FARMERS’ BEHAVIORAL RESPONSES TO SEASONAL
RAINFALL FORECASTS IN THE SAHEL-SUDAN
M. Carla RONCOLI and Keith T. INGRAM* (University of Georgia, USA)
Christine C. JOST and Paul H. KIRSHEN (Tufts University, USA)
Pascal YAKA (Direction de la Météorologie Nationale, Burkina Faso)
Introduction
Recent developments in climate predictions suggest that seasonal rainfall forecasts have
potential for alleviating the vulnerability of livelihoods to climate variability in the Sahel-Sudan
of West Africa, where most rural households depend on rainfed agriculture for food and income
(Hammer et al., 2001). Washington and Downing (1999:255) postulated that “climate forecasts
may indeed revolutionize resource management in Africa.” Still, much remains to be learned
about whether and how African farmers will understand and respond to scientifically derived
forecasts and what will be the social, economic, environmental impacts of farmers’ decisions that
are based on climate forecasts.
This paper reports on an experience of experimental dissemination of seasonal rainfall
forecasts in Burkina Faso undertaken in the context of the Climate Forecasting and Agricultural
Resources (C FAR) project.1 This comparison of responses of farmers in three different agroecological zones of the country illustrates the diversity of options and constraints that farmers
*
Corresponding author: Keith T. Ingram, Associate Professor, Department of Crop and Soil Sciences, The
University of Georgia, 1109 Experiment Street, Griffin, GA 30223-1797 USA. Tel: 1-770-412 4045; Fax: 1-770229-3215; E-mail: kingram@griffin.peachnet.edu.
1
The Climate Forecasting for Agricultural Resources (C FAR) project is an interdisciplinary research initiative that
aims at identifying opportunities and constraints to the application of seasonal rainfall forecasts to improve
agricultural production and livelihood security in West Africa. The project has been funded in part by the Office of
Global Programs of the U.S. National Oceanic and Atmospheric Administration and the U.S. Agency for
International Development. Tufts University and the University of Georgia jointly implement C FAR in partnership
with the Burkina Faso national meteorological service (DMN), the Burkina Faso national environmental and
1
face in the Sahel-Sudan region as they assimilate climate forecasts into the suite of information
that they use to establish their production strategies. Based on this case study, we have identified
key factors that can enhance farmers’ ability to respond optimally to forecasts and we
recommend priories for further research.
Research Context
Three sites were selected to represent the main agro-ecological zones and livelihood
configurations of Burkina Faso (see Map). Bouahoun village, near the town of Houndé, is located
in a commercial cotton production area in the Southwest. Bonam village, near the town of
Boulsa, is in a rainfed cereal farming area in the Central Plateau. In the Sahel we studied two
villages, Koria and Sambonaye, both near the town of Dori; they represent the relative extremes
along the continuum from agricultural to pastoral production systems, with agriculture
dominating the livelihood in Koria and pastoralism in Sambonaye. Rainfall characteristics of the
three sites are given in Table 1 and salient cultural and crop management characteristics are
given in Table 2.
In all sites agriculture depends largely on seasonal rainfall and is therefore a high-risk
endeavor. Producers from all sites concurred that drawing a livelihood from the natural
environment, whether by farming or by herding, has become increasingly arduous and risky
during the last two to three decades. Farmers reported a greater frequency of water-deficit years,
late onset of the rainy season, premature end of rains, and anomalous rainfall distribution. Crop
production strategies strive to manage risks and reduce losses by diversifying field locations,
types of cultivated soils, and crop and variety mixes. In all sites farmers emphasize cultivation in
agricultural research institute (INERA), and Plan International, one of the largest international non-governmental
development organizations operating in Burkina Faso.
2
lowland or upland fields according to their expectations of seasonal rainfall, but land shortage
limits this strategy, especially in the Southwest site.
In the Central Plateau farmers have responded to increased frequency of drought and
shortening of the rainy season by shifting from traditional varieties that have a long duration (120
to150 days) to those that have medium duration (70 to 90 days) or short duration (50 to 60 days).
