Evaluation of biomass enhancing

practices in the Yatenga region of

Burkina Faso
MSc Thesis – Plant Production Systems

Jori Bremer

October 2016
 Evaluation of biomass enhancing practices in the
Yatenga region of Burkina Faso
MSc Thesis Report

Name student: Jori Bremer

Registration number: 931025122040
Study: MSc Plant Sciences
Specialization: Natural Resource Management
Chair group: Plant Production Systems (PPS)
Code: PPS-80436
Period: September 2015 – October 2016
Supervision: dr. ir. Katrien Descheemaeker
MSc Wim Paas
Examination: dr. ir. Katrien Descheemaeker
dr. ir. Maja Slingerland

ii
iii
Preface
Over the course of the last year I learnt a lot about my thesis subject but also gained many new
insights fuelling my personal development. The trip to Mali has been an eye opener, as I found
myself face to face with a different world and new ways of thinking. The field work in Mali
confronted me with ‘real farmers’, which changed their static role on paper to real and live people.
This fed my enthusiasm to do my best at this thesis. I am very grateful towards all the respondents
who had patience and time for answering my questions, gave me new information and raised
interesting questions related to my project. In many cases, however, I also got questions about
what I was going to do and how I was going to help them. Those exchanges made me question my
role as a plant sciences student and made me wonder about how I/we can contribute to the
improvements in ‘the quality of life’.

Hereby I would like to thank Issa and Bakary, who made my stay at the guest house much more
comfortable and pleasant and who introduced me to all the nice places in Koutiala. Also many
thanks for the introduction to Malian cooking and for taking the time to make and drink tea, which
lead to nice exchanges teaching me about our cultural differences. After being in Koutiala for a
month, I stayed in Bamako for a week, where I was warmly welcomed by Fily in her family home.
Being her guest gave me the opportunity to discover how it is to live in Mali for a Malian family and
allowed me to be part of their daily routines. Here again the cooking amazed me and being able to
cook a ‘European’ meal in a pot on coals was a great experience. Fily also showed me around in the
headquarters of the IER (Institut d’Economie Rurale) and ICRISAT, which I really appreciated as it
showed me how research is performed in practice, in a different location from the university.

In this way I would like to thank the ILRI team in Burkina Faso, especially Viviane Yameogo and
Augustine Ayantunde, for their input. Thank you for collecting and arranging data and providing
information, without which it would not have been possible to perform this thesis. I am grateful for
the support from the WLE Biomass project, enabling data collection. A big thanks also goes to
CCAFS for the provision of household data.

Of course I owe much to everyone at A.M.E.D.D., especially Bougouna, Ousmane and Arouna, who
spent time and energy helping me in setting up the fieldwork and getting around in Koutiala. Sory,
Michel and Bakary have been a great help, without them it would have been much more difficult to
perform the interviews. I especially appreciated the discussions about farming in Mali and in the
Netherlands, where I learned that what is normal for me is often different from what is normal in
Mali, and the other way around. I am also grateful for the time taken to make sure we were all on
one line before starting the interviews and organising the fieldwork, this gave me confidence in
conducting my first ever interviews.

I also owe much to Mink, who has been a great help and has provided answers to my unending list
of questions related to the use of FARMSIM. Finally, I would like to thank Katrien and Wim, who
supported me throughout the whole thesis process. Thank you for supporting me in my wish to go
abroad, even with the impediments and the impossibility to go to Burkina Faso. This lead to an
experience which I will not soon forget. I am also grateful for all the help you gave during the
thesis and the great feedback I have received. It has helped me in going further, digging deeper
and improving my work.

iv
v
Summary
In the Yatenga region of Burkina Faso agriculture is the main source of livelihoods. However,
farmers are poor and vulnerable due to erratic rainfall. This leads to the need to produce food while
thinking of the maintenance or improvement of soil quality, to ensure future food production. For
the project ‘Realising the full potential of mixed crop-livestock systems in rapidly changing Sahelian
agro-ecological landscapes’ (WLE Volta and Niger Focal Region), several suggested interventions
are evaluated based on their impact on food and feed production and on soil fertility. The
adoptability of these interventions for farmers is also assessed, as this influences the reach and
impact of the project. The aim of this thesis was to build a decision-tool to help (among others)
extension agents in choosing the right interventions to promote per farm. Six farm types are
distinguished based on their resource endowment, as resource endowment influences farming
constraints and goals and alters the impacts of interventions on the farming system.

For the impact evaluation of the interventions, the NUANCES-FARMSIM (Nutrient Use in Animal and
Cropping systems – Efficiencies and Scales, FARM SIMulator) model was used, an integrated crop
livestock model adapted to African smallholder farming systems. The model was adapted to the
local situation (e.g. rainfall, soil and household characteristics) and a baseline scenario was run for
12 years. The different interventions were then simulated and the model was run again. The
indicators taken into account were total farm calorie production (averaged over the years),
livestock productivity (milk production and cattle weight, averaged over the years) and farm
average soil organic carbon (SOC, for the last simulation year). The outcomes for the different
interventions were compared to the outcomes of the baseline scenario to monitor changes. For all
farm types the outcomes for the different interventions followed the same trends, although the
magnitude of the impacts varied. This depended mainly on the crops grown and on the area of
farm land available per cattle head.

Farmer constraints, goals and attitudes were assessed through conducting interviews in the villages
Ziga, Ninigui and Thiou in the Yatenga region. Per type the strength of the constraints limiting
agricultural production (cropping land, pasture area, livestock, capital, education and technology)
were measured and combined to expert opinions on the inputs of these elements needed for the
implementation of interventions. From there it was possible to retrieve which interventions had the
least constraints limiting adoption and which interventions had the strongest constraints limiting
adoption. A difference in strength of present farming constraints was discerned based on resource
endowment. Farmers were also asked about their farming goals, which also influence adoption
potential. Farmer goals turned out to be dependent on resource endowment, similarly to the
farming constraints. It was hypothesized that farmer attitude (positive or negative view on
farming) would also influence intervention take-up. In this thesis it was however impossible to
discern clear attitude groups.

From the information collected in this thesis a decision tool was made to help in the process of
choosing the best suited interventions for different farm types. The following steps should be taken
for finding those interventions. Households should be classified based on resource endowment.
Then the constraints present for the household should be noted. This determines which constraints
might limit intervention adoption or which constraints should be alleviated for the implementation
of the interventions. The impact of the interventions on the farming system should then be found
(impact on SOC, food production and livestock productivity). With the information about farming
constraints and intervention impact, the user of the decision tool should consider which
intervention to promote and implement. Once one or several interventions are chosen it should be
checked whether the interventions lead towards the goal(s) of the farmer, as when this is the case
the rate of adoption of the intervention will be increased.

