W ater Resources D evelopm ent, Vol. 12, No. 3, pp.

277 296, 1996

The Im pacts of Clim ate Variability and Future Clim ate

Change in the N ile Basin on W ater Resources in Egypt

D ECLAN C ON W A Y & M IKE H ULM E

Climatic Research U nit, School of Environm ental Sciences, U niversity of East Anglia,
N orwich N R4 7TJ, U K

A BSTRACT This paper describes the application of hydrologic models of the Blue Nile
and Lake Victoria sub-basins to assess the magnitude of potential im pacts of climate
change on M ain Nile discharge. The models are calibrated to simulate historical observed
runoff and then driven with the temperature and precipitation changes from three
general circulation m odel (GCM ) climate scenarios. The differences in the resulting
m agnitude and direction of changes in runoff highlight the inter-model differences in
future climate change scenarios. A `wet case, `dry case and composite case produced
1 15 ( 1 12), 2 9 ( 2 9) and 1 1 ( 1 7) per cent changes in m ean annual Blue Nile
(Lake Victoria) runoff for 2025, respectively. These gures are used to estimate changes
in the availability of N ile water in Egypt by m aking assumptions about the runoff
response in the other Nile sub-basins and the continued use of the Nile W aters
Agreement. Comparison of these availability scenarios with demand projections for
Egypt show a slight surplus of water in 2025 with and without climate change. If,
however, water demand for desert reclamation is taken into account then water de cits
occur for the present-day situation and also 2025 with (`dry case GCM only) and
without climate change. A revision of Egypts allocation of Nile water based on the
recent low- ow decade-mean ows of the Nile (1981 90) shows that during this period
Egypts water use actually exceeded availability. The magnitude of `natural uctuations
in discharge therefore has very im portant consequences for water resource management
regardless of future climate change.

I ntrodu ction

Thi s paper presen ts an assessm en t of the sen sitivity of run off in the Nile Basin
to climate change. Thi s is achieved by using future clim ate chan ge scen arios
deri ved from gen eral circulation models (GCM s) to dri ve hydro logic m odels of
the Blu e N ile and Lake Victoria sub-basins. To gether these tw o su b-basins
contribute over 70% of M ain N ile discharge. A num ber of assum ption s are made
about the resp onse of run off in the other Nile sub-basins based o n thei r
similarities to either the Blu e Nile or Lake Victoria sub-basins. Chan ges in runoff
from all sub-basins are then used to deri ve an overall esti m ate of M ain Nile
discharge. The implications of these scen arios of M ain N ile discharge are then
discussed w ith referen ce to curren t `natural variability in run off, future avail-
ability of water, an d dem and for w ater by the dow nstream riparian, Egyp t.
0790-0627/96/030277 20 $6.00 1996 Jou rnals O xford Ltd
278 D. Conway & M . Hulme

The N ile Basin is a com plex hy drologic system draining an area of roughly
2
3 million km and providin g freshw ater to millions of people (Fig ure 1).
2
W ithin the Basin there are ve major lakes of over 1000 km surface area
(Lakes Victoria, Ed ward , Albert, Kyoga and Tana), vast areas of perm anent
and seasonal w etlands (the Sudd , Bahr el G hazal and M achar M arshes), ve
m ajor reserv oir dam s (H igh A sw an Dam (H AD), Roseires, Khashm el G irba,
Senna r and Jebel Aulia) and four im portant hydroe lectric pow er dam s (H AD,
Tis Isat, Finc haa and O wen Falls). The course of the Nile ow s from highla nd
region s w ith abundant moisture to lowland plains w ith arid conditio ns. Egyp t,
and to a lesser extent Sudan, is alm ost w holly dependent upon water that
orig inates from the upstream N ile Basin countries (U ganda, Ethiopia, Tanzania,
Keny a, Rwand a, Burundi, Zaire and now Eritrea ), w hich m akes the issue of
w ater reso urce m anagem ent and ownersh ip highly contentio us and of strategic
im portance (Waterbury, 1979, 1982; G odana, 1985; Anderson , 1992; H ow ell &
A llan, 1994).
Balancin g the availability of water with dem and for w ater is likely to become
a major economic and en vironm ental issue in m any African countries in the
future, and according to Falkenm ark:

By the year 2025, 1.1 billio n people, representing tw o-third s of the

population of the African contin ent, w ill be living in conditio ns w ith
severe water scarcity. (Falkenmark, 1991, p. 85)

The certainty of increased water dem and in the future contrasts with the
uncertainty of clim atically indu ced changes in w ater supply. Estim ates of future
clim ate change due to greenh ouse gas forcing alone w ith resp ect to the 1961 90
average (Hulme & Viner, 1996), are in the regio n of 0.6 C to 1.5 C increase in
global-mean temperature by 2025. A ssociated w ith this increase in tem perature
w ill be changes in precipitation w hich, together with changes in potential
evapotranspiration (PET ), m ay have signi cant im pacts upon riv er discharge.
Population in Egyp t is expected to double by 2025 (50 m illio n in 1987 to 99
m illio n in 2025) and the per capita supply of freshw ater in Egyp t is expected to
3 3
decrease from 922 m (1990) to 337 m by 2025 (Abu-Zeid & H efny, 1992). Th is
m akes it prudent to assess the sensitivity of runoff in the Nile Basin to potential
future changes in clim ate.
A num ber of papers have looked at the im plications of uctuations in Nile
discharge for w ater reso urces in Egypt, particularly since the prolonged period
of low ow s during the 1970s and 1980s (Dem issie, 1990; Abu-Zeid & Bisw as,
1991; C onway & Hulm e, 1993). The histo rical uctuations in N ile River dis-
charge have also been review ed by Shahin (1985), Evans (1990), Sutcliffe &
Lazenby (1990) and Said (1993). O nly a few stu dies, how ever, have attem pted to
evaluate the im pacts of future clim ate change on runoff in the Nile Basin. The
rst of these w as by K ite & W aititu (1981) w ho looked at the N zoia riv er, a
tributary of Lake Victoria, using the Sacram ento W atershed catchm ent m odel.
They analysed the effects on river ow and lake level of varying precipitation
and evaporation in a sensitiv ity analysis. H ulm e (1990) review ed the factors
affecting precipitation over the Nile Basin at different tem poral and spatial
scales. He presente d future changes in temperature and precipitation, based on
the results of a num ber of GC M experim ents, for the N ile Basin with a
discussion of their implications for N ile discharge. G leick (1991) also analysed
the vulnera bility of Nile runoff to clim ate change. H e applied W igley & Jones s
 Im pacts of Climate Variability and Change in the N ile Basin 279

Figu re 1. Th e Nile Basin , location of major sub-basins and G CM gri d boxes used
in the study (shaded areas, 5 3 5 ).

