As I mentioned earlier, all households in my sample engaged in crop cultivation.7 For some households, however, crop cultivation was not the main occupation. In almost fifty percent of the households in the sample, agriculture provided less than half the income in cash and kind.8
6 Kendall’s tau-b=0.18 with significance 0.09. When, more correctly, the correlation is calculated between
land tenure situation and overall vulnerability minus index 7 (land tenure), Kendall’s tau-b even decreases to 0.026 with significance 0.799 (not significant). I used Kendall’s tau-b to calculate the strength of the correlation because both variables are of ordinal scale and because many ‘ties’ (equal scores) occur (see De Vos 1983: 60-67).
7 Note that crop cultivation was not a precondition for being part of the sample.
In most of these households, other sources of livelihood were ‘miscellaneous’, however, and no single source was more important than crop cultivation.
The farming system in the research area is based on cereals, legumes and yams. Farmers make little use of external inputs. The average area farmed per household in the sample was 6.93 acres or 2.73 hectares. With an average household size of 6.19 consumer units (adult male equivalents, see appendix), the average acreage per consumer unit was 1.12 acres or 0.44 hectares. The average household size expressed in number of persons was 8.38. Thus, the average area cultivated per capita was 0.83 acres or 0.33 hectares. When the average acreage per consumer unit is calculated as the average of individual household scores, the average increases to 1.23 acres or 0.48 hectares. This is because smaller households in the sample cultivated relatively more land per consumer unit.9 The respective figure for acreage per capita is 0.92 (0.36 ha).
A few calculations can put these figures into perspective. If we assume an average yield of six hundred kilograms of grain equivalents per hectare,10 the harvest of an average farm household (0.48*600=288 kg/consumer unit and 0.36*600=216kg/cap) was just enough to meet minimal energy requirements.11 The acreage farmed per capita varied considerably12 between households, however, depending on availability and strength of household labour, access to land, crop mix13, tillage method, access to capital, access to communal and hired labour and the degree of livelihood diversification and ‘de-agrarianisation’. In the survey sample, thirty-eight out of sixty households cultivated less than the average acreage14 and, with the same assumptions as above, were likely not to be self-sufficient in their food produc-
9 Pearson’s r = -0.253 with significance 0.051. A possible explanation for this correlation is that in the survey,
both agricultural and non-agricultural activities were relatively underreported in larger households.
10 See below for an estimation of yields of different crops on different types of fields. The average yield levels,
as reported by the Ministry of Food and Agriculture for Lawra district in 1999, were substantially higher: maize: 900 kg/ha; millet: 800 kg/ha; guinea corn: 1000 kg/ha; groundnuts: 1120 kg/ha.The acting manager of the Nandom Agricultural Project (Mr. Stan, interviewed in Nandom Town, 26/10/1999) confirmed that the yield levels as reported by the ministry are not attained in the Nandom area. According to Rung-Metzger and Diehl (1993: 194-5), “when yield data are estimated through crop cuts, there is the general danger to calculate averages only on the basis of measured crop cuts. Thereby, very often crop failures are neglected. Thus, average crop yields are easily overestimated (...).” Runge-Metzger & Diehl (ibid.) compared measured
and recalled (by farmers) yield levels. The latter were found to be considerably lower, also when crop fail- ures were included in the measured yield levels. This indicates that yield estimations of farmers tend to be lower than the actual yield levels.
11 In the Ghana Living Standards Survey (GLSS), minimal energy requirements were set at 2900 kcal/adult
male equivalent per day (which is on the high side). If we assume that one kilogram of grains contains 3500 kcal, the minimal amount of grains needed per adult male equivalent is 302 kg. The minimal energy requirement in Ghana according to FAO is set at 1830 kcal per capita, which amounts to 190 kg. So according to the FAO standards for Ghana, an average household in the survey was able to meet food needs in 1999 (assuming a yield of 600 kg grain equivalents per hectare). According to the GLSS method, individual food production was just below the energy requirement (Ghana Statistical Service 2000b; Runge- Metzger & Diehl 1993: 198; Dietz et al. 2001b: 201; personal communication with Dr. B. Burlingame,
Senior Officer of the Nutrition Assessment and Evaluation department of FAO, 7/11/2001;). Note that post- harvest loss and seed requirement are not integrated in the analysis (see pathway of Egidius Dugyi in chapter eight for a more disaggregated calculation in an individual farm household).
