International Conference “Advances in Planning,...
International Conference
“Advances in Planning, Design and Management
of Irrigation Systems as Related to Sustainable Land Use”
CIE - ECOWARM
Leuven (Belgium), 14-17 September 1992
INFORMAL IRRIGATION: A SOLUTION FOR SAHELIAN COUNTRIES?
Some remarks from dcase studies in the Senegal river Delta
Pierre-Yves LE GAL
Centre de Coopération Internationale en Recherche
Agronomique pour le Developpement (CIRAD)
BP 5035 - 34032 MONTPELLIER Cedex (France)
and
Institut Senegalais de Recherches Agricoles (ISRA)
BP 240 - Saint-Louis (SénCgal)
ABSTRACT
Informa1 irrigation has been developing extensively in the Senegal River Delta since 1988.
This paper aims to analyse capacities and conditions for farmers and the private sector to replace
the state in irrigation development. It presents some results obtained from a two-year monitoring
of six farmer organizations. Despite a large variability of land area, equipment and organizJ.tion,
aimilar agronomie problems linked to the basic conception of schemes and lack of manpower lead
ta relatively low yields. The low or ne.gative profits have induced an increase of debts with the
credit bank. Farmers’ performances are lower on informa1 perimeters than on a large,-escale scheme
operated by same people. Farmers and the private sector have shown their capacity to initiate
cconomic projects given the right environment. But new design of perimeters, better definition of
everyone’s responsabihties and farmers’ training are needed to improve the present non-sustainable
situation.
1. INTRODUCTION
Development of irrigation is of great importance for Sahelian countries. A persistent drought
since 1973 and high population growth have worsened their food security balance. Mort-ver
increasing rural incomes Will be necess’ary to stem migration to the cities. Irrigation cari h.Q to
reach these targets by increasi.ng potential yields and controlling water. But irrigation is net ii
traditional technology along the Sahelian rivers where the greatest potential for its development is
found. For fifty years, large schemes have been installed by colonial,and then independant, :;htes
and managed under their control The results have been disappointing for the Sahel (Seck, !~993.;
Tricast & Blanck:, 1989) as well as elsewhere in sub-Saharan Africa (Arditi, 1986; Blanc-.Parnard,
1986; Adams, 1990). Several problems have been identified including high investment costs, poor
water effïciency and $elds, ami lack of farmer participation. Until now only 320,000 ha have been
developed in the Sahel ,, representing 11 percent of potentially irrigable land (Aviron-Violet & (xl.,
1991).
Faced with this situation and glolbal economic problems, governments and donors are t.rying
to fa\\rour small-scale and informal irrigation (which Will refer in the following text to perimeters
implemented and managed by farmers without state participation) and to transfer economic
functions from the state to farmers and the private sector (Adams & Carter, 1987; Le Moi.gne &
Barghouti, 1990). In doing SO they hope to extend land under irrigation while reducing global
production costs. This new policy was ‘adopted by Senegal in 1987 after having tested small-scale
irrigation for ten years with the help of state and non-govemmental organîzations (Diemer & Van
Der Iaan, 1987; Niasse, 1991). In the Senegal River Delta, this recent trend has induced a large
extension of informa1 irrigation. This paper presents technical and economic information
colllected
from six farmer organizations in 1990 and 1991 compared with the results obtained by the same
people on a large-scaie scheme in 1991 I After analyzing the structure, operation and performances
of these cases, we discuss the capacity of such farming systems to replace state intervention in
irrigation development and the necessary technical and institutional improvements,
2. THE CONTEXT
The Senegal River Delta covers 5000 km2 from Saint-Louis to Dagana. Hooded irrigated
rice i, the main component of present cropping systems as it is well adapted to the clayey antf salty
lowlands. Rice cultivation enjoys optimal growing conditions both during the ra!ny season from
July-Xugust to November-December and the hot dry season from February Warch to July ,
permitting annual double cropping. Rai.nfed crops have almost disappeared as anmal rainfall has
exceeded 300mm oniy six times since 1976 (Raes & Sy, 1992).
Irrigation development has followed two major stages since the introduction, under state
control, of the first schemes in a rural setting in 1957 (Le Gal & Dia, 1991), From 1965 to 1987
a parastatal Irrigation Authority, SAED (“Soc&? nationale d’amenagement et d’exploitation des
terres du delta du Senegal et des vallt?es du fleuve Senegal et de la Faléme”) was in charge of
the complete sector including design, implementation and maintenance of schemes, water
management, credit and input supplies, land preparation, paddy commercialization and milling, and
extension service. Farmer participation was only required from sowing to harvest and specifïc
organizations were set up to manage credit and input supplies between SAED and individual
farmers. During this period large-scale schemes (1000 to 2000 ha each) were implemented on
10600 ha, with total water control and electric-powered irrigation pumps (S AED, 1990). From
1975 small-scale schemes managed by farmer organizations (50 to 100 ha each) bave also been
developed on 2300 ha with total or partial water control, diesel-powered pumps, and SAED or non
governmental organization funds. Two dams have been built on the Senegal River: Diama frear
Saint-Louis which prevents ocean water flowing upriver and Manantali in Mali which controls
downriver flooding. They have been o:perating since 1988, making possible both total comrol of
the water level in the river, and annual double cropping whose extension is a major state objective
in the intensification of irrigated farming systems in the Senegal River Valley.
SAED has promoted rice cropping systems characterized by labour-saving techniques.
Dryland tillage has been always mechanized using tractors and poly-disc harrows. In addition,
broadcast seeding omo flooded fields, chemical herbicides and fertilizers bave allowed an inc:rcase
in the cultivated area per labour unit. ‘These techniques are still in use nowadays but traditional
manual harvesting has been p‘artly replaced by combine-harvesters since 1988 (Havard, 1990),
Despite average yields from 4 to 5 Tlha, high investment and operation costs have increased
costs of rice production. Local production cou1.d not compete with cheaper broken rice imported
from Asia and was highly subsidized for food security and self-sufficiency reasons. In 1987 the
state decided to limit its expenses by gradually giving up SAED functions in favor of the private
sector and farmer organizations. At first this evolution concerned credit, now managed by a semi-
private bank, CNCAS (“Caisse Nationale de CrCdit Agricole du Sénegal”), input and equipment
suppl!,, and water management on the large-scale schemes rehabilitated in 199 1 aftes 15 ta 20 ;years
of operation. The state has given a legal status recognized by CNCAS to farmer organizations
which manage perimcters and equipment, the “Groupement d’Interêt Economique” (GTE), lit has
a,lso taansfered land distribution to local structures called Rural Communities (“Communautées
Rurales”), constituted by farmers’ representatives. Once the privatization of rice milling is
completed in 1993 and alll of its schemes have been rehabilitated, it Will control only regionaLleve1
water management and research and extension services.