vi
vii
Contents
Preface ................................................................................................................................ iv
Summary ............................................................................................................................. vi
1 Introduction ................................................................................................................... 2
1.1 General introduction ................................................................................................ 2
1.2 Present challenges .................................................................................................. 2
1.3 Future challenges .................................................................................................... 2
1.4 Using a model to quantify the impact of interventions ................................................. 3
1.5 Need for adaptation of interventions to farm types ..................................................... 4
1.5.1 Resource endowment ........................................................................................... 4
1.5.2 Constraints ......................................................................................................... 4
1.5.3 Goals and attitudes .............................................................................................. 4
1.5.4 Examination of farm heterogeneity ........................................................................ 4
1.6 Problem Analysis ..................................................................................................... 5
1.7 Aim of study ........................................................................................................... 5
2 Materials and methods ..................................................................................................... 6
2.1 Description of the overall project .............................................................................. 6
2.2 Environment ........................................................................................................... 6
2.3 Overview of steps taken .......................................................................................... 8
2.4 Interventions to be reviewed ........................................ Error! Bookmark not defined.
2.5 Data sources .......................................................................................................... 9
2.6 Literature review of assessed interventions ................................................................ 9
2.7 Impact indicators ................................................................................................... 10
2.7.1 Food production ................................................................................................. 10
2.7.2 Livestock productivity ......................................................................................... 11
2.7.3 Soil Organic Carbon ............................................................................................ 11
2.8 Farm typology ....................................................................................................... 12
2.8.1 Decision tree ...................................................................................................... 12
2.9 Model study ........................................................................................................... 13
2.9.1 Model introduction .............................................................................................. 13
2.9.2 Modelling approach ............................................................................................. 13
2.10 Interview .............................................................................................................. 18
2.10.1 Conducting the interview ................................................................................. 18
2.10.2 Data analysis .................................................................................................. 18
2.10.3 Matching of constraints .................................................................................... 19
3 Results.......................................................................................................................... 22
3.1 Literature study ..................................................................................................... 22
3.1.1 Improvement of livestock forage .......................................................................... 22
3.1.2 Dual-purpose varieties ........................................................................................ 22

viii
 3.1.3 Drought tolerant crops ........................................................................................ 23
3.1.4 Fallow ............................................................................................................... 23
3.1.5 Mulching ............................................................................................................ 23
3.1.6 Crop residue management ................................................................................... 24
3.1.7 Manure management .......................................................................................... 24
3.1.8 Physical and chemical treatment of crop residues ................................................... 25
3.2 Typology ............................................................................................................... 25
3.2.1 Crops grown ...................................................................................................... 25
3.2.2 Manure and fertilizer use ..................................................................................... 26
3.2.3 External livestock feed ........................................................................................ 26
3.3 Modelling .............................................................................................................. 26
3.3.1 Baseline scenario ................................................................................................ 26
3.3.2 Impact of interventions ....................................................................................... 29
3.4 Interview .............................................................................................................. 32
3.4.1 Basic farm typology ............................................................................................ 32
3.4.2 Constraints ........................................................................................................ 33
3.4.3 Goals ................................................................................................................ 34
3.4.4 Attitudes............................................................................................................ 36
3.5 Expert opinion on constraint strengths ..................................................................... 39
3.6 Matching expert opinion to farm constraints .............................................................. 40
3.7 Decision tool .......................................................................................................... 44
3.8 Example ................................................................................................................ 45
4 Discussion ..................................................................................................................... 46
4.1 Setting of baseline scenario..................................................................................... 46
4.1.1 Crop .................................................................................................................. 46
4.1.2 Livestock ........................................................................................................... 46
4.1.3 Fertilizer ............................................................................................................ 46
4.1.4 Crop residues ..................................................................................................... 46
4.2 Outcomes baseline scenario .................................................................................... 47
4.2.1 Yields ................................................................................................................ 47
4.2.2 Food self-sufficiency ........................................................................................... 47
4.2.3 Livestock production ........................................................................................... 47
4.3 Impact of interventions........................................................................................... 48
4.4 Modelling considerations ......................................................................................... 49
4.4.1 Weather and soil data ......................................................................................... 49
4.4.2 Modelling limitations ........................................................................................... 50
4.5 Interview .............................................................................................................. 51
4.5.1 Basic farm characteristics .................................................................................... 51
4.5.2 Constraints related to type .................................................................................. 51
4.5.3 Goals related to type........................................................................................... 51

ix
 4.5.4 Constraints and goals related to gender ................................................................ 52
4.5.5 Attitude ............................................................................................................. 52
4.6 Considerations interviews ....................................................................................... 53
4.6.1 Overall considerations ......................................................................................... 53
4.6.2 Considerations constraints ................................................................................... 53
4.6.3 Considerations goals ........................................................................................... 54
4.7 Decision tool .......................................................................................................... 54
5 Conclusion ..................................................................................................................... 56
6 References .................................................................................................................... 58
7 Appendices .................................................................................................................... 64
7.1 Appendix I: Nutritional values for food and feed ........................................................ 64
7.2 Appendix II: Parametrising FARMSIM ....................................................................... 65
7.3 Appendix III: Interviewing protocol .......................................................................... 71
7.3.1 Introduction ....................................................................................................... 71
7.3.2 Aim ................................................................................................................... 71
7.3.3 Farmer Selection ................................................................................................ 72
7.3.4 Interview ........................................................................................................... 73
7.3.5 Checklist for procedures. ..................................................................................... 79
7.4 Appendix IV: Determination of expert opinion ........................................................... 80
7.4.1 Questionnaire ..................................................................................................... 80
7.4.2 Assessment of biomass enhancing interventions in Burkina Faso ............................. 80
7.5 Appendix V: Typology ............................................................................................. 82
7.6 Appendix VII: Impact of interventions ...................................................................... 83
7.7 Appendix VIII: Interviews ...................................................................................... 87

x
1
1 Introduction
1.1 General introduction
Agriculture is the principle source of livelihood in the Sahel, employing more than half of the
working population and contributing to close to 40% of the gross domestic product (Mortimore and
Adams, 2001). However, farmers in the Sahel are one of the poorest in the world (Zorom et al.
2013) and they also are vulnerable, due to the erratic rainfall and the frequent long dry spells,
creating an unpredictable natural production potential (the World Bank, 2013; Dreschel et al.
2005).

Farmers in the Yatenga province of Northern Burkina Faso are no exception. This province has
problems of environmental deterioration; a great part of the land is degraded or eroded (Critchley,
1991). Many households suffer from yearly hunger periods during the dry periods, lasting from 3 to
5 months (WLE, 2014).

1.2 Present challenges

Besides water as a limiting factor to agricultural production in Burkina Faso, low soil fertility and
nutrient limitation are also strong. It limits plant production and restricts the vegetation use of
water to 10-15% of the rainfall (Bationo & Buerkert, 2001; Bationo & Mokwunye, 1991; de Ridder
et al., 2004). Low soil fertility has been blamed to cause chronic hunger for 28% of the African
population and to lead to more than 50% of the population surviving on less than 1$ a day
(Bationo et al., 2007). The increasing population pressure leads to a reduction in the amount and
the duration of fallow periods, causing a shift to permanent cultivation and a following decrease in
soil quality. The use of external inputs is still marginal, with the yearly nitrogen plus phosphate
fertilizer use in Burkina Faso in the year 2010 averaging only 8 kg of N+P205 total nutrients per
hectare. In the same year this value was 120 and 227 kg N+P 2O5 total nutrients for France and the
Netherlands respectively (FAOSTAT, 2013). For these various reasons insufficient nutrients are
returned to the soil, resulting in soil depletion (de Ridder et al. 2004). Therefore the rate of soil
degradation in the Sahel is high (Mando & Stroosnijder, 1999).