12 Standard deviation: 0.765 acres; standard deviation/average: 0.62. 13 Some crops are more labour intensive than others.
14 The distribution was positively skewed (Sk = 2.166). A few farm households cultivated an amount of land
tion. The acreage cultivated per consumer unit correlated quite strongly and very significantly with overall vulnerability.15 Households with a large area under cultivation were relatively secure.
Photo 4
Woman sowing guinea corn with a baby on her back
Location of farms
People have fields around their compounds and fields at varying distances from their dwell- ings. The compound farms (‘sigman’) are usually upland farms, since people build their houses in relatively high places to protect against floods. On compound farms, the most common crops are guinea corn, maize, early millet, late millet and different vegetables like pumpkin, okra, bir16 and gyamboro. These vegetables are grown (by the women) between the grains, and are used as soup ingredients to accompany the staple foods. Occasionally, beans are sown on the compound farm, too. Some women have small17, fenced vegetable gardens around the house. These should not, however, be confused with the bigger and male-con- trolled dry season gardens.
Beyond the compound farm, the Dagara distinguish the farm of the settlement (‘puo’), from the farm beyond the settlement (‘wie’) and the uncultivated land in ‘the bush’ (‘kar’) (Tengan 2000: 228-233). It is not always clear where one ends and the other one starts, but the ‘wie’ is always beyond a natural or man-made boundary like a hill, river, or road. This classi- fication in ‘sigman’, ‘puo’, ‘wie’ and ‘kar’ does not coincide with the classification of land
15 Spearman’s rho = 0.588 with significance 0.000. The overall vulnerability index includes the acreage
cultivated per consumer unit. When the contribution of this individual index is isolated, the correlation is of course weaker: Spearman’s rho = 0.442 with significance 0.000). I used Spearman’s rho to calculate the strength of the correlation because the lowest scale variable (overall vulnerability index) is of ordinal scale, and because virtually no ‘ties’ (equal scores) occur (see De Vos 1983: 60-67).
16 Hibiscus subdariffa.
use types in compound farms, bush farms, far-away bush farms and fallow land, used in farming system research in northern Ghana (Runge-Metzger & Diehl 1993: 64). Far-away bush farms are “more than 30 km away from the villages” (ibid.). This type of farm does not exist in the research area. Only six out of sixty farmers reported to have farms at a distance greater than two miles from their dwellings. The most distant farm was six miles from the cultivator’s dwelling.
In my survey sample, a breakdown of farm sizes according to location and drainage situa- tion reveals that thirty-two percent of the land people cultivated lies in the direct vicinity of their compound (‘sigman’). These fields are mostly upland fields. The fields of the settlement (‘puo’) and beyond the settlement (‘wie’) can be subdivided into upland fields and lowland fields. Besides the compound fields, forty-two percent of the total land under cultivation concerned upland fields, and twenty-six percent concerned lowland and valley bottom fields. Most households have access to upland as well as lowland farms. In my sample, only eight households did not cultivate lowland farms.
Upland farms
In the upland farms of the neighbourhood (‘puo’) and beyond the settlement (‘wie’), millet and guinea-corn are the main crops, but legumes such as beans, bambara beans and ground- nuts are important here, too. The importance of maize decreases with increasing distance from the compound and with decreasing soil fertility. Towards the east of the research area (around Lambusie, in Sisala land), cowpeas are an important upland crop. A few farmers have recently started to experiment with soybeans. The most common inter-crop18 combinations are millet with beans, guinea corn with beans (cereal-legume), guinea corn with millet or maize (cereal- cereal) and groundnuts with bambara beans (legume-legume). A few farmers intercropped groundnuts and cereals, but this was not as common a practice as it is elsewhere in northern Ghana.19 In cereal-legume mixtures, beans are sown first and used as a cover crop to improve germination and reduce the impact of dry spells. Densities of bean plants are usually low, and very few farmers reported bean harvests of above a hundred kilograms. They quantify their bean harvest in bowls (2.5 kg) rather than bags (approximately 100 kg). Intercropping is an effective method to combat certain weeds and to reduce the variability of output in areas where rainfall is unreliable. Moreover, total output per hectare is often higher under inter- cropping regimes than under sole cropping (see Richards 1985: 66). Intercropping does not require much extra labour input when fields are tilled manually. Moreover, farmers indicated that intercropping is an effective way to deal with or even benefit from micro-variations in soil characteristics (patchwork agriculture).
