Transcript Climate Extremes Index (CEI) Case Study: Africa

Lezlie C. Moriniere, ATMO529 (Fall07)
Arid Land Resources Sciences / Global Change
Focus: Climate ChangeXevents Human Migration
Presentation Outline
 Scientific Motivation
 Introduction
 Terms
 IPCC
 Dataset
 Analysis Methods
 NOAA Standard
 Results
 Summary and Steps Ahead
Scientific Motivation: Climate Refugees?
1. CREATE
EXTREMES INDEX
USING CLIMATE
VARIABLES
3.ALIGN WITH
MIGRATION
STATISTICS TO
DETECT TRENDS
Global
Change
2. EXTRACT DATA ON
CLIMATE-RELATED
DISASTER EVENTS
Human
Migration
Natural Hazards
 Extreme
Events
What is an Extreme Event?
 “an event that is rare within the statistical reference
distribution at a particular place”
(IPCC, 2001)
 Rare: x ≤ 10th percentile or x ≥ 90th percentile
 4 attributes: rate of exceedence, mean excess, volatility,
clustering in time (Stephenson, 2002)
 Measures: scale parameter (β), percentile thresholds
empirical ranking
 Comfort in the Means vs. Intrigue in the Extremes
IPCC EXTREME EVENT Prediction:
FAR
Direction and
Phenomenon
TEMPERATURE
Aspect
Intensity
and Frequency
PRECIPITATION
Intensity and
Frequency
DROUGHT
Area
CYCLONES
Intensity
SEA LEVEL RISE
Frequency
Likelihood
↑
↑
↑
↑
↑
>90% probability
> 90 to 99% probability
66 to 90% probability
66 to 90% probability
66 to 90% probability
CRU TS 2.1
Global Climate Database
 East Anglia University’s Climate Research Unit (CRU):
Michael, T.D. and Jones, P.D., 2005. An improved method of
constructing a database of monthly climate observations and
associated high resolution grids. Int.J. Climatology 25: 693-712.
 Reformatted for ARCInfo:
CGIAR (Consultative Group for Intl. Agricultural Research),
Consortium for Spatial Information)
 Gridded 0.5°x0.5°, 11042 grids (Africa)
 9 climate variables (Tmx, Tmn, Precip,Wet, Tmp, Dtr,
Frs,Vap Cld): 102 years, monthly, 1901-2002
 Software used:
 Analysis/Figures: MatLab
 Map: ESRI ArcGIS
Analysis Methods
 Precipitation: Beta
 Produce SPI for Continent of Africa
 Local Significance
 Composites
 Climate Extreme Index
 2-tailed Exceedence per Variable
 Calculate Index
 Composites
 Local Significance
 All: Temporal Trends Spatial Trends
Climate Extremes Index (CEI)
NOAA (Policy)
U of A (Research Application)
 coterminous USA, 1910-present
 Africa + islands, 1901-2002
 Seasonal/Annual, 1° x 1 ° Grids
 Monthly/Seas./Ann., 0.5° x 0.5 ° Grids
 Arithmetic average of 6 indicators:
 Arithmetic average of 5 indicators:
PERCENTAGE of AREA
EXCEEDENCE :
1. ∑ (Max.Temperature HI ,
Max.Temperature LO )
2. ∑ (Min.Temperature HI ,
Min.Temperature LO )
3. ∑ ( PDSI HI , PDSI LO )
4. 2 * ( 1-day Precipitation HI )
FREQUENCY of TEMPORAL
EXCEEDENCE:
1. ∑ (Max.Temperature HI ,
Max.Temperature LO )
2. ∑ (Min.Temperature HI ,
Min.Temperature LO )
3. ∑ ( SPI HI , SPI LO )
4. ∑ ( Precipitation HI ,
Precipitation/Wetdays HI )
5. ∑( WetDaysHI , DryDays HI )
6. ∑ (Wind velocities^2)
5.
6.