1.3 Future challenges

Next to the existing challenges, it is expected that climate change will strongly affect Africa,
especially the drier regions such as the Sahel. Rainfall will become even more irregular, both in
amount and distribution (Barbier et al 2009). Panthou et al. (2014) found that the frequency of
extreme rainfall events has increased in the last decades. These changes have and will have
repercussions on ecosystem services, agricultural production and livelihoods (Mertz et al. 2009).

Combined with this, the population growth rate in the Sahel is one of the highest in the world,
reaching 3.1% in Burkina Faso (Haub and Kaneda, 2014). This population increase had led to a
decreasing per capita food production over the past thirty years, even though the overall food
production in the country has increased (Bationo et al., 2007).

The interplay and additional challenges of soil degradation, population growth and climate
change lead to the already scarce natural resources and biomass production being under
increasing pressure. There is a need to enhance food and feed production, while keeping soil
quality as optimal as possible.

There are several different ways for increasing food and feed production and promoting soi l
fertility. Potential avenues for improving these aspects are specific interventions aimed at
improving soil fertility (e.g. manure management), genetic improvements in crops to increase the
food and feed production, practices to reduce erosion, to optimize growing conditions, to conserve
water, etc. In this thesis the focus is on biomass enhancing activities because biomass production
serves many purposes in farming systems. It is important as a source of food, feed, fuel and fibre
and increasing biomass production could therefore lead to decreasing the hunger period and
increasing income.

2
Biomass is an important player in regulating and supporting ecosystem services (ES). Regulating
ES are benefits received by the regulation of ecosystem processes, such as pollination, erosion
control, water regulation. Supporting ES are necessary for providing the other ES and have only
indirect effect on people (such as soil formation processes which indirectly affect people through
improved food production). Biomass plays a role in these two ES as it provides organic matter
(OM) for the soil and it can act as a protective soil cover. Returning OM to the soil improves soil
fertility and structure through many different pathways (MEA, 2003), including the provision of
nutrients, the provision of habitat for soil organisms and the improvement of soil water holding
capacity. A greater soil cover decreases erosion and traps sediments, thus keeping nutrients in
place and available for uptake by plants (Zhang et al., 2007). Next to this a soil cover also
stimulates biological activity, changes soil temperature and reduces water evaporation (Mando &
Stroosnijder, 1999). On the long term, enhanced biomass production could therefore lead to
improved plant nutrient provision and better soil water retention (Power, 2010). This then
increases plant production and leads to improved food and feed availability.

Biomass is also an important player in the provisioning ecosystem services. Crop biomass
production leads to the availability of food and feed, while the biomass is also a source of raw
materials which can be used for construction and fuel (Valbuena, 2015). The presence of biomass
also affects the flow and the purification of fresh water, influencing the quantity and quality of fresh
water available. Forests and grassland can also be a source of plants used in medicine.

The optimal situation is attained when food and feed production are improved and the amount of
OM returned to the soil has increased. This means that both the provisioning and the regulating
and supporting ecosystem services would improve. However, trade-offs are often found between
maximizing provisioning services on the one hand and maximising regulating, cultural and
supporting services on the other hand (MEA, 2003 and Elmqvist et al., 2013). Therefore optimizing
food production in the short term might not lead to highest amount of biomass available for soil
cover and soil organic matter in the long term. At the same time, it is not impossible to improve
both the provisioning and the regulating ES, as cases are found where the implementation of new
agricultural techniques have led to the reduction of the trade-off between the two ecosystems
services (Elmqvist et al., 2013). Examples are the introduction of integrated pest management and
the implementation of integrated nutrient management (Pretty et al., 2006).

1.4 Using a model to quantify the impact of interventions

To be able to reduce the trade-off it is important to understand the role of biomass in farming
systems, as different interventions will affect the nature or the magnitude of the trade-off in a
different way (MEA, 2003).

One way to understand the impact of interventions is through the use of a model. A model can give
insight into the complexity and the conflicts in agro-ecosystems, through the simulation of different
system components, their interactions and the effects of processes at various levels (field, farm,
landscape) (McCown, et al., 1996). Not only information can be gained about the yearly impact of
the interventions, a model can also give insight about the long-term consequences of interventions
and the sustainability of a system. The level chosen for this thesis is the farm level, enabling the
following of on-farm food and feed production while keeping track of the soil fertility in different
fields. This makes it possible to simulate the impact of different management options.

A model can also be used to perform an exploratory research, by simulating the impact of potential
interventions. It hereby gives insight in the underlying reasons for observed effects, the key
constraint and opportunities in agricultural systems (Whitbread et al., 2010). Next to this the use
of a model can also directly engage farmers, inviting them in the design of farming system, leading
to the designing of on-farm experiments and changes in the farming practices. (Whitbread et al.,
2010). In addition to this a modelling exercise is fast once it is set up, compared to a field trial.

The NUANCES FARMSIM model (Nutrient Use in Animal and Cropping systems – Efficiencies and
Scales, FARM SIMulator) is a model adapted to the African smallholder farming systems, which can

3
help understanding the role of biomass in farming systems through simulation exercises. It is
possible to use NUANCES-FARMSIM to analyse the effect of various interventions within farming
systems.

1.5 Need for adaptation of interventions to farm types

Next to understanding the role of the biomass, it is also important to think about the drivers and
constraints which influence the practice of biomass enhancing measures, or how well interventions
are adapted to the context in which farmers operate. Any innovation must answer the goals of
farmers and their short-term will to increase income, food production and reduce risks (Bationo &
Buerkert, 2001). Two main paradigms exist on the explanation of the adoption of technologies
(Bidogeza, 2009):

- The economic paradigm. It states that resource endowment determines adoption of

interventions. In this paradigm no emphasis is put on the heterogeneity in preference
among farmers.
- The innovation diffusion paradigm. This paradigm puts weight on the role of information,
risk factors and social positions in the adoption of interventions. Households have different
roles; they can for example be innovators or followers and therefore more or less promptly
adopt new interventions. No emphasis is put on the practical ability of the farmers to
perform the intervention.

In this thesis elements from the two paradigms are used. Resource endowment is taken into
account in determining the adoption of interventions. Farmer constraint, goals and attitudes, which
correspond more to the diffusion paradigm are also considered.

1.5.1 Resource endowment

Land, livestock, capital and labour resources differ between farmers and gender. These differences
will determine soil fertility management practices, farm management and the type of investments
made (Defoer et al., 2000). Resource endowment will also influence the farmer’s perception of a
new intervention and play a role in the adoption of interventions. The role of biomass will also be
different on the different farms, as available resources (land, livestock, etc.) vary. Thus the
repercussions of interventions on food production, livestock productivity and SOC are unique. For
this reason it is necessary to tailor biomass interventions to specific contexts, by exploring the role
of biomass for farmers with varying resource endowments.

1.5.2 Constraints
It is also important to review farming constraints, defined here as the limitations present reducing
potential agricultural production. Where no limitations are present, the constraints restricting
adoption are the lowest and the adoption potential could be increased. Production constraints can
be felt at farm level (e.g. land shortage) or beyond the farm level (e.g. no access to seeds). The
farmer’s resources will partly define what the constraints for biomass production are, as the
constraints often depend on socio-economic conditions (Defoer et al., 2000). Examples of potential
constraints are capital, education, technology, land and livestock.

