Was the 2015 subpolar North Atlantic cold blob predictable?

Elizabeth Maroon, Steve Yeager, Gokhan Danabasoglu

NCAR, CGD, Oceanography Section

NCAR is sponsored by the

National Science Foundation

Anomalously cold subpolar North Atlantic SST in 2015

Was the cold blob predicted by the CESM Decadal Prediction Large Ensemble (DPLE)?

CESM DPLE (Yeager et al. 2018):

• 10-year-long, 40-member hindcastensembles initialized each November from 1954-2016

• Full field initial conditions from a Forced Ocean Sea Ice (FOSI) simulation driven by CORE* forcing

Using the DPLE hindcast ensemble initialized in November 2014 to assess if the peak cold anomaly in summer 2015 was predictable

First month of integration looks good…

2014 Nov subpolar SST anomaly (K)

None of the ensemble members capture the magnitude of the cold anomaly by summer 2015

Observed and DPLE SST trend in opposite directions in 2015

1. Does CESM capture subpolar anomalies of the same magnitude as observations?

2. Was this a rare event? Do we need more than 40 ensemble members?

Does the CESM Large Ensemble produce subpolar cold blobs of similar magnitude? Yes.

CESM Large

Ensemble (Kay et al.

2015) ensemble mean

(1920-2018)

Observed distribution

(1920-2018)

Spread of internal

variability from all

ensemble members

None of the DPLE members reproduce the observed persistently positive NAO in winter 2015

NA

O in

dex

time (monthly means)

Should we have expected 6 consecutive months of positive NAO conditions in any of the 40 ensemble members?

NAO index

frequency

NAO index

frequency

# of consecutive months with NAO >+1

# o

f events

Should we have expected 6 consecutive months of positive NAO conditions in any of the 40 ensemble members?

Add 20 more ensemble members? 40 ensemble members

NA

O index

# o

f D

PLE

mem

bers

subpolar North Atlantic SST anomaly (K)

JJA 2015

Add 20 more ensemble members?60 ensemble members

No new ensemble members capture cold blob or persistently

positive NAO behaviorN

AO

index

# o

f D

PLE

mem

bers

subpolar North Atlantic SST anomaly (K)

JJA 2015

Additional ensemble members from Nan Rosenbloom

• None of the CESM DPLE ensemble members initialized in 11/2014 capture the 2015 subpolar cold anomalies

• Highly positive NAO conditions contributed to the formation of the 2015 cold blob

• This period also had a once-in-the-historic record 6-consecutive months of NAO positive conditions

• Adding more 11/2014 DPLE ensemble members did not produce a cold blob or persistently positive NAO conditions

Acknowledgments: NSF OPP #1737377, OCE #1243015

Summary

ERA-Interim DPLE

No

ve

mb

er

De

ce

mb

er

Ja

nu

ary

Fe

bru

ary

Ma

rch

Surface heat flux

Blue: heat loss by ocean

Heat flux out of the subpolar ocean during December, January, and March was critical for the amplification of the cold blob (Duchez et al. 2016)

No

ve

mb

er

De

ce

mb

er

Ja

nu

ary

Fe

bru

ary

Ma

rch

ERA-Interim

DPLE ensemble mean

Surface heat flux

Blue: heat loss by ocean

No

ve

mb

er

De

ce

mb

er

Ja

nu

ary

Fe

bru

ary

Ma

rch

ERA-Interim

Was upper ocean heat content better predicted? No.

upper 295m ocean heat content

Both surface heat fluxes and advection contribute to negative heat tendency in early 2015

ocean heating

ocean cooling

FOSI upper 295m ocean heat budget terms

FOSI tendency has two negative anomalies in January and March 2015

upper 295m ocean heat tendency

ocean heating

ocean cooling

No DPLE members replicate the FOSI’s surface heat flux and advection anomalies in January and March.

tendency due to surface heat fluxes

tendency due to advection

Winter 2015