Sixth ICTP Workshop on the Theory and Use of Regional Climate Models

07- 18 MAY 2012

Trend and frequency of drought over Ethiopia using observational and model driven indices

T Zeleke1~2 F Giorgi2, G Mengistu1 GT Diro2

1 Department of Physics, Addis Ababa University, Ethiopia 2 Earth System Physics Section, The Abdus Salam ICTP, Trieste, Italy

STEP PhD student (AAU and ICTP)

Outline

Introduction Objective

Data and methods Results

Conclusion

Future work

Introduction •!•Drought is commonly defined as a deficiency of precipitation from expected climatological mean that extends over a season or longer period of time (Palmer 1965)

•!• Drought is one of the major weather related disasters and recent events over East Africa have demonstrated the continuing exposure to this natural hazard

•!• The impact of drought depends on the severity, duration and spatial extent of the rainfall deficit

•The aim of this work is: • To quantifies the severity, duration and the spatial extent of drought over this region

•Possible association with large scale phenomenon (SST, global circulation) and local forcing (land use change)

Data and methods

~GPCP and simulated RegCM4 rainfall are used to calculate standardized precipitation index (SPI) and Palmer drought severity index (PDSI)

~Trend empirical orthogonal function (TEOF), is employed to extract the dominant long term component (trend) and its spatial pattern in the raw SPI and PDSI data.

~Wavelet, and regression methods are applied to the dominant trend principal component (TPC) to analyze the periodicity and its trend

~Extreme drought percentile per sub period is also used to consolidate the wavelet analysis.

~Correlation and wavelet coherence (WTC) methods are also applied to see the possible linkage of TPC-SPI and SST.

~In addition to TEOF, homogenous regions, time series are used to see the characteristics of drought (Just to solidify the above methods)

Results The 1st Dominant Pattern of TEOF which explain the southern regions, exhibits significant dry trend over the southern regions After 1997/98 the southern regions seems to be drought area, even if 2006/07 showed small recovery from existing condition Frequent and intense ! Model driven PDSI high correlation with observation

-m 15 (I)

~

~ 10 .a ~ 5

a) PDSI (Observation) b) PDSI (RCM)

0.55

15 0.35

0.15 10

-0.05 5

-0.25 35 40 45 35 40 45

Longitude (deg) Longitude (deg)

c) PDSI of TPC-1, its wavelet and trend, corr with rem - 0.82 1 ~~----~------~------~------~------~------~~5

2

3

cn- 4 :a-C'IJ 5 ~ -a 6 0 7

~ a 9

10 .... . . .. ... . ......... -- ----- --- ----- --- ---- --- ----- ---

1

-1

-3

11 ~------~------~------~------~------~------~~ -5

83 87 91 95

Time

99 03 07

0.55

0.35

0.15

-0.05

-0.25

1

0.8

0.6

0.4

0.2

>- The same result is obtained using SPI >- 6-7 period signals are high significant power >- The pattern and time component are reproduced by the model

a) SPI-1 2 (GPCP)

35 40 4 5

Longitude (deg)

0 .45

0 .25

0.05

-0 .15

-0.35

-0.55

b) SPI-1 2 (RCM)

35 4 0 45

Longitude (deg)

c) SPI-1 2 of TPC-1, its wavelet and trend, corr with rem - 0.62 1 ~~--~------~------~------~------~------~~ 5

2 ········ ·· · ·· · · ~·······················:··········· · ···· · ······<············· · ·· ······ ·L······················'···················· ·· · 3 ..... . . .