(1985) theoretical model (based on the annual w ater balance) to thre e su b-basins
of the N ile Basin, the Upper W hi te N ile, Sobat and Blue Nile/Atbara. The model
produ ced a 50% redu ction in runoff in the Blu e N ile catchment du e to a 20%
decrease in precipitation. Abu-Zeid & Bisw as (1991) consi dered the im plications
of clim ate uctuations for water manage ment w ith an em phasis on A frica and
the N ile. Th ey stre sse d the uncertainties in volved in predicting future clim ate
change and that existing planni ng processe s and hydro logic m ethodologie s ne ed
to be im proved to deal w ith such challeng es. Th ey also em phasized the im port-
280 D. Conway & M . Hulme

ance of uctuations in riv er ow over the histo rical period for managing w ater
reso urces. Onyeji & Fisch er (1994) und ertook an econom ic analysis of potential
im pacts of clim ate change in Eg ypt. Th eir projections indic ate a decline in
self-suf ciency (agricultural and non-agricultural) and that clim ate change has a
num ber of potentially negative effects. Their analysis does not incorporate
clim atically indu ced changes in Nile supply, although this is addressed by
Strzepek et al. (1995). Other studies have concentrated on the potential im pacts
of sea-level rise in the Nile Delta (Millim an et al., 1989; Sestini, 1992). Th e m odels
used in this study w ere used by C onway et al. (1996) as part of an integra ted
m odelling approach to assess the im pacts of driving forces at different spatial
scales (global: clim ate change; regio nal: land-use change; riv er basin: w ater
reso urce managem ent policy) on runoff in the N ile Basin. To date this has been
the only attempt to utilize basin-sp eci c models of the Nile sub-basins w ith
clim ate change scen arios taken from GC M experim ents.

The N ile H ydro logic M odels

O w ing to the coarse spatial resolu tion of G CM output and the crude representa-
tion of runoff processes in G CM s most w ater resource impact stu dies use
region al w ater balance m odels w ith only the tem perature and precipitation
changes taken from G CM experim ents (Gleick, 1989). The hydrologic m odels
used for clim ate impact studies are often data intensive, yet m ost of the Nile
sub-basins have lim ited rain gauge coverage, few long-term tem perature
records, poor river gauge data since the early 1980s, and very scarce daily data.
The size and complexity of the Nile sub-basins, togeth er with the lack of data,
is a severe constraint to the developm ent of sophistic ated catchm ent models.
The development of less complex grid-based hyd rologic m odels that require
lim ited data inputs and run on m onthly time-steps is, how ever, feasible for some
of the Nile sub-basins. The N ile Basin may be divid ed into a number of
sub-basins with very different phy sical, clim atic and hydrological characteristics
(Conw ay & H ulm e, 1993). For im pact assessm ent studies it is therefo re necessary
to treat each sub-basin separately. This section brie y describes the two hydro-
logic models used for this study (see C onw ay, 1993, for a m ore detailed
description).
A model of the Blue N ile w as dev eloped along the lines of the hydro logic
m odelling approach used for the Am azon and Zam bezi rivers by Vorosm urty et
al. (1989; Vorosm urty & M oore, 1991). Th e model covers the catchm en t upstream
of the El Diem river gauge on the Sudan Ethio pia border, an area of roughly
2
170 000 km . In the model the catchm ent is divid ed into 10-minute latitude by
10-minu te longitu de grid cells. Estim ates of m ean m onthly precipitation and
evapotranspiration are calculated for all grid cells and fed into a sim ple m onthly
w ater balance. The runoff obtained for each cell is then summ ed to produce an
overall estim ate of catchm ent runoff (see Figu re 2 for an overv iew of the model
structure). The m odel has three adjustable parameters which are constant for all
grid cells except for the cells representing the Lake Tana and Dabus sw amp
subcatchm ents. The grid cells in these tw o subcatchm ents require different
param eter values in ord er to sim ulate the storage and delaying effect on runoff
from these hydrologically special areas. The m odel w as calibrated to reproduce
m ean monthly runoff at El Diem /Roseires (1951 87, extracted from W orld Bank,
1989; H urst & Philip s, 1933) and validated by its ability to sim ulate subcatch-
 Im pacts of Climate Variability and Change in the N ile Basin 281

Figu re 2. Structure of the water balance m odel of the Blu e N ile River Basin
(SW H C 5 Soil w ater holdi ng capacity).

m ent runoff and hi storical variations in Blue Nile runoff. Sim ulated m ean annu al
ru noff w as within 3% of the observed, although fairly sign i cant errors occurre d
in indi vidual ye ars. Consi dering the paucity of precipitation data for the sub-
basin these results are reasonable. Figure 3 show s the observ ed and simulated
m onthly runoff for the Blue Nile upstream of Roseires/El Diem for the period
between 1967 and 1986.
For Lake Victoria an existin g monthly water balance model of the lake
catchment was used. The model w as developed by the Institute of H ydrolo gy
(IH) based on earli er w ork by H ydro m et (Kite, 1981) to und erstan d the cause of
an unprecedented rise in the lake levels and out ows of the W hite N ile lake s
282 D. Conway & M . Hulme

Figu re 3. Observe d (bold lin e) and sim ulated (dashe d line ) m onth ly Blue Nile
3
discharge (km ), 20-year period, 1967 86.

that occurre d in the early 1960s (IH, 1984; Piper et al., 1986). The m odel w as
recen tly updated by Sene & Plinston (1994) and used for clim ate im pact stu dies
by C onway (1993). The m odel uses hi storical observations of over-lake precipi-
tation, lake tributary in ow s, lake evaporation, lake levels and lake out ow s.
The best peri od for data availability (prim arily lim ited by tributary in ow data)
w as betw een 1956 and 1978. For this period the m odel was able to reprodu ce
hi storical uctuations in lake levels usin g over-lake precipitation, tributary
in ow s and lake evaporation (Fig ure 4).

Figu re 4. Observe d (bold line) and sim ulated (dashe d line ) Lake Victoria out ow s,
23-year period, 1956 78.
 Im pacts of Climate Variability and Change in the N ile Basin 283

Fu ture C lim ate C hange

Climate Scenarios from General Circulation M odels

G C M s are the m ost sophisticated m ethod curre n tly available for esti mating the
future effects of increasing gre enh ouse gas con centration s on clim ate (Carter et
al., 1995). C oupled ocean -atm osphere G C M s can calculate th e tim e evolu tion of
tem peratu re plus hum idity, w ind, soil m oisture, ice an d oth er variables for the
land surface and atm osph ere. M odel calculations are based on th e law s of
ph ysi cs an d executed at w idely spaced points of a th ree-dim ension al gri d at a
resolution of approxim ately 5 by 5 latitude and longitude (som e G C M s now
operate at 2.5 by 3.75 resolution). C on trol experim ents are ru n to sim ulate
curre nt clim ate (1 3 C O 2 ) and two m ethods can be used to esti m ate the respon se
of future clim ate to gre en house gas forcing: equilibrium response and transien t
respon se. Th e equ ilibrium response of clim ate w ill be discusse d here since only
a few tran sient experim ents h ave been condu cted to date, and their results are
less w idely available and their use in im pact studies is associated with som e
aw kward m ethodological problem s (Carter et al., 1995). In equilibriu m mode, the
G C M is run w ith an abrupt increase in C O 2 con centration s, usu ally a doubling
from 300 ppm to 600 ppm (2 3 C O 2 ). The difference betw een th e 1 3 C O 2 an d the
2 3 C O 2 clim ate at each grid point is then used to represen t the equilibrium
clim ate ch an ge to a doubling of atm ospheric C O 2 (or C O 2 equivalent).
C lim ate scen arios from G C M s are not predictions. G C M s operate at various
resolutions, use different represe ntations of ph ysi cal processe s and hence pro-
duce different con trol clim ates and responses to increased C O 2 . C on trol run
sim ulations fail to reproduce accurately various features of curren t clim ate,
particularly at regional scales an d particularly w ith som e precipitation regi m es.
There is also uncertainty about the role of feedback m echanism s, such as clouds
or sea ice, in the respon se of the clim ate syste m to in creases in gre en house gases.
Fu rth erm ore, assum ptions need to be m ade th at land cover does not ch an ge in
the future and also, for exam ple, th at ocean circulation does not chan ge. Th e
G C M scenarios presen ted h ere are th erefore no m ore than plausible estim ates
of the direction and magn itu de of future clim ate ch ange . Th e level of uncer-
tainty in su ch results is, h ow ever, con siderably less for tem perature th an for
precipitation.

