Tropical Cyclones in a regional climate change...

Tropical Cyclones in a regional climate change projections with RegCM4 over Central America CORDEX domain

Tropical Cyclones in a regional climate changeprojections with RegCM4 over Central America

CORDEX domain

Gulilat Tefera [email protected]

Centre ESCER, University of Quebec at Montreal (UQAM), Montreal, Canada

Thanks toF. Giorgi, R. Fuentes-Franco, K. Walsh, G. Giuliani, E. Coppola

—May 2014


Introduction

Motivation

IntroductionI Tropical Cyclones (TCs) are one of the high impact extreme

events affecting many regions.

I Q. how do their characteristics (intensity, frequency, duration)change under greenhouse gas warming?

I Several studies with GCM w.r.t climate projections tend toagree on the characteristics of future TC activity on globalscale.

I The change in specific ocean basins are however moreuncertain, for example in Atlantic basin some found adecrease (e.g. Bengtsson et al., 2007) and others found theopposite (e.g. Oouchi et al., 2006), or found no significantchange or the result depends on the model configuration (e.g.,Murakami et al., 2010)

I This implying the need for further regional investigations


Introduction

Model set up

RegCM (4.3) set upI The model domain is following the CORDEX Central America

specification(15W - 145W and 25S-42N)

I Horizontal resolution is 50km, and vertically 18 σ levels

I Perfect Boundary Condition integration :

I ERA-Interim for Initial and LBCI BATS for land surface processesI Convection:Emanuel over ocean and Grell over land

I The scenario integrations: 4 ensemble runsI Two model configurations each driven by two CMIP5 GCMs

I The two configurations are (Grell-BATS) and (Eman-CLM)I The two CMIP5 GCMs are HadGEM2-ES and MPI-ESMI simulation period is 1970 - 2100 with GHG forcing from the

RCP8.5 pathway


Introduction

Model set up

Data and analysis periodI Observed TCs are obtained from the Unisys Web site

(http://weather.unisys.com/hurricane/)/

I The dataset consists of six hourly best-track data

I TCs with intensity corresponding surface wind of 17.5ms−1 orgreater are considered in the analysis

I Validation is performed for the ERA-Interim driven and for theGCM control (present day) runs.

I The analysis is carried out for 22 years of the present period(1982 - 2003)

I The future period considered is the last 22 years of the 21century


Introduction

Model set up

Some features of TCs


Introduction

Model set up

TC detection methodThe detection criteria are similar to Walsh et al., (2004) and are asfollows:

I The relative vorticity at 850hPa must be greater than 1.0x10−5 s−1

I There must be a closed pressure minimum within the radius of100km. This minimum pressure is then defined as the centerof the cyclone.

I The minimum surface pressure satisfying the above criteria isat least 2 hPa lower than the averaged surface pressure overthe surrounding 2X2 grid boxes.

I The 10m surface wind speed must exceed 17.5m s−1

I The minimum pressure at the center of the cyclone at thegenesis must be over ocean points where the SST is higherthan 260C .


Introduction

Model set up

TC tracking procedure

Tracking is performed on those grid points satisfying the detectioncriteria. The following procedure is used to create a trackingdatabase

1. For each detected cyclone point, a check is performed on eachsix hourly sample to see whether there are cyclones during the24 hour period within a radius of 6 X 6 longitude-latitude gridboxes.

2. If there are some cyclones detected satisfying the previouscondition, then the closest cyclone is chosen as belonging tothe same track as the initial one.

3. To qualify as a track, there must be more than one detectionwithin the next 24 hours and within a 6 X 6 longitude-latitudegrid boxes.


present day TC characteristics

from ’Perfect’ boundary run

ERA-Interim forced runs




Frequency and track

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks density from HURR dataset [1982:2003]

0

5

10

15

20

25

30

35

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks density from RegCM ERA−Eman run [1982:2003]

0

5

10

15

20

25

30

35

The model underestimate the TC frequency over Pacific but

overestimates over Atlantic

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks from HURR dataset [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks from ERA−Interim runs [1982:2003]

HURR data RegCM-ERA Interim




Inter-annual variability:

