Burkina Faso: Integrating r1 + Indigenous and Scientific Rainfall Forecasting : ~~2331 2 T _ his case study describes how Locality and livelihoods farmers of Burkina Faso predict seasonal rainfall and examines Bonam is a village located in the how their forecasts relate to scientific Namcntenga Province, one of the poor- F' -I ones. In recent years, meteorological est of fortymfive provinces in Burkina science has made enormous progress in Faso. Most of Bonam inhabitants are predicting climate. The realization Mossi, the dominant ethnic group in that sea surface temperatures (SSTs) the country, and draw, a livelihood from - _ influence global atmospheric circula- rain-fed farming of grain and legumi- tion enables scientists to formulate nous crops, combined with some live- forecasts of seasonal rainfall. These are stock production, petty trade, and la- l____________ qpresented as the probability of total bor migration to C6te d'Ivoire. Some J 4 8seasonal rainfall quantitv being in the Fulani pastoralists have also settled in above normal, below normal, or normal the area. compared with an average resulting Rain falls during a single season last- from analysis of thirtv- years series. In ing from May to October, characterized WNest Africa, seasonal rainfall relates to by extreme inter-annual and non-uni- the three months of July, August, and form distribution wvithin seasons. The September, during which 90 percent of long-term mean annual rainfall is 674 total annual rainfall occurs. mm. At the onset of the rainy season, Rather than conceiving local and sci- farmers decide what, when, and where entific knowledge as absolutely incom- to plant. Thev do so according to their iNo 39 patible, our research shows that farm- expectations for seasonal rainfall, striv- December 2001 ers are used to operating in multiple ing to minimize risk by combining the cognitive frameworks and that the are interested in receiving scientific fore- ~ reportsperiodically o casts because they perceive local fore- w eK) initiatives casts as becoming less reliable due to Ž-fca. It is published increasing climate variabilitv. But to 'Reg Pion's Knowledge and ~4ii1I~(e1)eras part of an evolving effectively convev scientific forecasts t K ter between eWold ~~~~between the World we need to understand how people NGO&m*tles NOOs, develop- think about rainfall, including how , fnstitutions and multilateral they perceive and predict variabilityx - ti:n. The iew expressed in Understanding local cultural models i^ Th~le are those of the authors 1 Understanding local cultural models IS beno ,e attributed to the essential for effectivelyT communicat- Group or its partners in ing research products and develop- l ateAwebpage on IRKis XO/ A1D t ment technology. , /jvwwwordbank,org/afr/ 2 water retention capacity of different soil types and field posi- NWomen also observe insect behavior at water sources and tions with the water requirements of various crops and crop in rubbish heaps outside compound walls. For instance, varieties. bugvar6 are black insects of the Orthoptera sp. that dig con- Farmers' forecasting knowledge encompasses shared and cave nests in rubbish heaps outside compounds. After the selective repertoires. Experienced (mostly elderly male) first rains, larvae emerge, filling the nests with dirt. WVomen farmers formulate hypotheses about seasonal rainfall by ob- said that they expect a good farming season if bugvare fill serving natural phenomena, while cultural and ritual special- their nests to the brim with dirt, which symbolizes a full gra- ists draw predictions from divination, visions, or dreams. nary. But, while signs are observed throughout the dry season, it is at the onset of the rains that farmers consolidate their ex- Shared forecasting knowledge pectations for rainy season. An early onset, especiallvwith a regular succession of rains that enable good crop establish- Environmental indicators that farmers use to predict sea- ment, is the most wvidelv considered indicator of a good sea- sonal rainfall become available for observation at different m i son. The number of times farmers must plant is also kcy in times of vear. Among the earliest and most widelv relied evaluating the nature of a season and predicting crop perfor- upon indicators are the timing, intensity, and duration oft ac.Freape ve dogtafce h eini ' ' < ~~~~~mance. For example, when a drought affected the region in cold temperatures during the earlv part of the dry season 1 ~~~~~~1997, farmers had known a food