Future Climate Scenarios for the Blue Nile and Lake Victoria Sub-basins

The results from three equilibrium G C M experim en ts w ere used for th is study .
The G eophysi cal Fluid Dyn am ics Laboratory (GFD L; W eath erald & M an abe,
1986) an d G odd ard Institute for Space Studies (GISS; H an son et al., 1984) were
chosen as exam ples of `dry and `wet cases, respectively to high ligh t the
inter-m odel differences in precipitation change s. The th ird case w as a com posite
scenario ge nerated from seven differen t G C M experim ents using a m ethod
originally developed by Santer et al. (1990) (see also H ulm e, 1994). The C om -
posite precipitation scenario was ge nerated from a weigh ted mean of seven
equ ilibrium G C M change elds (H ulm e, 1994). Each G C M w as given a w eigh t-
ing based on its ability to sim ulate curre nt precipitation. Th e tem perature
change s w ere given equal w eigh ts in th e com posite scen ario. In orde r to
introdu ce tim e dependen cy in to equilibrium GC M results the m ethod developed
by Santer et al. (1990) w as used. Th e 2 3 C O 2 2 1 3 C O 2 differen ces are divided
284 D. Conway & M . Hulme

T ab le 1. Ab solute change in tem perature and percentage change in precipitation

for thre e scenarios b ased on a 1 C increase in glob al m ean temperature

Tem perature ( C ) Precipitation (% )

G CM D JF M AM JJA SO N ANN D JF MAM JJA SO N ANN

B lue N ile:
C om posite 1.1 1.0 0.9 1.0 1.0 4.0 2 1.1 2.6 3.6 2.2
G FD L 0.7 0.7 0.7 0.7 0.7 2 3.0 2 2.1 2 1.1 2 2.4 2 1.9
G ISS, 1.0 0.9 0.6 0.9 0.8 12.0 14.4 6.4 1.9 7.4

L ake V ictoria:
C om posite 0.9 0.8 1.0 0.8 0.9 3.0 7.0 6.0 5.0 5.0
G FD L 0.6 0.7 0.6 0.7 0.7 3.0 2 3.0 5.0 2 3.0 2 1.0
G ISS 1.0 1.0 0.6 0.9 0.8 1.0 6.0 36.0 4.0 5.0

N ote: A nnu al and seasonal (D ecem ber February DJF , M arch M ay M A M , June A ugust JJA , Septem b er
N ovem ber SO N ) changes are show n. C hanges are show n for tw o grid points (Blue N ile, m ean 37.5 E,
12.5 N & 37.5 E, 7.5 N Lake V ictoria, 32.5 E, 2.5 S), and have been standardized b y each climate m od el
sen sitivity (s ee H ulm e, 1994).

b y the overall clim ate sensitivity of each GC M (e.g. 4.2 G ISS and 4.0 G FD L) to
produce a standardized change in tem perature/precipitation per degree Celsius
change in glob al-m ean tem perature. T he clim ate sensitiv ity is a term used to
describ e the equ ilib rium glob al-mean tem perature change for a doub ling of
atmospheric C O 2 a gure that varie s conside rab ly b etw een G C M experim ents.
For this study a rise of 1 C in glob al-m ean tem perature was chosen to represent
the realized tem perature rise b y the ye ar 2025 w ith respect to the prese nt (de ne d
as the average of 1961 90). T his increase is comfortab ly w ithin the range of
temperatures estim ated usin g a rang e of em issio ns scenarios, clim ate sensitivities
and igno ring the cooling effect of sulphate aerosols (Hulm e & Vine r, 1996).
T ab le 1 lists the seasonal (Decem b er Fe b ruary DJF, M arch M ay M A M , June
A ugust JJA, Septemb er N ovemb er SO N) and annual ab solute tem perature and
percentage precipitation changes produced b y the three scenarios for the m ean
of two grid b oxes overlying the Blue N ile (37.5 E, 7.5 N ; 37.5 E, 12.5 N) and one
grid b ox overl ying Lake Victoria (32.5 E, 2.5 S) (see Figu re 1). All thre e scenarios
show increases in temperature in all four seasons. T he G CM s produced a m uch
less hom ogene ous set of changes in seasonal and annual precipitation. C hange s
in annual precipitation range d from 2 2% (GFD L) to 1 7% (GISS) over the Blue
N ile. T he signi cant differences of m agnitude, and dire ction of change , for
precipitation highl ight the w ell-known disparitie s b etween G CM sim ulations of
precipitation (e.g. G ates et al., 1992). For the grid b ox overlying Lake Victoria
(32.5 E, 2.5 S) tem perature change s w ere sim ilar in m agnitu de to those in the Blue
N ile Basin, with GISS slightly w armer than G FD L. T he dry scenario (GFD L) onl y
produced a decrease in precipitation in M A M and SON and an overall annual
decrease of 1% .

Im pacts on R unoff in the Blue Nile and Lake Victoria Basins

GCM Scenarios

T o ge ne rate clim ate change elds for input to the Blue N ile model, the G CM
 Im pacts of Climate Variability and Change in the N ile Basin 285

T ab le 2. T he percen tage change in Blue N ile and Lake Victoria

mean annual runoff

% ch ange in annual runo ff

G CM C hang ed climate inputs B lue N ile Basin Lake V ictoria

C om posite PET and precipitation 1 0.7 1 6.9

G FD L ,, 2 8.6 2 9.2
G ISS ,, 1 15.3 1 11.8
C om posite PET on ly 2 6.0 2 9.2
G FD L ,, 2 4.1 2 6.9
G ISS ,, 2 3.9 2 8.8
C om posite Precipitation on ly 1 6.9 1 16.2
G FD L ,, 2 4.5 2 2.3
G ISS ,, 1 19.7 1 20.9

N ote: Three climate scenarios, w ith d ifferen t com binations o f PET and precipitation
chang es applied, each assum ing a 1 global m ean w arm ing.

