Soybean and Maize Zoning in West African Economic and Monetary Union — A simulation Approach

The West African Economic and Monetary Union (abbreviated as UEMOA from its name in French: Union Économique et Monétaire Ouest-Africaine) is an organization of eight West African countries: Benin, Burkina Faso, Ivory Coast, Guinea-Bissau, Mali, Niger, Senegal and Togo. This region suffers an agricultural yield gap mainly due to misplanning of crop zoning. Our study aimed to perform an agricultural zoning for maize and soybean in the UEMOA region based on (i) potential yield (carbon dioxide, temperature, photosynthetically active radiation, photoperiod and genotype), (ii) attainable yield, under high crop management and technology inputs and (iii) actual yield. Results show that the UEMOA region is very suitable for growing maize and soybeans; however, a large gap exists between attainable and actual yields. It is shown that the adoption of better management and technology strategies is a way to greatly increase local yields.


Introduction
UEMOA, the West African Economic and Monetary Union or in French: Union Économique et Monétaire Ouest-Africaine, is an organization of eight West African countries founded to promote the economy of participating countries: Benin, Burkina Faso, Ivory Coast, Guinea-Bissau, Mali, Niger, Senegal and Togo.It presents a GPD (Gross Domestic Product) of 146 billion dollars, and a population of about 100 million people (a mixture of ethnic groups and non-African ancestries), under a French colonization history and in the lowest quartile of the HDI (Human Development Index) (Wikipedia, 2014).
The UEMOA countries are, in general, relatively un-urbanized with an average of only 36% of the population living in cities; the minimum in Niger, with 18% and the maximum in Ivory Coast with 51%.All countries depend on food imports to supply their needs.In agricultural sector, the main problem is food security since the local agriculture employs low technology and the main objective is family subsistence.A pressing need exists for an increase of land productivity and employment of local labor force, aiming to ensure food security, as well as reducing the dependency of imports and external debts.One of the authors of this study completed a mission to evaluate the status of agricultural production of the area and suggest yield improvement mechanisms.
Estimates of agricultural yield can be expressed as potential yield, attainable yield and actual yield, and the respective yield gaps (Sentelhas et al., 2015).Potential yield depends solely on air carbon dioxide content, temperature, photosynthetically active radiation, photoperiod and genotype.The attainable yield is the yield achieved with best management practices on a given location.Both attainable and potential yields vary from season to season and year to year depending on meteorological factors.Understanding the existing yield gaps and identifying the most important constraints to achieving optimal yield allows prioritizing efforts in improving yield and profit.It also maximizes the return of investment in research and development.

Database
The applied methodology is based on established and known concepts of yield (Doorenbos & Pruitt, 1975), using geoprocessing tools, computational programming and spatial modelling, similar to those described in Mueller at al. (2018).Edaphic and climatic data were combined with information of soil cover, respecting protected areas or those with preservation interest, resulting in an estimate of the potential production of maize and soybean, integrated to the dynamics of local land occupation.
Using soil maps common to all UEMOA countries, a database was constructed in a SIG environment using the software TNTMIPS, containing: (a) a digital cartographic base (hydrographic and political division), (b) soil classes, and (c) history of weather data Data about the political division of the UEMOA territory was accessed from Thematic Mapping (2009).Soil attributes of the mapping units were based on FAO (2012), a classification that comprises natural soil fertility, texture, relief, soil depth, erosion susceptibility (slope, texture, and depth), drainage, stoniness and rockiness.Historical data of monthly and yearly rainfall, minimum, maximum and average monthly air temperature and solar radiation from the period between 1950 and 2000, were extracted from Hijmans et al. (2005a, b).These authors made a worldwide survey of the above-mentioned variables, eliminating stations that presented errors.Pedologic data, soil water storage capacity, standardization of attributes, soil names and other soil data were obtained from FAO (2012), based on the World Reference Base for Soil Resources (WRB).Data of altitude and slope were obtained from SRTM (2010).
The climatological water balances were established through a computational routine of integrated variable processing of rainfall, temperature, solar radiation and soil water storage.This procedure allowed the calculation of water deficit and excess, potential evapotranspiration (ET0) and actual evapotranspiration (ETa).
Soil quality indexes calculation was based on soil data according to the WRB classification (FAO, 2006).

