This chapter gives a brief summary of the computational steps and data involved in the district-wise AEZ crop productivity assessment followed in this study. Most of the elements touched upon here have been discussed in more detail in some of the other technical annexes which will be referred to as appropriate.
The crop-wise productivity assessment comprises a first appraisal of the detailed Kenya land resources inventory that has been compiled (see Technical Annex 1). It also yields the necessary information and analysis to tackle more complex district planning tasks involving food production targets, crop and livestock interaction, cash crop production objectives, fuelwood requirements, the need to preserve forest and wildlife resources, etc. Scenario formulation for district planning is a subject of a further report.
The procedures introduced here provide detailed estimates of rainfed arable land resources, optionally on a district basis, by province and of the national total. The analysis can be carried out separately for each of the specified levels of technology (low, intermediate and high). The presentation of results, Appendix B to D, concentrates on the intermediate level of inputs which appears to be most realistic in relation to analysis for the year 2000.
In the following we will recapitulate the data elements included in the land resources inventory and their relation to the crop models involved.
The climatic resources inventory of Kenya quantifies both heat and moisture conditions. The quantification of heat attributes has been achieved by defining reference thermal zones. As temperature seasonality effects of latitude are minor due to the equatorial location of Kenya the temperature zones are closely correlated to altitude ranges. To cater for differences in temperature adaptability between crops, nine thermal zones have been distinguished in the climatic inventory in accordance with the four temperature adaptability groups of crops, pasture and tree species discussed in Technical Annexes 3, 5 and 6. Each thermal zone, based on intervals of 2.5 degrees Celsius, relates to an altitude range of about 385 meters.
Quantification of moisture conditions was achieved through the concept of reference length of growing period (LGP days). Crop growth is considered possible if moisture supply from rainfall exceeds ½ of potential evapotranspiration. The moisture period regime has been inventoried by means of three complementary concepts:
- number of separate length of growing periods within a year, summarized as a historical profile of pattern of length of growing periods per year (LGP-Pattern). Twenty two such LGP Patterns are recognized in the inventory.
- the mean total dominant length of growing period, i.e. the sum of mean individual dominant and associated length of growing periods occurring during the year. Fifteen LGP zones, thirteen spanning 30 day intervals each, plus an all year-dry and all-year-humid zone, are distinguished.
- year-to-year variability of each length of growing period and the associated moisture condition.
The map of dominant LGP zones and the map of LGP-Pattern zones together with the associated information (in table form) on length and probability of occurrence of corresponding associated growing periods provide the historical profile of any mean total dominant length of growing period in any of the twenty-two LGP-Pattern zones. A detailed presentation of the climatic inventory, data sources and compilation, is contained in Technical Annex 1.
The Exploratory Soil Map of Kenya (Siderius and van der Pouw 1980; Sombroek, Braun and van der Pouw 1982) at a scale of 1:1 million was used to compile the land resources inventory underlying the present assessment. This soil map includes information on distribution and characteristics of soils, landform and geology/parent material.
The digitized map distinguishes 392 different soil mapping units. These mapping units describe soil associations or soil complexes composed of dominant soils, associated soils and inclusions (390 mapping units) or relate to water bodies and major urban areas (2 mapping units).
The productivity potential and suitability of different soil units within a mapping unit may vary widely. Therefore, a complete mapping unit composition table has been provided (van der Pouw 1983) containing percentage allocation of the mapping units by soil unit, slope class, soil texture and soil phases. The mapping unit composition table also contains information derived from the legend of the soil map regarding landform and geology/parent material. Soil unite have been defined in terms of measurable and observable properties of the soil, and specific clusters of such properties are combined into ‘diagnostic horizons’. The 123 soil units (plus 5 miscellaneous units) inventoried in the Kenya soil resources inventory and the diagnostic horizons are presented in Technical Annex 1.
Soil texture is quantified in terms of three major textural divisions (coarse, medium, fine) which have been further subdivided into seventeen textural classes related to contents of clay, silt and sand. The presence of coarse material in the soil profile has been inventoried separately from soil texture by means of a coarse material indicator distinguishing six types of stoniness.