Millet is the staple of choice in the Sahel and is grown mostly in sandy soils. On the other
hand, more farmers are planting medium-duration sorghum varieties, which ripen before local
millet varieties, especially in valley bottoms and on clayey soils that were previously used for
pasture. The expansion of agriculture into areas that were formerly pasture and declining rainfall
3
make it more difficult for Fulbe households to make a living on livestock production alone. As a
result, many Fulbe households combine transhumant pastoralism with some crop production
during the rainy season. At the same time, agriculturalists are diversifying into livestock
production, especially for the sale of milk where villages are close enough for access to daily
markets. In the Southwest, cotton farming has reduced the diversity of cropping systems,
although profits margins are too small for many farmers to abandon household food production.
Most farmers grow similar acreages of cotton and maize, with small fields of rice, sorghum,
peanut, sesame, and beans.
SOFITEX is a semi-private partnership of the Burkina Faso government and the French
textile industry that purchases, processes, and exports cotton. By channeling technology, inputs,
and credit, SOFITEX has enabled Southwest farmers to have far greater market orientation and
access. A few wealthy farmers in the Southwest have tractors and most farming households own
at least one plow and pair of oxen. In contrast, less than one-fourth of households in the Central
Plateau site own plows, while Sahelian agriculturalists do not use plows because the topsoil layer
is too shallow and sandy and the subsoil too rocky.
In all three agro-ecozones, livestock husbandry systems center on the herding of
ruminants. Although historically the Fulbe specialized in cattle husbandry, most family livestock
holdings now include cattle, sheep, and goats. In the Sahel some farmers also keep camels and
they produce small ruminants for either consumption or sale. They usually keep cattle for milk
production, although they may also sell cattle in response to climate or economic stress.
Although most Sahelians still carry out some form of migration with their animals, many farmers
in the Central Plateau and most in the Southwest now graze their animals in the vicinity of the
4
village all year round. Farmers in the Southwest also produce swine, which has become a source
of income for the household women who tend them.
All households in the Southwest site own at least one radio and several households own
battery-operated televisions. About one-third of households in the Central Plateau site and oneforth or fewer in the Sahel own radios. The Southwest site receives Radio Nationale du Burkina
(RNB), which is relayed from Bobo-Dioulasso where several private FM stations also operate.
The Central Plateau site receives RNB from Ouagadougou, although the transmission is
occasionally disturbed. A few private FM stations broadcast in the area but they have limited
span. The Sahel sites lay outside of the RNB broadcast range, but there is a foreign-sponsored
community FM station in Dori with a span of 100 km that is widely listened to. Better quality
radios can also tune into Mali and Niger radio stations.
Development, dissemination, and evaluation of the forecast
In 2000, C FAR organized the experimental dissemination of forecasts to farmers in its
three sites. A provisional forecast was presented at the Climate Outlook Forum, also known as
PRESAO (Prévisions Saisonnières pour l’Afrique de l’Ouest), a yearly conference since 1998.
The forum in 2000 brought together forecast producers and potential users from most West
African countries in early May, that is, before the onset of the rainy season, to develop a
consensus forecast for the region. Forum participants produce a consensus forecast by
combining the forecasts developed by each West African meteorological service with regional
forecasts developed by international meteorological centers, such as the Hadley Center and the
U.S. National Oceanic and Atmospheric Administration (NOAA). Forecasts are presented as the
probability of the rainfall being in the high, middle, or low tercile of long-term historical rainfall
data (Table 1). In West Africa forecasts predict total rainfall during July, August, and September,
5
the three-month period at the core of the rainy season during which at least 80% of annual rains
fall.
The 2000 Climate Outlook Forum was held in Ouagadougou on 8-12 May 2000. The core
of the program was a workshop, during which climate scientists interacted with a small group of
farmers from the C FAR sites and elicited their information needs and their views regarding the
potential uses of forecasts (Roncoli et al., 2000). The forecast developed at the 2000 Climate
Outlook Forum indicated a 40% probability that seasonal rainfall would be in the higher tercile,
40% for the middle, and 20% for the lower tercile in all three climatic zones of Burkina Faso.
