Nutrient balances at farm level in Machakos (Kenya), using a participatory nutrient monitoring (NUTMON) approach

ARTICLE IN PRESS

Land Use Policy 22 (2005) 13–22

Nutrient balances at farm level in Machakos (Kenya), using a participatory nutrient monitoring (NUTMON) approach L.N. Gachimbia,*, H. van Keulenb, E.G. Thuraniraa, A.M. Karukuc, A. de Jagerd, S. Nguluuc, B.M. Ikomboc, J.M. Kinamac, J.K. Itabaric, S.M. Nandwaa a National Agricultural Research Laboratories, Kenya Agricultural Research Institute, P.O. Box 14733, Nairobi, Kenya Plant Research International, Wageningen University and Research Centre, P.O. Box 16, Wageningen 6700 AA, The Netherlands c National Dryland Farming Research Centre, Kenya Agricultural Research Institute, P.O. Box 340, Machakos, Kenya d Agricultural Economics Research Institute, Wageningen University and Research Centre, P.O. Box 29703, Den Haag 2502 LS, The Netherlands b

Received 16 September 2002; received in revised form 12 May 2003; accepted 10 July 2003

Abstract A total of 74 farms were selected from Machakos, Mwingi and Makueni districts in Kenya, using participatory techniques and classified in three categories on the basis of soil fertility management (low level, medium and high level). Soil fertility management was monitored, using the NUTrient MONitoring methodology, which appears a suitable and appropriate tool for the diagnostic phase of Farming System Analysis and Design in Arid and Semi-Arid Lands of Kenya. The participatory inventory and monitoring procedures applied, involving the farmers in the analysis of their own situation, forced the farmers to think about the processes and flows associated with the nutrient balances and the associated consequences for the quality of their soil resource base. Farmers’ performance during the feedback, through farmers’ workshops, of the results of soil analyses and the discussion on possible solutions for the identified problems, creates confidence that in the participatory learning and action phase, farmers will be equally involved and willing to adopt suggested adaptations. The first results of the quantitative analyses of nutrient balances at farm level show that farm balances for NPK are negative. This is in agreement with earlier work in the high-potential areas of Kenya and elsewhere in East Africa, as well as those from semi-arid regions in West Africa. It has been shown that client characterisation, as an emerging component of a research approach towards design of sustainable agricultural production systems, helps in identifying major potential recommendation domains and related research and development problems and opportunities, and potential interventions on which to focus the research agenda. r 2004 Elsevier Ltd. All rights reserved.

Introduction The rapid increase in Kenya’s population has three major effects. Firstly, migration from rural to urban areas, because of job opportunities, leaving farming largely to the women and children. Secondly, outmigration from the high-potential areas to arid and medium rainfall areas (arid and semi-arid lands, ASAL), in search of new farmland. The introduction in the ASALs of crop production technologies, developed and adapted in the high-potential areas, appeared risky and *Corresponding author. Tel.: +254-2-4444256. E-mail addresses: [email protected] (L.N. Gachimbi), [email protected] (H.v. Keulen). 0264-8377/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.landusepol.2003.07.002

sometimes disastrous, because of their unfavourable agro-ecological conditions (FURP, 1988). Rainfall is low, unreliable and erratic, soils are fragile, low in fertility and because of their sandy texture, susceptible to erosion and leaching (Mutiso, 1991). Thirdly, the increased pressure on the land necessitates intensification of land use, often without the necessary external inputs, such as fertilisers, to sustain its productivity. Consequently, soil nutrient levels have declined, leading to reduced soil fertility and environmental degradation (Nandwa et al., 2000; Smaling, 1993; Ikombo, 1984). The resulting reductions in crop production and soil cover have aggravated the vicious cycle through further nutrient losses from erosion and leaching and further degradation of the resource base. Currently, ASALs are