This turnover of responsabilities has already induced large socio-economic changes: the
number of CIE exploded from 52 in 1986 to 1165 in 1989. This growth was accompanied by
increases in credit (Table 1) and equipment. For example the number of tractors went up frcrm 80
in 1985 to 175 in 1992, and pumps from 320 to 1510 during the same period for ail the Se.negal
River Valley. Rice production rose 60% between 1987 and 1990. An exhaustive survey conducted
in 1991 enumerates 5 14 informai perimeters in the Delta, covering 18000 ha and using 595 diesel-
powered pumps (Ba (SC Havard, 1992). As SAED has not implemented any new schemes since
1987, these data indica.te that 15700 new hectares have been put under irrigation in four years. But
during the same period CNCAS has noted a quick drop in the loan repayment rate which reached
only 58% in 1990-1991 and 17% for the rainy season 1991-1992 (Table 1). Although no detailed
statistics are available this diffrcult financial situation seems to be correlated with the development
of informa1 irrigation At a micro-econo,mic level our survey of six GIE provides some information
to discuss this region$t hypothesis and the future of informa1 irrigation in the Delta
Table 1: Evolution of loans distributed by CNCAS and repayment rate
(1 US$=300 FCFA)
-.-
--
--.l
“-”
---_..-l.
.---
--.-<“-
3. MATERIALS AN-D METHODS
The main objective of this survey has been to colle& preliminary technical. and economic
information about informa1 irrigation to design a research and development program adapted to the
problems encountered by farmers. Tt was carried out in 1990 and 1991 in ;i single village
representative of the local rice-based farming systems, Diawar, and was part of a global farming
system research program conducted sinc.e 1987. The village has 134 production U~$S, with. 34 GIE
in 1990 and 62 in 1992. Most of the farmers cultivate on different perimeters: they own one or
several plots from 2 :o 4 ha per farmer, on a large-scale scheme rehabilitated by SAED in 1990
and operated during the rainy season 1991. They are also members of one or more GIE ttihich
manage one or two perimeters from 5 to 100 ha.
Six GIE have ‘heen chosen according to their differences in land and labour size. Their
perimeters are dispersed within a radius of 30 km because of the lack of land closer to the village.
~11 of them are made up of native farmers whereas some GIE in the Delta gather people from
varions originsr civil servants, jobless young graduates, traders or religious chiefs.
Mostofthe
farmers are illiterate or unaware of collecting information on their own activities. Research has
therefore had to depend on a permanent ,monitoring of farmers’ technical activities over the cultural
cycle (field trips and farmers’ declarations) and a set of economic data from different sources: btlls,
SAED receipts from paddy sales and other records collected by farmers on notebooks, calendars,
etc. s Iand areas were estimated by farmers as well as production, counted in sacks. The destination
of sacks was also noted, including cost of hired combine-harvesters, stock for seeds and borne
consumption: religions gifts, and sales for debt repayment and cash. This work wa? complemented
by interviews with GlE leaders to better understand the decision-making processes and problcms
encountered Based on this information technical choices, yields and balance sheets have been
analysed. They have been compared with data collected on 12 farmers’ plots cultivated on the
large-scale scheme during the 1991 rainy season, using the same methods of survey. These people
are all members of the six GIE sample.
4. REXJLTS
Creation and operation of informa1 perimeters usually follow these different steps: (1)
creation of a GIE, (2) allocation of land by Rural Communities, (3) purchase of a pump and
creation of the perimeter (cash or credit), (4) request for a short-term loan to CNCAS for
operation, (5) commercialization of paddy to SAED or private traders and repayment of loan to
the CNCAS.
Structure and organization
The creation of a GIE is initiated by one to three leaders who obtain the start-up funds. TO
tope with labour demand as well as for social reasons they join together a variable number of
people (in the sample, from 3 to 39 members), mainly family related. In two GTE fiired manpower
is also used over the cropping season. Total and per unit labour acreage vary widety from one GIE
to another but their tnain characteristic of size and social structure is instability (Table II),
Depending on their economic returns, cash supplies and land resources, GIE modify the
implemented and cultivated acreage every year. They cari also manage several perimeters (fve
GTE) or give one up because of saliniby or weed infestation. People also tend to move from one
GIE t,o another when a budget detïcit occurs or when they want to create their oun GIE. Without
a cadastre these trends are difficult to quantify on a global scale.
Land distribution between GIE members and management follows different forms. In the
sample one of them has adopted an individual mode similar to the situation on large-scale schemes.
Memhers hold individual plots while the GIE assumes the collective management of credit, input
supplies and water. After harvest and before commercialization the expenses are estimated for each
member and collecte<d hy the GIE to repay the bank and other debts.
However, three GIE are managed on a collective mode. In these cases, leaders assume
management responsability for credi t, work organization, commercialization ~ and profit
distribution, while members contribute their labour for seeding, weeding and fertilization, Exyenses
and profits are calculated as a whole. Profit is then distributed among members according 90 a
Framework decided by leaders whereas hired manpower is paid on a fixed basis in cash or paddy.
Two GIE show intermediate situations where part of the land area is collectively managed to caver
general expenses and the rest is distributed among members. As in the coll.ective mode these
arrangements benefit m,ainly the leaders who manage collective profits and cari rcpay their iinitial
investments. But the leaders also assume most of the economic risk since other members may
abandon the GIE in case of failure.
Table II: Characteristics of land area and mannower (2-vear average on six GIE)
GIE
AK
YG SD
EM
MY
W
members
29
39
perimeters (n)
1
1
land area (ha)
77
102
area/member (ha)
2.7
2.6
1992
members
24
5
5
10
12
21
perimeters (n)
21
3
2 2
1
2
land area (ha)
83
17
71 34
43
184
area/member (ha)
3.5
3.4
14.2 3.4
3.6
8.8
area/perimeter (ha)
6
36 17
43
92
Irrigation and trop management
Technically the perimeters present limiting conditions for flooded irrigated rice cultivation,
To ccspe with farmerr;” low cash availability and lack of access to long-term loans, their design is
very simple and they are implemented ;at low cost, without previous soi1 and topography stildles.
They usually consist of one main canal and sometimes secondary canais designed according te the
natural topography arrd built by hired graders. Dikes and levees are not compacteti and there is no
leveling of plots. Instead, corrections of slope are attempted by dividing the plots but significant
differences in level cari still remain. Finally, no drainage network is implemented: some natural
depressions cari be used to receive drainage water but usually farmers just leave the flood water
to evaporate, This situation is particulary dangerous on the salty soils of the Delta (Loyer, 1989).
Pumps are placed on river banks anywhere between a few meters and two kilometers from
the perimeters. Pumping height does not exceed four meters. Engine powers vary from 45 to 70
hp and pump theorical overflow from 450 to 750 m3/h. Overflow per hectare varies from 1 I to
23 m3lhlha for multiple reasons. At first farmers choose their equipment according to mid-term
objectives of land area which cari differ from their current needs or means. They also bave
difficulties in fitting t’he characteristics of pumps and engines to overflow of canais ami water losses
which are usually unknown.
Cultural techniques adopted on these perimeters follow the general pattern described in
section 2. In Diawar harvest is mostly mechanized on informa1 and large-scale schemes as five
combine-harvesters belong to farmer organizations in the village. But the general management of
cropping systems on the six GIE is compromised by t.he high ratio of land to labour unit and the
frequent manpower shortages for a11 the operations still carried out by hand.