Soil phases indicate land characteristics which are not considered in the definition of the soil units but bear significance regarding the use, crop suitability and management of land. The nineteen soil phases recognized in this study relate to physico-chemical limitations of the soil, effective soil depth limitations and mechanical hindrance. Up to three soil phases have been combined resulting in some seventy different soil phase combinations occurring in the inventory.
Six basic slope classes, in twelve combinations, have been employed in the Exploratory Soil Map of Kenya. The six basic slope classes are: 0–2%, 2–5%, 5–8%, 8–16%, 16–30%, and >30%. These have been further refined by forming slope class associations and deriving mean slope quartiles for each of the slope classes. The slope association information, assembled in the form of a slope composition table, has been used to break up extents by soil units into several classes of ranges of slope gradients appropriate for matching to land utilization types and evaluation of soil erosion hazards.
The land resources inventory combines all the digitized Kenya map overlays that relate to climatic conditions, soil inventory, administrative units and selected properties of present land use, i.e. cash crop zones, forest areas, irrigation schemes, Tsetse infestation and game parks.
The compilation of the land resources inventory involves the application of the mapping unit composition table and slope composition table to the soil mapping unit overlay and the aggregation of the GIS raster information to agro-ecological cells. The steps involved include:
- creation of a GIS data base file
- sorting of GIS data base file and aggregation to GIS inventory file
- application of mapping unit composition table
- application of slope composition table
The land resources inventory for Kenya eventually contains unique records, holding eighteen fields each, with the following information:
The GIS map overlays are based on a grid of one square kilometer pixels covering a rectangle of 1085 by 900 kilometers, i.e. some 976,500 raster points per map (about 575 thousand of which within the country). The compilation sketched above produces a data file of about 91,000 unique (in terms of the combination of the considered land and climate characteristics) records comprising the Kenya land resources inventory. This level of detail ensures that each agro-ecological cell represents a fairly homogenous set of agro-climatic and soil physical conditions, a requirement crucial to adequately matching the cell properties to land utilization types. This will be discussed next.
A total of twenty-five crop species, divided into sixty-four crop types, are included in the assessment. The full list of crops and crop types is presented in Appendix A. Although the emphasis is on food crops, coffee, cotton, pineapple, pyrethrum, sisal and tea are considered, for their economic relevance to agriculture and to account for reported and/or projected land area occupied by these crops. The remaining nineteen crops are represented in terms of fifty-eight crop types to account for differences in ecotype adaptation and growth cycle within each crop species. In addition, pastures, comprising some thirty grass and legume species, and two classes of tree species, with and without nitrogen fixing ability, comprising some thirty-one fuelwood species, have been included in the assessment.
Crops have climatic requirements for photosynthesis and phenology both of which bear a relationship to yield. In the FAO AEZ methodology crops are classified into four climatic adaptability groups according to their fairly distinct photosynthesis characteristics. Each group comprises of crops of ‘similar ability’ in relation to potential photosynthesis.
For example, barley, oat, wheat, phaseolus bean and white potato have a C3 photosynthesis pathway and belong to adaptability group I. They are adapted to operate under cool conditions with less than 20 degrees Celsius mean daily temperature, i.e. an altitude above 1500 meters under Kenyan conditions. Mean daily temperatures of 10 to 12.5 degrees Celsius and below, i.e. the altitude range of 2700 to 3100 meters and above, are assumed to pose a risk of frost damage too great for successful cultivation of these crops. At altitudes between 1500 m and 2700 m (thermal zone 4, 5 and 6) the length of the growth cycle of these crops increases by 5 to 6 days for each 100 m increase in altitude.
Similarly, for each of the 64 crop types considered, for pastures and fuelwood species, the relation between crop growth cycles, thermal requirements and inventoried thermal zones in Kenya has been assessed and rated. Five suitability classes are employed, i.e. S1 to S4 and N, and the ratings apply to all three levels of input. A rating of S1 indicates that the temperature conditions for growth and yield physiology and phenological development are optimal and that thermal conditions do not hinder achievement of maximum yield potential. Ratings of S2, S3 and S4 indicate that temperature conditions for growth and development are sub-optimal resulting in suppression of yield potential by 25, 50 and 75 percent, respectively. A rating of N indicates that temperatures are not suitable for cultivation of the crop.