Ci) 4 :ro.. cu 5 ~ - 6 "C 0 7 "i:

&. 8 9

1 0 ..................... ·! ... .................... : ......... ·-·· ..... .... . ; .. .... .... ........ ..... : ... .... .... .... ..... .. ·' ... ..... .... ..... .... .. ~ ....... .. 11 ~----~------~------~------~------~------~~-5

83 87 9 1 95

Time

99 03 07

0 .45

0 .25

0.05

-0.15

-0.35

-0 .55

1

0 .8

0 .6

0 .4

0 .2

The 2nd Dominant Pattern of TEOF explain the northern regions, which superimposed very low frequency signals and increasing dry trend up to "'1998 /decreasing dry trend since "'1998 "'1982/83, "'1990, "'1994/95, "'1997, "'2002/03 and "'2008 drought years for northern regions. Whereas "'1985, "'1988, "'1991/92, "'2003 and "'2005 relatively wet years.

a) SPI-1 2 (G PCP) b) SPI-1 2 (RCM)

0 . 15 0 .15

-0.05 -0.05

-0.25 -0.25

-0.45 -0.45

-0.65 -0.65 35 40 45 35 40 45

Longitude (deg) Longitude (deg)

c) SPI-1 2 of TPC-2, its wavelet and trend, corr with rem - 0_83

1 5 1

2

3 3 0.8

cn- 4 "'-cu 5 1 0.6 !: 6

"'C 0 7 -1 0.4 ·;:: Q)

8 c.. 9 -3 0.2

10

11 -5 83 87 91 95 99 03 07

Time

a) 1983·1989 (PDSI) b) 1990·19911 (PDSI) a) 1983·1989 (SPI) b) 1990·1996 (SPI)

35 40 45 35 40 45 35 40 45 35 40 45

P =(number drought months/total months)*1 00 c) 1997·2003 (PDSI)

35 40 45

Longitude (deg)

SPI1 2 (1 983·1 989)

100 90 80 70 60 50 40 30 20 10 0

d) 2004·2009 (PDSI)

35 40 45

Longitude (deg)

SPI1 2 (1 990·1 996)

49 ~s 35 40 45 Mean of droughtoidaices

SPI1 2 (1 997 ·2003)

35 40 45 Longitude (deg)

0 .9

0.7

0.5

0.3

0.1

·0.1

·0.3

·0.5

·0.7

·0.9

SPI12 (2004·2009)

35 40 45 Longitude (deg)

c) 1997-2003 (SPI) d) 2004-2009 (SPI)

100 90

o; 15 so .. 70 ~ GO {l 10 50 :J 40 '.i

5 30 ... 20 10

0 35 40 45 35 40 45

Longitude (deg) Longitude (deg)

•During the 1st two heptads drought occurred more frequent and intense over the northern regions. •For the last two heptads the southern regions .. drought area whereas northern recover from drought

• Number of sever and extreme sever drought occurred years percentile and mean value of time series of drought indices indicate the same results

•Southern regions significant dry trend northern regions dominated by low frequency signals

CRU

35 40 45

L ongitude (deg)

0.035 0 .025 0 .0 15 0 .005 -0 .005 -0.0 15 -0.025 -0 .035 -0 .045 -0 .055

T emperature T PC

RCM

35 40 45

L ongitude (d eg)

0 .035 0 .025 0 .0 15 0 .005 -0 .005 -0.0 15 -0.025 -0.035 -0 .045 -0 .0 55

1983 1985 1987 1989 199 1 1993 1995 1997 1999 200 1 2003 2 005 2007 2009

Year

Homogeneous regions, gauge stations and altitude (m)

34 36 38 40 42 44 4 6 48 50

Longitude (deg)

~ .. ., Z: "C 0

'1: ., Q.

tii "' :!!. ., ... 3 ;;; ...J

0 Q.

1-

a) GPCP b) RegCM

15

-o.oo::; -o.oo::; -0 .015 -0.015 -0.025 -0.025 -O.Q35 -0.035 -0.045 -0.045

35 40 45 35 40 45

Longitude (deg) Longitude (deg)

Trend principal component of first dominant mode

-40L-----~--------~----------~--------~----------~------~

1985

reg - 1

87 91 95 99 03 07

reg - 8

87 91 95 99 03 07 Time (year)

1990 1995

Time

2000

reg- 3

87 91 95 99 03 07

reg -12

87 91 95 99 03 07 Time (year)

2005

Is there any association with global SST ??