1.5.3 Goals and attitudes

Two other important points in determining the suitability of an intervention to farmers are farming
goal (what the farmer intends to attain through farming) and the farming attitude (the farmer’s
disposition towards farming). They influence the adoption rate of the intervention as they
determine which practices are most relevant to farmers. For example, if crop yield production is
very important, measures which strongly increase soil organic carbon (SOC) but for which the
increase in yields is only slight and visible after several years will not be prioritized by the farmers
and adoption rates may be low.

1.5.4 Examination of farm heterogeneity

An important step in analysing the suitability of interventions for farmers is the examination of
farm heterogeneity. Failing in doing this can lead to the promotion and implementation of a ‘one-
size- fits-all’ intervention which does not fulfil the farmer’s requirements (Woolverton, 2014). In
the aim of creating biomass enhancing interventions, not only one solution should be manufactured

4
but a “basket of options”, should be proposed and reviewed. For each group of farmers the best
suitable options can then be suggested (Defoer et al., 2000).

1.6 Problem Analysis

There is an increasing pressure on biomass in the Yatenga province of Burkina Faso. The ‘basket of
options’ available for increasing biomass production has to be reviewed, to quantify the potential of
interventions to increase food production, livestock productivity and promote soil fertility. The
heterogeneity of farmers should be taken into account to select and promote the interventions best
suited for each farmer and with the greatest chance of adoption.

1.7 Aim of study

The aims of this project are the following:

- Understanding farm heterogeneity, through the creation of a farm typology

- Assessing the effects of interventions on SOC, food production and livestock productivity
(milk and livestock weight) for different farm types through a modelling exercise
- Understanding farming constraints, attitudes and goals with the help of an interview
- Developing a decision support tool to enable the matching of the best suitable and most
easily implemented interventions to farm types

5
2 Materials and methods
2.1 Description of the overall project
This thesis is situated within the project ‘Realising the full potential of mixed crop-livestock systems
in rapidly changing Sahelian agro-ecological landscapes’, WLE Volta and Niger Focal Region. This
project is led by the International Livestock Research Institute (ILRI), through a consortium with
several partners. The aim of this project is to: ‘investigate the gender gap and assess how women
and youth can contribute and benefit from improved biomass production and productivity of crop-
livestock systems’. It also investigated ‘the effect of improved biomass production and
management on soil and water related ecosystem services’. Within this project a lot of attention is
drawn towards the impacts of interventions on agricultural productivity, natural resource
sustainability, food security and livelihoods.

The role of Wageningen University (WUR) within this project is to assess the impacts and trade-offs
related to intervention adoption at different scales (farm scale, regional scale, etc.) through
modelling and participatory tools. This is done for two main aims: designing technologies adapted
to households and farmers and creating decision support tools.

Within the project, this thesis assesses the impact of interventions on food security, income and
soil quality through a modelling exercise, for the Yatenga region of Burkina Faso. Through
interviews the adoptability of the interventions for different farm types in this region is analysed.
From this information a decision tool is made to help extension agents in choosing the most
suitable interventions for different farmers.

2.2 Environment
The location of this project is in the Yatenga region, around the city Ouahigouya. This region is
situated in the northern part of Burkina Faso, a land-locked country in West Africa (Figure 1).

a b

Figure 1: Map of Africa with Burkina Faso in red (a) (source: History & Maps, 2015) and the map of
Burkina Faso with all major cities (b) (source: Bambara, 2010).

Burkina Faso itself is situated in the Sahel, the semi-arid region forming the transition between the
Sahara desert and the more humid savannas. This ecological zone extends from the Atlantic coast
in West Africa to Sudan in the east (de Ridder et al, 1982) and is delimited by the 100mm and
600mm isohyets, with rainfall strongly increasing from North to South. There is one rainy period
during summer, lasting two to four months. It usually starts in June and lasts until September
(Figure 2). This period is followed by extremely dry conditions during the rest of the year (de
Ridder et al, 1982).

6
 Monthly average 200
rainfall (mm)
150

100

50

Figure 2: Monthly average rainfall (mm) in Ouahigouya for the years 1981 to 1992. From FAO
weather station placed at 13.75 latitude, -2.42 longitude and an elevation of 336m.

In the study area, the year can be divided in three following three seasons (Figure 3): the rainy
season (June to October), the early dry season (November to January) and the late dry season
(February to May). These three seasons define biomass availability: in the wet season the crops
are grown and a lot of biomass is available in the rangelands, in the early dry season there is still
some vegetation present, while at the end of the dry season biomass availability is lowest and
there is a shortage of grazing resources and crop residues for animal feed. At that time in the year
the quality of the pasture is also lowest (Breman & de Ridder, 1991).

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure 3: Schematic representation of the seasons in Burkina Faso. Green represents the wet
season, orange the early dry season and red the late dry season.

Soils in the Sahel region range from sandy to sandy loam, soils with a good permeability. The
levels of organic matter, total nitrogen and cation exchange capacity are low, leading to inherent
low soil fertility (Bationo & Mokwunye, 1991). Most soils are also low in phosphorus. Table 1 shows
an example of a typical soil in the Ouahigouya region (A. Ayantunde, personal communication;
Leenaars et al., 2014)

Table 1: Typical soil properties for top soils in the Yatenga region.
Exch. K
Bulk density Top soil Clay Silt pH SOC Mineral Olsen P
(mmol(+)/k
(kg/m³) depth (m) (%) (%) H2O (%) N (%) (mg/kg)
g)
1550 0.2 10 10 5.2 0.8 0.08 3.93 4.8

Agriculture is important in Burkina Faso, as it represents around 30% of the countries’ GDP and
employs more than 90% of the working population (FAO, 2014). Most farms are small scale farms,
the area they occupy usually being less than 5 ha (FAO, 2014). The main crops produced are
sorghum, millet and maize as staple food and cotton and groundnut for cash crops (Mortimore and
Adams, 2001). Crops are grown once a year, during the rainy season. Yields are generally low.

7
The Yatenga region (with the capital Ouahigouya) has one of the highest population growth rates of
Burkina Faso. Until the 1970s, there was enough land available and it was possible for farmers to
expand agriculture on marginal land. The traditional agricultural practices caused erosion but long
fallow periods restored the soil fertility. Population pressure led to reduction and elimination of
fallows. Since the 1980s, to avoid the decline in agricultural production, agriculture is intensified
through the introduction of zaï and stone bunds, improving soil and water conservation
(Douxchamps et al., 2004). Thus far the efforts to enclose livestock to collect better manure and
reduce grazing pressure have not been successful (Ouedraogo et al., 1996). Therefore livestock
(cattle and small ruminants) graze in the vicinity of the villages throughout the year and feed
mainly from the rangelands (de Ridder et al., 1982). In the dry season, when there is feed
shortage in the rangelands, the livestock depend mainly on crop residues (Rattunde, 1998).
Livestock plays an important role as it serves multiple functions, such as milk provision, draft
power provision and the role of insurance in case of failed crop (Herrero et al., 2003).