1980 1985 1990 1995 2000−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5

2

2.5

Year

Standa

rdized

anoma

lyJASO Atlantic cyclone inter−annual variability

r = 0.57

HURRRegCM

−ERA

1980 1985 1990 1995 2000

−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5

2

2.5

Year

Standa

rdized

anoma

ly

JASO East Pacific cyclone inter−annual variability

r = 0.27

HURRRegCM

−ERA

RegCM4 realistically simulated the interannual variability TCs over North

Atlantic




Annual cycle

2 4 6 8 10 120

20

40

60

80

100

120

140

160

Month

Num

ber o

f TC

grid

s/ye

arTC frequency over Central America (140W −20W and 0N − 40N)

ObservedRegCM

−ERAint

RegCM4 realistically simulated the annual cycle of TCs




Intensity

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitud

e

JASO Cyclone windspeed from HURR dataset [1982:2003]

16

18

20

22

24

26

28

30

32

34

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitud

e

Maximum wind speed from RegCM−ERAInterim run [1982:2003]

16

18

20

22

24

26

28

30

32

34

50km is still coarse to simulate intense cyclones



present day TC characteristics from GCMs

Present day GCM forced runs




Impact of model physics on TC climatology

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks density from RegCM HadGEM−Eman run [1982:2003]

0

5

10

15

20

25

30

35

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

JASO Cyclone tracks density from RegCM MPI−Eman run [1982:2003]

0

5

10

15

20

25

30

35

Grell produces less TCs irrespective of the boundary forcing

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks density from RegCM HadGEM−Grell run [1982:2003]

0

5

10

15

20

25

30

35

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

JASO Cyclone tracks density from RegCM MPI−Grell run [1982:2003]

0

5

10

15

20

25

30

35

Emanuel Emanuel

Grell Grell

HadGEM forced run MPI forced run




Why HadGEM produced less TCs over tropical Atlantic

Longitude

Latit

ude

HadGEM mean JASO temperature [C] for [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

10

15

20

25

30

Longitude

Latit

ude

MPI mean JASO temperature [C] for [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

10

15

20

25

30

Longitude

Latit

ude

HadISST mean JASO temperature [C] for [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

10

15

20

25

30

HadGEM SST is colder over the tropical Atlantic


Future changes: Frequency, Intensity and Tracks and Lifecycle

Future Changes



Future Changes for JASO: TC Frequency

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

Future (RCP85) − Present for RegCM HadGEM−Eman run

−10

−8

−6

−4

−2

0

2

4

6

8

10

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

Future (RCP85) − Present JASO from RegCM MPI−Eman run

−10

−8

−6

−4

−2

0

2

4

6

8

10

A decrease in TC frequency over the tropical Atlantic

Longitude

Latitu

de

Future − Present HadGEM mean JASO wind shear

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

−10

−8

−6

−4

−2

0

2

4

6

8

10

Longitude

Latitu

de

Future − Present MPI mean JASO wind shear

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

−10

−8

−6

−4

−2

0

2

4

6

8

10

These might be related to the increase in the wind shear over the tropical Atlantic



Frequency vs stability (Temperature profile)

Longitude

Pressu

reMPI mean JASO temperature [C] profile RCP85 − Present

−80 −70 −60 −50 −40 −30

100

200

300

400

500

600

700

800

900

10002

2.5

3

3.5

4

4.5

5

5.5

6

Atlantic

Longitude

Pressu

re

MPI mean JASO temperature [C] profile RCP85 − Present

−130 −120 −110 −100

100

200

300

400

500

600

700

800

900

1000 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

East Pacific

more stable atmosphere in the future climate



Future Changes: Tracks and life cycle

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

JASO Cyclone tracks from MPI−Eman runs [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

JASO Cyclone tracks from MPI−Eman RCP85 runs [2078:2099]