ARTICLE IN PRESS 14

L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

characterised by low-productive and non-sustainable agro-ecosystems that seriously threaten the livelihoods of the farming population. The high population density has also been claimed to have contributed to an increase in soil and water conservation investments, use of farmyard manure as fertiliser, and crop diversification to reduce risks of complete failure (Tiffen et al., 1994). Livestock production, mainly dairying, is based on extensive and semiintensive grazing management systems, based on indigenous cattle that are resistant to local diseases and perform relatively well on the local low-quality feeds, while to a limited extent improved crossbred animals are kept in zero-grazing units. In Kajiado, livestock production is by far the most important activity, while arable farming is found in small pockets. The National Soil Fertility and Plant Nutrition (NSFPN) Research Programme aims at the development of improved sustainable agricultural production systems in the medium and low rainfall areas. Because of the crucial role of crop nutrients in sustainable agricultural development in the ASALs, the research programme focuses on the development of integrated soil and crop nutrient management techniques. Although land and labour productivity have increased significantly as a result of introduction of modern technologies in agriculture, they often have had only a marginal effect on practices of small farmers. Limited adoption of the technologies developed by research institutes by farmers, points to a strong and urgent need for more end-user driven research and development programmes. In response to this need, a multitude of new participatory methods and approaches have been developed in recent years (Uphoff, 2002). These have drawn on many well-established local traditions, and have put participation, action research and adult education at the forefront. Within the tradition of development programmes, projects and initiatives, these approaches represent a departure from standard practice. Extractive research (i.e. research in which farmers are asked to provide information (surveys, etc.) without ever receiving any feedback) is being complemented, and to a large extent replaced, by investigation and analysis by local people themselves. Methods are being used not only to improve communication between the local population and outsiders, but also to stimulate active involvement of the population in the analysis of its own conditions. The objectives of the project ‘Nutrient stocks and nutrient flows in Kenya’s arid and semi-arid lands (NUTSAL)’ were: (i) to identify, in close co-operation with the farmers, the major constraints faced by small scale farm households in the ASALs of Kenya, with special emphasis on nutrient balances; (ii) to identify, through participatory design and testing, alternative

production techniques that might alleviate these constraints, and contribute to adoption of more sustainable agricultural production systems. In this paper, the participatory nutrient monitoring component of the project is described, and some preliminary results are presented.

Materials and methods Methodology Introduction In the study, the participatory NUTMON-methodology (Van den Bosch et al., 1998a, 2001; de Jager et al., 1998a, b) was applied. In the methodology, two major phases can be distinguished: (i) diagnosis and analysis of existing farm and nutrient management systems, (ii) participatory technology development. In the participatory diagnostic phase the current state of natural resources and farm management are assessed with respect to their influence on resource flows and financial performance, in the context of their socio-economic environment (de Jager et al., 1998a, b). In this phase, a variety of tools is used, ranging from generally applied tools such as natural resource flow maps, transect walks and farmers’ soil maps, to a specially developed quantitative monitoring and analysis tool, the NUTMON toolbox (Van den Bosch et al., 2001), designed to assess nutrient flows and financial performance indicators. Monitoring takes place at plot and farm household level, where most of the decisions on nutrient management are taken. Influences of processes at lower spatial scales (for instance leaching and denitrification) are incorporated in the farm level approach through pedotransfer functions (Tietje and Tapkenhinrichs, 1993). At least two monitoring sessions are held each season, i.e. shortly after planting and at or shortly after harvest. Through collaboration among scientists, extensionists and farmers, as part of a research and development agenda, the diagnostic phase also serves to create farm household awareness of nutrient management aspects. The results of the diagnostic phase form the basis for the participatory technology development phase. In this phase, a range of technologies, based on integrated nutrient management, is tested as part of on-farm trials. Possible options for experimentation are identified and ranked with participating and other interested farmers. The performance of these technologies is assessed, as are their consequences for the resource flow and financial performance indicators. On the basis of these indicators, the most appropriate technological innovations can be selected. Farmers, extensionists and researchers are fully involved in this planning and implementation process together, and therefore, both existing local indigenous knowledge and scientific knowledge (or combinations of

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

the two) are incorporated in the process of technology development and testing. Participatory farm selection and inventory Farm selection. Farms are selected through participatory procedures, in which the local population, extensionists, scientists and other stakeholders, such as NGOs (de Jager et al., 2001), are involved. The actual selection procedure and criteria depend on the site characteristics and purpose of the study. Inventory. The NUTMON-methodology is based on a systematic collection of information from the farm household on farm characteristics, farm practices and farm management. This information, both quantitative and qualitative, is used to quantify flows of material (with emphasis on nutrients) and cash through the farm system. Information collection starts with the farm inventory, that is, in principle, repeated before each crop cycle. In the farm inventory, information is collected on the farm household, its composition, and its assets (land, capital goods, i.e. machinery and animals). The farm household is characterised in terms of available labour and consumer units, while the education level of the head of the household is also recorded. The available land resources are specified in terms of both Farm Section Units (FSU, land units with more or less ‘stable’ soil characteristics) and Primary Production Units (PPU, land units dedicated to production of a certain commodity in a given season). Capital goods are specified, such as hoe(s), plough(s), wheelbarrow(s), etc. Animals present at the farm (Secondary Production Units, SPU) are defined in terms of Animal Management Groups, i.e. groups of animals (generally) of the same species that are managed by the farm household as one unit in terms of feeding, confinement, grazing, etc. If relevant, within a Management Group, allowance can be made for different age groups, such as milking cows, heifers and calves. The presence of redistribution units, such as bomas (night corrals), manure heaps, and compost pits is recorded. In addition, the Household (HH) is defined, as the labour supply and consumption unit, Stock, representing a temporary store for staple crops (cereals and pulses), crop residues (for cattle feeding) and/or chemical fertiliser for future use and finally the ‘external world’ (EXT), consisting of markets, neighbours and/or other families, serving as a source of and/or destination for flows, that as externalities (not on-farm) are not monitored. During the inventory, a ‘farm sketch’ is produced that visualises the spatial distribution of the various units distinguished. Methods to qualitatively assess natural resource qualities and natural resource flows are an integral part of the farm inventory, implemented in close collaboration with farm household members. These methods include drawing farm soil maps, including local names and character-