2.3 Overview of steps taken

The flow chart in Figure 4 is a schematic overview of the methodology followed to create a decision
tool helping in the choice of the best suited interventions.

Firstly a short literature search was performed, to understand what the interventions consist of and
what possible impacts they can have on farming system, while also learning what possible
constraints could limit the adoption of the interventions.

Secondly a farm typology was used to study farm heterogeneity to get an overview of the situation
of farmers in the Yatenga region of Burkina Faso.

Thirdly the NUANCES-FARMSIM model was used to explore the impact of specific interventions for
different farm types. The impact of the interventions was scored on the following points: food
production, livestock productivity and soil organic carbon, used as an indicator for soil fertility.

Fourthly interviews were held in the Yatenga region to collect information about farmer’s
constraints, goals and attitudes. This information was matched to information received from
experts about the strength of constraints limiting the adoption of several interventions.

Lastly a decision tool was made, showing which interventions are most adapted to different farm
types. This information can then be used within the WLE-Volta project to match the interventions
to farmers but it can also be used by extension officers, to find the best interventions for different
farm types.

8
 Figure 4: Flow chart of the steps taken during the thesis.

2.4 Data sources

In this thesis, two main sources of data were used for the parametrization of the baseline
modelling scenarios and for the definition of the farm types. Each time one of the two databases
was used in this thesis, it was mentioned in the methodology.

The farm types were based on information from the Detailed Household Characterization Survey,
collected by CCAFS (CGIAR’s research program on Climate Change, Agriculture and Food Security).
This was done as data from the WLE-Volta project was released only later. The main aim of this
survey was to collect standard indicators as well as key livelihood indicators, with the aim of
standardizing analyses. The survey was held in July and August 2012, in the villages Barga, Titao
and Namissiguima in the Yatenga and Loroum states of Burkina Faso.

The input data for the baseline scenarios was mainly based on the household survey data made
available from within the WLE-Volta project (survey held in June and July 2015). The objectives of
the survey were to characterise the types of farms, describe the opportunities and constraints for
biomass production, as well as collecting information about the farmer’s perception of the effects of
the biomass enhancing intervention. The villages included in the data set are Ziga, Thiou, Ninigui
and Pogoro-silmimossin, all villages in the Ouahigouya district.

2.5 Literature review of assessed interventions

The selected interventions (Table 3) were assessed based on a short literature review. The
research was done in google scholar, Scopus and through Wageningen library, with the following
search terms: key words of the intervention, coupled to “impact”, “effect”, “Sub-Saharan Africa”,
“SOC”, “yield” and a combination of these keywords.

The literature information was summarized in a quick explanation of the intervention, its’ impacts
on the farming system and on the requirements in terms of labour, education, equipment and
access and availability of inputs.

9
Within the WLE-Volta project several interventions were offered (Table 3). Only a set of
interventions was selected for this thesis, based on the possibility to simulate the intervention in
the current version of FARMSIM.

Table 3: Overview of interventions in the WLE-Volta project, separated whether included in the
thesis or not.
Interventions included in the thesis Interventions excluded from thesis

Pasture planted with native or exotic species Establishment firewalls

Forage planted as live hedge Collection of wood for fire
Planting of trees with multiple uses Clearing and pruning
Planting of forage cover crop Planting of vegetation bands
Planting of improved, dual-purpose varieties Gardens for food
Planting of drought tolerant crops Irrigated forage
Fallow Corridor planting with crops in between rows or trees
Collection herbaceous biomass for animal feed Afforestation
Collection and storage crop residue for animal feed Communal forest set-up
Leaving crop residues on field for animal browsing Burning of residues in the field
Mulching Early fire
Physical treatment crop residue Assisted Natural Regeneration
Chemical treatment crop residues Composting
Grazing rotation
Deferred grazing
Collection herbaceous biomass for selling
Collection crop residue for selling
Use of crop residue for building

2.6 Impact indicators

Biomass in this thesis had two main purposes; the improvement of livelihoods and the
improvement of soil fertility. Indicators were chosen to measure the goals of livelihood and soil
fertility. Livelihood was measured through food production and livestock productivity and soil
fertility was measured through soil organic carbon (SOC) content. The impact of the different
interventions gave information about the most beneficial interventions for these indicators.
Hereunder different indicators used in this thesis are further explained and a recapitulation can be
seen in Table 4.

2.6.1 Food production

As there is a yearly shortage of food in the Sahel region, it is crucial to increase food production.

In the baseline there were two ways of expressing food production: crop production and food self-
sufficiency. Crop production (kg/ha) was followed from years 1 to 12, enabling the visualisation of
yield evolution over time. The fraction of food self-sufficiency was calculated by comparing calorie
production to calorie need on a yearly basis. Vitamins or proteins were not taken into account. The
amount of calories needed for the household during the year was calculated by multiplying the
number of household members by their yearly calorific needs. The following was assumed: boys
aged 1 to 10 needed 1 300 kcal per day, while girls of the same age required 1 200 kcal, adult men
required 2 200 kcal per day and women needed 1 800 kcal per day (HHS/USDA, 2010). Calorie
production was calculated by multiplying annual crop yields by the area of the crops grown and the
calorific values of the crop (Annex 1, Table 1 for the nutritional value of crops). Food originating
off-farm was not taken into account. When calculating the average value of food self-sufficiency
over the years the first year of simulation was not taken into account.

When comparing the outcomes of the interventions to the outcomes of the baseline scenario, the
average calorie production (over years 2 to 12) was taken into account. The average calorie
production under the intervention was divided by the average calorie production under the baseline
scenario and this outcome was multiplied by 100, giving the percentage of increase or decrease of
average calorie production under the interventions compared to the baseline scenario..

10
2.6.2 Livestock productivity
Livestock also suffer from the yearly drought through feed shortage (quality and quantity). Survival
of livestock depends on the amount and quality of feed available (pasture grass, external feed and
crop residues). Livestock bodyweight and milk production were analysed.

For the baseline scenario milk production (kg/year) and livestock body weight, summed over the
animals (kg body weight/month) were followed over the 12 simulation years.

When comparing livestock productivity under the different interventions to livestock productivity for
the baseline scenario, milk production was averaged over the years 2 to 12 and cattle weight,
summed over all the cattle head of the farm, was averaged over these 132 months. Average milk
production and average cattle weight under the interventions were divided by average milk
production and average cattle weight in the baseline scenario. The outcome was then multiplied by
100. This enabled to find the percentage of increase or decrease in milk production and weight gain
due to the intervention, compared to the baseline scenario.

2.6.3 Soil Organic Carbon

Soil organic carbon (SOC) is a component of soil organic matter, decomposed plant and animal
organic matter in the soil (Wang, 2013). SOC is important in chemical, physical and biological soil
fertility (Chan, 2010) as SOC increases soil fertility through improving plant nutrient availability,
soil structure, water holding capacity and providing food for soil organisms. Increasing SOC levels
has been shown to increase yields in many cases (Körschens et al., 1998). Therefore the
maintenance of SOC is an effective way to counteract land degradation while increasing food
production (Bationo et al., 2007). Carbon inputs depend mainly on biomass productivity and the
return of biomass to the soil. Major carbon outputs are mineralization, leaching, erosion and run-
off. At this moment the amount of SOM in soils of the Sahel is very low, ranging from less than 1%
to around 2% (Beal at al., 2015). The aim of implementing the interventions is to increase the SOC
by increasing the amount of organic matter going to the soil. Other possibilities to increase SOC in
soils, not explored here, are changing the quality of the OM input, placing the OM in deeper soil
layers and enhancing the soil physical production of the OM through the formation of complexes
(Gobin et al., 2011).