crisis loomed ahead by the (November-January). Farmers believe that intense cold (be- end of planting time (early August), half a year before official low 15 degrees C) during this time corresponds to abundant famine early-warning systems registered distress signals. rainfall during the rainv season and that if this cold period Some elders are also able to interpret constellation move- begins early or ends late, the rains will do likewise. Follow- begins eary r ndltethrinwlldolikwie.oov ments and lunar phases. For instance, thev consider the vis- ing the cold-dry period is a hot-drv period (February-April): ible phases of the moon, especially the full moon, to be morc intense heat at this time is also believed to predict good rainfall. likely to be dry than dark phases because moonlight exerts a Theal second most common forecasting indicator is thepro- force that prevents rain from falling. Stars also signify suit- Tsctiond mosfrut by commoin foaltreasting indicatoruis thepro- able planting periods for different fields and crops. The ap- duction of fruit by certain local trees, wvhich occurs between pearance of Souci (Pleiades) in early May indicates that it is April and June. According to farmers, good vields from trees April and Juecodigtamrsod srtime to prepare vallev bottom fields for planting. Shifts in such as tacanga (Buty,rospermum parkii) and sibgaI suchg assms tiaanga sBuprspedi umc park afavorableseasond sa t the position of the Budb Kutoega (Ursa Major) identify the (Anogeissus leiocarputs) predict a favorable season. On the satadedo h eidfrpatn ogu h per other hand farmers link abundant fruitstart and end of the period for planting sorghum. The appear- other hand, truit production bynobga ance of Tatba (Orion) in Julv coincides with heaxv rains (sa (Sclerocarya birrea) and sabtuluga (Lannea acida) trees to y t f m drought. ~~~~~~~~~~~nyanga) that favor maize planting. drought. Trees are also used as signs for the approaching of the rains or the presence of water. WNhen sibga begin fruiting and Specialized forecasting knowledge sabtuluga lose their leaves farmers know that they should get ready for planting. Kankanga is a fig-like tree that grows In contrast to the observation of environmental signs which where the water table is near the soil surface. Therefore, it is available to most farmers, divination and other spiritual signals herders where to dig wvells to water their cattle and practices are the prerogative of select groups or individuals. farmers w-here they can plant water-demanding crops, such as The most authoritative among them is the Tengsoba, the el- cotton. dest descendent of the clan that first settled the land. He Another indicator that becomes available at the onset of performs sacrifices to mediate between the living and the the rains is the water level in streams and ponds. If it remains ancestral and earth spirits that inhabit fetish sites, some of consistently high after the first rains, farmers believe that which influence the rains. Predictions are drawn from the the season will be favorable since heavy rains at the onset are behavior of sacrificed animals, how long it takes to fall, in believed to be a propitious sign. Herders who pasture ani- which direction it falls, and where the blood spills. The mals in the bush watch the nesting of small quail-like bird Tengsoba and other traditional specialists might also receive (known as koobre in More) and believe that when nests hang forecasts from ancestors or deities in the form of dreams or high on trees then the rains will be heavy; when nests hang visions. low, the rains will be scarce. 3 Murabouts ire spiritualists wvho situate themselves within their authority, especially among the youth. Climate vari- the Islamic tradition, although Islamic orthodoxy frowns on ability has also weakened farmers' confidence in local knowi- tlicir practices. These spiritualists range from Islamic clerics edge. Elders recalled that in the past theyvwere able to pre- (Limnairi) wvho are versed in the Koran to diviners who mix diet the rain onset so accurately that they could mobilize Islam with indigenous beliefs and practices. The Zambende family labor plant on dry soil, knowving that the rains would ceremony, which marks the beginning of the Muslim year, is soon follow, but now their sons refuse to go to the field until the key venue for maraibouts to issue their forecasts. it actually rains. MIa-rabouts base their predictions on the day of the week that But this does not mean that farmers