Weed control is a major technical problem. It depends mainly on the time of herbicide
application and water control in this ricr: cropping system (De Datta & Herdt, 1983). Farmers face
an accumulation of constraints: they often perform operations too late due to a lack of manpower;
they carmot, drain their plots before spra.ying as recommended when using propan arrd 2-43; they
reduce the amounts of herbicides to half of the suppliers’ recommended rate for economic rasons;
they cannot use flooding as a method of control without levelling. Furthermore elevated areas in
plots favor weed infestation. Efficiency of chemical fertilizers is also limited in this contexr..
Average yield comparison between the six GIE and the large-scale scheme, respectiveiy 3.4
and 5.6 T/ha, emphazise the difficulties of trop management on informal perimeters. This
difference is not due to input consumlption, rather similar in the two situations (Table III), It
appears to be mainly correlated with the specifïc natural conditions of petimeters (especially
salinity, topography, weed and bird control), the limited water control, and farmers” ability to adapt
to these new problenns. Furthermore average yields showed a 35% decrease on four perirneters
between 1990 and 1991, with respectively 3.7 and 2.4 T/ha. If further surveys confirrn this
tendency the future of informa1 irrigation in its present form would be at risk
Table HT: Comparison of input consumption between informa1 irrigation (2-year average on six
GTE) and large-scale scheme (1991 rainy season average on 12 plots)
seeils
Wha)
145
141
Crop calendar
The GIE surv+eyed have given Ipriority to the rainy season, consïdered as the “normal”’
season by farmers since it has been the only one available for thirty years (Table IV). The hot-dry
season or double cropping of rice are scarcely practiced for three reasons.: (1) farmers prefer not
to take the risk of harvesting under rain as sowing and harvest are often spread over three or four
weeks in a same GIE; (2) they have problems to get short-term loans in time to sow in Febru‘ary-
March; (3) they have difficulty managing the bottle-necks from the hot-dry season rice harvest to
the rainy season rice sowing, especially when they have to hire combine-harvesters and tractors.
Table IV: Time and duration of sowing and harvest (2-year average for six GIE)
TO avoid these problems and to increase cultivated land per labour unit, in 1991 farmers
adopteA an intermediate solution they term “inter-season” because rice is souri in May nnd
harvested in Gctober after the rains. For example in 1991-1992 two GIE used one perimeter during
the rainy season and another one during the inter-season, whereas their members were practising
double cropping on their own plot on the large-scale scheme. But many agronomie problems arise
in inter-season cropping: maturation of the 120 days-duration variety under rain, supplementary
irrigation requirements, losses to birds and poor quality of paddy. As farmers bave to wait fur the
rainy season to end and for combine-harvesters to be available, paddy is generally harvested 30 to
60 davs after its physiological maturity which lead to poor milling yields.
Economies
Due to higher expenses and lowler yields, the average net revenue per hectare of the GIE
surveyed is low compared with results o’btained on the large-scale scheme in the 1991 rainy sason
(Table V). Differences in expenses are mainly related to water costs, 250 and 170 US$/ha
respectively, and loar interests, 80 and 45 US$/ha respective&. On the large-scalc scheme water
rosts are tïxed per land unit but do not caver a11 the maintenance costs and depreciation of the large
initial investment whivh cari vary from 3,500 US$/ha for a rehabilitation to 13,000 US$/ha for a
new scheme. On informai perimeters farmers take ail the costs in charge: installation and
maintenance paid using short-term loans (100 US$/ha on average), pump depreciation (85 US$!‘ha
over three years) and operation (85 US$/ha).
Higher financral costs incurred by GIE result from inadequate cash assets relative to
seasonal needs, amounting to about 600 US$/ha. Their demands for short-term loans are often
over-sized to caver expenses that CNCAS does not normally finance (for examplc purchase of a
car) or to repay input suppliers who agree to lend them cash for the GIE counterpart funding
Tecluired for bank loans (15 % of the loan amount) (B&ières & al., 1991).
High production costs for paddy ‘(240 US$/T on average) are another consequence of thcse
poor economic and technical results. Paddy price Will be govemmentally frxed by rhe governrncnr
at 280 US$ per metric tonne until 1993, but Will then decrease to 250 US$/
privatized, and probably less for poor quality grains. The present performances of informai
perimeters could not sustain such a strategy. Moreover two GIE present negative profits after two
years of operation, respectively 13,500 US$ and 23,000 US$, and they have beern unable to yay
back their loans in time to CNCAS.
Table V: Comparison of economics results between informal irrigation (2-year averagc on six GiE)
and large-scale scheme (1991 rainy season average on 12 plots)
5. DISCUSSION
Farmers have used the conjunction of fïve factors to develop informal irrigation: (1) land
resources free of traditional tenure constraints, managed by their own representatives, and easily
irrigable along the ri>Jers, have been alvailable. (2) Water resources have been adequate and
available throughout tbe year. (3) A financial service has been operating with large amounts of
money and few coercve rules. (4) Farrner organizations have been able to adope a legal status
recognized by CNCAS . (5) Labour-saving cultural techniques have allowed a large cultivated land
area per labour unit.
GIE strategies and performances
Farmers have developped informa1 irrigation at first to extend their cultivated land a.nd to
prevent installation of urban people and iagro-business companies. Being short of capital and labour,
they have adopted extensive strategies of land occupation but their technical choices
(mechanization) and eonstraints (pumping, partial water control) have resulted in large expenses
and poor conditions of production.
Profitability of these operations appears insufficient or even negative. Yields and revemtes
are clearly lower than performances obtained on the large-scale scheme. However this comparison
would be different if initial investment and rehabilitation of large-scale schemes, still supported by
the state, were includ;d in farmers’ expenses. For example state involvement was estimated at. 85%
of the total cost of production for rice paddy in 1982 (Engelhard & Ben Abdallah, 1986). No recent
study has been available since the turnover of responsabilities from the state to farmers.
Three causes cari be suggested to explain the poor results of informa1 irrigation: the
technical limitations trfperimeters, the difficulties farmers have controlling, over large areas, weeds
and fertilization which mainly determine yields, and the lack of rigour in economic management”
Farmers are iaced with a difficult challenge if they want to keep their labour-saving but
costl‘l techniques. Possibilities of reducing expenses without decreasing yields are Iimitecl in this
farming system. (1) ‘Nater consumption is not really controlled due to poor design of perimeters.
(2) Replacing combine-harvesters by hired labour would save only 40 US$lha (respectivell:y 170
US$/ha and 130 US$/ha for 4 Tlha yield) with present prices of paddy and mechanized harvesting
(15 % of production) To fit the production cost of paddy with price reduction and stabilization or
to increase farmer rIzvenue, farmers slhould change their economic behaviour towards intensive
strategies where high yields and double cropping Will replace land occupat.ion objectives. For
example 6.5 T/ha yic:ld would be necessary to rank with the present production COS~ of the large-.
scale scheme withour increasing expenses of the 6 GIE surveyed.
This first diagnosis gives some suggestions in designing a research and development
progr-am on informa1 irrigation in the Delta. The main technical challenge is to design surface
irrigation fitting m?.rltiple objectives: total water control; easy maintenance; adaptation to
mechanization and double cropping; and costs compatible with farmers’ economic capacity and
available long-term loans, which meams between 1500 and 3000 US$/ha, Fcological problems
which necessitate pl&annual studies, like the risk of salinization, rising water-table, and fertility
drop, need special attention since the future of informal irrigation is linked to them. Ether
probiems, not specifïc to informa1 irrigation, must have suitable solutions: water and cropping
management, and development of technical and economic Decision Support Systems.
