This study was conducted in the Centre Cameroon region, specifically in the production basins of Bokito and Awae. The Bokito Basin, located between latitudes 3.965 - 5.041 N and longitudes 10.466 - 11.771 E, belongs to the forest-savannah transition agro-ecological zone, which is characterized by bimodal rainfall, with an average annual precipitation of 1350 mm, an average annual temperature of 25˚C and an altitude ranging between 210 and 1212 m [22] . This basin is made up of four sub-divisions namely, Bokito, Bafia, Kiiki and Ombessa, with Bokito being the largest in terms of area. The Awae Basin is located between latitudes 3.618 - 4.106 N and longitudes 11.650 - 12.268 E, and includes the sub-divisions of Awae and Mengang. This basin belongs to the bimodal rainfall forest agroecological zone, with an average annual precipitation of 1500 mm, an average annual temperature of 24˚C and an altitude raging between 544 and 1168 m above sea level [23] .

Climate data were obtained from the Bafia weather station for the Bokito production basin, and online through the Climatic Research Unit gridded Time Series (CRUTS) platform for the Awae production basin [24] . The climate characteristics of interest were average annual precipitation, average annual temperature and average annual relative humidity, which enabled a climate assessment for pineapple [6] [7] .

Pedological data are based on previous studies conducted in the Bokito and Awae Basins. For this purpose, these data were extracted from dissertations, published articles [25] [26] [27] [28] [29] . Only those of the Awae Basin were obtained from the camsodat 0.1 database [30] . For the first 5 profiles in the Awae Basin, CEC was estimated by the Pedo-Transfer Function (PTF), Equation (1) for Ferraslsols [31] . The selected profiles were chosen on the basis of the topo sequence, after projecting onto a map and creating a topographic profile. This was done in order to take into account the diversity of soils and to avoid redundancy of the same soil type. The soil types found in the Awae Basin are deep A-Bo-C Ferrasols and those in the Bokito Basin are deep A-Bt-C Acrisols with Gleyic properties, variable depth A-Bt-C Nitisols and medium depth A-(Bw)-C Cambisols [29] [32] [33] [34] .

The processing of data related to the soil characteristics such as texture, Organic Carbon (OC), Cation Exchange Capacity (CEC), pH, Electrical Conductivity (EC) and Exchangeable Sodium Percentage (ESP) follows the procedure described by [35] .

The assessment of land suitability for pineapple cultivation of the three types of land use observed in the Centre region [36] was done according to the Fuzzy-AHP method. This method is described as more accurate than the FAO parametric method following Storie and/or square root modalities [11] [15] [37] [38] [39] [40] [41] .

Based on the land characteristics for rainfed pineapple [7] , the Sigmoid and trapezoidal Fuzzy functions (Table 1) were used to determine the constraint score of each land characteristic [11] [42] . This score is contained in the standard range of 0 to 1, with 0 representing extremely severe limitations and 1 no limitations for the characteristic of interest. For this interval, five categories, S1-0 suitable without any limitations, S1-1 suitable with low limitations, S2 moderately suitable with moderate limitations, S3 marginally suitable with severe limitations and N unsuitable with very severe limitations, were defined by FAO [6] . After recalculating the clay and silt contents along the profile, only the clay content value was used to determine the texture score [43] .

Land assessment is a multi-criteria method requiring an understanding of the impact of each of these criteria in contributing to crop yield. As a matter of fact, each land characteristic is a criterion that influences crop yields. However, at different levels, the judicious choice of the weight of each of these criteria makes it possible to bring the estimate as close as possible to the observation. [44] defines a procedure to take into account the impact of each criterion by comparing

For the sigmoid function, a, b and c represent the lower bound with very severe limitation and a score of 0, severe limitation with a score of 0.5 and no limitation with a score of 1 respectively. For the trapezoidal function, a and d represent the lower and upper bound with very severe limitations and a score of 0, b and c are the interval for which there are no limitations with a score of 1.

them two by two, i.e. the importance of a criterion a versus b. This comparison is done using a quantitative grading scale (Table 2). The validity of this procedure lies in the Consistency Ratio (CR) which must be less than 10% to be accepted [44] . The consistency ratio is given by Equation (1). The closer the consistency ratio is to 0.000, the better the consistency of the weighting matrix. This step provides the weight for each of the characteristics.

RC = CI/RI (1)

where CI = Consistency Index given by Equation (2); RI = Random consistency Index given by (Table 3).

CI = ( λ max − n ) / ( n − 1 ) (2)

where λ_max = maximum eigenvalue of the judgment matrix; n = order of the matrix or number of characteristics assessed.

The final land suitability index is given by Equation (3), and is the result of additive linear combination models [12] [15] .

IA = ∑ i = 1 n w i × A i (3)

where IA is the final suitability, W_i is the weight of factor i, which is calculated from the matrix AHP, A_i is the degree of contribution of each land characteristic with i ranging from 1 to n (Table 4). With the values of W_i and A_i both between 0 and 1, the land index is therefore between 0 and 1, with 0 indicating unsuitability and 1 perfect suitability. The final suitability is assessed on the basis of the intensity of the limitations (Table 4), which is similar to that of [45] . In this work, climate and soil are considered to have the same weight and there is no need to construct a two-dimensional matrix [44] .