•The 1st TPC and global SST •The equatorial, northern Atlantic and ENSO events are correlated to the drought over the southern Ethiopia •Same result also found carr of homg-regions with SST.

Trend EOF of SPI-1 2 corr with SST (annual) 0.45

80 ~ ~--· ......... ··- -~ o · -

0.35

60

0 .25

40

0.15

20

0 .05

~ 0 ::::J -~ ___.

-20

-0.05

-0.15

-40

- 0 .25

-60

-0.35

-80

-0.45 -150 -100 -50 0 50 100 150

Longitude (deg)

·The 2nd dominant TPC (explain northern regions), more correlated with northern and central Atlantic and the north west Indian ocean . •

SPI-1 2 TPC2 corr with SST (annual)

0 .75

8 0 0.65

0 .55

0 .45

0.3 5

0.25

0. 15

0.05

-0.05

-0. 15

-0.25

-4 0 -0.35

-0.45 -60

-0.55

-80 ... ... ..... , .... ..... ... ..... ... ... ; .. ..... ... ..... .... .... , .... .... .... ..... ... ... ; .. .... ....... .. ... ... . :· ····· ··· ····· ··· ····· :·· ····· ········ ···· ···: ····· ··· ··· -0.65

-0.75 - 150 - 100 -50 0 5 0 100 15 0

Longitude (deg)

• Have common high power, which is ~2 and ~4-6 years band and coherent in the period from --1997-2001. • Similar results are found using Palmer drought severity index (PDSI). •Drought in/out phase with the east/west for 4 years period signals.

VVavele1: coherence o1' East: Paci'fic SST and TPC-SPI

1

0.9

0.8

0.7

0.6

0 .5

0.4

0 .3

0.2

0 . 1

1985 1990 1995 2000 2005

Time (year)

~ CG <b ~ -.:::::J <> -ii

0....

~ CG

~ -.:::::J <> -ii

0....

- ,, i ~ 10

~ "

0.25

0.5

1

2

4

a

a3

0.25

0.5

1

2

4

a

a3

Obs (1990)

3 5 40 45

Obs (2007)

3 5 40 4 5

Longtlude(deg)

a7

a7

Indian and Atlantic

91 95 99 03 07

Atlantic and ENSC>

91 95 99 03 07

ACM(1990)

0,,

0

-0.~

- I

35 40 4 5

RCM(2 007)

'·" '"tit 10

" 0 ,,

0

-0.5 35 40 4 5

Longltude(d e g)

1

o.a

0.6

0.4

0 . 2

1

o.a

0 .6

0.4

0.2

0.25

E <> = ~ I--en en

0.5

1

2

4

a a3

3

2

1

0

-1

- '" i ~ 10

~ 5

Indian and E-pac

a7 91

35 40 4 5

Obs(1991!1)

35 40 45

Longtlude(de g)

95 99 03

3

2

07

ACM(1997)

35 4 0 45

RCM(1 991!J)

3 5 40 45

Longltude(deg)

1

o.a

0.6

0.4

0.2

•Not only selected ocean basin like SIOD, Atlantic, Indian etc may have significant relation with drought So, association with drought require further work.

Summery •!•During the recent decade, there has been an increase in frequency and intensity of drought over the southern regions of Ethiopia

•!• This drought can be due to both local and remote forcing.

•!•ENSO and Equatorial Atlantic SST anomalies seem to be the candidates for the remote forcing

•!•EI Nino and +IOD mostly(1982, 1994, 1997, 2006) occurred at the same year but 2007 La Nino occurred with +IOD (wet over Ethiopia)

Future work

We will examine the projection of drought using different RegCMx output Scenarios.

Thank you!

Thank very much again all ESP group Specially Filippo and Gulilat