This distribution was presented to the participating farmers by laying out differently colored slips
of paper (20 green, 20 yellow, and 10 red), and then placing them in a bowl and randomly
selecting one slip of paper. The selection was repeated several times to show that the low
probability scenario could occur, though with lesser frequency.
The seasonal rainfall forecast was updated at the end of June, with minor changes (35%
higher, 45% middle, 20% lower tercile). In mid-July a team composed of C FAR collaborators
from the meteorological service and the national agricultural research institute of Burkina Faso
(INERA) traveled to the C FAR sites to present the updated forecast to the communities. The
information presented at the village meeting reiterated what had been presented during the
Climate Outlook Forum. The forecast dissemination team informed the communities that their
studies predicted that seasonal rainfall in 2000 would likely be in the ‘middle range.’ In
particular, they expected rainfall to be less than that of 1999 (a heavy rainfall year) but more than
that of 1997 (a drought year). The team also explained that the forecast is ‘probabilistic,’
meaning that there is a margin of uncertainty, that forecasts are for ‘zones’ and do not predict
rainfall in a specific village or a field, that forecasts relate only to July, August, and September
6
rainfall, and that forecasts cannot predict the date of onset or the end of the rainy season, which
are the most important parameters for farmers (Ingram et al., 2002).
Between mid-January and the end of February 2001, a research team composed of an
agronomist, an anthropologist, a veterinarian, and a water management expert, returned to the
research sites to assess the impact of forecast dissemination. The team spent 7-10 days in each
site doing semi-structured interviews and focus groups to understand what rural producers had
understood, how they had used the information in their production choices, and how it could be
better explained and delivered to farmers. From 5 to 15 group and individual interview sessions
were conducted in each village, including samples of villagers who had and had not attended the
meeting in July during which the forecast was disseminated. About half of the farmers
interviewed in the Sahel and Central Plateau sites and two-thirds of farmers interviewed in the
Southwest site had attended the forecast dissemination meeting.
Some farmers also reported hearing about the forecast on the radio during local or
national radio broadcasts that had reported on the Climate Outlook Forum. But generally farmers
were unclear about the source of the information: some mentioned the ‘Météo,’ others
agricultural extension or SOFITEX.
In fact, the rains did not perform as predicted2. Seasonal rainfall in Dori and Boulsa was
in the lower tercile, that is, the lowest probability outcome of the forecast (Table 1) with the
seasonal rainfall in Boulsa being the lowest amount of the previous 40 years. The severity of
water deficit differed among zones and even among and within villages within the same zone.
This variability in water deficit reflects the spatial variability that characterizes Sahel-Sudan
rainfall, as well as small-scale differences in soil types and field locations that affect plant growth
7
and crop performance. Drought was most severe in the Central Plateau, which less than half of
the long-term average seasonal rainfall, and in the Sahel, where the season began well but then
had dry periods during key stages of crop growth. In the Southwest, crops suffered from water
deficit periods and a premature end of the rains, but the situation was less critical than in the
other zones. By January, most households in the Sahel and in the Central Plateau reported that
they were rationing daily intake of food, selling livestock and other assets to buy grain, and
pastoralists had begun taking their herds south earlier.
Farmers’ responses to the 2000 forecast
A key question was how farmers would react to the ‘failure’ of the forecast to predict
the actual rainfall situation. In fact, few farmers seemed surprised or upset that the low
probability event occurred despite being deeply troubled by the prospects of food insecurity.
Farmers’ assessment of the ‘accuracy’ of the forecast was largely a function of how they had
understood the information provided, which did not always coincide with the message that the
team had tried to deliver. Rather, farmers interpreted the forecast according to what they
themselves expected on the basis of their own experience and observations, particularly
observations during the onset of the rainy season. Farmers consider the nature of the onset and
the way rainfall patterns follow during planting to be the most reliable predictors of how the
rainy season will unfold (Roncoli et al., 2002).
Generally, farmers expect a favorable season if rains begin early and from the South.
