcing: Wind Stress

Heat Flux

Data:

T(z)&S(z)

SST

SSH

(40)

WMO IOC

40

NEW GODAS

• Based on MOMv4, the new GODAS has a resolution of 1/2

^o

x 1/2

^o

(1/4

^o

in the tropics). It uses a z-coordinate in the vertical and has 40 levels. The domain is global and includes an ice model.

• GODAS has a 3D variational assimilation scheme.

• Assimilation data are temperature profiles (XBT, Argo, TAO, TRITON, PIRATA, RAMA), salinity profiles (Argo, synthetic profiles derived

from a seasonal T-S relation), TOPEX/Jason-1 Altimetry and SST (Reynolds’ daily 1/4

^o

SST).

• GODAS can be forced by surface flux products (wind stress, heat flux

components) in “stand alone” mode or be coupled to an atmospheric

model.

(41)

WMO IOC

41

New CFS at NCEP

GODAS 3DVAR Ocean Model

MOMv4 fully global

1/2

^o

x1/2

^o

(1/4

^o

in tropics) 40 levels

Atmospheric Model GFS (2007)

T382 64 levels

Land Model Ice Mdl SIS LDAS

GDAS GSI

6hr

24h r

6hr

Ice Ext

6hr

Climate Forecast System

(42)

WMO IOC

42

GODAS compared with independent surface drifter velocities

MOMv4 based GODAS 1/2

^o

resolution Global

MOMv3 based GODAS 1

^o

resolution Quasi-global

AOML surface drifter based climatology Independent Lumpkin et al.

GODAS has eastward flow on equator in Indian Ocean

Drifters show stronger flow in western Boundary and

Southern Ocean

The agreement is very good given that GODAS does not directly assimilate velocity

observations and the drifter velocities are derived from the lagrangian motion of the

drifters.

(43)

WMO IOC

43

MOMv4 based GODAS 1/2

^o

resolution Global

MOMv3 based GODAS 1

^o

resolution Quasi-global

AOML surface drifter based SST climatology

Independent data (Lumpkin et al.)

GODAS compared with surface drifter velocities

(44)

WMO IOC

44

CFSR

GODAS

CFSR - GODAS

CFSR SSH has a higher correlation with Altimetry SSH than GODAS SSH in the tropical Indian Ocean and extratropical Oceans, but CFSR has a lower correlation in the tropical Atlantic, which is probably associated with drifts in subsurface

temperature and salinity and erroneous overturning

circulation cells in the tropics

(see next slide).

(45)

WMO IOC

45

Climate Dynamics of Eastern and Southern Africa

• Present the basic atmospheric features of southern and eastern Africa climate

• Place the climate dynamics of the region in a global context as seen from NCEP

reanalysis-1

• See if the NCEP CFS Reanalysis (CFSR)

is able to capture the atmospheric features

(46)

WMO IOC

46

Northern and southern hemisphere seasons DJF, austral summer

JAS, boreal summer

Emphasis on the large-scale flow and precipitation fields

Climate Dynamics of Eastern and

Southern Africa

(47)

WMO IOC

47

Low level, 850 hPa geopotential heights, DJF 1979-2008

Atmospheric Circulation Atlantic-Indian Ocean Basin

Subtropical

highs

(48)

WMO IOC

48

Low level, 850 hPa geopotential heights, DJF 1979-2008

Thermal lows

Atmospheric Circulation

Atlantic-Indian Ocean Basin

(49)

WMO IOC

49

Low level, 850 hPa geop. heights & winds,

DJF 1979-2008

(50)

WMO IOC

50

Low level, 850 hPa geop. heights & winds, DJF 1979-2008

RA-1 CFSR

(51)

WMO IOC

51

Reanalysis-1

Low level, 850 hPa geopotential heights & winds

Southern Hemisphere Summer DJF 1979-2008

Northern Hemisphere Summer JAS 1979-2008

(52)

WMO IOC

52

CFSR

Low level, 850 hPa geopotential heights & winds

Southern Hemisphere Summer DJF 1979-2008

Northern Hemisphere Summer JAS 1979-2008

(53)

WMO IOC

53

What is the relationship between the large-scale

circulation and the precipitation climatology?