perceive scientific in- marks the first day of Zaimbenle. Each day is associated with formation as a threat to local culture. On the contrary, be- a different prophet and the year that follows is characterized cause they perceive local forecasts to have become less reli- by svymbolic events in the life or time of that prophet. Proph- able, farmers are keenly interested in alternative sources of ecies are written in Arabic texts along with instructions for information. Rather than being static and uniform, their cog- ritual offerings and other measures to stave off inauspicious nitive landscape already incorporates a plurality of knowl- events. edge frameNvorks. Local forecasting combines empirical ob- LTnlike the Tengsoba and the marabouts who can only issue servations and spiritual insights that draw from a variety of predictions and offer intercessory prayers, there are other religious traditions. Farmers mix local and introduced prac- specialists, known as sa tatta, who claim direct command tices and technologies and families pragmatically combine over the rains. But their powers are feared as rainmaking is modern medicine and local treatments when one of their coonsidered to endanger both practitioners and the commu- members is sick. It is the imposition of any one knowlcdge nity. Invoked rains are believed to be mostly violent down- system as representing the totality of truth that is resented pours (saraogo) accompanied by heavy wiind, sharp thunder, and resisted. anLd lightning that cause damage to crops, houses, and ani- Hence, scientific information must be presented in xvays mals. For this reason, a government cloud-seeding project at that conform to cultural notions concerning the nature of the beginning of the 1998 rainy season caused anxiety rather knowledge, its production and validation, and its relation- than rclief among farmers in Bonam. ship to societv. Local systems of thought emphasize the par- Some spiritualists foresaNv the 1997 drought and their re- tial nature of human understanding, which means that no one spouses have implications for their potential role in diffusing source of information is considered as having the entire pic- scientific forecasts as wvell. The Tengsoba admitted having ture. Local forecasts systems rely on a range of indicators bad dire premonitions during the drv season, but he did not that become available to different people and at different reveal them. other spiritualists confirm that they tend not times. Farmers do not resolve contradictions among indica- to publicize dire predictions because to do so "would be like tors into a cogent scenario. Discrepancies among forecasts or launching a curse" against those under their authority; Ver- between forecasts and outcomes are explained in terms ot balizing negative forecasts would reifv them into an inevi- the diversity of ecological niches and cropping systems. The table outcome by voiding any possibility for supplication and probabilistic nature of the forecast may also be explained by negotiation with the spirits. The failure to alter dire predic- reference to the uncertain nature of destinv as the outcome tions may also undermine the credibility of spiritualists re- of negotiations betveen the living and the spirits and, ulti- sponsible for interceding between the lixing and the spirits. mately, of the arbitrary will of the latter. Furthermore, a bleak forecast may also discourage people There are several aspects of method and content where from farming and induce them to migrate, undermining the local and scientific knowledge of forecasting converges. For social order on which the elders' authority rests. example in formulating predictions, spiritualists use ap- proaches reminiscent of scientific practice. Zctmbende prophecies derive from the exegesis of textual material and Integrating local and scientific forecasts from consultations among spiritualists, wvho then officially communicate thiemn to the lay public. In the dloinaiii of env i- Bonam farmers recognize that both society and climate have ronmental knowvledgie farmer forecasts resemble scientific undergone significant changes in the last few decades. Tradi- ronmentad knowledge,iamer forecastsema blsentific tional leaders lament that formal education, monotheistic religions, and modernization ideologies have diminished natural phenomena. The generation ot knowledge from ob- servation is consistent with cultural learning styles whereby 4 clhildren learni froni vatching adults rather than through ver- distribution