change s from eight adjacent grid b oxes (30 E 45 E; 0 N 20 N) w ere interp o-

lated to a 10-minute reso lution using a G aussian space- ltering method. T he
gridd ed elds of seasonal change s in precipitation and temperature were
used to drive the Blue N ile model. T he m odel w as run for 24 months to
sim ulate mean monthl y runoff for the calib ration period and the rst 12
m onths of the sim ulation w ere discard ed to allow soil m oisture and lagge d
runoff to adjust to the ne w clim ate. Seasonal change s w ere applied to each
m onth in their resp ective seasons. Sim ulations w ere perform ed with b oth PET
and precipitation change d, with onl y PET changed , and w ith onl y precipi-
tation change d. T he method of conve rting tem perature change s into percent-
age change s in PET w as derive d b y Budyko (1982) as used b y N ash & G leick
(1991). T he relationship assum es a 4% change in PET per 1 change in tem-
perature. A sim ilar approach was followed for the Lake Victoria m odel. For
each scenario the seasonal change s in tem perature and precipitation for the
grid b ox overlyin g Lake Victoria w ere applied to all ye ars in the lake -level
sim ulation.
T he differences b etw een the ob served runoff and the runoff und er a per-
turb ed climate were calculated and are liste d in T ab le 2 expressed as percent-
age change s in annual runoff. As expected, for b oth sub -b asin s, the dry and
w et G CM scenarios produce decreases and increases in out ow s, respectively.
For the Blue Nile and Lake Victoria w ith the composite scenario the losses
due to increased tem perature were offset b y the increase in precipitation.
Fig ure 5 show s the optim ized and perturb ed m ean monthl y runoff for the
Blue Nile after applying the PET and precipitation change s from the three
G CM scenarios. T he large st changes occurre d in July, August and Septem b er
b ut there w ere no signi cant change s in the annual cycle of monthl y runoff.
T he re was also no signi cant change in the seasonality of out ow s from Lake
Victoria (Figure 6). T he results with onl y PET and onl y precipitation change d
highl ight an interesting difference b etween the two sub -b asins: change s in
PET have a gre ater impact on runoff in Lake Victoria than in the Blu e Nile
m odel owing to the large expanse of open water in the lake sub -b asin.
286 D. Conway & M . Hulme

Figu re 5. O ptim ized and GC M scenario m ean monthly runoff for the Blue Nile
based on 1951 87 runoff at El Diem: optimized; com posite; - - - - GFD L;
GISS.

Sensitivity of Runoff to Clim ate Change: Hypothetical Changes in PE T and Precipi-

tation

A range of hypothetical change s in PET and precipitation were used to dri ve the
sub-basin m odels to exam ine their sensi tivity furthe r. PET w as change d by 6 4%
and precipitation change d in 5% steps from 2 25% to 1 25% . The anom alies
w ere applied equally to all m onths. Th e results, expresse d as percentage change
in annual ru noff, are show n for the Blue N ile in Fig ure 7. Change s in precipi-
tation produ ce large r change s in ru noff than change s in PET. The runoff
response is greater than the precipitation anom aly: a 10% increase in precipi-
tation causes a 34% increase in runoff and a 4% decrease in PET causes an 8%
increase in runoff. The relationshi p betw een change in precipitation and change
in runoff is rough ly line ar (but not 1:1) dow n to a 10% redu ction in precipitation.
A 25% incre ase in catchm ent precipitation caused a 90% incre ase in ru noff and
a 25% decrease caused a 70% reduction in run-off. In contrast, for Lake Victoria
(not show n) a 10% increase in precipitation causes a 31% incre ase in runoff and
a 4% increase in PET causes an 11% decrease in runoff. Thi s gre ater sensi tivity
 Im pacts of Climate Variability and Change in the N ile Basin 287

Figu re 6. Observe d and G CM scen ario m ean monthly Lake Victoria out ow s
based on 1956 78: observe d; composite ; G FD L; - - - - GISS.

of ru noff to changes in precipitation than to change s in PET has also been found
in other m odelli ng stu dies (for example N emec & Schaake , 1982) and theoreti-
cally by W igley & Jone s (1985).

Im plications of N atural Variability and Fu ture C limate Change for W ater

R esou rces in Egypt

Impact of `Natural Variability on W ater Supply

Flu ctuations in decade-m ean M ain Nile discharge of up to 6 20% have occurre d
during this cen tury (Conway & Hulme, 1993). Thi s `natural variability in the
discharge of the M ain N ile has had im portant conse quences for w ater resou rce
m anagem ent in Eg ypt (A bu-Zeid & Abdel-Dayum , 1992). Thi s section presents
an analysi s of the availability of w ater and demand for water in Egy pt projected
for the ye ar 2025. Two Nile ow scenarios are used; one based on the long-te rm
m ean Nile discharge (1900 59) curre ntly used for w ater reso urce m anage ment,
and an extrem e low - ow case based on the decade-m ean discharge (1981 90).
Three projections of future water availability and demand for w ater were
drawn up by the Egyp tian M ini stry of Public W orks and W ater R esources as
part of a long-term water resource manage m ent plan. The m id-case is used he re
(S3, see A bu Zeid & Hefny, 1992). Table 3 show s a breakdown of curre nt (1990)
and projected (2025) w ater availability and demand in Egyp t unde r this proje c-
tion w hi ch is based on the follow ing assu mptions:
completion of the Jongle i canal;
incre ases in groundw ater extraction;
incre ased reu se of agri cultu ral drainage water;
eld-irri gation ef ciency im proved to 65% ;
potable water-u se ef ciency im proved to 80% (curre ntly 50%);
improved control of w ater transport and distribution;
the 1959 Nile W aters A gre em ent to provide the legal basis for w ater plannin g;
no allowance for potential increases in abstractions from the upstre am
riparians.
288 D. Conway & M . Hulme

Figu re 7. Change in mean annual Blu e Nile runoff w ith a range of hypothetical
PET and precipitation change s: PET 0%; - - - - PET 1 4% ; PET 2 4% .

Two assum ptions about clim ate are im plicit in the proje ction of w ater avail-
ability:

a return to the mean annual ow volu me based on the period betw een 1900
and 1959;
no change in availability or dem and ( 6 ) as a result of clim ate change due to
incre asing concentrations of gree nho use gases (or, ind eed, any other cause).

The rst tw o colu mns in Table 3 show that there will be a ne balance between
the future availability of w ater and dem and for w ater. It should be noted that
some of these proje ctions are based on assu mptions that w ill not necessarily
hold true and the estimates of future dem and are rather conse rvative. If the land
reclam ation policy is fully implem ented it w ill lead to a shortfall in su pply of 0.9
3 3
km by 2025 (based on a demand of 1 km per 67 000 ha; Abu Zeid & H efny,
1992). It is clear therefore that the proportion of w ater now allocated for
agri cultu ral purposes will have to decrease as non-agri cultu ral demand s in-
crease in the future.
Table 4 shows the division of water betw een Egypt and Sudan accordi ng to
3
the N ile W aters Agre ement of 1959, based on a m ean annual ow of 84 km for
the period 1900 59. Also show n is an equivalent breakd ow n based on the m ean
ow for the period betw een 1981 and 1990. A revision of water availability and
dem and proje ctions based on these gu res make s the curre nt level of dem and
exceed availability (see gu res in bold, rst colum n of Table 3). W ith regard to
3
1990 there is a water su rplus of 4.3 km based on the 1900 59 m ean annu al
 Im pacts of Climate Variability and Change in the N ile Basin 289

T ab le 3. Curre nt (1990) and proje cted (2025) water availab ility and dem and
3
scenarios for Egyp t (km )

Cu rren t climate Ch anged climate

Scenarios for 2025
1990 S3 C om p (T o nly) GFDL (T only) GIS S (T on ly)