GIE socio-economic environment
By financing farmers and the private service sector, CNCAS has greatly sustained the
development of informai irrigation. It has been flexible for loan attributions and repayments, but
it ha5 not invested in profitability evaluation and economic monitoring of the operations fananeed.
Tt does not know exastly which perimet’ers have been unprofitable and why. This survey shows thal
farmers have taken advantage of these features to collect money they were needing but did no1
have. and that some 3f them have met problems to pay back their debts. Similar results on a lasger
numher of GIE coul’d explain the present critical position of CNCAS.
How has C&AS evaluated its risks while financing informal irrigation is still unknawn.
But present problems and relationships; between GIE and CNCAS suggest that it should improve
its knowledge of its customers’ projects, and be involved in their economic management through
advising and analyses of results.
State implication in the development of informa1 irrigation has been very limited since ‘Rural
Comrnuni ties has been in charge of land distribution, and farmers have not received advice adapted
to the complex problems they have encountered. Difficulties collecting surface drainage water on
the Delta salty soils, ‘equity in regional land and water distribution, and protection of areas for
ecological reasons, suggest that the state should conceive and apply legislation which specifies the
frame in which informa1 irrigation cari be developed. Training, extension and rcsearch facilities
should also be providied under its control, with better involvement of farmers and the private xector
in the structures in charge of these functions.
6. CONCLUSION
The example of the Senegal River Delta shows that, glven the right environment, farmers
and the private sector cari take: initiatives to replace the state in assuming economic functions and
developing irrigation. This first result is important for Sahelian countries where irrigation has
traditionally been synonymous with state management. But the problems encountered already in
the short-term both ?y the credit bank: and a sample of GIE, and the poor results of informa1
irrigation compared with a large-scale scheme, underline the fragility of the current situation, Two
priorities cari be drawn from this ex:perience: the need for better definition and division of
responsabilities betwc:cn farmers,the state and the private sector, and the need for training farmers
confronted with complex technical and economic problems. Research has an important role to play
in creating sustainab’ie and profitable solutions adapted to this new conte.xt, and to describe and
understand its rapid evolution.
ACKNOWLEDGEMENTS
I would like to thank Mohamed Diop for collecting data, Michaël Dingkhun, Tom Randolph
and Dirk Raes for comments on the draft of this paper, and Sally Moffat for correcting the proofs.
REFERENCES
Adams W.M., 1990 ‘How small is beautiful? Scale, control and sucess in Kerryan irrigation”.
World Development. Vo1.18, n”10, pp. 1309-1323.
Adams W.M., Carter R., 1987. ‘Small scale irrigation in sub-Saharian Africa’. Proaress in
Phvsical Geography. ‘Vol.11, pp. l-27.
Arditi C,, 1986. ‘Quelques réflexions socio-economiques sur la riziculture irriguce dans le Nord-
Cameroun (SEMRY 1 et SEMRY II)‘, Documents Svstèmes Agraires n”6. 435-444.
Aviron-Violet J., Ibo B.D., Soumaila A., Van Steekelenburg P.N.G., Waldsteirr A., 199 1 1 ‘Le
developpement des cultures irriguees clans le Sahel. Contrainte de la politique des irrigations et
strategie paysanne. R;3-pport de synthèse’. OCDE-CIISS, 216 p,
Ba T , Havard M., !992. ‘Les groupes motopompes et les périmétres prives du Delta du fleuve
Senégal’. ISRA. 4q p.
BtTliéres J-F., Havarc3 M., Le Gal, P.Y., 1991. ‘Le financement de l’agriculture irriguCe dans le
Delta du fleuve S&éga.l: Intérêts et dCrives du credit bancaire’. Paper presented at the
International CIRAD - O.S.IJ. - U.01. Workshop “Finance and Rural Development in West
Africa” . Ouagadougou. 2 l-25 Octobre 199 1.
Blanc-Pamard Ch., 1986. ‘Autour du riz, le difficile face-à-face des paysans et de la. SQMALAC
dans ‘Za cuvette du lac Alaotra (Hautes Terres centrales de Madagascar)‘. Documents Svst&mes
A@res n”6. 461-479
De Datta S, K , , Herdt ‘S” W “, 1983. ‘Weed control technology in irrigated rice’. In : Proceedings of
the Conference on “Weed Control in Rice”. 31 August - 4 September 1981. IRRI-JWSS,
Diemer G., Van Der Iaan E.., 1987. ‘]L’irrigation au Sahel’. Ed. Karthala-CTA. 226 p.
Engelhard Ph.) .Ben Abdallah T., 1986. ‘Enjeux de l’après-barrage’. ENDA - Ministère de la
Coopt%ation. 632 p.
avard M., 1990. ‘Etude et &a.luation des systèmes m&z$Cs: l’exemple du delta du fleuve
SénCgal’ . Cahiers de ta Recherche-DCveloppement n”28, 17-32.
IX Gal P.Y. ~ Dia I., 1991. ‘Le dCsen,gagement de 1’Etat et ses conséquences dans le Deita du
Fleube %nCgal’. In “La vallée du fleuve SénCpal. Evaluations et oerwectives d’une d&ennie
d’am&nagements (1980-19901” B. Crousse, P. Mathieu, S.M. Seck (E~S). Karthala, Paris. pp% 160-
174.
Le Moigne G. Y Bargbouti S., 1990. ‘Irrigation in sub-Saharan Africa: the development of @clic
and private systems’ World Bank Technical Paper n”123. 99 p.
Loyer J-Y,, 1989. ‘Les sols salés de la basse vallée du fleuve S&ga,l. Caract&isation,
distribution et évolution sous cultures’, Etudes et Thèses. ORSTOM. 137 p.
Niasse M., 1991. ‘h:s @rimètres irrigu& villageois vieillissent mal. Les paysans se d&engagent-
ils en même temps que la SAED?’ In “La val& du fleuve SCnCPal. Evaluations et oersoectives
d*une dCcennie d’aménagements (1980-1990)” B. Crousse, P. Mathieu, S.M. Seck [E~S).
Karthala, Paris. pp. 97-115.
Raes D,, Sy B., 199’2,. ‘Analyse de la pluviomCtrie et de l’&apotranpiration de la zone du Delta
du fleuve Sgnégal’. Projet Gestion de l’Eau, Bulletin Technique n”5. SAED”
SAED, 1990. ‘La SAED a 25 ans’. SAED/CSE.
Seck S.M., 1991. ‘Sur la dynamique de l’irrigation dans la vallée du fleuve‘. In “La val& du
fle”ue Sénégal.- Evaluations et perspectives d’une dt?cennie d’amenagements [1980-1990)” B,
.-L .-.L-----1.
Crousse, P. Mathieu, S.M. Seck (E~S). Karthala, Paris. pp. 17-43.
Tricart J., Blanck J. P., 1989. ‘L’Office du Niger, mirage du développement au Mali’. ,4nnales
de GéograDhie, n”549, pp. 567-587.