of seasonal rainfall, but the integratim of scieni- bal instructioni or asking questions. tific forecasts with local knowledge might allowv ornle infer- Meteorologists could build on local understanding of the ences in this regard. For example, the abnormally ieavv rains relationship betwveen temperatures and rainfall to cxplain that fell in ,JuIv and August 1999 could have beer predicted the technical aspects of scientific forecasts based on sea sur- by combining farmers' predictions of delayed onsct with the face temperatures. Farmers' interpretations of wvind patterns scientific forecast for above-normal seasonal raint ill. also recognize the ocean as the origin for rain. I)uring the dry season, farmers expect w inds to blow westwvard, that is, to go to the ocean to pick up wvater, and then return blowving Conclusions cast\ard at the onset of the rain' season. Farmers predicted Our findings show that neither the experiencs f bridging and explained drought from the absence of such winds. knoN-ledge swstcms nor the concepts of ralinf'all f<.)rccastiing But farmers' forecasts diverge from scientific ones in im- portant wavs, particularly the scale and parameters the)' ad- are alien to the farmers of Burkina Faso. Local for casts con- dress. Unlike scientific forecasts, w-hich are formulated in verge wvith scienrtifi es in soemeaspects oft cointent iiand referenice to "zones," the production and the application of 'mod, btalso diere in t racticall significanceo local forecasts are deeply localized. They derive from an inti- and moral meanings These contrtsts challenge science On mate interaction with a microenvironment whose rhythms tw fronts. are intertwined with thecclesffamlOn the one hand, the specificity of local foreec st parain- are intertwilned xvith the evecles of familv and c ommunits- life. eters urges science to be more responsive to farm ers' infor- It is not the generic sibga or tcmanga that farmers usuallv con- sider in predicting r . bmation needs and more relevant to the livelihood (lecisions sider in prediuting rainfall, buit specific trees near their .. , .....................thcv facc. O)n the other hand, local systems ot' ecxp _rtise aind home or their fields that the) might have observed over a lifetime, leadership tightly link knowledge and social responsibility, NW'hile scientific forecasts hinge Ott estimates of total sea- calling for scientists to be moreaware ofandaccountable for the impacts of the hnowvledgc they produce anid provide to sonal quantity, farmers evaluate seasons in terms of types and time of rainfall. For instance, they recognize that the same users. This responsibility propels Us beyond the task ol integrat- amount of rainfall can Icad to different production outcomes if it occurs as stt nyauga (prolonged but consistent rain that ing local and scientific kno-ledge in formand coiitent fore- . . . . s 1 , l ~~~~~~casts. to face the challenge of addlressing consequ,_,nces and leaves the soil moist for several days) or as s(arCtogo (local- context of their use. In particular, the provision Oi informa- ized th.inderstorm accompanied by N-iolent wind). NNater- tioti needs to be integrated with appropriate intcrvcntions deiicit periods that occur during establishment or headilng gotiate a mitigated our- that bolster farmers' albility- to negoit ii.tdot willt cause more damage to crops than those that occur dur- come of predicted scenarios. Scientists, policvmlahers, do- ing other crop groxvth stages. The time of onset and of termination, marking the dura- nors, and development practitioners must work together to tion of the risisu asaetdevise consistent and sustainable approaches to imnproving nlon of the rains, IS such a salient parameter that a forecast ot thfexiit ' ~~~~~~~~~~~~the flexcibilitv, of local production svsterns anld the -esilincye an "above average" seasonal rainfall is inxvariablv understood of livelihood security for the resoturce-limited farmntrs of the bv Bonam farmers as predicting a longer season. Currently,1 1 1. 1Sudano-Sahel region. science is unable to reliably predict either the duration or Cyte' e. SBurgugUgflO i - 1110 This article was written by CarlaRonco4 KeithInWram, PalKfrshMen and(htne Jost of the Clmate Forecastingfor Agriultural Resources (CFAR) project annthe University of Geor- gia and Tufts University andfunded by the Aifine aRofsPro ra ft Na O adApheri n - tion. A longer version isfordtcoming in Soceety NarlReei 3Fo mm n ona C Roncoli at: (404) 524-8833 or croncoligaes.g e .edu