Our approach, activities and achievements in climate researchfrancesco/IRI_REVIEW... · 2013. 7. 1. · Mid-Feb 2012 Plume of Model ENSO Predictions OBS FORECAST IRI/CPC Historically

!e International Research Institutefor Climate and Society

Our approach, activities and achievements

in climate research

Andrew W. Robertson and IRI Climate Group


Approach

• Use-inspired climate forecasting and climate science research

• Collaborative climate and interdisciplinary research

• Participatory capacity building research


What we mean by “Use-inspired climate forecasting and climate science

research”

➡ R&D needed to produce salient & credible climate information for better real-world management of climate related risks/climate change adaptation

(counter examples: deterministic seasonal forecasts and raw downscaled CMIP projections)


Examples


Operational Products:Net Assessment & ENSO Outlook

• Multi-institutional 2-tier seasonal forecasting system since late 1990s

• Currently 2 AGCMs run in house, 4 outside, plus CFSv2

• Parametric multimodel combination/calibration

• Large set of additional GCM products run each month for research & partner use


Seasonal forecast “Net Assessments”

Net assessments were visited 52,000 times in 2012 - “Global Public Good”

Is this the extent of reliable seasonal forecast information?


Forecasting the full distribution:“Flexible” Forecasts Maprooms

STD RAINFALL

FRE

QU

EN

CY


30

12

24

12

24

24

10

24

10

FORECAST SST SCENARIOS

TROP. PACIFIC: THREE

(multi-models, dynamical and statistical)

TROP. ATL and INDIAN

(2 and 3 multi-models) EXTRATROPICAL

(damped persistence)

GLOBAL ATMOSPHERIC

MODELS

COAPS(FSU)

ECHAM4.5(MPI)

CCM3.6(NCAR)

GMAO(NASA)

COLA2

GFDL

ForecastSST

Ensembles3/6 Mo. lead

PersistedSST

Ensembles3 Mo. lead

IRI DYNAMICAL CLIMATE FORECAST SYSTEM

POSTPROCESSING

MULTIMODELENSEMBLING

PERSISTED GLOBAL

SSTANOMALY

2-tiered OCEAN ATMOSPHERE

30

modelweighting



ENSO QUICK LOOK February 16, 2012 A monthly summary of the status of El Nino,La Nina and the Southern Oscillation, or “ENSO”, based on NINO3.4 index (120-170W, 5S-5N)A majority of the ENSO prediction models call for weak La Nina conditions during the February-March pe-riod, transitioning to neutral conditions during the March-May period with the most likely time of dissipationoccurring in early April.

Official Early-Feb CPC/IRI Consensus Forecast1

JFM FMA MAM AMJ MJJ JJA JAS ASO SON0

10

20

30

40

50

60

70

80

90

100

Time Period

Prob

abili

ty (%

)ENSO state based on NINO3.4 SST Anomaly

Neutral ENSO: −0.45oC to 0.45oC

2012 2012

Mid-Feb IRI/CPC Plume-Based Forecast2

FMA MAM AMJ MJJ JJA JAS ASO SON OND0

10

20

30

40

50

60

70

80

90

100

Time Period

ENSO state based on NINO3.4 SST AnomalyNeutral ENSO: −0.45oC to 0.45oC

2012 2012

El NiñoNeutralLa Niña

Historical Sea Surface Temperature Index

1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012−3

−2

−1

0

1

2

3

←NINO

3.4

Inde

x (o C)

Time Period

NDJ Jan JFM FMA MAM AMJ MJJ JJA JAS ASO SON OND-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Nino

3.4

SST

Anom

aly

(OC)

DYN AVGSTAT AVGCPC CON

Dynamical Model:

Statistical Model:

NCEP CFSv2 NASA GMAO NCEP CFS JMA SCRIPPS LDEO AUS/POAMA ECMWF UKMO KMA SNU ESSIC ICM ECHAM/MOM COLA ANOM MetFRANCE COLA CCSM3 CS-IRI-MM GFDL CM2.1 CMC CANSIP

CPC MRKOV CDC LIM CPC CA CPC CCA CSU CLIPR UBC NNET FSU REGR UCLA-TCD

2011 2012

Mid-Feb 2012 Plume of Model ENSO Predictions

OBS FORECAST

IRI/CPC

Historically Speaking

El Nino and La Nina events tend to develop during the period Apr-Jun and they:- Tend to reach their maximum strength during Dec-Feb- Typically persist for 9-12 months, though occasionally persisting for up to 2 years- Typically recur every 2 to 7 years

1Official: Based on a consensus of CPC and IRI forecasters, in association with CPC/IRI ENSO Diagnostic Discussion.2Unofficial: Purely objective, based on regression, using equally weighted model predictions from the plume.