“Advances in Planning, Design and Management
of Irrigation Systems as Related to Sustainable Land Use”
CIE - ECOWARM
Leuven (Belgium), 14-17 September 1992
INFORMAL IRRIGATION: A SOLUTION FOR SAHELIAN COUNTRIES?
Some remarks from dcase studies in the Senegal river Delta
Pierre-Yves LE GAL
Centre de Coopération Internationale en Recherche
Agronomique pour le Developpement (CIRAD)
BP 5035 - 34032 MONTPELLIER Cedex (France)
and
Institut Senegalais de Recherches Agricoles (ISRA)
BP 240 - Saint-Louis (SénCgal)
ABSTRACT
Informa1 irrigation has been developing extensively in the Senegal River Delta since 1988.
This paper aims to analyse capacities and conditions for farmers and the private sector to replace
the state in irrigation development. It presents some results obtained from a two-year monitoring
of six farmer organizations. Despite a large variability of land area, equipment and organizJ.tion,
aimilar agronomie problems linked to the basic conception of schemes and lack of manpower lead
ta relatively low yields. The low or ne.gative profits have induced an increase of debts with the
credit bank. Farmers’ performances are lower on informa1 perimeters than on a large,-escale scheme
operated by same people. Farmers and the private sector have shown their capacity to initiate
cconomic projects given the right environment. But new design of perimeters, better definition of
everyone’s responsabihties and farmers’ training are needed to improve the present non-sustainable
situation.
1. INTRODUCTION
Development of irrigation is of great importance for Sahelian countries. A persistent drought
since 1973 and high population growth have worsened their food security balance. Mort-ver
increasing rural incomes Will be necess’ary to stem migration to the cities. Irrigation cari h.Q to
reach these targets by increasi.ng potential yields and controlling water. But irrigation is net ii
traditional technology along the Sahelian rivers where the greatest potential for its development is
found. For fifty years, large schemes have been installed by colonial,and then independant, :;htes
and managed under their control The results have been disappointing for the Sahel (Seck, !~993.;
Tricast & Blanck:, 1989) as well as elsewhere in sub-Saharan Africa (Arditi, 1986; Blanc-.Parnard,
1986; Adams, 1990). Several problems have been identified including high investment costs, poor
water effïciency and $elds, ami lack of farmer participation. Until now only 320,000 ha have been
developed in the Sahel ,, representing 11 percent of potentially irrigable land (Aviron-Violet & (xl.,
1991).
Faced with this situation and glolbal economic problems, governments and donors are t.rying
to fa\\rour small-scale and informal irrigation (which Will refer in the following text to perimeters
implemented and managed by farmers without state participation) and to transfer economic
functions from the state to farmers and the private sector (Adams & Carter, 1987; Le Moi.gne &
Barghouti, 1990). In doing SO they hope to extend land under irrigation while reducing global
production costs. This new policy was ‘adopted by Senegal in 1987 after having tested small-scale
irrigation for ten years with the help of state and non-govemmental organîzations (Diemer & Van
Der Iaan, 1987; Niasse, 1991). In the Senegal River Delta, this recent trend has induced a large
extension of informa1 irrigation. This paper presents technical and economic information
colllected
from six farmer organizations in 1990 and 1991 compared with the results obtained by the same
people on a large-scaie scheme in 1991 I After analyzing the structure, operation and performances
of these cases, we discuss the capacity of such farming systems to replace state intervention in
irrigation development and the necessary technical and institutional improvements,
2. THE CONTEXT
The Senegal River Delta covers 5000 km2 from Saint-Louis to Dagana. Hooded irrigated
rice i, the main component of present cropping systems as it is well adapted to the clayey antf salty
lowlands. Rice cultivation enjoys optimal growing conditions both during the ra!ny season from
July-Xugust to November-December and the hot dry season from February Warch to July ,
permitting annual double cropping. Rai.nfed crops have almost disappeared as anmal rainfall has
exceeded 300mm oniy six times since 1976 (Raes & Sy, 1992).
Irrigation development has followed two major stages since the introduction, under state
control, of the first schemes in a rural setting in 1957 (Le Gal & Dia, 1991), From 1965 to 1987
a parastatal Irrigation Authority, SAED (“Soc&? nationale d’amenagement et d’exploitation des
terres du delta du Senegal et des vallt?es du fleuve Senegal et de la Faléme”) was in charge of
the complete sector including design, implementation and maintenance of schemes, water
management, credit and input supplies, land preparation, paddy commercialization and milling, and
extension service. Farmer participation was only required from sowing to harvest and specifïc
organizations were set up to manage credit and input supplies between SAED and individual
farmers. During this period large-scale schemes (1000 to 2000 ha each) were implemented on
10600 ha, with total water control and electric-powered irrigation pumps (S AED, 1990). From
1975 small-scale schemes managed by farmer organizations (50 to 100 ha each) bave also been
developed on 2300 ha with total or partial water control, diesel-powered pumps, and SAED or non
governmental organization funds. Two dams have been built on the Senegal River: Diama frear
Saint-Louis which prevents ocean water flowing upriver and Manantali in Mali which controls
downriver flooding. They have been o:perating since 1988, making possible both total comrol of
the water level in the river, and annual double cropping whose extension is a major state objective
in the intensification of irrigated farming systems in the Senegal River Valley.
SAED has promoted rice cropping systems characterized by labour-saving techniques.
Dryland tillage has been always mechanized using tractors and poly-disc harrows. In addition,
broadcast seeding omo flooded fields, chemical herbicides and fertilizers bave allowed an inc:rcase
in the cultivated area per labour unit. ‘These techniques are still in use nowadays but traditional
manual harvesting has been p‘artly replaced by combine-harvesters since 1988 (Havard, 1990),
Despite average yields from 4 to 5 Tlha, high investment and operation costs have increased
costs of rice production. Local production cou1.d not compete with cheaper broken rice imported
from Asia and was highly subsidized for food security and self-sufficiency reasons. In 1987 the
state decided to limit its expenses by gradually giving up SAED functions in favor of the private
sector and farmer organizations. At first this evolution concerned credit, now managed by a semi-
private bank, CNCAS (“Caisse Nationale de CrCdit Agricole du Sénegal”), input and equipment
suppl!,, and water management on the large-scale schemes rehabilitated in 199 1 aftes 15 ta 20 ;years
of operation. The state has given a legal status recognized by CNCAS to farmer organizations
which manage perimcters and equipment, the “Groupement d’Interêt Economique” (GTE), lit has
a,lso taansfered land distribution to local structures called Rural Communities (“Communautées
Rurales”), constituted by farmers’ representatives. Once the privatization of rice milling is
completed in 1993 and alll of its schemes have been rehabilitated, it Will control only regionaLleve1
water management and research and extension services.