18 “Intercropping – the planting of different crops in the same field during the same season – is used
synonymously with ‘mixed cropping’. Where a crop is planted and harvested, and followed by further crops in the same year, it is usual to speak of ‘sequential cropping’. The term ‘relay cropping’ is used where these sequences overlap. Intercropping, sequential cropping and relay cropping are thus distinguished from ‘sole cropping’ – the planting of one crop per field per season, and ‘monoculture’ – the planting of a single crop in the same field for a succession of seasons, or indefinitely.” (Richards 1985: 63-64). Intercropping has several advantages: less year-to-year variation in output; weed suppression and higher total energy output (Al- Hassan et al: 135).
Fixed crop rotations as described by Adolph et al. (1993: 52-3) for the Upper West Region were not reported in the research area, but some farmers rotate millet and guinea corn to combat striga (a weed). Some farmers grow groundnuts or bambara beans when a field is exhausted.20 The following year, these fields will be sown with cereals again. Sole cropping of millet, guinea corn and groundnuts occurs most on upland ‘bush farms’.
Lowland farms
The principal crops on lowland farms are rice and yams. Farmers cultivate yams on mounds with rice in between. In addition, rice is grown as a sole crop on separate fields. Some people also sow maize, guinea corn and (early) millet on small mounds with rice in between and a few farmers cultivate sweet potatoes. The cultivation of lowland farms gains importance with increasing population pressure and declining fertility of upland soil. In areas with high pressure on the land, the most fertile soil occurs down the slope (see Runge-Metzger & Diehl 1993: 65; Ruthenberg 1980: 127). Many farmers in the Nandom Area indicated that increased use of lowland farms was also a response to more unreliable rainfall and drier conditions. When a drought strikes and upland crops fail, lowland fields may still yield well. This is due to the difference in the drainage situation. Increased lowland cultivation is a typical example of adaptation to altered conditions. It is, however, difficult to determine to which altered conditions people have primarily adapted their farm practices. What is the driving force behind this change in land use: increased population pressure or climate change? I cannot answer that question unambiguously.21 People adapt their livelihoods to a set of altered condi- tions, and it has been my aim to also take the non-climatic factors into account in order to avoid climatic determinism.
Dry season gardens
On the valley floors, one also finds the commercial dry season gardens where tomatoes, okra, garden-egg, pumpkin (-leaves), bean (-leaves), lettuce and hot peppers are the most important crops. In some dry season gardens, people also grow cassava, sweet potatoes, plantain and cabbage. Dry season onion farming has not gained much importance in the Nandom area, but in the Lambusie area, three out of eight respondents did cultivate onions.
In 1999, dry season gardening was a source of income to eleven households in the sample. Four more households did have a dry season garden, but did not cultivate it in the year under investigation. The gardens are spatially concentrated in the village of Burutu, just south of Nandom Town, and in Lambusie, six kilometres east of Nandom Town. Besides topography (access to valley bottoms) and infrastructure (dams and dugouts), market access is an impor- tant determinant in the decision to start a dry season garden. In the most remote villages, no dry season gardens were encountered. On Sundays, but if necessary also on weekdays, Burutu gardeners sell their produce at the Nandom market. Gardeners from Lambusie sell their
20 Groundnut plants fix nitrogen in the soil.
21 There is no significant correlation between ‘distance to Nandom Town’ (as a proxy for population pressure)
and the percentage of land cultivated in lowlands (Pearson’s r = 0.026 with significance 0.41). In the villages near Nandom Town, where population pressure on the land is higher, people do not tend to cultivate a larger part of their total farm in lowland areas.
produce at several markets, including more distant ones. Their onions are less perishable than the tomatoes that are mainly grown in Burutu. For a more detailed account of dry season gardening, see ‘pathway’ of Suurib Kyoo (Van der Geest 2002a).