2 4 6 8 10 12 140

200

400

600

800

1000

Days

Numb

er of

even

ts

Histogram of cyclones life cycle MPI−Eman runs

PresentFuture

The frequency of TCs particularly over tropical Atlantic decreases,

however those will get reduced in number are the short lived ones



changes in Annual cycle

Annual Cycle: Atlantic vs East Pacific

2 4 6 8 10 120

20

40

60

80

100

120

140

160

Month

Num

ber o

f TC

grids

/year

TC frequency over North Atlantic

ObservedRegCM

−ERAint

MPI−Eman

HadGEM−Eman

MPI−Eman−RCP85

HadGEM−Eman−RCP85

2 4 6 8 10 120

20

40

60

80

100

120

140

160

Month

Num

ber o

f TC

grids

/year

TC frequency over East Pacific

ObservedRegCM

−ERAint

MPI−Eman

HadGEM−Eman

MPI−Eman−RCP85


I MPI run suggests an increase in the frequency over Pacific, whereas almost the opposite in HadGEM run

I Over the Atlantic basin, HadGEM2 shows a shift in the TCs season whereas MPI favours the reduction

2 4 6 8 10 120

20

40

60

80

100

120

140

160

180

Month

Num

ber o

f TC

grid

s/ye

ar

TC frequency over Central America (140W −20W and 0N − 40N)

ObservedRegCM

−ERAint

MPI−Eman

HadGEM−Eman

MPI−Eman−RCP85




changes in Annual cycle

Future Changes: Intensity

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Future − Present maximum wind speed of cyclones from RegCM HadGEM−Eman run

−5

−4

−3

−2

−1

0

1

2

3

4

5

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

Future − Present maximum wind speed from RegCM MPI−Eman run

−5

−4

−3

−2

−1

0

1

2

3

4

5

East Pacific- the highest change in the wind intensity

15 20 25 30 35 400

50

100

150

200

250

300

Wind speed ms−1

Num

ber o

f TCs

Histogram of max. wind speed for HadGEM2 − Eman

PresentFuture

15 20 25 30 35 400

100

200

300

400

500

600

700

Wind speed ms−1

Num

ber o

f TCs

Histogram of max. wind speed MPI−Eman

PresentFuture

A decrease in the weak TCs but an increase in the frequency of the

strongest TCs


Sensitivity to the removal of SST threshold from the detection criteria

The impact of using SST threshold in the detectionprocedure on the differences in tropical cyclone formationbetween current and future climate.

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latitu

de

JASO Cyclone tracks from MPI−Eman runs [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

LongitudeLa

titude

JASO Cyclone tracks from MPI−Eman RCP85 runs [2078:2099]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks from MPI−Eman WOSST−WTMPF [1982:2003]

−140 −120 −100 −80 −60 −40 −200

5

10

15

20

25

30

35

40

Longitude

Latit

ude

JASO Cyclone tracks from MPI−Eman RCP85 WOSST−WTMPF [2078:2099]



Impact of SST threshold on Intensity

15 20 25 30 35 400

100

200

300

400

500

600

700

Wind speed ms−1

Numb

er of TC

s

Histogram of max. wind speed MPI−Eman

PresentFuture

15 20 25 30 35 400

200

400

600

800

1000

1200

Wind speed ms−1

Numb

er of e

vnts

Histogram of max. wind speed MPI−Eman WOSST−WTMPF

PresentFuture



Impact of SST threshold on life cycle

2 4 6 8 10 12 140

200

400

600

800

1000

Days

Numb

er of e

vents

Histogram of cyclones life cycle MPI−Eman runs

PresentFuture

2 4 6 8 10 12 140

200

400

600

800

1000

1200

Days

Numb

er of e

vents

Cyclones life cycle MPI−Eman WOSST−WTMPF runs

PresentFuture


Conclusions

ConclusionsI TC climatology is found to be very sensitive to the convection

scheme, and Grell scheme produces very few TC counts.

I HadGEM2-ES forced RegCM4 produces unrealistic TC counts overthe tropical Atlantic because of a cold SST bias there.

I The MPI-ESM forced RegCM-Emanuel configuration reproducedwell the TC climatology for the present climate and proven to bethe best combination.

I Results under the RCP8.5 scenario suggest that the futurefrequency of the most intense TCs will increase but the number ofweek and short lived TCs will decrease.

I The overall frequency of TCs decreases in the tropical Atlantic andcoastal regions of the Pacific but increases over Central Pacific andnorthern Atlantic.

I Removing the SST threshold from the detection criterion does notchange the result substantially.

Tropical Cyclones in a regional climate change...

Documents

Transcript of Tropical Cyclones in a regional climate change...