(54)

WMO IOC

54

Intertropical Convergence Zone (ITCZ) South Indian Convergence Zone (SICZ)

Austral Summer DJF Precipitation

(55)

WMO IOC

55

Austral Summer DJF Precipitation

(56)

WMO IOC

56

Austral Summer Precipitation Characteristics

• Synoptic systems form along the SICZ, an area favorable for tropical-temperate troughs (TTTs) to develop. These systems produce much of the rainfall over southern Africa in the summer.

• TTTs account for much of the precipitation

variability on the synoptic time scale with

northwest to southeast oriented cloud band

that span southern Africa and the southwest

Indian Ocean

(57)

WMO IOC

57

Boreal summer: July-August-September

(58)

WMO IOC

58

Monsoon Trough

Indian Monsoon

West African Monsoon

Somali Jet

Reanalysis 1

Low level, 925 hPa geop. heights & winds, JAS

1979-2008

(59)

WMO IOC

59

CFSR RA-1

Low level, 925 hPa geop. heights & winds, JAS

1979-2008

(60)

WMO IOC

60

Somali Jet Characteristics

• Also called the East African Jet (EAJ)

• Likely the strongest and most sustained low-level wind patterns in the world (mean maximum wind speeds of 25 m/s – observed up to 50 m/s)

• Low-level wind speed maximum at 1.5 km msl

• Accounts for 65% of global cross equatorial mass flux

• Links African and Asian rainfall systems

(61)

WMO IOC

61

Indian monsoon upper-level high

Tropical

Easterly Jet

Upper level heights and winds 200 hPa JAS

Reanalysis 1

(62)

WMO IOC

62

RA-1 CFSR

Upper level heights and winds 200 hPa JAS

(63)

WMO IOC

63

Northern Summer JAS Precipitation

(64)

WMO IOC

64

Summary Preliminary

Assessment of CFSR over Africa

• For both austral and boreal summers, the subtropical ridges

appear to be shallower in CFSR than in RA1, while the monsoon troughs are stronger

• The upper level jet streak in the Indian monsoon is placed farther north in the CFSR, while the upper level equatorial low is deeper.

• The precipitation positive biases in the Gulf of Guinea region during the boreal summer and in southern Africa during the

austral summer appear to be related to the strong thermal lows and winds. CFSR did not seem to improve the negative

precipitation biases in the Sahel north of the Gulf of Guinea

region.

(65)

WMO IOC

65

CPC Resources

• http://www.cpc.noaa.gov/

• http://www.cpc.noaa.gov/products/african_desk/

• http://www.cpc.noaa.gov/products/analysis_monitoring/

enso_advisory/index.shtml

• http://www.cpc.ncep.noaa.gov/products/GODAS/ocean_br

iefing_gif/global_ocean_monitoring_current.ppt

(66)

WMO IOC

66

NCEP Data Access

Daily RA2 and monthly CFSR

http://nomad1.ncep.noaa.gov/ncep_data/

Daily CFSR

www.ncdc.noaa.gov

(67)

WMO IOC

67

African Desk Training Program

• Sponsored by the US National Weather Service as part of the contribution to the WMO Voluntary Cooperation Program (VCP)

• African Desk established in 1994 and hands-on-raining began in 1995.

Since then the African Desk has hosted more than 90 residents from over 30 countries in Africa

• Objective: Build human resource capacity in weather and climate at NMHSs in Africa

• Professional meteorologists with at least 4-year university degree and 3 years professional experience are eligible to apply. Candidates must be nominated by the WMO PR of the country of origin

• African Desk offers 12 slots per year: six in each weather and climate

desks. Trainees arrive at staggered intervals of 2 months. The duration

of the training is 4 months

(68)

WMO IOC

68

African Desk Training Curriculum

• Climate Desk

– Modes of variability: ENSO and MJO – Seasonal rainfall predictions

– Week1&2 outlooks – Climate diagnostics

– Climate monitoring and hazard assessments – Satellite rainfall estimates

• Weather Desk

– Use of ensembles in operational weather forecasting – Model performance evaluation and forecast verifications – Case studies

– Regional modeling: NWS Weather Research Forecast (WRF)

(69)

WMO IOC

69

Future Plan

• Expand CPC International Desks

– International Monsoon Desk – Begin with China and expand to include SE Asia later

– Off-site 1-2 week trainings in all ocean basins:

Indian Ocean, Vietnam, June 2009; Mediterranean, Turkey, July 2010; Atlantic, summer 2011; Pacific, summer 2012

• Challenges

– National resources

– Implementation plan

(70)

WMO IOC

70