Crop yields
Crop yields are highest on compound farms, in the lowland and valley floor farms and in the heavily manured and intensively cultivated dry season gardens.22 Compound farms are rela- tively fertile because household waste, human waste, animal droppings (in the dry season) and animal dung from the kraal (in the wet season) supply the soil with nutrients. Few farmers carry manure to their more distant fields. The compound farms are cultivated continuously.23 Although yields are relatively high on the compound farms, good yields are not guaranteed. In 1999, for example, seventeen out of forty-five households24 estimated the yields on their compound farms to be below 500 kilograms per hectare. This figure excluded the harvest of vegetables used as soup ingredients. Nineteen farmers estimated their compound farm yields to be between 500 and 1000 kilograms per hectare, and nine farmers estimated it at over 1000 kilograms per hectare.
In the lowland and valley floor farms, relatively high yields are mainly due to the higher fertility of the soil and higher labour input. Lowland soil is usually heavier and difficult to till. Yam and rice cultivation is more labour intensive than the cultivation of the staple crops millet, guinea corn and maize. In my sample, forty-six out of sixty households cultivated rice. This contradicts the agricultural production data of the Ministry of Food and Agriculture. Those data suggest that rice production is negligible in the Lawra District. On average, the farmers in my sample sowed about ten to fifteen percent of their cultivated land with rice. In 1999, a wet year25, eighteen farmers estimated their rice yields at over a thousand kilogram per hectare.26
Thirty-three households in the sample cultivated yams. Yield levels were usually reported in the number of harvested tubers. The average reported yield was 241 tubers, with an average field size of approximately a quarter of an acre. On the Nandom market, one average-sized yam tuber yielded about 1,000 cedis in 1999. These figures suggest that market-oriented yam production could be a very profitable activity. The potential monetary revenue of yams per
hectare can be estimated at over 2 million cedis27, which is far above the yields of grain
22 No yield estimates were recorded for dry season gardens. 23 That is, every year in the wet season; not year-round.
24 The sample size was 60, but 15 respondents indicated that they didn’t know how much they harvested from
their compound farms and that they found it very difficult to estimate. For the remaining forty-five house- holds, farm size and output (usually in bowls or bags) were estimated. I left the estimation of farm sizes to my interpreters/assistants (Festus Lankuu for Dagara households and Victor B. Damian for Sisala house- holds) who both had an agricultural extension background.
25 Humid conditions are favourable for rice production.
26 Fifteen farmers estimated their rice yields at between five hundred and one thousand kilogram and five at
lower than five hundred kilograms. For eight rice farmers no yield data were recorded. Rice yields were higher than the yields of all other grains and legumes. Note that rice yield data are relatively reliable as compared to millet and sorghum yields because, after harvesting, rice grains are stored in bags of 100 kilograms. The grains of millet and guinea corn are left on the heads and stored in the traditional granaries.
27 One forth of an acre is roughly one tenth of a hectare. The yield per hectare would thus be over 2,000 tubers
crops.28 Only a few farmers reported yam sales, however. Presently, yams are cultivated on less than two percent of the total cropped area. Yams are clearly a ‘niche’ crop, and the soil characteristics of the research area will probably not allow a great expansion of yam produc- tion in the future. An exception may be the largely uncultivated stretch of lowlands along the Black Volta River.
Good maize and guinea corn yields can be achieved in the lowlands, too. Heavy rainfall and floods, however, can easily cause total failures of these crops, as was the case in many lowland farms in 1999. This makes lowland maize and guinea corn yields more variable. Moreover, the required labour input is higher than on upland farms because lowland maize and guinea corn are sown on mounds, whereas on upland farms, these crops are sown ‘on the flat’. In drought years, crop yields are usually higher in the lowlands than in the uplands. In that sense, lowland cultivation is a good on-farm insurance against drought.
In the upland ‘bush farms’, yields are relatively low. Cultivation on these farms, especially on the ‘wie’ (farm beyond the settlement), is supposed to shift, but fallows have shortened or have been abandoned due to land shortage, especially in the vicinity of Nandom Town. The loss in soil fertility is not sufficiently compensated for by increased input of nutrients. Chemi- cal fertilisers have never really gained much importance in the local farming system. In recent