Forecasting the Onset of El Nio Based on GPCP Precipitation:

1
Forecasting the Onset of El Niño Based on GPCP
Precipitation Validation and Future Work Scott
Curtis Joint Center for Earth Systems
Technology University of Maryland, Baltimore
County NASA/Goddard Space Flight Center
Robert F. Adler NASA/Goddard Space Flight Center
George J. Huffman SSAI/NASA/Goddard Space Flight
Center
Curtis, S., G. J. Huffman, and R. F. Adler,
2002 Precipitation Anomalies in the Tropical
Indian Ocean and Their Relation to the Initiation
of El Niño, Geophysical Research Letters,
29(10), 1441, doi10.1029/2001GL013399.
2
DATA Precipitation, Surface Winds, and SST
GPCP Pentad Precipitation (Xie et al. 2003)
GPROF SSM/I Pentad Precipitation (2.5, global,
1979-present) NCEP Daily (avg. to pentad)
Reanalysis 1000mb Winds (2.5, global,
1979-1999) TRMM Daily (avg. to pentad) 3B42
Precipitation (Adler et al. 2000) (1, tropics,
2001-2002) TRMM Daily (avg. to pentad) SST
(0.25, tropics, 2001-2002) Acknowledgement
TMI-SST data were produced by Remote Sensing
Systems and sponsored by NASAs Earth Science
Information Partnerships (ESIP) QuikSCAT 12
Hourly (avg. to pentad) Sea Surface Winds (0.5,
near-global, 2001-2002) Acknowledgement Data
obtained from the NASA / NOAA sponsored data
system Seaflux, at JPL through the courtesy of W.
Timothy Liu and Wenqing Tang
3
The Indian Ocean and El Niño Longitude-time
sections of Equatorial 850mb zonal wind anomalies
from daily forecast values (Krishnamurti et al.
2000, Figure 1)
1991-92
1994-95
1993-94
1997-98
observational summaries suggest important
teleconnections of these zonal winds over the
near-equatorial southern Indian Ocean and the
subsequent onset of the El Niño in the equatorial
Pacific Ocean (Krishnamurti et al.
2000). However, no uniform response in MJO in the
eastern Indian Ocean has been related to El Niño
(Fink and Speth 1997, Vincent et al. 1998).
4
GPCP Pentad Precipitation and NCEP 1000mb Winds
(1979-1999)
Curtis et al. (2002)
Hastenrath et al. (1993), Saji et al. (1999),
Webster et al. (1999) (autumn during El Niño)
Eastern Equatorial Indian Ocean (autumn-winter
before El Niño)
5
Results from Curtis et al. (2002)
1. Greatest relationship with El Niño found in
an index of the gradient of precipitation
anomalies (SE minus NW) 2. Two Dimensional
Solution A) Interannual average (2-7 year
filter) of the gradient is a positive
maximum B) Followed 3-4 months later by
significant intraseasonal variability in the
gradient (30-60 day power from a wavelet
analysis) 3. These conditions only occurred
before El Niños (1982-83, 1986-87, 1991-92,
1994-95, and 1997-98)
6
Summary
Eastern Indian Ocean
Western Pacific Ocean
Enhanced westerlies over several months
Air-sea interaction (Kelvin wave)
El Niño development
MJO ?
Wind has been well observed (on sub-monthly time
scales) from buoys in the Pacific, but not the
Indian Ocean. However, precipitation/cloud
patterns (observed from satellites over the past
two decades) offer an indirect measure of global
wind variations (eg. the MJO in the Indian Ocean).
7
An Experimental Onset Predictor Index

• 2-7 year filter was replaced with trailing 37
  pentad (half-year) average
• Ignore the 30-60 day power for pentads with
  6-month trailing mean lt 0
• Real-time updates beginning in 2001
• Index exceeded positive forecast conditions for
  the first time on Jan. 21-25 2002 (calculated in
  early Feb.)
• Based on past events, we called for onset to
  occur 6-9 months from January
  (July-October)

8
Trailing 37-pentad average (mm day-1)
30-60 day power (mm day-1)2
9
Nino3.4
4.33
4.17
3.58
2.92
6.25
30-60 day power (mm/d)2
Validation
Nino3.4 (C)
LEGEND
Black miss Blue 1-4 months Green 5-8
months Pink 9-12 months Red 12 months
10
(No Transcript)
11
Future Work

• Precipitation is not THE answer. We are
  currently investigating intraseasonal variations
  in wind, SST, and rainfall from satellites in the
  equatorial Indian Ocean sector (96-97 01-02)
• Extension of the analysis back in time
• The western Pacific
• Extended-MJO time scales (60 days)
• Communication, dialogue, and feedback

12
Air-sea coupled MJO (30-day variability in 2002)
15-day difference in daily pentad
averages Precipitation -30 to 30
mm/day Zonal winds -15 to 15 m/s SST -2.5
to 2.5 C Start Jan 6-10 2002 minus Dec
22-26 2001 End Feb 15-19 2002 minus Jan 31 -
Feb 4 2002 See also Harrison and Vecchi 2001
Woolnough et al. 2000 Wang and Xie 1998
13
Gridded Global Historical Climatology Network
Precipitation Anomalies October 2000
Intraseasonal variability from monthly data
Cocos (Keeling) Islands, Australia
0 mm
Darwin, Australia
4 zero-line crossings required 60-day
oscillations, with max/min at mid-month
14
10 out of the 13 strongest El Niños (Jan-Feb
SOI) were preceded (lt 12 months) by
oscillations Darwin precipitation also shows a
similar relationship