W ater sources:
River N ile 55.5 55.5 56.4 (51.5) 50.1 (52.7) 64.7 (52.7)
49.9 49.9 50.7 (46.3) 45.0 (47.4) 58.3 (47.3)
Jon glei C anal 2.0 Com pletion is not certain
D eep groun dw ater 0.5 3.5 U n likely to be affected on this timescale
N ile V alley and D elta 2.6 3.6 Aq ui er recharge rates m ay change. Salt water
groun dw ater incursion in the N ile Delta
A gricultural drainage water 4.7 5.0 D irect CO 2 effects on crop w ater use. H igher
irrigation w ater losses due to increased PE T
T reated sew age water 0.2 2.0

T otal 63.5 71.6 72.5 (67.6) 66.2 (68.8) 80.8 (68.8)

57.9 66.0 66.8 (62.4) 61.1 (63.5) 71.4 (63.4)

W ater dem and :

A griculture (p resent area) 49.7 46.6 L ikely to increase
M unicipal and ind ustrial 7.7 10.8 ,,
N avigation 1.8 0.3 ,,

T otal 59.2 57.7 57.7 57.7 57.7

W ater surplus/de cit: 4.3 13.9 14.8 (9.9) 8.5 (11.1) 23.1 (11.1)
2 1.3 8.3 9.1 (4.7) 3.4 (5.8) 16.7 (5.7)
W ater dem and fo r desert
reclam ation 5.2 14.3 14.3 14.3 14.3
W ater surplus/de cit 2 0.9 2 0.4 0.3 ( 2 4.4) 2 5.8 ( 2 3.2) 8.8 ( 2 3.2)
2 6.5 2 6.0 2 5.2 ( 2 9.6) 2 10.9 ( 2 8.5) 2.4 ( 2 8.6)

N ote: S3 is future dem and w ithout climate chang e. Clim ate change scenarios taken from this study; see
text. N orm al gures are b ased on the m ean fo r 1900 59, gures in bo ld are based on the extrem e low
d ecade-m ean discharge b etw een 1981 and 1990. T on ly 5 tem perature ch ange only.
S ou rce: A d apted from Ab u Zeid & H efny (1992, p. 45).

3
discharge , and a surplus of 13.9 km proje cted for 2025. W he n the availab ility
gu re is b ased on the 1981 90 m ean, curre nt demand exceeds availab ility b y
3 3
1.3 km and the projected surplus for 2025 is red uced to 8.3 km . T his de cit
has so far b een met b y store d w ater in the HA D reservoir and additional w ater
from Sudans allocation w hich has not b een fully utilized. C urre ntly, there is
no clear evide nce to sugge st that a return to highe r rainfall in the Sahe l of
A frica, and b y association Blue N ile ow s, is any m ore like ly than a continuation
of the present conditions (H ulme & Kelly, 1993). Recen t Nile ow s have,
how ever, returne d to ne arer the long-te rm m ean and, following a wet ye ar, the
1994 ood level was highe r than average and H AD reservo ir levels surpasse d
the level of 173 m for the rst time since 1987 (Anon, 1993). Neverthe less the
m agnitu de of `natural uctuations in discharge clearly has very important
consequences for w ater resource m anage m en t regardl ess of future clim ate
change .
290 D. Conway & M . Hulme

T ab le 4. Allocation of w ater accord ing to the Nile W aters Agreem ent 1959, b ased
on the origina l mean ann ual ow (1900 59) and the recent decade-m ean (1981 90)
3 3
V olum e (km ) 1900 59 % V olum e (k m ) 1981 90

E gypt 55.5 75 49.9

Sudan 18.5 25 16.6
H A D reservoir evaporation 10.0 10.0
3
M ean ow (k m ) 84.0 76.5

Implications of Climate Change due to the Enhanced Greenhouse Effect for M ain Nile
D ischarge

A gainst the b ackgro und of high `natural variab ility and a ne b alance b etw een
future water availab ility and dem and, climatically indu ced change s in M ain Nile
discharge w ill furthe r increase the uncertainty of this vital natural reso urce. T he
im pacts of clim ate change on runoff have b een calculated for the tw o large st
contrib uting N ile sub -b asins ( . 70% of M ain Nile discharge ). In orde r to
transl ate these changes in term s of M ain N ile discharge a nu mb er of assum p-
tions ab out the response of the other sub -b asins were m ade b ased on their
characteristics and their sim ilaritie s to eithe r the Blue N ile or Lake Victoria
sub -b asins. T hese are describ ed in Conw ay et al. (1996), and sum marized in
T ab le 5. Annual m ean discharge s (b ased on the period 1900 59 to allow
com parison w ith the scenarios in T ab le 3) w ere calculated for all sub -b asins that
contrib ute to the M ain N ile . T he percentage change s in m ean runoff produced
b y the m odels in each of the clim ate scenarios w ere then applied to these
discharge volum es in the manne r describ ed in T ab le 5. T o assess the im plications
of these changes for Egypt the resulting M ain N ile discharge s were calculated
and div ide d accord ing to the Nile W aters A gre ement (see T ab le 4). A fter
3
sub tracting 10 km for evaporation at Aswan the total w as split 75% for Egypt
and 25% for Sudan to represent their allocations of water unde r a future ow
scenario. T ab le 3 lists the resulting ow allocations for Egyp t togethe r w ith
potential im pacts on the other w ater sources in Egypt and potential increases in
dem and for water for land reclam ation.
T he thre e climate-change scenarios produced changes in M ain Nile discharge
and hen ce Eg ypt s allocation of Nile w ater of 1 0.9 km , 2 5.4 km , and 1 9.2
3 3

3
km (com posite, GFD L, G ISS, resp ectively, T ab le 3 top gures in last three
colum ns) b y 2025. W ithout the water require ments for land reclamation none of
the scenarios leads to an estim ated de cit b etween w ater availab ility and
dem and b y 2025. W ith land reclam ation, however, the G FD L scenario produces
a water de cit, and if the 1981 90 ow s are now typical of the `natural
b ackgro und ow in the ab sence of gre enho use gas-indu ced climate change then
sub stantial de cits appear for b oth the com posite and GFDL scenarios ( gures
in b old). W ith onl y change s in PET applied (results sh ow n in parenthe ses, T ab le
3), all scenarios produce fairly sub stantial reductions in discharg e. T he se reduc-
tions are on the conservative sid e b ecause no allow ance has b een made for
in creased losse s to evaporation from rive r and reservoir surfaces in the sem i-arid
and arid parts of Sudan. In add ition to this a 1 C increase in tem perature in
U pper Egyp t would increase losse s to evaporation from the HA D reservoir b y
 Im pacts of Climate Variability and Change in the N ile Basin 291

T ab le 5. Annual mean discharge s and losse s to evaporation for the Nile sub -b asin s
3
(km )

V olum e of w ater
con tribution o r loss
3
Sub-b asin 1900 59 (km ) Perturbation applied

L ake V ictoria 1 20.78 % change prod uced b y Lake V ictoria m od el

E quatorial Lakes 1 5.99 N o chang e
Sudd 2 12.60 The percentage historical loss is used ( 2 47.1% )
See Sutcliffe & Pa rks (1987)
B ahr el Gh azal 0.00 N o chang e
Sobat 1 13.48 % chang e produ ced by Blue N ile m od el is
applied to half the Sob at discharge
C entral Sudan 2 3.15 N o chang e
B lue N ile (El Diem ) 1 50.22 % ch ange produced by Blue N ile m odel
B lue N ile (K hartoum ):
E vap. 2 2.45 N o chang e
D inder 1 1.09 % ch ange produced by Blue N ile m odel
R ahad 1 2.99 % ch ange produced by Blue N ile m odel
A tbara 1 12.43 % ch ange produced by Blue N ile m odel
N orthern Sudan 2 3.00 N o chang e

T otal 85.78 Scaled to 1900 59 m ean discharge;

84.00

N ote: All m eans based as far as possible on the period 1900 59 used for the N ile W aters A greem ent.
T he treatment of each sub-b asin in order to calculate the im pact of clim ate chang e on M ain N ile
d ischarge is sum m arized.