15

istics of the soils and nutrient flow maps, visualising flows of nutrients within the farm and over the farm boundaries. Quantification of nutrient stocks, flows and balances Stocks. The soil nutrient stock of a farm is defined as the total quantity of a nutrient present in the top 30 cm of the soil profile, from where crops usually take up the major part of the nutrients (Van den Bosch et al., 1998a). These stocks include dissolved ions, nutrients in organic matter, adsorbed to the solid phase or in stable inorganic components. To assess the size of these stocks per FSU for the macro-nutrients NPK and carbon, between 10 and 25 samples per farm are analysed, depending on farm size and heterogeneity of its soils. Flows. During monitoring, material flows among farm units and over the farm boundaries are quantified through interviews with farm household members. This refers to purchase of chemical and/or organic fertilisers, sales of crop and animal products, and ‘redistribution’ flows through grazing, filling of compost pits and/or application of farmyard manure. Quantitative information on part of these flows (purchase of fertiliser, sale of maize grain) is supplied directly by the farm household members; other flows that can not easily be quantified such as intake of animals during grazing and manure excretion of different types of animals, are estimated by means of a simple sub-model, that calculates intake and excretion on the basis of animal type, grazing time and forage availability (Van den Bosch et al., 2001). The material flows are converted into nutrient flows through assigning specific nutrient contents to each of the components. N, P and K contents of these ‘nutrient carriers’ are determined in the laboratory. A final group of flows (atmospheric deposition, gaseous losses, leaching and erosion) is quantified on the basis of measurable site characteristics (e.g. rainfall, soil texture), using pedo-transfer functions (Van den Bosch et al., 1998a). Balances. On the basis of the quantified flows, the NUTMON toolbox calculates balances at farm and activity level for specified time periods, i.e. growing season, calendar year. The toolbox includes a database, a static calculation model and a user interface that facilitates data entry, data manipulation and presentation of the results.

NUTSAL study area Biophysical The study area comprises parts of Machakos, Mwingi, Makueni and Kajiado districts, particularly the regions classified as Agro-Ecological Zones (AEZ) 4

ARTICLE IN PRESS 16

L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

and 5 (Jaetzold and Schmidt, 1983). These regions are characterised by low, unreliable, highly variable rainfall, bi-modally distributed, resulting in two growing seasons. Average annual rainfall varies between 500 and 700 mm in AEZ5, and between 600 and 800 mm in AEZ4, mainly depending on altitude. Soil depth depends on parent material and slope; soils are generally low in organic matter and deficient in N and P. Available K is generally adequate (FURP, 1988; Ikombo, 1984). Low infiltration rates and susceptibility to sealing make the soils vulnerable to erosion. Most of the rains occur in relatively heavy showers at the beginning of the growing season when land is bare, which aggravates the risk of soil erosion (Mutiso, 1991; Pagiola, 1996). Farming systems in the region are characterised by cultivation of maize, beans, cowpea, pigeon pea, cassava, cotton and sunflower and livestock keeping (cattle, sheep and goats). Land preparation, planting and cultivation is mainly done by oxen, while in the drier areas the use of hand hoes and digging sticks prevails.

Table 1 Characteristics of the six selected clusters Cluster (district)

Annual rainfall (mm)

Agroecological zone (AEZ)a

Farming system

Kionyweni (Machakos)

500

LM4

Matuu (Machakos)

600

LM5

Kasikeu (Makueni) Kibwezi (Makueni)

700

UM4

562

LM5

Kiomo (Mwingi)

600

LM5

Enkorika (Kajiado)

500

LM5

Cross-bred cattle, maize, beans and fruit trees Local cattle, irrigated agricultural (okra, tomatoes, cowpeas, pigeon peas) maize, beans and sorghum Maize, pigeon peas, beans and cowpeas Irrigated agricultural (okra, pigeonpeas, cowpeas, sorghum Maize, beans, sorghum, millet, pigeon peas Maize, beans, free range pastrolism

a

Agro-ecological zonation by Jaetzold and Schmidt (1983).