This turnover of responsabilities has already induced large socio-economic changes: the
number of CIE exploded from 52 in 1986 to 1165 in 1989. This growth was accompanied by
increases in credit (Table 1) and equipment. For example the number of tractors went up frcrm 80
in 1985 to 175 in 1992, and pumps from 320 to 1510 during the same period for ail the Se.negal
River Valley. Rice production rose 60% between 1987 and 1990. An exhaustive survey conducted
in 1991 enumerates 5 14 informai perimeters in the Delta, covering 18000 ha and using 595 diesel-
powered pumps (Ba (SC Havard, 1992). As SAED has not implemented any new schemes since
1987, these data indica.te that 15700 new hectares have been put under irrigation in four years. But
during the same period CNCAS has noted a quick drop in the loan repayment rate which reached
only 58% in 1990-1991 and 17% for the rainy season 1991-1992 (Table 1). Although no detailed
statistics are available this diffrcult financial situation seems to be correlated with the development
of informa1 irrigation At a micro-econo,mic level our survey of six GIE provides some information
to discuss this region$t hypothesis and the future of informa1 irrigation in the Delta
Table 1: Evolution of loans distributed by CNCAS and repayment rate
(1 US$=300 FCFA)
-.-
--
--.l
“-”
---_..-l.
.---
--.-<“-
3. MATERIALS AN-D METHODS
The main objective of this survey has been to colle& preliminary technical. and economic
information about informa1 irrigation to design a research and development program adapted to the
problems encountered by farmers. Tt was carried out in 1990 and 1991 in ;i single village
representative of the local rice-based farming systems, Diawar, and was part of a global farming
system research program conducted sinc.e 1987. The village has 134 production U~$S, with. 34 GIE
in 1990 and 62 in 1992. Most of the farmers cultivate on different perimeters: they own one or
several plots from 2 :o 4 ha per farmer, on a large-scale scheme rehabilitated by SAED in 1990
and operated during the rainy season 1991. They are also members of one or more GIE ttihich
manage one or two perimeters from 5 to 100 ha.
Six GIE have ‘heen chosen according to their differences in land and labour size. Their
perimeters are dispersed within a radius of 30 km because of the lack of land closer to the village.
~11 of them are made up of native farmers whereas some GIE in the Delta gather people from
varions originsr civil servants, jobless young graduates, traders or religious chiefs.
Mostofthe
farmers are illiterate or unaware of collecting information on their own activities. Research has
therefore had to depend on a permanent ,monitoring of farmers’ technical activities over the cultural
cycle (field trips and farmers’ declarations) and a set of economic data from different sources: btlls,
SAED receipts from paddy sales and other records collected by farmers on notebooks, calendars,
etc. s Iand areas were estimated by farmers as well as production, counted in sacks. The destination
of sacks was also noted, including cost of hired combine-harvesters, stock for seeds and borne
consumption: religions gifts, and sales for debt repayment and cash. This work wa? complemented
by interviews with GlE leaders to better understand the decision-making processes and problcms
encountered Based on this information technical choices, yields and balance sheets have been
analysed. They have been compared with data collected on 12 farmers’ plots cultivated on the
large-scale scheme during the 1991 rainy season, using the same methods of survey. These people
are all members of the six GIE sample.
4. REXJLTS
Creation and operation of informa1 perimeters usually follow these different steps: (1)
creation of a GIE, (2) allocation of land by Rural Communities, (3) purchase of a pump and
creation of the perimeter (cash or credit), (4) request for a short-term loan to CNCAS for
operation, (5) commercialization of paddy to SAED or private traders and repayment of loan to
the CNCAS.
Structure and organization
The creation of a GIE is initiated by one to three leaders who obtain the start-up funds. TO
tope with labour demand as well as for social reasons they join together a variable number of
people (in the sample, from 3 to 39 members), mainly family related. In two GTE fiired manpower
is also used over the cropping season. Total and per unit labour acreage vary widety from one GIE
to another but their tnain characteristic of size and social structure is instability (Table II),
Depending on their economic returns, cash supplies and land resources, GIE modify the
implemented and cultivated acreage every year. They cari also manage several perimeters (fve
GTE) or give one up because of saliniby or weed infestation. People also tend to move from one
GIE t,o another when a budget detïcit occurs or when they want to create their oun GIE. Without
a cadastre these trends are difficult to quantify on a global scale.
Land distribution between GIE members and management follows different forms. In the
sample one of them has adopted an individual mode similar to the situation on large-scale schemes.
Memhers hold individual plots while the GIE assumes the collective management of credit, input
supplies and water. After harvest and before commercialization the expenses are estimated for each
member and collecte<d hy the GIE to repay the bank and other debts.
However, three GIE are managed on a collective mode. In these cases, leaders assume
management responsability for credi t, work organization, commercialization ~ and profit
distribution, while members contribute their labour for seeding, weeding and fertilization, Exyenses
and profits are calculated as a whole. Profit is then distributed among members according 90 a
Framework decided by leaders whereas hired manpower is paid on a fixed basis in cash or paddy.
Two GIE show intermediate situations where part of the land area is collectively managed to caver
general expenses and the rest is distributed among members. As in the coll.ective mode these
arrangements benefit m,ainly the leaders who manage collective profits and cari rcpay their iinitial
investments. But the leaders also assume most of the economic risk since other members may
abandon the GIE in case of failure.
Table II: Characteristics of land area and mannower (2-vear average on six GIE)
GIE
AK
YG SD
EM
MY
W
members
29
39
perimeters (n)
1
1
land area (ha)
77
102
area/member (ha)
2.7
2.6
1992
members
24
5
5
10
12
21
perimeters (n)
21
3
2 2
1
2
land area (ha)
83
17
71 34
43
184
area/member (ha)
3.5
3.4
14.2 3.4
3.6
8.8
area/perimeter (ha)
6
36 17
43
92
Irrigation and trop management
Technically the perimeters present limiting conditions for flooded irrigated rice cultivation,
To ccspe with farmerr;” low cash availability and lack of access to long-term loans, their design is
very simple and they are implemented ;at low cost, without previous soi1 and topography stildles.
They usually consist of one main canal and sometimes secondary canais designed according te the
natural topography arrd built by hired graders. Dikes and levees are not compacteti and there is no
leveling of plots. Instead, corrections of slope are attempted by dividing the plots but significant
differences in level cari still remain. Finally, no drainage network is implemented: some natural
depressions cari be used to receive drainage water but usually farmers just leave the flood water
to evaporate, This situation is particulary dangerous on the salty soils of the Delta (Loyer, 1989).
Pumps are placed on river banks anywhere between a few meters and two kilometers from
the perimeters. Pumping height does not exceed four meters. Engine powers vary from 45 to 70
hp and pump theorical overflow from 450 to 750 m3/h. Overflow per hectare varies from 1 I to
23 m3lhlha for multiple reasons. At first farmers choose their equipment according to mid-term
objectives of land area which cari differ from their current needs or means. They also bave
difficulties in fitting t’he characteristics of pumps and engines to overflow of canais ami water losses
which are usually unknown.
Cultural techniques adopted on these perimeters follow the general pattern described in
section 2. In Diawar harvest is mostly mechanized on informa1 and large-scale schemes as five
combine-harvesters belong to farmer organizations in the village. But the general management of
cropping systems on the six GIE is compromised by t.he high ratio of land to labour unit and the
frequent manpower shortages for a11 the operations still carried out by hand.