They prepare for drought if the onset is late and rains comes from the North. Onset
characteristics that farmers reported to indicate a good season include planting rains that last
several hours during the night, leave the soil moist for several days, and are no more than 3-4
2
The National Meteorological Service estimated the forecast to have been accurate in the past in 35% of the cases in
8
days apart. Therefore, in the Sahel, where the rains had begun relatively early (in the third week
of June) and were followed by good planting rains, respondents recalled that the visiting
scientists had predicted a favorable season (‘good’ or ‘not bad’). Some farmers believed the team
had predicted that it would rain until the end of the season and that it would rain throughout the
region. A few farmers also understood that rains would be ‘not abundant,’ ‘not many,’ ‘less than
in 1999.’ Most farmers expected rains to be at least ‘average’ (Table 3).
The most common responses to the forecast reported by Sahel farmers pertained the
choice and size of fields (Table 3). About half of the farmers interviewed who had received the
forecast mentioned increasing the area they cultivated by 20-30% to plant sorghum and other
crops, which they expected to do well under wetter conditions. In one-third of the interviews,
respondents reported rehabilitating old clayey fields (clairières). One-third of farmers also
reported investing in seed and planting more secondary crops, such as maize, peanut, cowpea,
sesame, and sorrel. Usually only women cultivate these crops, but because the 2000 season
looked promising some men planted them as well.
Other responses to the forecast pertained to labor allocation, choice of crops and varieties,
and manure application. Farmers described how every time it rained they rushed to the fields,
especially clayey ones, and cultivated intensively to optimize infiltration and transplant to fill
gaps. A few farmers planted longer duration varieties of sorghum and peanut, whose taste they
prefer but which require more rain than shorter duration varieties. A few farmers mentioned
increasing manure application, especially in sandy soils, to take advantage of the expected good
rains.
which forecast and actual rainfall were compared.
9
In the Central Plateau, where the rainfall onset was three weeks late, farmers generally
retained the perception that the rains would be less than average (‘not too much,’ ‘not heavy,’ or
‘less than in 1999’). Only one farmer understood the forecast to be for ‘good’ rains. Half of the
farmers interviewed recalled that the team had predicted that the season would be ‘not long’ and
marked by ‘not many big rains,’ which would be widely spaced. Most farmers reported that the
team had predicted that the rainy season would last about 90 days, an apparent misunderstanding
of the fact that the forecast refers only to the months of July, August, and September. In the
Central Plateau 90 days is a short but viable farming season.
Half of the farmers in the Central Plateau responded to their understanding of the forecast
by selecting shorter duration varieties, especially for sorghum, which is the main staple crop and
includes a wide repertoire of local varieties (Roncoli et al., 1999). Farmers planted a
combination of 50-day Pisnu variety in upland fields and 70-day Piswopoi variety and 90-day
Pielga in lowland fields. As the season progressed, if Pielga or Piswopoi failed to establish in
lowland fields, farmers replaced it with Pisnu, despite the fact that in lowland fields Pisnu may
suffer from water logging. As in the Sahel, Bonam farmers also made adjustments in terms of
where and how much to plant. Half of the farmers who reported using the forecast did so through
their choice of fields, planting lowland fields that they had abandoned because of flooding in
1999.
In the Southwest, where rains had begun one week late but had continued regularly,
two-thirds of respondents interpreted the forecasts provided at the village meeting or on the radio
as predicting ‘average,’ ‘adequate,’ or ‘sufficient,’ although not abundant or excessive seasonal
rainfall. Most farmers interviewed also remembered the team had predicted rainfall to be less
than 1999 when the Southwest had suffered devastating floods, which farmers were obviously
10
hoping would not occur again. One-third of farmers understood that the season would be ‘good.’
Farmers who had received the SOFITEX forecast during the farmer forum or from the radio
broadcast expected ‘abundant’ rainfall.
Choice of crop and crop varieties and timing of farming tasks were the dominant
responses to expectations for less than abundant, but adequate, rainfall in the Southwest. Half of
the farmers who had received a forecast modified their choice of crops and crop varieties, some
selecting sorghum and maize varieties with shorter growth durations, others replacing peanut or
cowpea with sesame, which can mature even after the rains stop. Some farmers planted their
lowland fields with sorghum rather than maize or rice, which are less drought resistant. But one
farmer, who had initially decided not to plant rice that year, reversed his decision when he heard
the forecast, which he understood to predict good rains.
