Sub-Seasonal Predictions at NCEP/CPC · 2015. 11. 27. · associated with the slowly evolving parts...

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 1/25

Sub-Seasonal Predictions at NCEP/CPC

Arun Kumar

Climate Prediction Center

arun.kumar@noaa.gov

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 2/25

• Climate Prediction Center (CPC) “…delivers real-time

products and information that predict and describe

climate variations on timescales from weeks to years

thereby promoting effective management of climate risk

and a climate-resilient society”

• Operational responsibilities

– Real time climate monitoring

– Extended-range climate outlooks

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 3/25

• Operational climate outlook products for

surface temperature and precipitation

– Week1 (6-10 day average) : Daily

– Week 2 (8-14 day average) : Daily

– Monthly mean : Twice a month

– Seasonal mean : Once a month

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 4/25

Week 2 (8-14 day average) Outlooks

Surface Temperature Precipitation

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 5/25

Seasonal Outlooks

Surface Temperature Precipitation

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 6/25 6

Forecast Uncertainty

Minutes

Hours

Days

1 Week

2 Week

Months

Seasons

Years

Fore

cast

Lea

d Ti

me

Warnings & Alert Coordination

Watches

Forecasts

Threats Assessments

Guidance

Outlook

Benefits

NWS Seamless Suite of Forecast Products

NCEP CPC

No official NWS outlooks for the Week 3-4 time period Gap in services for a period when climate related information could help decision makers

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 7/25

Additional Background

• Initiation of an activity targeting the Week 3-4 time period

was made a major goal in the updated CPC 5-year

strategic plan based on discussion at CPC, stakeholder

feedback, etc.

• NWS and NOAA leadership as well as the Office of

Science Technology Policy (OSTP) at the White House

urged making development in this area a high priority

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 8/25

Challenges in Filling the Forecast Gap

• Declining influence from atmospheric initial conditions

(die off curves)

• Time average is short to benefit from the signal

associated with the slowly evolving parts of the climate

system (e.g., SSTs)

• Consequently, the Week 3-4 time range are likely to

suffer from low predictability and prediction skill

• Important to understand this limitation and manage

expectations

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 9/25

Annually averaged RPSS for week1 – week4 surface temperature forecast (94-06) Weigel et. al. (2008), MWR

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 10/25

RPSS for week1 – week4 surface temperature forecast (94-06) Weigel et. al. (2008), MWR

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 11/25

Skill of monthly mean for different lead time Kumar et al. (2011), Climate Dynamics

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 12/25

Sources for Predictability

• Madden-Julian Oscillation (MJO)

• Modes of extratropical patterns of variability

(PNA, NAO, Blocking)

• Stratosphere – Troposphere teleconnection

• Soil moisture; Snow

• ENSO; Local/coastal SST anomalies

• Climate Trends

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 13/25

Development of Weeks 3-4 Outlook at CPC

• Forecast tools

– Empirical

– Dynamical

• Consolidation of various forecast guidance

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 14/25

Development of Weeks 3-4 Outlook at CPC

• Empirical tools utilizing historical observations

– MJO-ENSO phase model: Current strength and

phase of ENSO and MJO as well as trends (Johnson

et al., 2014, Weather and Forecasting)

– Constructed analogue tool: Matching analogues of

200-mb streamfunction of the past to the current

conditions. Past cases are objectively weighted.

– Coupled Linear Inverse Model (C-LIM)

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 15/25

MJO-ENSO Phase Model Retrospective Skill

Weeks 3+4 Heidke Skill Score from combined effects of ENSO+MJO+Trend Johnson et al., 2014, Weather and Forecasting

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 16/25

Constructed Analog Retrospective Skill

Temperature Precipitation

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 17/25

Development of Weeks 3-4 Outlook at CPC

• Dynamical models

– Forecasts from CFSv2, ECMWF, JMA

– Model data is bias corrected and calibrated

based on available reforecasts

– Plans in FY16 to include Environment Canada

and NCEP GEFS

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 18/25

CFS Retrospective Forecast Skill

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 19/25

CFS Retrospective Forecast Skill

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 20/25

Experimental Week 3 & 4 Product

• The experimental product is 2-class (above or below-average)

temperature and precipitation outlook maps for the favored category

of two-week mean temperature and two-week total accumulated

precipitation

• The target is a combined two week outlook for Weeks 3-4 in the

future

• Outlook maps depict probabilities for the favored category

• Released every Friday

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 21/25

Orange: Above average temperatures favored Blue: Below average temperatures favored Equal Chances (EC): Equal odds for above/below

Green: Above average precipitation favored Brown: Below average precipitation favored Equal Chances (EC): Equal odds for above/below

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 22/25

A: 50% B: 50%

A: 30% B: 70%

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 23/25

Comments

• Need to have a better coordination for operational forecasts from

different centers (e.g., scheduling; hindcast period etc.)

• Projects like S2S and NMME will help provide better estimates for

average predictability, and manage user expectations

• Expectation is that skill is going to be low (in extratropics, noise for

shorter time averages is high compared to the predictable signal

from various sources)

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 24/25

1976

1981

1995

2007

2015

ECMWF Workshop on Sub-Seasonal Predictability –November 2015 25/25

Thanks!