For the baseline scenario SOC (%) was followed over the 12 simulation years. To get the value of
average SOC for a farm, the SOC of the different fields was multiplied by the area of the fields,
summed across the fields and the result was then divided by the total area of the farm.

When comparing the SOC outcomes for the baseline scenario with the SOC outcomes of the
interventions, only the SOC value for the last year of simulation were taken into account (year 12),
as SOC is a long-term impact indicator. The average SOC for the farm was calculated for this year,
for the baseline scenario as well as for the intervention. The value of SOC for the intervention was
divided by the value of SOC for the baseline scenario and the outcome was multiplied by 100. This
gave the percentage increase or decrease in SOC of the intervention after 12 years of simulation.

Table 4: Indicators used in the analysis of the impact of interventions.

Category Food production Animal production Soil quality
Impact indicator Calories produced Animal body weight and Soil Organic Carbon
milk production

Unit for the Annual crop yields (kg/ha). Annual animal body weight Annual SOC on farm
baseline scenario Annual food self-sufficiency (kg), summed across (%).
(fraction). animals. Annual milk
production (kg).

Unit when Gain or loss of calorie Gain or loss of livestock Gain or loss in farm
comparing to the production (%), averaged over productivity (%), averaged SOC (%), averaged
base line scenario 11 simulation years over 11 simulation years. over 11 simulation
years.

11
 2.7 Farm typology
With data collected by CCAFS a farm typology was made by W. Paas.

The resulting structural typology was based on resource endowment (Alvarez et al., 2014). The key
variables used to separate farmers in different classes were the following:

- Household size (the number of household members per farm; “the household include
people who live and share meals at least one season per year and income generated with
farming.” (Quiros, 2013))
- Total area (total area of farm fields, hectare)
- Cereal area (area of the farm planted with cereals, hectare)
- Number of cattle (zebu, cow, camel)
- Number of small ruminants (sheep, goat, etc.)

This resulted in 6 different farm types. Type 5 did not possess livestock, while type 6 possessed the
most cattle. As according to the farmers cattle is very important, type 5 was considered the least
resource endowed farm type and type 6 was considered the wealthiest type. Type 1 had a small
amount of land and livestock, type 2 had a greater area of land and type 4 had some animals but
only little land. Type 3 was average with some land and some animals.

2.7.1 Decision tree

The information about the types was used to make a decision tree, enabling the classification of the
farms in the right types (Figure 5). The first criterion used was the number of cattle, as this
enabled the segregation of the farm types 5 and 6 from all other farm types.

Presence of cattle or small

ruminants?

No Yes

Type 5 More than 20 head cattle?

Yes No

Type 6 ≤ 6 small ruminants?

Yes No

≥ 2.75 ha total land area? ≤ 4.5 ha total land area?

Yes No Yes No

Type 3 Type 1 0 or 1 head of cattle 0 or 1 head of cattle?

Yes No Yes No

Type 3 Type 4 Type 3 Type 2

Figure 5: Decision tree for the classification of farmers in the right type.
12
2.8 Model study
2.8.1 Model introduction
The impact of introducing the technologies was assessed for each farm with the use of the
NUANCES FARMSIM model (Nutrient Use in Animal and Cropping systems – Efficiencies and Scales,
FARM SIMulator). The nuances framework is developed specifically for understanding of the
complexity of African farming and for discovering of best-fit technologies (van Wijk et al., 2009).
NUANCES-FARMSIM is an integrated crop livestock model, adapted to the African smallholder
farming systems.

NUANCES FARMSIM is a model constituted of three sub-systems:

- FIELD (Field-scale resource Interactions, use Efficiencies and Long-term soil fertility
Development), which calculates the crop production per season (depending on soil fertility
and inputs), simulates water and macronutrient dynamics and monitors indicators of
resource degradation (e.g. SOM and soil erosion). Crops and soil types can be defined, as
well as the rainfall and the nutrient inputs into the system.
- LIVSIM (LIVEstock SIMulator), which is a livestock productivity model simulating animal
production (body weight, milk production) based on breed-specific potential, feed
requirements and actual feed availability. Different animal types can be chosen (breed,
weight, age, sex, etc.). The model has a time step of a month.
- HEAPSIM (HEAP SIMulator), which simulates the nutrient cycling through manure
collection, storage and application. This module also has a time-step of a month.

A user interface is used to insert the input conditions. The different interventions are simulated
with the help of the model and the impact on food production, livestock productivity and SOC is
followed over the seasons. A duration of 12 years is chosen to be able to see the evolution in the
outcomes.

2.8.2 Modelling approach

A baseline scenario was made for each farm type: this was the starting point situation. The model
had to be adapted to the environmental situation in the Ouahigouya district (rainfall), to the soil
properties and the different farm types were constructed. This baseline scenario was the situation
to which the impact of the different interventions was compared.

2.8.2.1 Input data

2.8.2.1.1 Monthly rainfall

Rainfall data were from a meteorological weather station in Ouahigouya (latitude 13.57°, longitude
-2.42° and elevation 336m). The total dataset contained information about rainfall over 60 years
(1950-2010) but only the data from 1981 to 1992 was used, as this was the most complete. The
associated rainfall data can be seen under Appendix II, Table 1.

2.8.2.1.2 Soil characteristics

Two sources describing soil quality were used:

- Personal communication by A. Ayantunde, about soils in the proximity of Ouahigouya, in

the villages Sabouna and Ziga
- Africa Soil Profile Database version 1.2., by ISRIC. Two representative soil samples, close
to Ouahigouya, are chosen (reference numbers BF2877_RPROFIL40 and BFVALSOL_CN-47)
(Leenaars et al., 2014).

The soil bulk density was determined according to soil texture. Soil texture was between a sandy
loam and a loamy sand, consisting of 10% silt and 10% clay (USDA, n.d., a). This led to a bulk

13
density of approximately 1,550kg/m³ (USDA, n.d., b). The chosen soil parameters for this thesis
are shown in Table 1.

2.8.2.1.3 Livestock characteristics

It was chosen to work with the cattle breed Méré, as this breed is kept in West Africa. The Méré
breed is a cross between Zébu Peulh (Fulani, Bos indicus) and N’Dama (Baoulé, Bos taurus)
(Sanogo, 2011). This breed is used for traction, milk, manure, as investment and as insurance in
case the crops failure. FARMSIM parametrization for this breed can be found in de Ridder at al.
(2015) and in Appendix II, Table 2.

2.8.2.1.4 Livestock ownership

From the WLE Volta database it was found that in small herd mostly bulls were present. When the
herd size increased the fraction of females in the herd also increased Table 6 represents the typical
herd composition used to generate livestock input data for the model. If the farm owned only 1
head of cattle, this was the animal with ID number 1. When the farmer owned 5 head of cattle, the
simulated cattle had the ID’s 1, 2, 3, 4 and 5. If more than 10 cattle heads were owned, the table
was repeated, until the required number of animals was reached.

