Intensification of Summer Rainfall Variability in the

1
Intensification of Summer Rainfall Variability in
the Southeastern United States in Recent Decades
Hui Wang1,2, Wenhong Li1, and Rong Fu1,3 1Georgia
Institute of Technology 2NOAA Climate Prediction
Center 3University of Texas at Austin
NOAAs 33rd Climate Diagnostics and Prediction
Workshop and CLIVAR Drought Workshop October
2024, 2008, Lincoln, Nebraska
2

• Highlights
• Intensified SE rainfall variability in recent 3
  decades.
• Consistent with the shift of rainfall PDF.
• Closely tired to the Atlantic SST variability.
• Certain predictability based on tropical SST.

3
Background Summer droughts repeatedly struck the
Southeast in recent decades.
2007 Southeast drought Worst in 100 years
Lake Lanier main water source for Atlanta

• Questions
• Has the rainfall variability been intensified in
  recent decades?
• What might cause the intensification?

4
Previous works

• Stahle Cleaveland 1992 decade-long spring
  rainfall extremes have been a prominent feature
  of SE US over the past 1000 yrs.
• Mo Schemn 2008 model simulations show
  seasonally varying ENSO forcing on SE US
• Seager et al. 2008 model simulations show weak
  connection with SSTA, thus low predictability

5
Data Precipitation CPC U.S. Unified
Precipitation (19481998, 0.25o x 0.25o) CPC
realtime U.S. Daily Precipitation Analysis
(19992007) SST NOAA Extended Reconstructed
SST (ERSST v3, 2o x 2o) Wind and height
NCEP/NCAR Reanalysis (850 and 200 hPa, 2.5o x
2.5o) Period JJA 19482007, 60 years
6
JJA Rainfall Variability
Part 1
Southeast Large mean rainfall High variability
Data 19842007 CPC
7
Change in SE Rainfall Variability
Normalized JJA Precipitation Anomaly
1 stdv0.64 mm/day
Year
1st 30 years 19481977 2nd 30 years 19782007
Wet and Dry (gt 1 Stdv) 4 11
Contribution to Precipitation Variance 32 68
8
Standardized Precipitation Index
SE SPI (3-month)
SE SPI (3-month)
4 11
1 3
Moderately to extremely wet/dry
Severely to extremely wet/dry
2nd 30 yrs
1st 30 yrs
SE SPI (9-month)
SE SPI (9-month)
Moderately to extremely wet/dry
5 14
1 7
Severely to extremely wet/dry
9
Shift of Rainfall PDF
2nd 30 years Decrease of frequency and
intensity in dry summers. Increase of frequency
and intensity in wet summers.
60 yrs 1st 30 yrs 2nd 30 yrs Climatology Wet
(5) Wet (5) Dry (5) Dry (5)
2nd 1st clm 1st 2nd
Frequency
2nd 1st clm 1st 2nd
2nd 1st clm 1st 2nd
5
1
8
12
15
11
Bootstrap test
Difference at 11 significance level
Difference at 5 significance level
Intensity
10
Warm Season Rainfall
Midwest floods
Southeast and Midwest out of phase
11
Composite Drought-related Large-scale
Circulation Regression vs. SE precip index JJA
19482007
Contour u200 Shading z200
LLJ
Shading prcp Vector wind 850 Contour div 925
12
Change in Zonal Wind Variability at the Jet-Steam
Level
1st 30 yrs
2nd 30 yrs
13
Co-variability of SE Rainfall with
SST Homogeneous correlation map
Part 2
ENSO
PCF SST SVD Mode 1
ATL Zonal Mode
ATL SST SVD Mode 2
ATL SST SVD Mode 1
Warming Trend
14
200-hPa Zonal Wind Variance 1st 30 yrs vs. 2nd
30 yrs
1st 30yrs
2nd 30yrs
ENSO SST ATL Zonal Mode Warming
Trend
15
SVD Time Series
ENSO
R 0.63
Variance
ENSO 1st 30yrs 2nd 30yrs
SST 51 49
US P 35 65
SST US precip
ATL zonal Mode
R 0.58
Variance
Zonal mode 1st 30yrs 2nd 30yrs
SST 39 61
US P 37 63
Warming
R 0.60
Variance
Warm-ing 1st 30yrs 2nd 30yrs
SST 35 65
US P 39 61
16
Reconstruction of SE Precipitation Index
Total (R0.92) ATL ZM (R0.87) Warming
(0.63) ENSO (R0.39)
Correlation with OBS
Multilinear regression based on SVD precipitation
time series.
ENSO dominant
17
Part 3
Empirical Forecast System
1. Project SST (OBS or FCST) onto the SVD SST
pattern SST projection
coefficient 2. Based on SVD SST US
precipitation relationship, precipitation
projection coefficient 3. Based on SVD
precipitation pattern,
forecast of precipitation anomaly
Cross-Validation Anomaly Correlation Hindcast
vs. OBS
2 ATL modes
ENSO
Hindcasts JJA 19482007 rainfall based on
observed SST
18
Forecast of Precipitation Anomaly for JJA
2008 Based on Observed SST
Forecast
Observation
Dry
Dry
Wet
Wet
19

• Summary
• SE summer precipitation exhibits higher
  interannual variability in recent 30 years.
• The shift of rainfall variability is linked to
  tropical Atlantic SST variability, suggesting
  certain predictability of SE summer rainfall
  based on tropical SST.