∑( WetDays HI DryDays HI )
∑ (Wind velocities^2)
Step 1: Max. Monthly Temp
Africa_CEI
Tmx_m
5.1 - 24
25 - 28
29 - 31
32 - 34
35 - 38
Step 2: Min. Monthly Temp
Step3: SPI (Drought and Moisture)
Severe Sahelian droughts 
• 1910-1914
•Mid1 970s
•Mid 1980s
Africa SPI
Africa_CEI
SPISUMm_HI
-1.9 - -0.69
-0.68 - -0.28
-0.27 - 0.067
0.068 - 0.48
0.49 - 1.4
Africa_CEI
SPI_m
Africa_CEI
SPI_m-1.2 - -0.51
-0.5
-0.12
-1.2 -- -0.51
-0.11
0.13
-0.5 - --0.12
0.14
-0.11 -- 0.4
0.13
Africa_CEI
SPISUMm_LO
-2.8 - -1.1
0.14 -- 0.4
0.41
0.92
0.41 - 0.92
102 Year Monthly SPI
-1 - -0.54
-0.53 - -0.099
-0.098 - 0.35
0.36 - 1.3
Step4: Precipitation & Intensity
Precip:
Winter Beta
Precip:
Summer Beta
Step5: Wet/Dry Days
Africa_CEI
Africa_CEI
Wet_aa
Wet_aa 0.29 - 31
0.2932- 31
- 67
32 -68
67- 110
68 -120
110- 150
120160
- 150
- 320
160 - 320
Africa_CEI
Africa_CEI
Dry_aa
Dry_aa
44 - 210
44 - 210
220 - 260
220 - 260
- 300
270270
- 300
- 330
310310
- 330
340340
- 360
- 360
CEI
 Composite:
 High >21%
1967,1968,1970,1974,
1975,1976,1995
 Low <18%
1925,1927,1940,1943,1
944,1948
CEI: Century and Seasonal Means
CEI: Winter
CEI: Spring
Africa_CEI
CEI CEI_m
cru_af_grd1.img
0 - 0.12
CEI_m
0 - 0.117
0.13 - 0.18
0.117 - 0.184
0.19 - 0.26
0.184 - 0.2608
0.27
0.2608 - 0.3706
- 0.37
0.3706 - 1.6
0.38 - 1.6
CEI: Summer
CEI: Fall
Results: Local Significance
 CEI: are the High and Low
Composite Years
significantly different?
 HyØ Winter/Summer:
Rejected, Yes
 2 tailed Ttest, P Values:
Summer: =0.022-0.026
Winter: =0.046-0.052
 Area >90 ci: Summer : 9
grids, Winter: 16 grids
 What contributes most
to the CEI?
 SPI
 HyØ Winter/Summer:
Rejected, Yes
 2 tailed Ttest, P Values:
Summer: =0.11-0.14
Winter: =-0.18—0.15
 Area >90 ci: Summer : 20
grids, Winter: 4 grids
 Max Temp:
 HyØ Winter/Summer:
Rejected, No
SPI:
Winter
-1.4 - -0.55
-0.54 - -0.21
Difference
-0.2 - 0.049
Hi-Lo/2
0.05 - 0.33
SPI:
Summer
-1.1 - -0.35
Difference
-0.34 - -0.047
-0.046 - 0.23
Hi-Lo/2
Africa_CEI
Africa_CEI
SPIwindif
SPIsumdif
0.34 - 1.5
0.24 - 0.56
0.57 - 1.9
Ttest results
 SPI Winter
Africa_tt.img
SPIT_win
<-.90
-0.9 to 0.9
>0.90
Summary
 SPI and other variables complement each other
 Different perspectives on extremes
 Africa Xtremes beg confirmation and monitoring
 Details are lost:
 Africa: huge and heterogeneous
 Many confounding factors and widely varying climatic
influences on the continent:
 Hadley Cell Circulation
 Mid-latitude Circulation
 Ever-mobile ITCZ
 El Nino Southern Oscillation, and NAO
Steps Ahead
 Master statistics specifically for extremes…
 Data acquisition: cyclone, 2002+
 Spatial disaggregating (latitude or country )
 Field Significance
 Global
 Triangulation:
 Disaster Events (lag time?)
 Human Migration