ENSO pages were visited 99,000 times

in 2012


New IRI/CPC Joint ENSO Forecasts

IRI recently began working with CPC to issue a joint ENSO forecast.

Benefits to CPC

• Expanded forecast lead time

• Co-branding on IRI/CPC ENSO plume

• Having a partner in drafting the monthly ENSO

Diagnostic Discussion

Areas of Future Collaboration

• Developing and testing better ENSO SST

thresholds• Creating a more flexible

ENSO climatology• Investigating the utility of

other ENSO indices besides the 3-month SST average in

the NINO3.4 region


GCM Products made at IRI


Forecast Development+

Diagnostics & Modeling

• New US NMME and coupled model collaborations

• Forecast calibration/MME methodologies

• Verification metrics development

• Downscaling and forecast tailoring


NMME Data Services


Coupled GCMs in IRIDLReal-time + Hindcasts

• NMME (CFSv2, CCSM3, GFDL, NSIPP, Env-Canada)

• UKMO-GloSea5

• POAMA (Australian BoM)

• ECMWF (SST)


120 1255

10

15

(a) r(GCM − MOS)

0

0.2

0.4

0.6

0.8

120 1255

10

15

(b) r(RCM − MOS)

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.80

20

40

60

80

100

Correlation threshold

% o

f sta

tions

> th

resh

old

(c) Correlation values

GCMRCM

Figure 7: Anomaly correlation coe�cient between AMJ seasonal average PAGASA rainfall

data and MOS corrected simulations of (a) GCM, (b) RCM simulated precipitation (only

positive values are plotted). Panel (c) shows the percentage of stations exceeding a given

correlation value, with the 95% one-tailed statistical significance value indicated as vertical

line.35

120 1255

10

15

(a) r(GCM, Obs)

0

0.2

0.4

0.6

0.8

120 1255

10

15

(b) r(RCM, Obs)

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.80

20

40

60

80

100

Correlation threshold

% o

f sta

tions

> th

resh

old

(c) Correlation values

GCMRCM

Figure 5: Anomaly correlation coe�cient between AMJ seasonal average PAGASA rainfall

data and (a) GCM, (b) RCM simulated precipitation interpolated to station locations (only

positive values are plotted). Panel (c) shows the percentage of stations exceeding a given

correlation value, with the 95% one-tailed statistical significance value indicated as vertical

line.

33

Dynamical (RCM, 50km) Statistical (MOS)

Downscaling: Dynamical or Statistical?

Anomaly correlation skill for April–June season

Robertson et al. (2013, MWR)GCM T42 simulations w/obs SST, 1977–2004


Predictability of “Weather within Climate”

seasonal totalrainfall

frequency

Choose the analysis to perform: PCR or CCA

SELECTING THE ANALYSIS

y = Ax + b

JJAS rainfall correlation skill (ECHAM4-CA: made from June 1)


1900 1910 1920 1930 1940 1950 1960 1970ï3

ï2

ï1

0

1

2

3

4

5Seasonal Amount

1900 1910 1920 1930 1940 1950 1960 1970ï3

ï2

ï1

0

1

2

3

4Frequency of Occurrence

1900 1910 1920 1930 1940 1950 1960 1970ï4

ï2

0

2

4

6Daily Mean Intensity

Why is frequency more predictable than intensity?