Yield and Land Suitability Estimates
Attainable crop yield was calculated considering a standard agronomic reference of high-input technology.The estimates of actual yield were based on the socio-economic and infrastructural conditions of the region, generating a low-input yield in which the difficulty of access to agro-chemicals or financial resources are taken into account, thus corresponding to a low level of employed technology.
The evaluation of the land suitability is a process of determining the soil aptitude for specific uses and refers to a set of basic principles and concepts that can be applied worldwide.It requires a comparison between different soil types, obtained benefits and existing needs.At a quantitative stage, levels of suitability were established: high, medium, low and not suitable.The criteria of classification can be found in Tables 1, 3 and 4. A production environment was defined as the sum of the interactions of soil attributes (fertility, drainage, texture and depth) and of the relief, which directly influence crop production.For this study, the methodology for production environment classification proposed by Prado (2007) was adapted to the characteristics of the available data for the region, combining suitability with respect to slope and soil type.Edaphic information was combined to relief information, and the combined quality of soil-relief was used as a logic key for suitability.Each soil presents class of land suitability for annual crops based on land slope S (Table 2).In this way, a Soil Quality Index (SQI) was calculated as the geometric mean of soil depth, drainage and combined texture-structure, each ranked from 1 to 3 (Table 1): Where, a 1 is the score for soil depth, a 2 for drainage and a 3 for texture-structure.According to the estimated values of SQI for each soil order, four classes were established for suitability levels: Not suitable (SQI ≤ 1.3), Low (1.3 < SQI < 1.8), Medium (1.8 ≤ SQI < 2.5), and High (SQI ≥ 2.5).
Table 1.Classification of the land suitability (high, medium, low and not suitable) per soil order, with the correlation between the World Reference Base for Soils Resources-WRB (FAO, 2006), based on the soil quality index (SQI): soil depth (a 1 ), drainage (a 2 ) and texture-structure (a 3 ) Land slope was obtained based on the land elevation model in 90 m (SRTM, 2010).Starting from the functions in the GIS (Geographic Information System) environment and with the utilization of algorithms from eight neighboring localities, it was possible to calculate the slope in percent, as shown in Table 2.The conjunction of SQI, which considers the land suitability, with the slope suitability (Table 2), yielded a general classification of the production environment, subdivided into nine classes (Table 3).The assimilation of CO 2 (ε, kg ha -1 h -1 ) by plants for the gross production of carbohydrate (η, kg ha -1 d -1 of CH 2 O) is related to the fraction of the photosynthetically active radiation (Ω, J m -2 s -1 ), according to: The CO 2 assimilation by C 3 (soybean) and C 4 (maize) plants correspond to the CO 2 concentration of 340 ppm (Co, ppm), can be described (Goudriaan, 1982) as: Where, Cx is the current atmospheric carbon concentration (assumed as Cx = 385 ppm), Co is the base carbon concentration, and p is the crop reflection coefficient (J J -1 ).δ and β are empirical parameters, δ = 0.48 kg[CO 2 ] ha -1 h -1 (J m -2 s -1 ) -1 and β = 0.8 for C 3 species and β = 0.4 for C 4 species, according to Penning de Vries et al. (1983).λ (kg ha -1 h -1 ) is an auxiliary variable calculated according to: Where, T ( o C) refers to air temperature; and w 0 , w 1 and w 2 to parameters linked to air temperature (Table 4) (Goudriaan, 1982).
Table 4. Empirical parameters depending on air temperature to obtain auxiliary variable λ (Goudriaan, 1982) Temperature range (in °C) Considering the gross mass of carbohydrates produced on a daily average basis for the whole growth cycle, estimated from the number of degree-days from plant emergence to flowering (GD f , ºC d) and from the length of the reproductive phase (D R , d) knowing the average photoperiod of the cycle (H, h d -1 ) and the average leaf area index of the cycle (I LA , m 2 [leaf] m -2 [soil]), the total carbon productivity (Ψ, kg[CO 2 ] ha -1 [soil]) can be estimated using the equations: ) is the carbon dioxide assimilation, Tb (°C) is the base temperature for crop development and T (°C) is the average air temperature.Considering the very low latitudes, photoperiod H was set to 12 h.
To transform Ψ in dry matter of the different plant organs (root, stalk or stem, leaf and grain), some corrections (C RMC and C PAR maintenance and growth respiration and photosynthetic active radiation coefficients, respectively) were used according to the principles presented by De Wit (1965,1978,1982), developed to estimate the potential yield of a crop through the available energy of the considered locality and experimental data presented by Doorenbos and Kassam (1979), calibrated for a wide range of climatic conditions.