Potential yields when considering moisture related constraints, i.e. water stress, weeds, pests and diseases, workability, for individual LGPs were estimated according to the AEZ method (FAO 1978–81) for all crops (excluding the cash crops, coffee, cotton, pineapple, pyrethrum, sisal and tea; see below), pastures and fuelwood species. Yields refer to single crops which act as building blocks in the formulation of annual cropping patterns and crop rotations.
All annual crops are matched to individual component length of growing periods. The LGP-Pattern evaluation takes into account the probability weights associated with the historical pattern of occurrence of length of growing periods. This provides a measure of the variability of the yields, quantifying potential yields under average, best and worst historical moisture conditions.
Perennial crops, including cassava, banana, oil palm and sugar cane, are evaluated with respect to mean total dominant length of growing period with yield potentials adjusted according to LGP-Pattern group to account for moisture stress where appropriate.
For the cash crops, coffee, cotton, pineapple, pyrethrum, sisal and tea, for which agro-climatic yield potential has not been estimated, an allocation rating table relating crop suitability to LGP and LGP-Pattern has been formulated to identify appropriate climatic zones for consideration in the crop productivity and resource optimization algorithms.
The above procedures apply to the climatic inventory on all soils except Fluvisols, where intensity and duration of flooding governs potential cultivation rather than local precipitation. In addition, cultivation of these soils is normally confined to post-flood periods. For these reasons, Fluvisols were rated separately for all crops, combining agro-climatic and agro-edaphic considerations.
To assess the suitability of soils for crop production, soil requirements of crops must be known and compared to chemical soil properties and physical limitations imposed by landform, such as slope, texture and coarse material. The basic soil requirements of crop plants can be summarized under broad topics related to internal and external soil characteristics:
- soil temperature regime
- soil moisture regime
- soil aeration regime
- natural soil fertility regime
- effective soil depth for root development and foothold
- soil texture regime
- soil salinity and toxicity harmful to crop growth
- soil slope and topographic characteristics
- occurrence of flooding
- soil accessibility and trafficability
Many soil characteristics, e.g. natural fertility, salinity, pH, gypsum content, etc., can be defined in a range that is optimal for a given crop, a range that is critical, and a range that is unsuitable under present technology (see Technical Annex 3, and FAO 1978–81, 1980). These relationships have been applied in the matching of the inventoried soil units with the soil requirements of crops.
The soil unit evaluation, expressed for each crop in five suitability classes, S1 to S4 and N, relates soil properties to the degree to which crop requirements can be met under a given management practice, i.e. level of inputs. Where appropriate, the soil unit rating is modified according to limitations implied by texture evaluation, stoniness and phase evaluation. For further detail refer to Technical Annex 4.
Limitations imposed by slope are taken into account in a threefold manner. As explained earlier, the original mapping units are split up, with regard to the inventoried slope class, into uniform slope ranges according to a slope composition table. Seven slope ranges have been considered as appropriate for LUT matching: 0–2%, 2–5%, 5–8%, 8–16%, 16–30%, 30–45% and >45%. For example, mapping units with an attached slope class code 7, slope symbol CD, 5–16 %, are assumed to involve slope ranges of 2–5%, 5–8%, 8–16% and 16–30% with 5, 20, 70 and 5 percent of the mapping unit falling into the respective slope ranges.
Secondly, a slope-cultivation association screen defines those slope ranges which are permissible for cultivation in relation to crop type, land use practice and level of inputs.
The computation of the land productivity assessment also involves the calculation of potential topsoil loss as well as relating topsoil loss to productivity loss. In the analysis, topsoil loss is estimated for each permissible crop combination and level of inputs using a modified Universal Soil Loss Equation (see Technical Annex 2). In the specification of the USLE, slope gradient is an important explanatory variable in estimating topsoil loss. Yield loss is associated to the estimated topsoil loss through a set of linear relationships accounting for soil properties (susceptibility to erosion), climatic conditions (regeneration capacity of topsoil) and level of management.