Results Farmer management groups and farm selection Clusters in ASAL were selected on the basis of agroecological zone, farming systems, and population density, as soil fertility maintenance problems are more pronounced in densely populated areas. In the six selected clusters, Kionyweni and Matuu (Machakos), Kasikeu and Kibwezi (Makueni), Kiomo (Mwingi) and Enkorika (Kajiado) both rainfed conditions, representative for the majority of the farming systems in these districts, and irrigated agriculture, which is practised on a limited scale (Matuu and Kibwezi), are represented (Table 1). A 1-day village meeting (baraza) was held in each of the clusters, attended by farmers, assistant-chief and village elders, researchers and extension agents, in which the objectives of the project, its activities and expected outputs were presented. This was followed by farm selection using the ‘participatory learning and action research’ (PLAR) approach (Defoer, 2000). Farmers provided criteria, used to characterise appropriate soil fertility management practices for sustained crop production. The criteria identified most frequently (Table 2), indicated that the farmers perceived good soil fertility management in terms of practices related to crop, soil and water management. In dryland farming, inorganic fertiliser application was not perceived as important, as was observed elsewhere (Freeman and Omiti, 2003; ICRA/KARI, 1996) due to low returns and the high risk of crop failure (due to soil

Table 2 Practices identified by farmers as criteria for appropriate soil fertility management Application of animal manure Deep and appropriate tillage (kukilia muunda) Optimum crop spacing Construction of bench terraces for soil and water conservation Planting in pure stands to avoid competition for water Timely and early planting Early weeding Ridging, especially when weeding with oxen (kumbiia muthanga) Crop rotation Keeping ‘good’ selected seeds (those that use their own or local seeds) or use of appropriate varieties (those that use seeds from outside) Knowledge and proper use of fertiliser Proper planning of farm activities Effective crop protection (chemical) Keeping farm records

moisture deficiency). For example, in Kasikeu, inorganic fertiliser was listed last, because of unfavourable experiences with fertiliser supplied as part of a famine relief package. Subsequently, the meeting classified all farmers in three categories on the basis of the criteria identified. The high level soil fertility management group consisted of those using at least three-quarters of the practices, the medium level group about half, the remainder being classified in the low level management group. Finally, the meeting selected ten farmers from each of the three categories considered capable and willing to participate

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

in the NUTSAL study, some of which were not present at the meeting. Farm inventory Results are presented for the clusters Kasikeu, Kibwezi and Kionyweni for the long rains 1999 as an example of the type of results generated with the NUTMON-model (Van den Bosch et al., 2001; de Jager et al., 1998a, b; Van den Bosch et al., 1998a, b). Farm characteristics Average farm size for the Kasikeu, Kibwezi and Kionyweni clusters is 3.1, 3.3 and 2.2 ha, respectively, of which 1.9, 1.8 and 1.7 ha were utilised for crop production. Crops grown include maize, bean, pigeon pea, cowpea, and fruit trees in Kasikeu and Kionyweni, and mainly vegetables in Kibwezi. The non-cropped area includes pasture. In all three clusters, dairy animals

17

were kept, with on average 5.2 and 4.5 Tropical Livestock Units per farm household in Kasikeu and Kionyweni, respectively, and 0.5 in the irrigated cluster Kibwezi, where only some families keep livestock. In all three clusters most farmers keep poultry, sheep, goats and donkeys. Nutrient flow map During the farm inventory, the concept of nutrient flows through the farm system is extensively discussed with farm household members, after which jointly these possible flows are visualised in a farm nutrient flow map (Fig. 1). The various farm components, i.e. RUs, PPUs, SPUs, HH and Stock, as well as EXT are identified, and the flows of nutrients among these components are indicated by arrows. This nutrient flow map is used during the monitoring sessions to identify flows that have taken place during the monitoring period.

Fig. 1. Example of a schematised farm nutrient flow map.

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

18

Soil characteristics From the farm section units identified by each farmer by its local name, soil samples were analysed for soil chemical and physical properties in the laboratory. In Table 3, an example is given of a complete set of soil characteristics.