Weed control is a major technical problem. It depends mainly on the time of herbicide
application and water control in this ricr: cropping system (De Datta & Herdt, 1983). Farmers face
an accumulation of constraints: they often perform operations too late due to a lack of manpower;
they carmot, drain their plots before spra.ying as recommended when using propan arrd 2-43; they
reduce the amounts of herbicides to half of the suppliers’ recommended rate for economic rasons;
they cannot use flooding as a method of control without levelling. Furthermore elevated areas in
plots favor weed infestation. Efficiency of chemical fertilizers is also limited in this contexr..
Average yield comparison between the six GIE and the large-scale scheme, respectiveiy 3.4
and 5.6 T/ha, emphazise the difficulties of trop management on informal perimeters. This
difference is not due to input consumlption, rather similar in the two situations (Table III), It
appears to be mainly correlated with the specifïc natural conditions of petimeters (especially
salinity, topography, weed and bird control), the limited water control, and farmers” ability to adapt
to these new problenns. Furthermore average yields showed a 35% decrease on four perirneters
between 1990 and 1991, with respectively 3.7 and 2.4 T/ha. If further surveys confirrn this
tendency the future of informa1 irrigation in its present form would be at risk
Table HT: Comparison of input consumption between informa1 irrigation (2-year average on six
GTE) and large-scale scheme (1991 rainy season average on 12 plots)
seeils
Wha)
145
141
Crop calendar
The GIE surv+eyed have given Ipriority to the rainy season, consïdered as the “normal”’
season by farmers since it has been the only one available for thirty years (Table IV). The hot-dry
season or double cropping of rice are scarcely practiced for three reasons.: (1) farmers prefer not
to take the risk of harvesting under rain as sowing and harvest are often spread over three or four
weeks in a same GIE; (2) they have problems to get short-term loans in time to sow in Febru‘ary-
March; (3) they have difficulty managing the bottle-necks from the hot-dry season rice harvest to
the rainy season rice sowing, especially when they have to hire combine-harvesters and tractors.
Table IV: Time and duration of sowing and harvest (2-year average for six GIE)
TO avoid these problems and to increase cultivated land per labour unit, in 1991 farmers
adopteA an intermediate solution they term “inter-season” because rice is souri in May nnd
harvested in Gctober after the rains. For example in 1991-1992 two GIE used one perimeter during
the rainy season and another one during the inter-season, whereas their members were practising
double cropping on their own plot on the large-scale scheme. But many agronomie problems arise
in inter-season cropping: maturation of the 120 days-duration variety under rain, supplementary
irrigation requirements, losses to birds and poor quality of paddy. As farmers bave to wait fur the
rainy season to end and for combine-harvesters to be available, paddy is generally harvested 30 to
60 davs after its physiological maturity which lead to poor milling yields.
Economies
Due to higher expenses and lowler yields, the average net revenue per hectare of the GIE
surveyed is low compared with results o’btained on the large-scale scheme in the 1991 rainy sason
(Table V). Differences in expenses are mainly related to water costs, 250 and 170 US$/ha
respectively, and loar interests, 80 and 45 US$/ha respective&. On the large-scalc scheme water
rosts are tïxed per land unit but do not caver a11 the maintenance costs and depreciation of the large
initial investment whivh cari vary from 3,500 US$/ha for a rehabilitation to 13,000 US$/ha for a
new scheme. On informai perimeters farmers take ail the costs in charge: installation and
maintenance paid using short-term loans (100 US$/ha on average), pump depreciation (85 US$!‘ha
over three years) and operation (85 US$/ha).
Higher financral costs incurred by GIE result from inadequate cash assets relative to
seasonal needs, amounting to about 600 US$/ha. Their demands for short-term loans are often
over-sized to caver expenses that CNCAS does not normally finance (for examplc purchase of a
car) or to repay input suppliers who agree to lend them cash for the GIE counterpart funding
Tecluired for bank loans (15 % of the loan amount) (B&ières & al., 1991).
High production costs for paddy ‘(240 US$/T on average) are another consequence of thcse
poor economic and technical results. Paddy price Will be govemmentally frxed by rhe governrncnr
at 280 US$ per metric tonne until 1993, but Will then decrease to 250 US$/
privatized, and probably less for poor quality grains. The present performances of informai
perimeters could not sustain such a strategy. Moreover two GIE present negative profits after two
years of operation, respectively 13,500 US$ and 23,000 US$, and they have beern unable to yay
back their loans in time to CNCAS.
Table V: Comparison of economics results between informal irrigation (2-year averagc on six GiE)
and large-scale scheme (1991 rainy season average on 12 plots)
5. DISCUSSION
Farmers have used the conjunction of fïve factors to develop informal irrigation: (1) land
resources free of traditional tenure constraints, managed by their own representatives, and easily
irrigable along the ri>Jers, have been alvailable. (2) Water resources have been adequate and
available throughout tbe year. (3) A financial service has been operating with large amounts of
money and few coercve rules. (4) Farrner organizations have been able to adope a legal status
recognized by CNCAS . (5) Labour-saving cultural techniques have allowed a large cultivated land
area per labour unit.
GIE strategies and performances
Farmers have developped informa1 irrigation at first to extend their cultivated land a.nd to
prevent installation of urban people and iagro-business companies. Being short of capital and labour,
they have adopted extensive strategies of land occupation but their technical choices
(mechanization) and eonstraints (pumping, partial water control) have resulted in large expenses
and poor conditions of production.
Profitability of these operations appears insufficient or even negative. Yields and revemtes
are clearly lower than performances obtained on the large-scale scheme. However this comparison
would be different if initial investment and rehabilitation of large-scale schemes, still supported by
the state, were includ;d in farmers’ expenses. For example state involvement was estimated at. 85%
of the total cost of production for rice paddy in 1982 (Engelhard & Ben Abdallah, 1986). No recent
study has been available since the turnover of responsabilities from the state to farmers.
Three causes cari be suggested to explain the poor results of informa1 irrigation: the
technical limitations trfperimeters, the difficulties farmers have controlling, over large areas, weeds
and fertilization which mainly determine yields, and the lack of rigour in economic management”
Farmers are iaced with a difficult challenge if they want to keep their labour-saving but
costl‘l techniques. Possibilities of reducing expenses without decreasing yields are Iimitecl in this
farming system. (1) ‘Nater consumption is not really controlled due to poor design of perimeters.
(2) Replacing combine-harvesters by hired labour would save only 40 US$lha (respectivell:y 170
US$/ha and 130 US$/ha for 4 Tlha yield) with present prices of paddy and mechanized harvesting
(15 % of production) To fit the production cost of paddy with price reduction and stabilization or
to increase farmer rIzvenue, farmers slhould change their economic behaviour towards intensive
strategies where high yields and double cropping Will replace land occupat.ion objectives. For
example 6.5 T/ha yic:ld would be necessary to rank with the present production COS~ of the large-.
scale scheme withour increasing expenses of the 6 GIE surveyed.
This first diagnosis gives some suggestions in designing a research and development
progr-am on informa1 irrigation in the Delta. The main technical challenge is to design surface
irrigation fitting m?.rltiple objectives: total water control; easy maintenance; adaptation to
mechanization and double cropping; and costs compatible with farmers’ economic capacity and
available long-term loans, which meams between 1500 and 3000 US$/ha, Fcological problems
which necessitate pl&annual studies, like the risk of salinization, rising water-table, and fertility
drop, need special attention since the future of informal irrigation is linked to them. Ether
probiems, not specifïc to informa1 irrigation, must have suitable solutions: water and cropping
management, and development of technical and economic Decision Support Systems.