The yield was estimated according to the FAO yield table for rainfed agriculture [35] . For this purpose, each bound of the yield interval of the pineapple crop without any constraints was multiplied by the value of the land index to define the interval corresponding to the determined suitability.

Pineapple crop yield is strongly influenced by the seedling density. However, several authors argue that density does not influence the fruit mass [46] [47] . For these reasons, a second estimate was made to take into account the increase in density and its contribution to yields with the advanced research on pineapple production intensification Equation (4):

Y = N P R × ( I T × M F C ) (4)

where

Y = yield in t/ha; NP_R: number of fruits harvested per unit area; M_FC: the average fruit mass for the given cultivar

The fruit mass of the smooth Cayenne variety used in the two production basins of the Centre region is between 1.5 and 4.5 kg. Minor soma clonal variations have resulted in variants with average fruit masses of 1 to 2 kg and 2 to 4 kg for the smooth Cayenne. That is, an average range of 1.5 to 3 kg/fruit, which was considered in this work [48] [49] . The number of fruits harvested in the Awae Basin ranges from 25,180 to 36,628, and in the Bokito Basin from 230 to 1390 fruits, these limits were considered in the yield assessment [36] .

The climate limitations imposed on the 6 land units in the Awae production basin are due to the annual precipitations ranging from 1546 to 1577 mm, which is moderate. The Bokito production basin has no climate constraints to pineapple production (Table 7). The base saturation rate V% has very severe limitations for all identified land units. This is because the sum of exchangeable bases is low, there is imbalance in the cation balance and the CEC is high. These soils can handle large amounts of fertilizer because of their low reservoir. The high clay content in these soils gives very severe limitations for pineapple quality for all identified land units; however, its Ferralsols structure improves infiltration and does not lead to temporary waterlogging situations. The limitations imposed by pH are severe, for land units from 2 to 6, phosphorus fixation by aluminum starts for pH ≤ 5.5 and increases for even lower values, the pH values for these units are in the range of 4.2 to 4.8 (Table 8). Temporary waterlogging imposes extremely severe constraints to pineapple cultivation on 99% of the land units in the Bokito production basin. Despite having 62.5% of favorable textures, these soils are dominated by micropores that reduce water infiltration and lead to this temporary saturation condition of 24 to 120 hours. 37.5% of these soils have very severe permanent limitations imposed by the (clay) texture. 50% of the soils have very severe limitations imposed by the organic carbon content (≤0.8%) and 25%

AW: for the Awae production basin; BK: for the Bokito production basin; indices 1 to 8 each represent a soil unit in each of the production basins.

AW: for the Awae production basin; BK: for the Bokito production basin; indices 1 to 8 each represent a soil unit in each of the production basins. CF: Coarse Fragment; CEC: Cation Exchange Capacity; V%: Base Saturation; OC: Organic Carbon; EC: Electrical Conductivity; ESP: Exchangeable Sodium Percentage; SW: Soil Water Saturation; Text: Texture.

have moderate limitations for the same characteristic. In addition, CEC imposes very severe limitations on 50% of the soils and moderate limitations on 25% of the soils. These soils are poor and cannot handle large amounts of fertilizer, the reservoirs are more than 50% full, only 1% of these soils have extremely severe limitations for base saturation. Also, 1% of these soils have moderate limitations due to the percentage of coarse fragments on the surface. This makes the soil preparation time increase as a result of the drudgery imposed by these coarse fragments (Table 8).

Pineapple has a moderate suitability S2sf in the production basin. Constraints come from texture (physical fertility), base saturation rate and pH. Expected yields range from 32 to 38 t/ha. The soil in the Bokito Basin is suitable S1wf for 25% of the soils and moderately suitable 12.5% S2sf, 37.5% S2wf and 25% S2wsf for 75% of the soils. Expected yields range from 34 to 43 t/ha (Table 9). Constraint correction (IT_PO) for base saturation, pH and exchangeable sodium percentage will improve the land index of all soils in the Awae Basin without changing the suitability class. The dominant constraint here is texture, which cannot

AW: for the Awae production basin; BK: for the Bokito production basin; indices 1 to 8 each represent a soil unit in each of the production basins. UP: Pedological Unit; IC: Climate Index; SI: Soil Index; ITA: Current Land Index; ITPO: Potential Land Index; RdtA: Current Expected Yield; RdtPO: Potential Expected Yield; YA (1.5): Current expected yield Equation (4) for a fruit mass of 1.5 kg; YPO (1.5): Potential expected yield Equation (4) for a fruit mass of 1.5 kg; YA (3.0): Potential expected yield Equation (4) for a fruit mass of 3.0 kg; (3.0): a - b in YA (1.5) and YPO (1.5) are calculated for the min and max numbers of harvested plants 25,180 - 36,828 found in the Awae Basin and 230 - 1390 in the Bokito Basin respectively.

be corrected. The correction of temporary water saturation, organic carbon, pH and exchangeable sodium will increase the land index and make 87.5% of the soil suitable for pineapple cultivation. These corrective measures are important; however, their influence on yield remains low. The use of high-yielding cultivars or more productive variants of the same cultivar will increase yields (Table 9). The two methods of estimating yields gave overlapping intervals, so it is possible to use this second method to determine yields with intensification-type farming practices.