Where, Γ (kg[wet botanical seed] ha -1 ) is the potential productivity, E CDM (kg[dry matter] kg Another great challenge related to agricultural activities is the informality of the Land Registry.Land ownership is commonly assigned to the population by customary use, under the influence and agreement of local leaders, but there is no documentation or centralized records to ensure legal certainty in the transfer of ownership of the land.Due to the resulting small size of the properties it is difficult to implement efficient agricultural management over larger areas.The result is a large yield gap, i.e., the actual yield is only a fraction of potential and attainable yields. From the point of view of technological adoption, FAO (2015) data based on the response of local farmers to questionnaires show that the consumption of agricultural inputs in UEMOA member countries is very low.
Regarding fertilizers, there are only the data of total consumption by countries, which hinders the understanding of the adopted technology level by each crop.However, based on the average productivity presented for maize cultivation (Table 5) it appears that there is very low technological input in the cultivation of this grain.
The fertilizer volumes consumed by the countries of UEMOA in 2012 can be seen in Table 6.The consumption of nitrogen, phosphorus and potassium across the African continent represent 2.7%, 3.0% and 1.8% of the world fertilizer consumption respectively.In Brazil alone, domestic consumption is 3.6%, 9.4% and 16.15% of the world total.
The three largest consumers of fertilizers in Africa are Egypt, South Africa and Ethiopia, and within UEMOA it is Mali, which consumes 2 to 4% of all fertilizer consumed on the African continent.
The world production of soybeans in 2013, according to FAO (2015) was 276 million tons; the United States was the largest producer, accounting for 32% of this total.Africa produces less than 1% of the world total, and imports of grain are close to the total production (IITA, 2014b), about 2.2 million tons in 2013.The largest producer of soybeans in the African continent is South Africa, followed by Nigeria, both account for 62% of all soybeans produced in this continent.l and ha -1 , yield levels reach 1,000 kg ha -1 , and below that area 2,000 kg ha -1 .The average actual yield of soybean in Brazil is 2,920 kg ha -1 and in Africa about 1,800 kg ha -1 .
Considering the relative land use as a gauge of economic activity, Burkina Faso is the country with the highest utilization of the territory with agricultural activities (65%) followed by Senegal (56%).Other countries have less than 50% of the territory covered by agricultural activities, and UEMOA, in general, only 18%.It is important to remember that Niger and Mali have 72% and 65% of their territories covered by dunes or areas without vegetation; in general, the territory of the UEMOA has 50% of its territory in the same condition.
Despite the excellent weather conditions prevailing in approximately half of the territory formed by the UEMOA, the actual yields of the maize and soybean crops are considered low, which conflicts with the simulated high potential and attainable yields.Apparently the reduced use of technologies is the major factor in this regard, explained by the low level of economic and social development of the region, the incipient supply chain of inputs and inadequate technology combined with the low purchasing power of local farmers and a lack of regulatory policies to implement agricultural activity in the region.
The simulated actual yield levels in this study are quite relevant and comparable to the yield data collected by FAO (2015) in each country, however, it is important to mention that the attainable yield levels, although realistic, are still far from the reality of this region.
The territory of the countries belonging to UEMOA, Mali, Niger, Senegal, Guinea Bissau, Burkina Faso, Ivory Coast, Togo and Benin are presented in relation to agricultural zoning: (i) about 50% of the area can be considered as not suitable for maize and soybeans, primarily due climatic and pedologic restrictions, as well as environmental, or even being occupied by urban centers; (ii) from the point of view of pedologic evaluation, these countries are in relatively flat areas of higher altitude and present restrictions for agriculture due to the climatic conditions of extreme aridity.Regarding soil classification and land suitability: there is large spatial variability; (iii) the average values of actual yield for maize and soybean are lower than the attainable yields indicating that the main local problem is the low level of technology; (iv) for maize and soybeans, the highest attainable yield in the UEMOA region lies in the southern part of latitudes below 15°N; and (v) local government should support programs that ensure to small farmers access to better technologies and market models that could better remunerate agricultural activities.

Table 2 .
Classes of land suitability for annual crops based on land slope S

Table 6 .
-1 [carbohydrate]) is the efficiency of converting carbohydrate in dry matter, HI (kg[dry botanical seed] kg -1 [dry matter]) is the harvest index, and u (kg[water] kg -1 [wet botanical seed]) is the seed water content.N, P 2 O 5 and K 2 O fertilizer consumption in 2012 in the UEMOA countries, in Africa and in Brazil(FAO,  2015) )

Table 7 .
Soybean yield, harvested area and average productivity of the countries of the UEMOA,Africa and  Brazil, for 2013 (FAO, 2015)