3 3
0.4 km (assum ing annual losses to b e 10 km and a 4% change in PET per
degre e change in tem perature).

C onclusions

T his paper has describ ed the application of m ethods to assess the potential
im pacts of clim ate change on runoff in the Nile Basin. H ydro logic m odels of
the Blue N ile and Lake Victoria sub -b asins w ere drive n with future clim ate
scenarios derive d from a num b er of GC M experim ents and a range of hypo-
thetical clim ate change s. T hre e scenarios were chosen for the clim ate sensitivity
study: a `dry case (GFD L), a `w et case (GISS) and a composite case ge ne rated
from a w eighted mean of seven G C M experim en ts. Both PET and precipitation
change s were applied and also PET change s alone b ecause of the highe r
uncertainty associated w ith the precipitation change s. T he clim ate perturb ations
w ere applied to the m odel sim ulations of mean m onthly runoff (no conside ra-
tion in this stu dy was giv en to change s in inte rannual clim ate variab ility). T he
thre e scenarios w ith b oth PET and precipitation change s applied (PET only)
produced the follow ing change s in annual runoff: Blue Nile 2 8.6% ( 2 4.1% ),
1 15.3% ( 2 3.9% ), and 1 0.7% ( 2 6.0% ) w ith the dry, wet and com posite
scenarios, resp ectively; Lake Victoria 2 9.2% ( 2 6.9% w ith onl y PET change d),
1 11.8% ( 2 8.8% ), and 1 6.9% ( 2 9.2% ) for the dry, wet and com posite scenarios,
resp ectively. Sensitivity analysis sh ow ed that runoff was slightl y less (m ore)
sensitiv e in the Lake Victoria Basin to change s in precipitation (PET ) than runoff
292 D. Conway & M . Hulme

in the Blue Nile Basin. Changes in PET had a greater impact in Lake Victoria
than in the Blue Nile, although they w ere not as signi cant as changes in
precipitation.
To discuss how these nd ings relate practically to Egy ptian w ater m anage-
m ent it is rst necessary to review the uncertainties in the analysis:

It is not yet possible to pred ict future clim ate change w ith a know n degree of
con denc e. There are large uncertainties in the clim ate scenarios derived from
GC M s, particularly in the precipitation changes. The scenarios presente d here
are therefore not pred ictions and to highlig ht the uncertainties examples of dry
and wet futures have been used. It is therefo re not yet possible to produce
reliable estim ates of future Nile discharge taking into account the effects of
clim ate change; a range of possibilities must be considere d.
Because of their different hydro logic characteristics the N ile sub-basins w ill
have varying resp onses to any given change in clim ate. Because of the lack of
data for som e sub-basins sim ple assum ptions have been m ade about their
resp onse to clim ate-change scenarios based on histo rical evidenc e and their
sim ilarity to either the Blue N ile or Lake Victoria sub-basins.
The hydro logic models are being used for conditio ns for which they w ere not
calibrated; the m odels do not incorporate changes in vegetation or soil
properties that m ay change in resp onse to clim ate change; they also do not
account for changes in stom atal resista nce w hich will affect the water-use
ef ciency of vegetation; the clim ate inputs to the m odels are based on
empirical relationship s (temperature and PET) that may no longer be valid in
a changed clim ate; and alterna tive hyd rologic models m ay produce different
resp onses to the same changes in clim ate because of differences in their
structure and parameter values (Leavesley, 1994). For these reasons, some of
the sensitivity obtained here m ay therefo re be an artefact of the model itself
rather than a realistic resp onse to clim ate change.
There is uncertainty concerning the future availability of w ater from non-Nile
sources and a high level of uncertainty in forecasting socioeconom ic change in
dem and for w ater.

N otw ithstand ing these uncertainties the results of this analysis highligh t the
following :

`N atural variability is critical for w ater supply in Egyp t because the balance
betw een the availability of w ater and dem and for water is very ne. It is
important to de ne the baseline Nile ows since these vary considera bly
dependin g upon w hich period they are taken from . It is shown here that,
based on the recent low- ow decade-m ean ow s of the Nile, Egyp ts current
water use actually exceeds current availability. This w ater de cit has only
been m aintained by extra water from Sudan s unused allocation and storage
in the H AD reservo ir. Such a situation cannot continue inde nitely if Nile
ows rem ain at their present level and w ould at som e stage call for an
updating of the 1959 agreem ent based on m ean discharge for, say, the 1961 90
period.
Population in Egypt is expected to double by 2025 (50 m illio n in 1987 to 99
3
million in 2025) and the per capita quota of fresh water w hich was 922 m in
3
1990 is expected to fall to 337 m by 2025. The potential for future clim atically-
indu ced changes in supply could exacerbate this situation. U sing the Blue Nile
 Im pacts of Climate Variability and Change in the N ile Basin 293

and Lake Victoria m odels, together with som e sim ple assum ptions about the
other N ile sub-basins, the following changes in Egyp ts allocation of Nile
water w ere obtained: 1 0.8 km ( 2 3.6 km ), 2 4.9 km ( 2 2.6 km ) and 1 8.4
3 3 3 3

km ( 1 2.6 km ) w ith both PET and precipitation changes (PET only) applied
3 3

from the com posite, GFD L and G ISS G C M scenarios, resp ectively. These
runoff scenarios are for 2025, based on mean annual Nile ows for the
period 1900 to 1959, and assum e a 1 global mean warm ing with regard to
1961 90.
Historical evidenc e and the low runoff ratios of the Blue N ile Basin and m ost
of the other N ile sub-basins suggest that runoff w ill have a fairly high
sensitivity to changes in precipitation and temperature. The results from w ater
balance models of the Blue N ile and Lake Victoria con rm this expectation.
None of the scenarios suggested any changes in the seasonal distrib ution of
runoff. Th e hy drologic m odel results show that changes in tem perature are
more im portant in the Lake Victoria sub-basin than in the Blue N ile sub-basin
ow ing to the large expanse of open w ater.
W ith the vast areas of sw amp and open water in the Upper W hite N ile Basin
and the sem i-arid condition s and prev alence of irriga ted agriculture on the
part of the dow nstrea m riparians, changes in tem perature and PET would
have a m ajor im pact upon sub-basin runoff and may well be more critical than
the direc t im pacts of changes in precipitation.