a major crop nutrient source: N is highly mobile in the soil–plant system and thus susceptible to losses in the form of leaching and volatilisation. Hence, a large proportion of the N in animal manure is lost before and during application, and therefore hardly contributes to the soil nitrogen store. P on the other hand, is far less susceptible to losses, and a substantial proportion of that element from animal manure therefore contributes to the soil store. Potassium is abundantly available in all soils (Fig. 2c), with values varying between 2 and 4 times the ‘adequate’ levels. Organic carbon levels in the soil (Fig. 2d) are again variable, with a factor 2.5 between the lowest and the highest value, but all well below the level considered ‘adequate’ from an agronomic point of view (Mehlich et al., 1964). If accumulated P indeed originates from animal manure, substantial amounts of animal manure must have been applied, which, however, does not show up in the C-content. This probably indicates that the organic components in the animal manure have been largely decomposed, under the relatively favourable conditions for microbial action, i.e. relatively high temperatures and sufficient moisture in the topsoil (De Ridder and Van Keulen, 1990). Following presentation of these results to the farmers, a number of sub-groups comprising 10 farmers each were formed to discuss possibilities for improving soil fertility management, an essential first step in PLAR. Concurrently, each farmer was also presented with the detailed data for each of his/her FSUs (Fig. 3), allowing comparison and discussion during the sub-group meetings. Farmers’ sub-group reports mentioned addition of farm yard manure or chemical fertilisers, terracing and/ or spreading nutrient-rich termite mound or ant hill soil to their land, or crop rotation, as possible practices to improve soil fertility. Innovative techniques were not suggested, but the discussions provided a solid basis for further development of appropriate technologies.

Soil nutrient status and feedback to farmers To provide a basis for the design of alternative production techniques, farmers’ conceptual awareness was required on nutrient deficiency and its consequences for crop growth and production. Awareness creation was carried out jointly by farmers, extensionists and researchers in meetings in each of the clusters. First, the concept of nutrient deficiency in plants was illustrated, using leaves with clear deficiency symptoms. Farmers could easily recognise a range of symptoms associated with ‘poor soil fertility’, such as reddish purplish leaves, low-quality maize cobs, stunted crops, low soil cover, hardening of the soil, yellowish green leaves and weak roots. However, the relations between individual nutrient element deficiencies and visible deficiency symptoms could not be identified by the farmers. Subsequently, the results of the soil analyses for N, P, K and C were discussed using graphs, showing average values for each farm and for each soil fertility management group. To put the measured values in perspective, they were expressed as a percentage of the value identified by the extension service as ‘agronomically adequate’ (Mehlich et al., 1964). As an example, the data for the ‘high level management’ group in Kionyweni, representative for the majority of the results obtained, are discussed here. Fig. 2a shows that, even in this high level soil fertility management group, soil-N values are all well below 50% of the ‘adequate’ level, with little variation among the farm households. The variability in soil P-levels (Fig. 2b) is much larger (a factor 3.5 between the lowest and the highest value) and, on average, the values are above 50% of the ‘adequate’ level, with more than half the values exceeding the 75% point. Such a combination of N- and P-values is typical for mixed farming systems, where soil fertility management to a large extent revolves around animal manure as

Monitoring results In Kasikeu, rainfall during this period was 169 mm, compared to a long-term average of 742, whilst in Kionyweni it was only 29 mm, compared to the long-

Table 3 Farmers’ soil classification and chemical and physical characteristics Farmer name: Location:

Ndambuki Kyalo Kionyweni

Soil type (farmers’ classification)

Texture

pH

OM (%)

P-Mehlich (mg/kg)

N (%)

K (cmolc/kg)

Sand (%)

Silt (%)

Clay (%)

Nthangathi Nthangathi na Kitune

Sandy clay loam Sandy loam

6.0 6.6

0.62 0.27

22 5

0.06 0.05

0.14 0.26

64 78

10 12

26 10

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

19

Fig. 2. (a) Kionyweni cluster nitrogen report, (b) Kionyweni cluster phosphorus report, (c) Kionyweni cluster potassium report, (d) Kionyweni cluster carbon report.

Fig. 3. Example of an individual farm soil sample report.

term average of 350. In Kibwezi, although rainfall was highly deficient (0 mm versus 562 long-term average), the monitored PPUs were all irrigated, so that the drought did not affect crop performance. Yields and nutrient balances Total stocks of N, P and K in the (upper 30 cm of the) soil, calculated from the chemical analyses of the soil samples, were between 2000 and 7000 kg/ha for N,

1000–12000 for P (available P only) and 6000–18000 for K. This illustrates one of the problems in soil fertility research: even though the soils are deficient in N and P, large quantities may still be present. Averaged for all farms, negative balances were found for N, P and K in all three clusters (Table 4). Average partial balances for N, P and K, consisting of farm gate nutrient flows in farm inputs and products (Van den Bosch et al., 2001) are positive for Kasikeu (except for N) and Kionyweni, but remain negative for Kibwezi. This difference implies that the negative overall balances are mainly due to losses through so-called ‘hard-toquantify’ flows (leaching, gaseous losses and erosion, Van den Bosch et al., 2001), whose rates are difficult to control by the farmer. The unfavourable weather conditions, and the associated poor crop performance (or complete failure) are illustrated by the very low quantities of nutrients extracted in economic crop products and crop residues (including grazing in the field); for Kasikeu and Kionyweni (rainfed environments) 4.0 and 1.1 for N, 0.7 and 0.1 for P and 4.7 and 1.0 for K, all in kg/ha, respectively. For Kibwezi, where mainly irrigated vegetables were grown in this period, the values are 5.2, 0.9 and 5.1 kg/ha, respectively.