1900 1910 1920 1930 1940 1950 1960 1970ï3

ï2

ï1

0

1

2

3

4

5Seasonal Amount

1900 1910 1920 1930 1940 1950 1960 1970ï3

ï2

ï1

0

1

2

3

4Frequency of Occurrence

1900 1910 1920 1930 1940 1950 1960 1970ï4

ï2

0

2

4

6Daily Mean Intensity

Rainfall at individual stations and station-average (NW India)

rainfall frequency mean intensity on wet days

Rainfall occurrence frequency is more spatially coherent than intensity.

Moron et al. (2007, JCL)


Figure 15. Cross validation correlation skills in rain-season rice production of all ecosystem 0

at regional level (a) and provincial level (b). WWV plus zonal wind anomalies over the west equatorial Pacific(1980-2007) in DJF were used as predictors of MLR.

43 |

KoideMonday, 1:14 AMAdded Text

KoideJul 3, 10:03 AMDeleted: 6

Figure 10. Cross validation correlation skills in dry-season rice production of all ecosystem 0

at regional level(a) and provincial level (b). Seasonal precipitation anomalies in OND over 0N-25N, 110E-130E forecasted with ECHAM4.5-MOM (1981-2007) on 1st June were used as predictors of CCA. Provinces without rice production data are shaded.

38 |

KoideThursday, 11:05 PMAdded Text

KoideThursday, 11:05 PMDeleted: 4 4

KoideJul 3, 10:03 AMFormatted: 0 inch Left Indent

Jan–Jun (Dry Season)

from prev. Jun 1

Predictability of Philippines Rice ProductionBased on CGCM hindcasts

Jul–Dec (Rainy Season)

from prev. Mar 1

ACC Skill of (a) Regional & (b) Provincial Production

1980–2007

Koide et al. (2013, JAMC)



y = Ax + b

!e International Research Institutefor Climate and Society!e International Research Institute

for Climate and Society

AGRHYMET Training (2012) on predictability of agroclimate and hydrological quantities

(incl. monsoon onset dates & river discharge)

SST Onset Date

Map of covariability between May SST and Rainfall Onset Dates



y = Ax + b


Timescales decomposition

Greene, Goddard & Cousin (2010, EOS)


Research: Models say that IF the subtropical North Atlantic warms up more than the global tropics, then the

Sahel could get wetter…

PastGreen: end 20th century – pre-Industrial

Blue: end – beginning 20th century

Future:Yellow: mid-21st (A1B) – end 20th

Red: end 21st – end 20th

A. Giannini

!e International Research Institutefor Climate and Society (Greene, et al. 2012, WRR)

Stochastic Simulation Framework


Funding and research partners

• Federal science agencies – individual PI’d grants (NOAA, NASA, NSF, DOE, ONR)

• Cooperative Agreements with NOAA and USAID

• Climate service/development projects


Linkages• CLIVAR/US-CLIVAR

– Scientific Steering Group, Co-Chair (L. Goddard)– Process Studies and Model Improvement panel (US) (A. Giannini)– Predictability, Prediction and Applications Interface panel (US)(A. Barnston)– Variability of the African Climate System (S. Mason)

• WCRP– Working Group on Regional Climate (S. Mason)– Working Group on Seasonal to Interannual Prediction (A. Robertson)– WCRP-WWRP/THORPEX Subseasonal to Seasonal Prediction Project planning group, Co-Chair (A. Robertson)

• WMO– Expert Team on Climate Services Information System (CCl), Chair (S. Mason)– Joint Working Group on Forecast Verification Research (S. Mason)

• NOAA Modeling, Analysis, Prediction & Projection (MAPP) Task Force– L. Goddard, A. Barnston, B. Lyon, S. Mason, M. Tippett,


Red Cross - IRI example

Opportunity to use information on multiple time scales


What’s next?

• New flexible seasonal forecast MME Net Assessment based on partnerships with climate modeling centers in US and internationally

• Pan-timescale climate forecast information from sub-seasonal to decadal

• Robust in-house climate research (forecasting, diagnostic, modeling) in collaboration with LDEO, CU, NOAA, universities, regional centers, NMHSs ...

• Tailored product development, esp. using IRIDL


Thank You!

Documents

Our approach, activities and achievements in climate researchfrancesco/IRI_REVIEW... · 2013. 7. 1. · Mid-Feb 2012 Plume of Model ENSO Predictions OBS FORECAST IRI/CPC Historically