THE BOUNDARY FORCING MECHANISMS AND PREDICTABILITY ASPECTS OF SEASONAL DROUGHTS AND FLOODS IN THE GR

1
THE BOUNDARY FORCING MECHANISMS AND
PREDICTABILITY ASPECTS OF SEASONAL DROUGHTS AND
FLOODS IN THE GREATER EASTERN AFRICA BY MUTEMI,
J.N, L.A. OGALLO UNIVERSITY OF NAIROBI P.O.
BOX 30197 NAIROBI KENYA CONFERENCE ON AFRICAN
DROUGHT OBSERVATIONS, MODELING, PREDICTABILITY,
IMPACTS. ICTP, TRIESTE-ITALY, 02 06 JUNE
2008. Acknowledgements/Contributions NBRP, UiB
Bejerknes Climate Centre for facilitating
acquisition of gauge data University of Nairobi
Kenya for enabling research
2
OUTLINE

• Impacts of prolonged rainfall deficits in Eastern
  Africa Motivation
• Importance of Regional Scale Processes
• Some Results
• Summary Remarks

3
Recent Drought Impacts in Eastern AfricaKenya
Ethiopia
1999/2000 In Kenya
Humanitarian agonies due to food crop failure
drought related livestock deaths due to
droughts of 1998-2001 in Greater Eastern Africa
(Kenya Ethiopia).
Ethiopia 1984 Woldu's daughter Birhan became the
poster child of the 1984 crisis when her image
featured in a shocking report by the BBC's
Michael Buerk..
4
MOTIVATION Seasonal rainfall forecasts are
crucial to welfare of communities and
regional economies whose GDP is dependant
of rain-fed agriculture. Using seasonal rainfall,
communities in the region are able to
produce a substantial amount of their
basic food, pastoral communities have pasture
water Water reservoirs and river systems
have supplies, e.g. for urban and power
generation Ability to forecast with sufficient
skill time lead is the only means of
providing quantitative info. that can
directly be used for advance preparedness
impact mitigation strategies, especially issues
of food security and water resources.
5
Processes of regional climate and how well they
are formulated represented, either explicitly
or parameterized in climate models (Global
Regional) is likely to be as source of
improvement in climate simulations for
prediction applications or climate change
projections at regional and local scales
across Africa. E.g. How well is the processing
involving the Ocean-Atmosphere feedback such
as the Indian Ocean Zonal Mode (Meyers et al.,
2007, Webster et al., 1999, etc). When we
remember the basic role of global oceans,
especially tropics in the climate system and that
only tropical Pacific (Nino 3, 3.4 SSTs)
is predictable with reasonable accuracy and lead
time, then we realize that climate models are
useful tools needing continuous improvement.
6
Annual cycle of rainfall is important to obtain
with reasonable skill in both predictions
and even climate change projections..
From CRU Data sets
Annual cycle from gauge observations
variability with ENSO, IND ATL SST boundary
forcing
7
OVER ETHIOPIA. Gauge data needs lots of
clean-up quality control This result expected
to improve substantially, but already,
predictability signals are evident in these
preliminary results Pending ?? ?? does this
result capture/ re-produce all the major extremes
in JJAS precipitation (Ethiopia droughts in
particular)?
8
OVER THE SUDAN
-ENSOIOZM significant Possible good lead time for
fcst.
-But see impact of IOZM
Precipitation over the SUDAN linkage with ENSO,
co-ENSOIOZM, and IOZM is a good result. Station
data sets may need closer look, possible
identification and correction of errors, but in
general, we only need to study its year to year
differences (to assess predictability). On
Atlantic ocean component of it, the
categorization of Atlantic SST modes done is
under improvement, and thus the ATL only
scenarios for SUDAN will change significantly in
the final results.
9
OVER UGANDA. Some data may need clean-up
quality control This may improve slightly
10
OVER KENYA. Local forcing (IOZM) outweighs
external (PAC ENSO) in consistency with GCM
studies, e.g. Goddard Graham (1999) Data
needs filtering to use those with best data
length and minimal gaps. Refining of ATL
SSTano. categories This result expected to
improve substantially, but already,
predictability can be seen. ?? Does this result
capture/ re-produce the major (or all) of
the precipitation extremes for the
two-seasons MAM OND over various
sectors of the country and with what
time leads, for real-time prediction
11
OVER TANZANIA. Tune-up of ATL SSTs
modes. Result may change, especially when
analysed on a spatial-temporal mode
12
OVER RWANDA. Some data may need clean-up
quality control
IOZM signal overrides PAC ENSO
13
OVER BURUNDI. Tune-up of ATL SSTs
modes. Result may change slightly.
14
Some modeling resultsA regional climate
phenomenaEquatorial
Eastern African (dry DJF) Southern
Africa
(Wet DJF) reverses during drought
events
15
Summary
In Eastern Africa, droughts frequent climatic
extremes, mainly in form of poor/ below
normal season rainfall. Their occurrence has
sever socio-economic impacts across the
region, and recovery from droughts impacts is
hard in many regions. Observational data
sets, especially precipitation can reveal useful
local and regional scale details of climate
dynamics that can lead to improved
predictability It is relatively easy to
determine how low the seasonal precipitation
needs to be to as an alert to adverse
drought impacts such as crop failure and
depletion of water resources. In this work, we
have shown that drought signals over Equatorial
Eastern Africa, encompassing the Ethiopia
and the Sudan are identifiable with respect to
SST anomalies involving Internal Ocean
dynamics over Indian Ocean (IOZM) and for
seasonal precipitation extremes over Eastern
Africa, the ENSO factor appears to be a
remote, and therefore of secondary importance as
a mechanism of interannual variability. It
is paramount that climate models, both global and
regional have sufficient regional/local
scale process representation for improvement of
skill in climate simulations for prediction
purposes and climate change projects. However,
advantageous use of current models and their
outputs, raw and/or processed (with
smoothing procedures such as MOS) are valuable
end-user products.
16
Acknowledgements

• Sponsors of this workshop EMC/NCEP, ICTP, ECMWF
  and their partners in supporting weather
  climate science.
• Conference directors for your effort in sourcing
  for resources to host this meeting and reaching
  out to us to come and exchange scientific ideas
  in this forum.
• To our home/affiliate institutions, especially
  the University of Nairobi, Kenya for granting me
  permission to participate in this conference as a
  researcher of the institution.
• THANKS TO ALL