It is at least prudent for water resource planners to consider the potential of

clim atically induced changes in w ater availability, whatever their direction.
`Natural uctuations in Nile ows are already a problem for w ater managem ent
and future changes m ay w ell w orsen the problem . Research into how to deal
w ith uctuations is necessary and Egyp ts policy toward s facing low ow s is
very important. A t present, forecasting of ow s is entirely based on discharge
data at upstream gauges (Abu-Zeid & A bdel-D ayem , 1992). A new forecasting
system is being develo ped in association w ith the USA using satellite estim ates
of precipitation in the headwater catchm ents (Schaake et al., 1993). This w ill aid
w ater m anagem ent decisions (prim arily releases from the HA D reserv oir) on a
year-to-year basis.
Future policy tow ards the managem ent and develo pm en t of water reso urces
is also im portant. The planned land reclamation schem e w ill require a heavy
com m itm ent of w ater that will redu ce exibility of supply in other areas. Th is
is at a tim e w hen the incorporation of exibility in planning is im portant.
A bu-Zeid & Hefny (1992) suggest that new sources of w ater are likely to be
treated w astew ater and agricultural drainage water. One other option is dem and
m anagem ent based on im proved allocation and econom ic pricing of w ater
(LeM oign e et al., 1992). W hitting ton & M cClelland (1992) rev iewed the econom ic
rate of return on investm ent in land reclam ation and found that it was probably
signi cantly overestim ated. As future demand for water for non-agricultural
uses increases it w ill becom e increasingly necessary for w ater to be transferre d
from the agricultural sector and for food production to be substituted for
im ports. Egypt has m oved from self-suf ciency in food during the 1970s to a
position w here roughly half the country s food req uirem ents are im ported
(Onyeji & Fischer, 1994).
Finally, it is very likely that the upstream riparians w ill increase abstractions
w hich w ill place a further constraint on w ater supply dow nstrea m in the m id-
294 D. Conway & M . Hulme

to long-term future (A bate, 1994). H istorically the legal and institu tional aspects
of w ater managem ent between the N ile Basin countries have been strain ed and
subject to much controversy (W aterbury, 1987). Future water reso urce develop-
m ent and the potential of clim atically induced changes in supply w ill increase
the need for cooperation and coordina ted planning. A t present cooperation and
com m unication in the N ile Basin is being promoted through a series of inter-
national conferences, started in 1993 in Asw an, to be held once a year in each
m ember country (Shady et al., 1994). Such m easures now are vital to en sure that
the increasin g pressu re on water reso urces in the future can be met w ith a
planned and ef cien t solution rather than by crisis m anagem ent.

A ck now ledgem ents

This paper origina ted as part of a PhD thesis funded by the Stock holm Environ-
m ent Institute.

R eferences
A no n (1994) R iver N ile ow s abunda ntly this year, E gyptian M ail, 9 O ctober 1994.
A bate (1994) W ater Resources Development in E thiopia (R eadin g, M A , Ithaca Press).
A bu -Z eid , M . & A bd el-D ayem , S. (1992) E gypt program m es and policy options for facing the low
N ile ow s, in M .A . A bu-Z eid & A .K. Bisw as (Eds) Clim atic Fluctu ations and W ater M an agem ent
(O xford, Heinem ann ), pp. 48 58.
A bu -Z eid , M . & Bisw as, A .K. (1991) Som e m ajor im plications of climatic uctuations on w ater
m anagem ent, In ternational Jou rnal of W ater Resources D evelopm ent 7, pp. 74 81.
A bu -Z eid , M . & Hefn y, K . (1992) W ater-resources assessment and m anagem ent in Eg ypt d uring
con ditions of stress, W M O Bulletin, 41, pp. 35 46.
A nd erson, E.W . (1992) The political and strategic signi cance of w ater, O utlook on A griculture, 21, pp.
247 253.
B udyk o, M .I. (1982) The Earths Climate: Past and F utu re, International Geop h ysics Series, V ol. 29
(N ew York, A cadem ic Press).
C arter, T.R., Pa rry, M .L., Harasaw a, H . & N ishioka, S. (1995) IPC C Technical G uidelines for A ssessin g
C lim ate Change Im pacts an d A daptation s (Lond on, W M O , U N EP).
C onw ay, D . (1993) T he dev elopm ent of a grid-based hy drologic m o del o f the Blue N ile and the
sen sitivity of N ile River discharge to climate ch ange, u npublished Ph D thesis, Un iversity of East
A n glia, U K .
C onw ay, D . & H ulm e, M . (1993) Recent uctuations in precipitation and runoff over the N ile
subbasins and their im pact on M ain N ile d ischarge, Clim atic Change, 25, pp. 127 151.
C onw ay, D., K rol, M ., A lcam o, J. & H ulm e, M . (1996) Future availability of w ater in Egy pt: the
interaction of global, regional and b asin-scale driving forces in the N ile Basin, A mb io, 25, pp.
336 342.
D em issie, M . (1990) A nalysis of d rough t in Ethiopia based on N ile River ow record s, in The
S tate-of-the-A rt of H ydrology an d H ydrogeology in the A rid an d Sem i-arid A reas of A frica, P roceedings of
the Sahel Forum (Illinois, IW R A), pp. 159 168.
E vans, T. (1990) H istory of N ile ow s, in: P.P. How ell & J.A . A llan (E ds) The N ile, Resource E valuation ,
R esource M anagem ent, H ydropolitics an d L egal Issues (L ondo n, SO AS -R GS), pp. 5 39.
Falkenm ark, M . (1991) Rapid population grow th and w ater scarcity: the predicam ent of tom orrow s
A frica, in: K . D avis & M .S. B ernstam (Eds ) Resources, Environm ent, Popu lation: Present K now ledge,
F utu re O ptions, Pop ulation and D evelopm en t Review (O xford , O xfo rd U niversity Press), pp. 81 94.
G ates, W .L., M itchell, J.F.B., B oer, G.J., C ubasch , U . & M eleshko, V.P. (1992) C limate m odellin g,
climate prediction and m odel validation, in: J.T . H ough ton, B.A . Ca lland er & S.K. V arney (Eds)
C lim ate Change 1992: the S upplementary Report to the IPC C Scienti c A ssessment (Ca m b ridge,
C am bridge U niversity Press), pp. 97 134.
G leick, P.H . (1989) C limate chang e, hy drology, and w ater resources, Reviews of G eophysics, 27, pp
329 344.
G leick, P.H . (1991) The vulnerability of run off in the N ile Basin to climatic chan ges, E nvironm ental
P rofession al, 13, pp. 66 73.
 Im pacts of Climate Variability and Change in the N ile Basin 295