No information was available about age or weight of the cattle, therefore it was attempted to
represent a typical herd composition. The maximum weight of a male Méré adult is of 400kg, while
for a female this is of 300kg. When making the table it was assumed cattle had not reached
maximum weight.

The average amount of livestock available in the farm was defined by the average number of cattle
in the farm type.

Table 6: Characteristics of the cattle in a typical herd of 10 cattle head.

Cattle ID 1 2 3 4 5 6 7 8 9 10
Sex (Female/Male) M M F F F F F F F M
Age (year) 6 4 6 4 2 7 5 3 1 5
Weight (kg) 300 300 250 250 150 250 250 200 100 300
Lactation (yes/no) x x n n n n y n n x
Gestation (yes/no) x x n n n n n n n x
Pregnancy (month) x x - - - - - - - x
Time since calving x x 24 12 - 36 12 - - x
(month)
Lactating (month) x x - - - - 12 - - x

2.8.2.1.5 Livestock feed

No information was available about the amount of biomass grazed by the livestock in the WLE-
Volta or in the CCAFS databases. The parameters for the quality of grass were adapted for a
decrease in pasture quality in the early dry season and an even greater decrease in pasture
quality in the late dry season (Descheemaeker, personal communication) (Appendix II, Table 4).
The quality of the pasture declined in the dry season; dry matter content increased, but
metabolisable energy (ME) and crude protein (CP) contents decreased.

Livestock lose weight during the dry season, due to a lower availability and quality of feed, but a
part of the weight gained during the wet season and should enable the cattle to survive during the
dry season (Sanogo, 2011; de Ridder et al., 1991). From this information for all farm types a
“basic” amount of feed was set (pasture + concentrates + roughage, Table 5) so that the seasonal
weight change was present. This basic amount of feed was given throughout the year. In the early
dry season 80% of crop residues are also given as feed and during the late dry season the
remaining 20% of crop residues remaining are offered.

14
Table 5: Feed given to bulls, calves, default cows and gestating of lactating cows throughout the
year.
Feed category Gestating and
Bulls Calves Default cows
lactating cows
Concentrates (kg/head/day) 1 0.5 0.5 0.75
Pasture (kg/head/day) 2 2 2 2
Roughage (kg/head/day) 3 2.5 2.5 2.5

2.8.2.1.6 Crop characteristics

The crops used in this modelling exercise were millet, sorghum, maize, groundnut and cowpea. The
characteristics of these crops can be seen in Appendix II, Table 3.

2.8.2.2 Parametrization of different farm types

One virtual farm was built for every farm type, with the information about the area of cropping
land and cattle ownership (average values of farm type). Cattle ownership was already known, as
these are the values for the average of the farm type. To complete the baseline scenario, the
following information was retrieved from the WLE database:

- Crops grown and area of crops grown,

- Livestock ownership
- Manure use,
- Fertilizer use,
- Crop residue use,
- Off-farm livestock feed fed,
- Livestock grazing quality

The following paragraphs explain how these model settings were defined.

2.8.2.2.1 Crops grown (use of WLE database)

The following steps were taken to calculate the averages for the areas of crops grown in the
different farm types. The data used comes from the WLE-Volta survey.

1- The frequency of the cultivation of crops by farmers within a type was calculated. Only
when crops were grown by more than 50% of the farmers were they included in the
baseline scenario (except for sesame, as this crop cannot be simulated in FARMSIM).
2- The average area of cultivation for all retained crops was calculated for the farmers
growing the crops.
3- With information about the average crop areas the area ratio of every crop was calculated.
For example, if the average areas found were 2ha of sorghum and 1ha of millet, the ratio
was of 2:1. This means that 66% of total farm land was planted with sorghum and 33%
was planted with millet.
4- The total cropping area of the farm was the average farm area for the farm type. Together
with information about the area ratio of each crop, the area of each crop was calculated.
For example, an average farm area of 4ha. Continuing the previous example, 66% of the
farm was covered with sorghum and 33% of the farm was covered with millet. Therefore in
total 2.66ha was under sorghum cultivation and 1.33ha was under millet cultivation.

2.8.2.2.2 Manure and fertilizer use (use of CCAFS database)

For the calculation of fertilizer and manure use, the frequency of the fertilizer and manure use
within the different farm types was calculated, using data from the CCAFS survey. When less than
50% of farmers within a farm type used manure or fertilizer, it was assumed that manure and
fertilizer were not applied in the baseline scenario. In cases where 50% or more of the farmers
used manure or fertilizer, the average quantity applied was calculated through averaging the
amount of fertilizer applied by farmers using fertilizer, for the specific crop. This amount was then
used in the baseline scenario. In the baseline scenario it was assumed manure was kept in the
open air.

15
No information was available about the way manure is kept (e.g. use of a plastic sheet to cover the
heap).

2.8.2.2.3 Crop residue use

For farms with and without livestock it was assumed that all crop residues were used as livestock
feed. This was due to the fact that there is pressure on the crop residue resource, which is used as
livestock feed, fuel, construction material, can be sold, etc. (Powell et al., 2004; Bationo and
Mokwunye, 1991). Therefore there are little to no crop residues available as soil amendment. Even
for farmers without livestock the crop residues can be eaten, as freely-roaming livestock from
neighbours has access to crop residues left on fields.

The quality of the crop residues was based on the data available in the model and from the PhD
thesis of O. Sanogo (2011).

2.8.2.2.4 External feed

Livestock feed in FARMSIM was composed of crop residues, pasture and external forage and
concentrates. The use of the crop residues was explained in the paragraph above. Information
about the feeding of livestock with external forages and concentrates during the three seasons was
available in the WLE database. The proportion of farmers feeding their livestock with external feed
was calculated for every season. If more than 50% of the farmers fed the livestock with external
feed, it was included in the baseline scenario.

No information was available about the composition of concentrates or the frequency or quantity of
the concentrates fed. Therefore the following composition was assumed (Sanogo, 2011):

a. Dry matter (DM) 916 g/kg

b. Metabolisable energy (ME) 12.3 (MJ/kg DM)
c. Crude protein (CP) 276 k/kg DM
d. Dry matter digestibility (DMD) 710 g/kg DM

2.8.2.3 Simulation of interventions

The baseline scenario described above was altered so as to simulate the interventions. The way in
which this was done is explained in the following paragraphs.

2.8.2.3.1 Improvement of livestock forage

2.8.2.3.1.1 Increasing grass quantity available

This intervention was built to simulate a potential effect of pasture planted with native or exotic
species. Pasture quantity available to the livestock was increased by adding 1kg grass/head/day to
the cattle diet throughout the year in a first instance, followed by the addition of 2kg
grass/head/day to cattle diet throughout the year in a second instance.

2.8.2.3.1.2 Improvement of grass quality

This intervention was also simulated as a potential effect of pasture planted with native or exotic
species. Pasture quality was improved through the increase of 10% and later 20% of the ME and
CP content of grass.