The techniques outlined so far comprise the suitability assessment part of the land productivity model. All three assessments, the climatic suitability, the edaphic suitability and the evaluation of slope limitations are carried out for each agro-ecological cell to estimate crop performance. Land that is reserved for other uses, e.g. forest areas and game parks, can - but need not - be excluded from the analysis. The results of the suitability assessment in Kenya are presented in chapter 3 in the form of generalized maps, Figure 3.1 to Figure 3.3, for crops, pastures and fuelwood species, respectively. Information by district and province can be found in Appendix B.
Six suitability classes have been defined relating average cell suitability to maximum attainable yield. Classes C1 to C5 relate to average attainable yields of >80%, 60–80%, 40–60%, 20–40%, and 5–20% compared to maximum yields. Note that extents in suitability class C5 will not be considered among the viable crop options, but have been included here to indicate the scope of production in very marginal areas. A sixth class accounts for areas that are entirely unsuitable or allow only for less than 5 percent of maximum yield. Data for the non-suitable class are not explicitly included in the tabulated results. The assessment in Appendix B assumes that cells marked as belonging to an irrigation scheme, forest zone or game park cannot be considered for rainfed production of crops, pastures or fuelwood species.
Consequently, this information is used to determine crop productivity potential. To this effect further considerations have to be taken into account including multiple cropping, in space and time, fallow requirements to maintain soil fertility, and production stability constraints to reflect consequences of rainfall variability.
Multiple cropping refers to the intensification of arable land use, both in time and space. The principles of yield increases resulting from a better use of time with crops in sequence is complementary to increases arising from a more efficient use of space with crops in mixture.
Sequential cropping is possible in areas where climatic conditions, temperature and moisture supply, permit crop growth beyond the duration of one crop. This mostly applies to areas in Kenya with a length of growing period above 210 days. The sub-humid and humid zones account for only 12 percent of Kenya's land area but some 60 percent of the rainfed arable land resources. The algorithms implemented for land productivity assessment explicitly construct and evaluate all feasible sequential crop combinations, both multiculture and monoculture, by matching individual crop cycle requirements to the relevant component length of growing periods implied by the LGP-Pattern. In this Technical Annex, however, we present arable land estimates and production potential for individual crops only (including sequential monoculture of crop types belonging to the same crop species).
Intercropping increments, i.e yield advantages from practices like strip, alley and relay cropping, are assumed to increase with length of growing period. No yield advantage is adjudged in LGPs with less than 120 days. Also, it is stipulated that advantages would be largest with mixtures where the individual component crops are rated as very suitable. The relative contribution of intercropping is assumed to be most pronounced under low level of inputs. No increment is applied under high level of inputs, since the mechanization requirements associated with this management level are generally regarded as being incompatible with intercropping practices. For wetland rice, sugarcane, banana, oil palm, pastures and fuelwood species no intercropping increments were applied.
To make the assessment more realistic, the model includes, as a scenario variable, a parameter stating the required minimum level of production stability. As the objective function of the planning scenarios is usually formulated in terms of average production conditions, i.e. yields attainable under different patterns of growing periods weighted by their historical profiles of occurrence, the production stability constraint stipulates that i each agro-ecological cell the selected activities would satisfy the condition that the minimum production levels, i.e. output in climatically worst years, would lie within a tolerable range to the best alternative under worst conditions. The trade-off is between loss i average productivity and risk aversion. The tolerance parameter can be varied with level of inputs. In the results presented, a production stability constraint has been imposed only when determining total arable land potential, but not when quantifying productivity potential of individual crop types. The estimates of potentially arable land by crop type, listed in Appendix D, therefore possibly include areas where production may not be feasible when imposing a production stability constraint.
The AEZ assessment is concerned with sustainable agricultural practices. In their natural state, many soils cannot be continuously cultivated without undergoing degradation manifesting itself in decreasing crop yields and deterioration of physico-chemical soil properties. The assumptions on fallow requirements incorporated into the model are formulated for four main groups of crops - cereals, legumes, roots and tubers, banana and sugar cane - and relate to soil unit, thermal zone and moisture regime. For details refer to Technical Annex 4. In the results, Appendix B to D, however, fallow requirements were not deducted in calculating crop-wise potentially arable land and implied production.