GIE socio-economic environment
By financing farmers and the private service sector, CNCAS has greatly sustained the
development of informai irrigation. It has been flexible for loan attributions and repayments, but
it ha5 not invested in profitability evaluation and economic monitoring of the operations fananeed.
Tt does not know exastly which perimet’ers have been unprofitable and why. This survey shows thal
farmers have taken advantage of these features to collect money they were needing but did no1
have. and that some 3f them have met problems to pay back their debts. Similar results on a lasger
numher of GIE coul’d explain the present critical position of CNCAS.
How has C&AS evaluated its risks while financing informal irrigation is still unknawn.
But present problems and relationships; between GIE and CNCAS suggest that it should improve
its knowledge of its customers’ projects, and be involved in their economic management through
advising and analyses of results.
State implication in the development of informa1 irrigation has been very limited since ‘Rural
Comrnuni ties has been in charge of land distribution, and farmers have not received advice adapted
to the complex problems they have encountered. Difficulties collecting surface drainage water on
the Delta salty soils, ‘equity in regional land and water distribution, and protection of areas for
ecological reasons, suggest that the state should conceive and apply legislation which specifies the
frame in which informa1 irrigation cari be developed. Training, extension and rcsearch facilities
should also be providied under its control, with better involvement of farmers and the private xector
in the structures in charge of these functions.
6. CONCLUSION
The example of the Senegal River Delta shows that, glven the right environment, farmers
and the private sector cari take: initiatives to replace the state in assuming economic functions and
developing irrigation. This first result is important for Sahelian countries where irrigation has
traditionally been synonymous with state management. But the problems encountered already in
the short-term both ?y the credit bank: and a sample of GIE, and the poor results of informa1
irrigation compared with a large-scale scheme, underline the fragility of the current situation, Two
priorities cari be drawn from this ex:perience: the need for better definition and division of
responsabilities betwc:cn farmers,the state and the private sector, and the need for training farmers
confronted with complex technical and economic problems. Research has an important role to play
in creating sustainab’ie and profitable solutions adapted to this new conte.xt, and to describe and
understand its rapid evolution.
ACKNOWLEDGEMENTS
I would like to thank Mohamed Diop for collecting data, Michaël Dingkhun, Tom Randolph
and Dirk Raes for comments on the draft of this paper, and Sally Moffat for correcting the proofs.
REFERENCES
Adams W.M., 1990 ‘How small is beautiful? Scale, control and sucess in Kerryan irrigation”.
World Development. Vo1.18, n”10, pp. 1309-1323.
Adams W.M., Carter R., 1987. ‘Small scale irrigation in sub-Saharian Africa’. Proaress in
Phvsical Geography. ‘Vol.11, pp. l-27.
Arditi C,, 1986. ‘Quelques réflexions socio-economiques sur la riziculture irriguce dans le Nord-
Cameroun (SEMRY 1 et SEMRY II)‘, Documents Svstèmes Agraires n”6. 435-444.
Aviron-Violet J., Ibo B.D., Soumaila A., Van Steekelenburg P.N.G., Waldsteirr A., 199 1 1 ‘Le
developpement des cultures irriguees clans le Sahel. Contrainte de la politique des irrigations et
strategie paysanne. R;3-pport de synthèse’. OCDE-CIISS, 216 p,
Ba T , Havard M., !992. ‘Les groupes motopompes et les périmétres prives du Delta du fleuve
Senégal’. ISRA. 4q p.
BtTliéres J-F., Havarc3 M., Le Gal, P.Y., 1991. ‘Le financement de l’agriculture irriguCe dans le
Delta du fleuve S&éga.l: Intérêts et dCrives du credit bancaire’. Paper presented at the
International CIRAD - O.S.IJ. - U.01. Workshop “Finance and Rural Development in West
Africa” . Ouagadougou. 2 l-25 Octobre 199 1.
Blanc-Pamard Ch., 1986. ‘Autour du riz, le difficile face-à-face des paysans et de la. SQMALAC
dans ‘Za cuvette du lac Alaotra (Hautes Terres centrales de Madagascar)‘. Documents Svst&mes
A@res n”6. 461-479
De Datta S, K , , Herdt ‘S” W “, 1983. ‘Weed control technology in irrigated rice’. In : Proceedings of
the Conference on “Weed Control in Rice”. 31 August - 4 September 1981. IRRI-JWSS,
Diemer G., Van Der Iaan E.., 1987. ‘]L’irrigation au Sahel’. Ed. Karthala-CTA. 226 p.
Engelhard Ph.) .Ben Abdallah T., 1986. ‘Enjeux de l’après-barrage’. ENDA - Ministère de la
Coopt%ation. 632 p.
avard M., 1990. ‘Etude et &a.luation des systèmes m&z$Cs: l’exemple du delta du fleuve
SénCgal’ . Cahiers de ta Recherche-DCveloppement n”28, 17-32.
IX Gal P.Y. ~ Dia I., 1991. ‘Le dCsen,gagement de 1’Etat et ses conséquences dans le Deita du
Fleube %nCgal’. In “La vallée du fleuve SénCpal. Evaluations et oerwectives d’une d&ennie
d’am&nagements (1980-19901” B. Crousse, P. Mathieu, S.M. Seck (E~S). Karthala, Paris. pp% 160-
174.
Le Moigne G. Y Bargbouti S., 1990. ‘Irrigation in sub-Saharan Africa: the development of @clic
and private systems’ World Bank Technical Paper n”123. 99 p.
Loyer J-Y,, 1989. ‘Les sols salés de la basse vallée du fleuve S&ga,l. Caract&isation,
distribution et évolution sous cultures’, Etudes et Thèses. ORSTOM. 137 p.
Niasse M., 1991. ‘h:s @rimètres irrigu& villageois vieillissent mal. Les paysans se d&engagent-
ils en même temps que la SAED?’ In “La val& du fleuve SCnCPal. Evaluations et oersoectives
d*une dCcennie d’aménagements (1980-1990)” B. Crousse, P. Mathieu, S.M. Seck [E~S).
Karthala, Paris. pp. 97-115.
Raes D,, Sy B., 199’2,. ‘Analyse de la pluviomCtrie et de l’&apotranpiration de la zone du Delta
du fleuve Sgnégal’. Projet Gestion de l’Eau, Bulletin Technique n”5. SAED”
SAED, 1990. ‘La SAED a 25 ans’. SAED/CSE.
Seck S.M., 1991. ‘Sur la dynamique de l’irrigation dans la vallée du fleuve‘. In “La val& du
fle”ue Sénégal.- Evaluations et perspectives d’une dt?cennie d’amenagements [1980-1990)” B,
.-L .-.L-----1.
Crousse, P. Mathieu, S.M. Seck (E~S). Karthala, Paris. pp. 17-43.
Tricart J., Blanck J. P., 1989. ‘L’Office du Niger, mirage du développement au Mali’. ,4nnales
de GéograDhie, n”549, pp. 567-587.