RESULTS AND DISCUSSION

Outputs from BAMnut Model - Crop Biomass and POD Models

The outputs of multiple simulations of the model for both biomass and pod yield have been classified into four representative ranges of suitability as shown in Table 2. Although these categories are arbitrary, they help to simplify the analyses and provide some basis for comparisons between regions. The levels, defined as very suitable (VS), suitable (S), moderately suitable (MS) and unsuitable (US), are based on reported pod yields for bambara groundnut at different locations in Africa. For example, the VS category with pod yields greater than 3000 kg ha-1 corresponds with reported pod yields of up to 3870 kg ha-1 in Zimbabwe (Johnson, 1968). Similarly, the MS category corresponds with the typical farmers' yields of 650-850 kg ha^-1 in Africa reported by Stanton et al. (1966). Regions producing pod yields below 300 kg ha^-1 are defined as unsuitable (category US) for bambara groundnut cultivation. Crop failure has not been included as it obviously has the definition of no yield.

Figures 10 and Figure 11respectively show the predicted biomass and pod yield of bambara groundnut across the world. The different suitability ranges defined in Table 2 are shown in different colours to indicate the geographical distributions of each classification.

Figures 10 and 11 show that there is potential for bambara groundnut production in many parts of the world beyond its current distribution with suitable areas with potential in America, Australia, Europe and Asia as well as Africa. In fact, locations within the Mediterranean region show the highest predicted biomass, often exceeding 8.5 t ha^-1.

Figures 12.1-12.11 and 13.1-13.11 provide a more detailed analysis of bambara groundnut biomass and pod yield predictions based on continent and country scales. The combination of Figures 10 and 11 as biomass and pod yield ranges with Figures 12.1-12.11 and 13.1-13.11 as the percentage of suitable arable lands for major parts of the world provides an integrated assessment of the potential production of this crop.

Figure 12.1describes the potential biomass production for each continent. Of these, it is interesting to note that almost 75% of South America and 25% of Asia, two regions of the world not previously associated with bambara groundnut, are classified in the VS, S or MS ranges, i.e. as at least moderately suitable for bambara groundnut cultivation. The comparable figure for Africa, where almost all bambara groundnut is currently grown, is 65%. It is also noteworthy that Europe, which has less than 10% of land that is classified as at least moderately suitable for cultivation, has the largest proportion (3%) within the VS category.

TABLE 2

Classification of suitability ranges
for predicted biomass and pod yield of bambara groundnut

Note: kg ha^-1 refers to pod or biomass per crop and may not be restricted to one crop per year.

Figures 12.2-12.11 provide more detail on the biomass predictions for specific countries within each region. Figure 12.2 shows that there is relatively little productive potential for bambara groundnut in Algeria, the Libyan Arab Jamahiriya and Egypt. In contrast 43% of Tunisia and 36% of Morocco are within the VS category for potential biomass.

Most areas in Central Africa (Figure 12.3) can be classified as moderately suitable. Although 45% of areas of Chad and 40% of the Sudan show crop failure, the whole of the Central African Republic can be classified within at least category MS for cultivation. Predictions indicate that bambara groundnut may be unproductive in some areas of southern Africa Figure 12.4). For example all of Lesotho is classified as unsuitable. In contrast, more than 90% of Botswana, more than 57% of South Africa and nearly 70% of Namibia are classified within at least category MS. All of Swaziland falls within the S and MS categories of suitability.

Figure 12.5 shows that, with the exception of Djibouti, all of East Africa is classified within categories S and MS for biomass production of bambara groundnut. Virtually the whole of West Africa (Figure 12.6) is classified within category MS. Surprisingly, there are no countries in this region that are within the VS or S categories. Nevertheless, complete crop failure is unlikely.

Despite its historical absence from this part of the world, a general assessment of the American continent (Figure 12.7) shows that there are many areas in this region that have high potential biomass productivity for bambara groundnut. In all countries in South America, with the exception of Argentina, more than 75% of the land area is within the S and MS categories of suitability. In contrast, 87% of the USA is classified as unsuitable for cultivation.

Within Asia (Figure 12.8), all of India, the Lao People's Democratic Republic, Philippines, Viet Nam, Sri Lanka, Malaysia, Cambodia, Indonesia, Thailand, and Bangladesh fall within categories S and MS.

Much of the Near East (Figure 12.9) is classified as unsuitable for cultivation of bambara groundnut, mainly because of low rainfall. However, for those countries with a more Mediterranean climate, there are potentially highly productive areas. For example 100%, 80% and 75% of Cyprus, Israel and Lebanon respectively fall within category VS of suitability.