About half of the farmers in the Southwest reported adjusting the timing of plowing and
planting to ensure crops were established as early as possible and were able to ripen fully in case
the rains stopped prematurely. One farmer planted directly after plowing without ridging to save
time, while another stopped planting in early July in order to have time to thoroughly hoe and
weed the fields he had planted thus far. Two farmers who had also heard radio broadcasts of the
forecast early enough, reported changing the orientation of ridges in ways that helped retain
rainwater in the fields. Some farmers who said that they expected less than optimal rainfall,
bought or applied less fertilizer, pesticides, and herbicides partly to reduce their debt burden in
case of poor yields, and partly because they expected weed and pest infestation to be less severe
with low rainfall.
Applying the forecast to field selection was less feasible in the Southwest than in the
other sites because cotton and maize fields account for much of the cultivable land near the
11
village. About one-third of farmers who had received forecasts expanded cultivation in lowland
fields and reported that others, who had abandoned their lowland fields because of the 1999
floods, brought them back into cultivation upon hearing the forecast. In contrast, one farmer who
had heard the SOFITEX radio broadcast for ‘abundant rainfall,’ abandoned negotiations for
borrowing a lowland field and planted on higher ground. He estimated that his maize production
from the upland field was 30% less than he would have obtained from the lowland field.3
This evaluation shows that farmers can and do incorporate information concerning
rainfall, as understood by them, into their production decisions, although the limited size and
purposive nature of the samples precludes our ability to estimate what proportion of the farmers
who heard the forecast did so. Yet the repertoire of responses mentioned above does not
correspond to the entire range of possible adaptive strategies (see Ingram et al., 2002). The fact
that the forecast dissemination team visits to the C FAR villages occurred in mid-July meant that
farmers who heard the forecast then for the first time were unable to use the information in ways
that would have called for prior action. By July it would have been too late for farmers to
implement certain strategies, such as borrowing new fields, renting tractors, reorienting ridges,
acquiring seed of suitable varieties, collecting and applying manure, or to order and ordering
chemical inputs.
Above all, resource availability constrained how farmers responded to the forecast.
Availability of land suited to various rainfall scenarios, especially lowland fields, which retain
more water, is among the most limiting factors. Availability of lowland fields is becoming
scarcer and more difficult to borrow because most owners cultivate them every year as
3
He estimated a loss of four to five 100-kg bags of maize, which could have fed his family for several weeks, and
corresponded to a value of 40,000 Fr. CFA in terms of maize prices at the time of the interview.
12
‘insurance’ against drought regardless of rainfall expectations simply to maintain their
occupation rights.
Soil type also constrained choice of crops and crop varieties. For instance Koria has few
lowland or clayey fields, which are particularly suited to sorghum. Farmers in the Central
Plateau reported lack of soil fertility and low soil moisture retention capacity as the most severe
constraints on their ability to respond to an expected drought. Cotton farmers in the Southwest
were particularly affected by the decision of SOFITEX to partially or entirely withhold credit for
inputs for GPCs (Groupement des Producteurs de Coton) who had defaulted on previous loans.
On the other hand, farmers explained that, because cotton was the only reliable source of cash
and because SOFITEX was the only source of credit for inputs for both cotton and maize, they
could not grow less cotton even if they wanted to.
Southwest farmers also mentioned labor shortage as a limiting factor because the
demands of cotton restrict the allocation of labor to other crops or activities. Central Plateau
farmers, where only one-third of households have access to animal traction, also mentioned labor
shortages (Roncoli et al., 2001). Without access to sufficient labor, farmers may not be able to
plant short duration sorghum varieties, such as Pisnu, because they require careful weeding.