G oda na, B.A . (1985) A fricas Shared W ater Resources (L ondon , Frances Pinter).
H ansen, J., Lacis, A ., Rind , D ., Russell, L., Stone, P., Fung , I. Ruedy, R. & Lerner, J. (1984) Clim ate
sen sitivity analysis of feed back m echa nisms, in: J. Hansen & T . Takahashi (E ds) Climate P rocesses
an d Clim ate Sensitivity, G eop hysical M on ograph N o . 29 (W ashington, D C, A m erican G eop hy sical
U n ion), pp. 130 163.
H ow ell, P .P. & A llan, J.A . (1994) T he N ile, M anaging a S carce Resource (C am bridge, C am bridge
U n iversity Press).
H ulm e, M . (1990) G lobal climate ch ange and the N ile Basin, in: P.P. H ow ell & J.A . A llan (Eds) The
N ile, Resource Evaluation, Resource M anagement, H ydropolitics an d L egal Issues (L ondo n, SO A S-RG S),
pp. 59 82.
H ulm e, M . (1994) Regional clim ate ch ange scen arios based on IP CC em issions projections with som e
illustrations for A frica, A rea, 26, pp. 33 44.
H ulm e, M . & K elly, M . (1993) E xploring the linkages betw een climate ch ange and deserti cation,
E nvironm ent, 35, pp. 4 11, 39 45.
H ulm e, M . & V iner, D . (1996) A climate scen ario for the tropics; sub m itted to C lim atic Change.
H urst, H .E. & Ph ilips, P. (1933) T he N ile Basin , V ol. IV : T en-day M ean and M onth ly M ean D ischarges
of the N ile and its Tributaries (C airo, M inistry of Pu b lic W orks, Phy sical D epartm ent).
In stitute of Hydrology (IH ) (1984) A Review of the Hy drology of L ake V ictoria (Lon don , O DA and
W allingford , IH ).
K ite, G .W . (1981) R ecent ch anges in level of Lake V ictoria, Hy drological Sciences Bu lletin, 26, pp.
233 243.
K ite, G.W . & W aititu, J.K . (1981) Effects of Changing Precipitation an d E vaporation on N zoia River Flow s
an d Lake V ictoria L evels, a contribution to the W orld C lim ate Prog ram , report, 5 pp. (G enev a, W orld
M eteorological O rganisation).
L eavesley, G.H. (1994) M od elling the effects o f clim ate change on w ater resources a review , Clim atic
C hange, 28, pp. 159 177.
L eM oigne, G., B arghouti, S., Feder, G., G arbus, L . & Xie, M . (1992) Coun try Experiences w ith W ater
R esources M anagem ent: Econ om ic, Institutional, T echnological, and Environm ental Issues (W ashington,
D C , W orld B ank).
M illiman, J.D ., Broadus, J.M . & Gable, F. (1989) E nvironm ental and econom ic im plications of rising
sea level and sub siding deltas: the N ile and Beng al exam ples, A mb io, 18, pp. 340 345.
N ash, L.L . & Gleick, P.H . (1991) Sensitivity of stream ow in the C olorado basin to climatic ch anges,
Jou rnal of H ydrology, 125, pp. 221 241.
N em ec, J. & Schaake, J. (1982) Sen sitivity of w ater reso urce systems to clim ate variation, H ydrological
S cience Jou rnal, 27, pp. 327 343.
O n yeji, S.E. & Fischer, G. (1994) A n econo m ic analysis o f potential clim ate chang e in E gypt, G lobal
E nvironm ental Change 4, pp. 281 299.
Piper, B.S., Plinston , D.T. & Sutcliffe, J.V . (1986) Th e w ater b alance o f L ake Victoria, H ydrological
S cience Jou rnal, 31, pp. 25 37.
Said, R. (1993) T he River N ile. G eology, Hy drology an d U tilizatio n (O xford , Elsevier).
Santer, B.D ., W igley, T.M .L., Schlesinger, M .E . & M itchell, J.F.B. (1990) Developing C lim ate S cenarios
from Equilibrium G CM results, M ax-Pla nck In stitut fu r M eterolo gie Report N o. 47 (H am burg,
M ax-Pla nck Institute), 29 pp.
Sch aake, J.C ., R agan, R.M . & V anb largan, E.J. (1993) GIS structure for the N ile River foreca st project,
in: K . K ovar & H.P. N achtneb el (Ed s) A pplication of G eographic Inform ation S ystems in Hy drology an d
W ater R esources M anagement, IA H S Pu b . no. 211 (W allingford , IA HS), pp. 427 431.
Sene, K .J. & Plinston , D .T. (1994) A review and update of the hy drology of Lake V ictoria, H ydrological
S cience Jou rnal, 39, pp. 47 63.
Sestini, G . (1992) Im plications of climatic changes for the N ile Delta, in: L. Jeftic, J.D. M illiman &
G . Sestini (Ed s) Clim atic Change and the M editerranean (Lond on, E dw ard Arn old), pp. 535 601.
Shady, A .M ., A d am , A .M . & M oham m ed , K .A. (1994) The N ile 2002: the vision tow ard coop eration
in the N ile B asin, W ater In ternational, 19, pp. 77 80.
Shahin, M . (1985) H ydrology of the Nile Basin (Am sterdam , Elsevier).
Strzepek, K .M ., C hibo O ny eji, S., M agdy Saleh & Yates, D .N . (1995) A n assessment of integrated
climate change im pacts on Eg ypt, in: K .M . Strzepek & J.B . Sm ith (Eds ) A s Clim ate Changes:
International Imp acts and Im plication s (Ca m bridge, C am bridge U niversity Press), pp. 180 200.
Sutcliffe, J.V . & Lazen by, J.B.C. (1990) Hydrological data requirements for planning N ile m anage-
m ent, in: P.P. How ell & J.A . A llan (Eds ). The N ile, Resource Evaluation, Resource M anagement,
H ydropolitics an d L egal Issues (Lond on, SO A S/RG S), pp. 107 136.
296 D. Conway & M . Hulme

Sutcliffe, J.V . & Pa rks, Y.P. (1987) H ydrological m odelling of the Su dd and Jon glei C anal, H ydrological
S cience Jou rnal, 32, pp. 143 159.
V orosm urty, C .J. & M oore, B. (1991) M odelling b asin-scale hy drology in support of physical climate
and global b iogeochem ical studies: an exam ple using the Zam bezi R iver, in: E.F. W ood, (E d.) Lan d
S urface A tm osphere Interaction s for Clim ate M odelling (D ordrecht, N etherlands), pp. 271 312.
V orosm urty, C .J., M oore, B., G race, A .L., G ildea, M .P., M elillo, J.M ., Peterson , B.J., R astester, E.B . &
Steudler, P.A. (1989) Con tinen tal scale m odels of water balance and uvial transport an appli-
cation to South A m erica, G lobal Biog eochemical Cycles, 3, pp. 241 265.
W aterbury, J. (1979) The H ydropolitics of the N ile Valley (S yracuse, Syracuse U niversity Press).
W aterbury, J. (1982) Riverains an d L acustrin es: Tow ard International Cooperation in the Nile B asin,
D iscussion Pa per N o. 107 (Prin ceton , N J, Princeton U niversity, Research Prog ram in D evelopm ent
Studies).
W aterbury, J. (1987) L egal and institutional arrang em ents for m anaging w ater resources in the N ile
B asin, W ater Resources D evelopm ent, 3, pp. 92 103.
W eatherald, R.T. & M anabe, S. (1986) An investigation of cloud cov er chang e in response to therm al
forcin g, Clim atic Change, 8, pp. 5 23.
W h ittington, D . & M cC lelland, E . (1992) C oop eration in the N ile Basin, W ater In ternational, 17, pp.
144 154.
W igley, T .M .L. & Jones, P.D . (1985) In uences of precipitation ch anges and direct CO 2 effects on
stream ow , Natu re, 314, pp. 149 152.
W o rld B ank (1989) Sub-Sah aran A frica H ydrological A ssessment S udan , W orld Bank-U N DP , R AF /87/
030.