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

20

Table 4 Farm N, P, K-flows (kg/ha) for Kasikeu, Kibwezi and Kionyweni, long rains 1999 Kasikeu N

Kibwezi P

K

N

Kionyweni P

K

N

P

K

Mineral fertiliser Organic fertiliser (incl. grazing) Atmospheric deposition Biological nitrogen fixation Total in Farmgate in

1.8 4.7 1.7 1.8 10.0 6.5

0.9 0.8 0.3 0 2.0 1.7

0 5.7 1.1 0 6.8 5.7

3.4 2.5 0 0.2 6.1 5.9

0.6 0.5 0 0 1.1 1.1

0 2.7 0 0 2.2 2.7

1.2 4.9 0.8 0.6 7.5 6.1

0.8 0.8 0.1 0 1.7 1.6

0.2 5.6 0.5 0 6.3 5.8

Crop products Crop residues (incl. manure export) Leaching Gaseous losses Erosion Human excreta Total out Farmgate out

1.7 2.3 9.1 1.6 1.4 3.3 19.4 4.0

0.4 0.3 0.1 0.1 0 0.8 1.7 0.7

1.0 3.7 0.8 0 1.1 0.9 7.5 4.7

4.6 0.6 3.7 0.3 0 1.8 11.0 5.2

0.8 0.1 0 0 0 0.5 1.4 0.9

4.4 0.7 0.2 0 0 0.3 5.6 5.1

0.4 0.7 4.4 2.0 0.4 4.6 12.5 1.1

0.1 0 0 0 0 1.2 1.3 0.1

0.2 0.8 0.9 0 2.8 1.0 5.7 1.0

9.4 2.5

0.3 1.0

0.7 1.0

4.9 0.7

0.3 0.2

3.4 2.4

5.0 5.0

0.4 1.5

0.6 4.8

Total balance Farmgate balance

Table 5 Yields and (macro-)nutrient balances for all maize/bean plots (1) and maize/bean plots harvested (2) in Kionyweni and Kasikeu, long rains 1999 (numbers in brackets indicate number of plots in the analysis) PRIVATE

Kionyweni1 (96)

Kionyweni2 (15)

Kasikeu1 (39)

Kasikeu2 (24)

Crop yield of first crop (in kg/ha) Crop yield of second crop (in kg/ha) Gross margin (in Ksh per ha) N balance (in kg/ha) P balance (in kg/ha) K balance (in kg/ha)

114 23 2730 2 2 1

241 147 12 262 13 2 28

386 219 12 655 19 2 16

568 331 18 504 21 3 19

o!– –oH2>N-flowso/H2>– –>

At plot level, the moisture deficient conditions during the growing season are evident (Table 5). In Kionyweni, only 15 out of a total of 96 monitored plots sown to maize/bean mixtures were harvested, in Kasikeu, 24 out of 39. For the harvested plots yields were extremely low, with values of 241 and 568 kg/ha for maize in Kionyweni and Kasikeu, respectively, and 147 and 331 kg/ha for the bean component. The low yields/crop failures are also reflected in the very low values of the nutrient flows. For the 15 plots harvested in Kionyweni, the average nitrogen balance equals 13 kg/ha. The spatial variability, associated with differences in soil properties and weather conditions, and in crop management, was substantial. The standard deviation of the nitrogen balance was 13 kg/ha, with minimum and maximum values of 40 and +3 kg/ha, respectively. On average, 2 kg/ha of N as inorganic fertiliser was applied to the plots, 4 kg/ha N in the form of animal manure, 8 kg in animal faeces and urine during grazing, while it was estimated that biological fixation contributed 3 and

deposition 1 kg/ha (Table 5). In addition to the manure, 9 and 3 kg/ha were exported from the field in the form of economic crop products and crop residues, respectively, while the estimated loss through soil erosion was 3 kg/ ha. The phosphorus and potassium balances (detailed data not shown) showed essentially the same picture (Table 6).