Conclusions
Farmers can and do respond to forecasts by enacting various strategies. Most responses
are minor modifications to a highly diversified and risk-averse production system rather than
drastic changes that seek to maximize yields or profits. The most common responses enacted in
2000 pertained to choice of fields, crops, and crop varieties, timing of agricultural tasks, and
application of manure and other inputs. Because many factors converge in shaping farming
decisions it is difficult to identify direct causal links between forecasts provided and behavioral
13
outcomes. Pastoralists in the Sahel and elsewhere did not use forecasts to support livestock
management decisions because those decisions tend to follow rather than anticipate the rains.
Farmers’ ability to respond optimally to a forecast was constrained by several factors.
Some of these relate to the nature of the information provided, others relate to the context of
applying the information, that is, what options and resources are available for farmers to use the
information in beneficial ways. As others have noted, lack of production resources limited the
flexibility and feasibility of adaptations to climate variability and applications of climate
information (Nelson and Finan, 2000; O’Brien et al., 2000; Valdivia et al., 2000). In all three
sites, farmers complain that land is becoming less available and less productive. In particular
lowland fields, which can sustain plant growth for longer periods without rain, are becoming
object of increasing competition among farmers and between farmers and herders. Draft animals
and equipment, viable seed for short duration varieties, chemical or organic fertilizer, and
agricultural credit are other key factors that may help farmers translate information into
successful adaptation.
There is still a strong disconnect between the parameters and scale of information that
farmers need and those of scientific forecasts (Broad and Agrawala, 2000; Stern and Easterling,
1999). Farmers are most interested in a local (down-scaled) forecast of the duration and
distribution of seasonal rainfall. Predictions of the dates of onset and end of the rains, and
predicting water deficit periods during the season are more important to farmers than are
measures of total seasonal precipitation. Although there has been some success in forecasting the
onset of the rains from wind shears associated with the northward movement of the intertropical
convergence zone, accurate forecasts of rainfall distribution and end are beyond the ability of
scientific models.
14
Farmers need the forecasts as early as possible in order to enact appropriate responses and
are willing to sacrifice some accuracy in favor of timeliness. Farmers said that they would prefer
having a provisional forecast in May or June rather than having a more accurate updated forecast
in July. Farmers would benefit most from forecasts that are delivered at least one month and
preferably two months before the start of the rainy season, but they can still benefit from later
forecasts since they stagger planting over June and July.
Further research is needed on how best to convey the probabilistic nature and the
limitations of the forecast to farmers (Patt, 2000). Farmers understand climate information in
terms of what they are most interested in knowing and what they expect on the basis of their own
experience. In some cases, their ability to interpret a forecast in the context of their own
predictions mitigated the possible negative impacts of the 2000 forecast, which only showed a
20% probability that seasonal rainfall would be below normal though that was indeed the
outcome. In other cases farmers did not grasp the probabilistic aspect of the forecast or were
misled by inconsistencies in the messages delivered through the various media and therefore
made choices that resulted in some loss. Given that most farmers of the Sahel-Sudan operate with
a slim margin above subsistence needs, errors can have dire consequences.
Yet farmers understood that even science cannot provide accurate predictions of rainfall
and that, given its highly degree of variability, farming will remain a risky undertaking. Farmers
perceive this variability to have increased in the last 20-30 years, in ways that are outpacing their
own ability to adapt. Therefore they are interested in receiving forecasts and other information
that can complement their own, particularly if forecasts are developed and disseminated in ways
that respects and build on their own experience and worldview. There is evidence that repeated
exposure to climate forecasts enables farmers to better understand them (Patt, 2002). A key
15
question that the C FAR project will address in its future research is how to package forecasts
with technical advice and interventions that can enhance its value and which intermediaries, such
as extension services, NGOs, media, and farmer leaders) can best help farmers understand and
integrate the information into their production decisions.
References
Agrawala S and K Broad. 2000. “The Ethiopia food crisis - uses and limits of climate forecasts.”
Science 289, pp. 1693-1694.
Hammer GL, JW Hansen, JG Phillips, JW Mjelde, H Hill, A Love, A Potgieter. 2001. “Advances
in Applications of Climate Predictions in Agriculture.” Agricultural Systems, 70, pp. 515-533.