The low potential productivity for much of Europe (Figure 12.10) is similar to that of the Near East but, for Europe, it is mainly low temperatures rather than soil moisture deficits that limit growth and development to only part of the year. However, in the Mediterranean countries of southern Europe there is considerable potential for cultivation. For example, 71% of Portugal, 60% of Greece, 57% of Spain and 49% of Italy are within categories VS and S. Where these appropriate environments are coupled with improved cultivation techniques there is considerable scope to achieve high yields of bambara groundnut (an indigenous African legume) within countries of the European Union.

Similarly, in Australia and Oceania (Figure 12.11) there are significant areas that are classified as moderately suitable for bambara groundnut cultivation. For example 62% of Australia is within categories S and MS.

Figures 13.1-13.11 provide the pod yields that correspond to the total biomass predictions in Figure 11. With the exception of Europe (6.5%), at least 10% of the land area of each continental region (Figure 13.1) can provide bambara groundnut pod yields within category S (1000-3000 kg ha^-1). As much as 30% of Africa falls within this category.

Within Africa (Figures 13.2-13.6) the largest potential yields occur in South, East and Central Africa. For example, 100% of Swaziland, 98% of Uganda, 95% of Zambia, 89% of Burundi, 84% of Zimbabwe and 79% of the Central African Republic have yield predictions within category S.

For America, (Figure 13.7) all of El Salvador, Honduras, Cuba and the Dominican Republic fall within category S for pod yield. Countries such as Uruguay, Panama, Haiti, Paraguay, Venezuela and Ecuador have land areas that are entirely within categories S or MS.

Within Asia (Figure 13.8), all of the Lao People's Democratic Republic, Malaysia, the Philippines, Sri Lanka, Viet Nam, Bangladesh, Indonesia, Cambodia and Thailand fall within categories S and MS. However, the country with the largest proportion of its area in category S is Tajikistan with more than 70% in category S.

In the Near East (Figure 13.9), all of Cyprus, 80% of Israel, 75% of Lebanon and 70% of the Syrian Arab Republic are within category S.

For Europe (Figure 13.10), Portugal, Greece, Spain and Italy have the greatest productive potential for bambara groundnut, with as much as 62% of Portugal in category S and more than 50% of the land area in each country within categories S and MS. There is even a small potential for bambara groundnut in France with over 12% of land area in categories S and MS.

Australia (Figure 13.11) has 60% of its land area within categories S and MS.

At this stage it is important to note the following limitations to the above analysis of potential pod yields in bambara groundnut. First, the methodology takes no account of specific soil types. Although the model requires inputs that depend on the physical characteristics of soil e.g. available water, there is no attempt to assess productivity in terms of soil classification for contrasting locations. Second, there is no allowance for the effects of pests and diseases on the capture and conversion of environmental resources and allocation to pod yield. Third, many bambara groundnut landraces have a specific daylength requirement for pod filling, i.e. allocation to yield will only begin at a particular daylength. In assessing potential productivity at each location, this daylength requirement has not been included.

This limitation is likely to restrict the period of the year when maximum crop yields can be achieved and its influence becomes more significant at locations with greater annual variation in daylength, i.e. those progressively further from the equator.

In practice this limitation can be overcome by inputting the daylength requirements for any particular landrace in relation to productivity at any specific latitude.

From the above analysis it is difficult to identify a uniform ecophysiological niche for bambara groundnut. However, based on model predictions, a combination of uniform distribution of rainfall during the growing season linked with relatively cool temperatures in terms of tropical species that extend the growing season appear to result in the highest potential pod yields.

Evaluation

As mentioned earlier, the spatial patterns of simulated yield can improve production estimates and highlight areas that are most vulnerable to drought (Carbone et al., 1996). However the main limitations in such analyses are not only the limited availability of climate and soils data, which preclude the use of the more sophisticated simulation models, but also the lack of observed or reported crop yields. This problem is common to all underutilised crops where a common system of yield reporting and evaluation is lacking. It is therefore extremely difficult to verify predicted yields of underutilised crops against reliable figures for actual on-farm yields achieved at similar locations. Moreover, predictions in this study are simply based on agro-ecological potential and ignore the effects of management. In some cases the actual yields are based on farmers fields (unreliable and unreplicated) and in others on experiments with management inputs. Thus, at this stage, it was not possible to construct a comprehensive comparison between predicted and reported figures.

Based on the above and on the scarcity of data available, Table 3 shows a simple comparison between predicted and reported pod yield values.

TABLE 3

Comparison between predicted and reported pod yield values (kg ha^-1)