Discussion and conclusions The NUTMON methodology appears a suitable and appropriate tool for the diagnostic phase of Farming System Analysis and Design in Arid and Semi-Arid Lands (ASAL) of Kenya. The participatory inventory and monitoring procedures applied in that phase, involving the farmers in the analysis of their own situation, forced farmers to think about the processes and flows associated with their nutrient balances and the consequent deterioration of their soil resource base

ARTICLE IN PRESS L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22 Table 6 Yields and (macro-)nutrient balances for all maize/bean plots (1) and maize/bean plots harvested (2) in Kionyweni and Kasikeu, long rains 1999 (numbers in brackets indicate number of plots in the analysis) PRIVATE Mineral fertilizer Organic fertilizer Grazing Atmospheric deposition Biological fixation Crop products Crop residues Manure Leaching Gaseous loss Erosion

Kionyweni1 Kionyweni2 Kasikeu1 Kasikeu2 (96) (15) (39) (24) 1 8 6 0 0 3 2 6 2 0 0

2 4 8 1 3 9 3 13 3 0 1

1 6 3 2 4 13 6 3 7 0 3

1 8 4 2 5 19 8 5 7 0 2

(Martin and Sherington, 1997; Haverkort et al., 1991). It appeared that farmers in these dryland areas are well aware of the precarious conditions of their soil resources. They recognise that in the current farming systems insufficient replenishment of soil nutrients takes place, so that soil quality gradually declines. Conducting soil sampling and nutrient monitoring activities with farmers considerably increased the insight in the cause of the soil nutrient depletion with farmers as well as with extensionists and researchers. The NUTMON method therefore provides a forum for integration of local traditional knowledge and more mainstream knowledge, and adoption of more appropriate technologies. Farmers’ involvement during the feedback, in farmers’ workshops, of the results of soil analyses, especially comparing different soil fertility management techniques applied by farmers and the different results in yield, nutrient balance and financial returns, proved an extremely appropriate concept to identify promising technical changes. The ensuing discussion, on possible solutions for the identified problems, creates confidence that in the participatory learning and action phase, farmers will be equally involved and willing to adopt suggested adaptations. The results of the quantitative analyses of nutrient balances at farm level show that farm balances for NPK are negative. Due to the ecological circumstances and the high subsistence level of the farms, the negative values are relatively low. However, the yield levels are in general low, with regular complete crop failures. An increase in soil fertility will definitely contribute to an increased income from farming and food security levels. The negative values confirm the observations during earlier work in the high-potential areas of Kenya and elsewhere in East Africa (de Jager et al., 2001; Van den Bosch et al., 1998b; Braun et al., 1997), as well as those from drier regions in West Africa (Mokwunye et al., 1996). These results, although they should be judged

21

with caution, as rainfall was below the long-term average, indicate that it is absolutely necessary to carefully examine the farming systems before drawing conclusions on their performance too easily (Zaal and Oostendorp, 2002). Moreover, the negative balances appear to be mainly the result of losses that cannot be easily controlled by the farmer, i.e. erosion, denitrification and leaching. The results of this diagnostic phase have found their way in a comprehensive Participatory Learning and Action Research (PLAR) programme in which appropriate technical innovations are tested with the same group of farm households. A similar participatory monitoring process is applied in evaluating and where necessary adapting and changing the innovations tested. Given the harsh ecological and economic conditions no spectacular results can be expected in the short-term, instead a slow but steady process of increasing sustainable output, maintaining the soil fertility and developing knowledge of farmers and researchers has taken off.

Acknowledgements The authors gratefully acknowledge European Union/Kenya Agricultural Research Institute, Agricultural Livestock Research Support Programme II Project number 6 ACP KE 061 (KE 6003/001) for the funding, Director KARI for permission to publish this paper and finally the Extension officers in the Ministry of Agriculture and Rural Development, Machakos District.

References Braun, A.R., Smaling, E.M.A., Muchugu, E.I., Shepherd, K.D., Corbett, J.D., 1997. Maintenance and improvement of soil productivity in the highlands of Ethiopia, Kenya, Madagascar and Uganda: an inventory of spatial and non-spatial survey and research data on natural resources and land productivity. African Highland Initiative Techn. Rep. Series No. 6, ICRAF, Nairobi, Kenya. Defoer, T., 2000. Moving methodologies. Learning about integrated soil fertility management in sub-Saharan Africa. Ph.D. Thesis, Wageningen University, Wageningen, The Netherlands, 189pp. de Jager, A., Nandwa, S.M., Okoth, P.F., 1998a. Monitoring nutrient flows and economic performance in African farming systems (NUTMON). I. Concepts and methodologies. Agriculture, Ecosystems and Environment 71, 37–48. de Jager, A., Kariuki, F.M., Matiri, M., Odendo, M., Wanyama, J.M., 1998b. Monitoring nutrient flows and economic performance in African farming systems (NUTMON). IV. Monitoring of farm economic performance in three districts in Kenya. Agriculture, Ecosystems and Environment 71, 81–92. de Jager, A., Onduru, D., Van Wijk, M.S., Vlaming, J., Gachini, G.N., 2001. Assessing sustainability of low-external-input farm management systems with the nutrient monitoring approach: a case study in Kenya. Agricultural Systems 69, 99–118.