Ingram, KT, MC Roncoli, and P Kirshen. 2002. “Opportunities and Constraints for Farmers of
West Africa to Use Seasonal Precipitation Forecasts with Burkina Faso as a Case Study.”
Agricultural Systems, in press.
Nelson DR and TJ Finan. 2000. “The Emergence of a Climate Anthropology in Northeast
Brazil.” Practicing Anthropology, 22:6-10.
O’Brien K, L Synga, K Naess, R Kingamkono, B Hochobeb. 2000. Is Information Enough? User
Responses to Seasonal Climate Forecasts in Southern Africa. Report to the World Bank, AFTE1ENVGC. Adaptation to Climate Change and Variability in Sub-Saharan Africa, Phase II.
Patt A. 2002. “Effective Seasonal Climate Forecasting: Examining Constraints for Subsistence
Farmers in Zimbabwe.” Global Environmental Change: Human and Policy Dimensions, under
review.
Patt A. 2001. “Understanding Uncertainty: Forecasting Seasonal Climate for Farmers in
Zimbabwe.” Risk Decision and Policy, 6, pp. 105-119.
Roncoli, MC, KT Ingram, and PH Kirshen. 2002. “Reading the Rains: Local Knowledge and
Rainfall Forecasting among Farmers of Burkina Faso.” Society and Natural Resources, 15, pp.
411-430.
Roncoli, MC, KT Ingram, and PH Kirshen. 2000. “Can Farmers of Burkina Faso Use Rainfall
Forecasts?” Practicing Anthropology, 22 (4), pp. 24-28.
Roncoli, MC, S Bahadio, S Boena. 1999. The Role of Rainfall Information in Farmers’
Decisions: Ethnographic Research in the Central Plateau (Burkina Faso). Technical Report,
16
Climate Forecasting and Agricultural Resources Project, University of Georgia, Athens,
September.
Valdivia C, C Jetté, R Quiroz, JL Gilles, S Materer. 2000. Peasant Households Strategies in the
Andes and Potential Users of Climate Forecasts: El Nino of 1997-1998. Paper presented at the
American Association of Agricultural Economics meetings, Tampa.
Stern PC and WE Easterling. 1999. Making Climate Forecasts Matter. National Academy Press,
Washington.
Washington, R, T Downing. 1999. “Seasonal Forecasting of African Rainfall: Prediction,
Responses, and Household Food Security.” The Geographical Journal, 165 (3), pp. 255-274.
17
Table 1: Total 2000 seasonal rainfall, long-term average rainfall, and rainfall tercile limits for C FAR sites.
Location
2000 rainfall
40-year average
Lower tercile
Middle tercile
Upper tercile
(mm)
1960-2000 (mm)
Min / Max (mm)
Min / Max (mm)
Min / Max (mm)
Dori
276
371
215 / 330
343 / 417
419 / 540
Boulsa
210
466
210 / 399
403 / 490
493 / 977
Bouahoun
562
562
362 / 544
547 / 664
672 / 885
Source: Direction de la Météorologie Nationale.
Table 2. Salient cultural and crop management characteristics for C FAR sites.
Location
Ethnic groups (indigenous)
Principal crops
Tillage
Cash source
Dori
Fulbe > Rimaibé = Bella = Gurmanché
Millet > Sorghum > Peanut
Hand hoe
Livestock / milk
Boulsa
Mossi > Fulbe
Sorghum > Millet >Peanut > Rice
Hand hoe > Bullock plow Surplus grain
Cotton > Maize > Rice >
Bullock plow > Tractor
Bouahoun Mossi > Bwa > Fulbe
Cotton
18
Table 3. Farmer perceptions of and responses to the 2000 seasonal rainfall forecast.
Location
Understanding of forecast
Most common responses to forecast (% of farmers interviewed)
Dori
Average
Increase area planted (50%); Rehabilitate abandoned fields (33%); Purchase seed for
secondary crops (33%)
Boulsa
Less than average; 90-day duration
Plant shorter duration varieties (50%); Increase planting in lowland fields (33%)
Bouahoun
Average; sufficient
Plant shorter duration crops and varieties (50%); Plow for early establishment (50%)
19