ARTICLE IN PRESS 22

L.N. Gachimbi et al. / Land Use Policy 22 (2005) 13–22

De Ridder, N., Van Keulen, H., 1990. Some aspects of the role of organic matter in sustainable intensified arable farming systems in the West-African semi-arid-tropics (SAT). Fertilizer Research 26, 299–310. Freeman, H.A., Omiti, J.M., 2003. Fertilizer use in semi-arid areas of Kenya: analysis of smallholder farmers’ adoption behavior under liberalized markets. Nutrient Cycling in Agroecosystems 66, 23–31. FURP, 1988. Fertilizer Use Recommendation Project (Phase I). Main report, Methodology and Inventory of Existing Information, Ministry of Agriculture, National Agricultural Laboratories, Nairobi, Kenya, 115pp.+Ann. Haverkort, B., van der Kamp, B., Waters-Bayer, A., 1991. Joining Farmers’ Experiments: Experiences in Participatory Technology Development. ITP, London, UK. ICRA/KARI, 1996. A participatory study of farmers’ constraints, opportunities and research needs in the Hilly Masses of Eastern Kenya. Working Document Series 54, International Centre of Development Oriented Research in Agriculture (ICRA)/Kenya Agricultural Research Institute (KARI), National Dryland Farming Research Centre, Katumani. Ikombo, B.M., 1984. Effects of farm yard manure and fertilizers in the semi-arid areas of eastern Kenya. East African Agricultural and Forestry Journal 44, 266–274. Jaetzold, R., Schmidt, H., 1983. Farm Management Handbook of Kenya. Vol. II/C, East Kenya. Natural Conditions and Farm Management Information, Kenya Ministry of Agriculture and German Agricultural Team (GTZ), Nairobi, Kenya. Martin, A., Sherington, J., 1997. Participatory research methods— implementation, effectiviness and institutional context. Agricultural Systems 55, 195–216. Mehlich, A., Bellis, E., Gitau, J.K., 1964. Fertilizing and liming in relation to soil chemical properties. Scott Laboratories, Department of Agriculture, Nairobi, Kenya. Mokwunye, A.U., de Jager, A., Smaling, E.M.A., 1996. Restoring and maintaining the productivity of West African Soils: Key to sustainable development. IFDC Misc. Studies No. 9, Lom!e.

Mutiso, S.K., 1991. Rainfall and environmental profile. In: Mortimore, M. (Ed.), Environmental Change and Dryland Management in Machakos District (Kenya 1930–1990). Overseas Development Institute (ODI) Working Paper 53, London, UK, pp. 71–83. Nandwa, S.M., Onduru, D.D., Gachimbi, L.N., 2000. Soil Fertility Regeneration in Kenya. In: Hilhorst, T., Muchena, F.M. (Eds.), Nutrients on the Move. Soil Fertility Dynamics in African Farming Systems. IIED, London, pp. 119–132. Pagiola, S., 1996. Price policy and returns to soil conservation in semiarid Kenya. Environmental and Resource Economics 8, 255–271. Smaling, E.A.M., 1993. An agro-ecological framework for integrated nutrient management with special reference to Kenya. Ph.D. Thesis, Wageningen Agricultural University, Wageningen, 250pp. Tietje, O., Tapkenhinrichs, M., 1993. Evaluation of pedo-transfer functions. Soil Science Society of America Journal 57, 1088–1095. Tiffen, M., Mortimore, M., Gichuki, F., 1994. More People, Less Erosion: Environmental Recovery in Kenya. Wiley, Chichester. Uphoff, N. (Ed.), 2002. Agroecological Innovations. Increasing Food Production with Participatory Development. Earthscan Public. Ltd., London, 306pp. Van den Bosch, H., de Jager, A., Vlaming, J., 1998a. Monitoring nutrient flows and economic performance in African farming systems (NUTMON). II. Tool development. Agriculture Ecosystems and Environment 71, 49–62. Van den Bosch, H., Gitari, J.N., Ogaro, V.N., Maobe, S., Vlaming, J., 1998b. Monitoring nutrient flows and economic performance in African farming systems (NUTMON). III. Monitoring nutrient flows and balances in three districts in Kenya. Agriculture Ecosystems and Environment 71, 63–80. Van den Bosch, H., Vlaming, J., Van Wijk, M.S., de Jager, A., Bannink, A., Van Keulen, H., 2001. Manual to the NUTMON Methodology. Alterra/LEI, Wageningen University and Research Centre, Wageningen, The Netherlands. Zaal, F., Oostendorp, F.H., 2002. Explaining a miracle: intensification and the transition towards sustainable small-scale agriculture in dryland Machakos and Kutui districts, Kenya. World Development 30, 1271–1287.