2.8.2.3.1.3 Increasing external forage in the form of legumes

Forages planted as live hedge and the planting of trees with multiple uses were assumed to add
legumes to the cattle diet. Therefor these interventions were simulated through the addition of
legumes to the cattle diet. Three different levels of legume feeding were assumed: the addition of
0.5kg legume/head/day, 1kg legume/head/day and 2kg legume/head/day.

2.8.2.3.2 Dual-purpose varieties

Two different scenarios were constructed, where either cowpea or sorghum were used as dual-
purpose crops.

16
Firstly the area of cowpea was increased, as cowpea is a dual-purpose crop of itself. The area of
the farm under cowpea cultivation was increased from 0% to 25%, 50%, 75% and 100% total
farm area.

Secondly the crop parameters of sorghum were altered in two different ways, to mimic a dual-
purpose sorghum variety:

- Increased potential crop residue yield by 10% in a first run and by 20% in a second run,
while keeping the grain yield at the original level. This was achieved through increasing
light determined yields while reducing the harvest index. The increase of crop residue
production by 10% and 20% was chosen as increased stover production ranged from 0-
75% (Bossuet, 2010)
- Crop residue quality was increased by increasing CP and ME content of the sorghum forage
by 10% in a first run and 20% in a second run. As the idea was to change crop residue
quality not only the forage given to the livestock should increase in ME and CP, also the
crop residues produced on farm should become higher in quality (as setting forage quality
only influences the feed given to the livestock and not the crop characteristics). This was
accounted for by increasing the minimum and maximum amount of N in crop stover 10%
and then 20%. This increased crop residue quality but reduced the quantity of grain and
crop residues produced.

2.8.2.3.3 Drought tolerant crops

To increase drought tolerance of crops water conversion efficiency was altered, in a first run by
10% and in a second run by 20%. This was done for both sorghum and millet (in different runs),
as they are currently two existing drought-tolerant crops (CGIAR, n.d.).

2.8.2.3.4 Fallow
Fallow was simulated through the growing of crops with a harvest index (HI) of 0 (no grain
production) and 100% of biomass staying on the field. Two crops that were used for the simulation
of fallow were sorghum and cowpea.

First 10% and then 20% of the farm areas was converted to this fallow system. These proportions
were chosen because there was no information available about the farm surface fallowed or the
frequency of fallowing. Also, a general trend in sub-Saharan Africa is the reduction in fallows, which
even sometimes completely disappear (Sanogo, 2011).

This gave rise to four scenarios:

- 10% fallow with sorghum as fallow crop

- 20% fallow with sorghum as a fallow crop
- 10% fallow with cowpea as fallow crop
- 20% fallow with cowpea as a fallow crop

The area of the farm which was not converted to fallow was used to grow the crops. The areas over
which the crops were grown kept the same ratios in the fallow interventions as in the baseline
intervention (e.g. if 50% of land was under sorghum cultivation in the baseline scenario, also 50%
of land not under fallow was under sorghum cultivation during the fallow intervention).

2.8.2.3.5 Mulching
Mulching was simulated through the addition of the crop residues to the field. It was therefore
similar to the situations where crop residues were left on the field. The impact of mulching legume
residues (cowpea and groundnut) and cereal residues (sorghum, maize and millet) was analysed.

2.8.2.3.6 Crop residue management

The amount of crop residues allocated to the livestock was reduced, leading to more crop residues
being left on the fields. The following fractions of crop residues were used:

- 100% crop residues left on field, 0% used as livestock feed

17
 - 75% crop residues left on field, 25% used as livestock feed
- 25% crop residues left on field, 75% used as livestock feed
- 0% crop residues left on field, 100% used as livestock feed

2.8.2.3.7 Manure Management: Collection and storage and application

Manure collection was varied from 0% to 100%, by steps of 25% (0%, 25%, 50%, 75%, and
100%). The amount of collected manure applied was assumed to be 100%, with the amount of
manure spread proportional to the size of the field. In a first simulation exercise manure was kept
in an open heap, while in a second simulation exercise manure was kept in a heap covered by a
plastic sheet.

2.8.2.3.8 Physical and chemical treatment of crop residues

The crude protein and metabolisable energy content of all crop residues was increased by 10% and
then 20%. This situation was therefore similar to the sorghum dual-purpose intervention, with the
difference that not only sorghum but all the crop residues increased in quality.

2.9 Interview
2.9.1 Conducting the interview
Interviews were conducted to understand the constraints, goals and attitudes of farmers. These
were also the three subjects which made up the different parts of the interview.

Interview contents:

- Firstly broad questions about the constraints in agricultural production were asked. Then
more emphasis was put on land shortage, labour shortage, livestock shortage, capital
shortage and a shortage in education and technology. The farmer stated if he or she
thought the constraint was present and how strong this constraint was. The strength of the
constraint was scored on a Likert-type scale of importance from 1 to 5, from ‘very low’ to
‘very strong’ importance. An explanation was asked to clarify the given answer.
- Then the farmer was asked about his or her goals in agriculture. Specific statements were
made and the farmer had to state if he or she agreed to the statement and how strong his
or her feeling of agreement was. An explanation for the answer was asked. As for the
constraints the answers were scored on a Likert-type scale, ranging from 1 to 5, from ‘not
important’ to ‘very important’. The score 3 stood for a neutral answer.
- Lastly the farmer was asked about his or her attitude towards agriculture. This was done
through the grading on a Likert-types scale of answers on statements (scale of 1 to 5, from
total disagreement to total agreement, with 3 a neutral answer).

The interview was conducted in three villages in the region of Ouahigouya: Pogoro-silmimossin,
Ninigui and Thiou. When possible, four farmers from every type, randomly selected per village,
were interviewed.The detailed protocol for the interview can be found in Appendix III. This contains
the set-up of the interview, the questions asked and the selection of the farmers.

2.9.2 Data analysis

The first step was to verify if the farms were still situated within the right farm type (WLE Volta
data was collected in 2015 and this interview took place in 2016). If this was not the case the
farms were re-classified in the current farm type.

2.9.2.1 Analysis of association between type or gender with constraints and goals
A fisher exact test was used to discover if there were significant differences in the proportion of
respondent considering the elements as a constraint, between the types and per gender. The same
test was used to find if the proportions of answers given for the goals were different depending on
type and gender. A fisher test was chosen as sample sizes were small and the Likert type ranking
led to the obtaining of categorical answers (Ott & Longnecker, 2015). A significance level of 0.05
was chosen.

18
2.9.2.2 Analysis of farmer attitude
The distribution of the answers over the scale was calculated and discussed. The weighted
averages were given.

Then an agglomerative hierarchical cluster analysis (HCA) was performed, with an average linkage
within groups method (as linkage criterion) to construct a dendogram (Köhn, & Hubert, 2006). This
means the clusters were arranged in such a way that the average distance between all the
individuals in the resulting cluster was minimised. From this dendogram clusters in farmer attitude
were defined.

2.9.3 Matching of constraints

2.9.3.1 Calculating strength of agricultural constraints

Firstly the percentage of farmers within a type who thought the resource was a constraint was
calculated and reported.

Secondly the average weighted importance of the constraint was calculated (when the resource
was not seen as a constraint the score was not taken into account). For example, if 20% of people
say the importance is 3 and 40% say the importance is 4 and 40% say the importance is 5, the
weighted average will be 0.2*3+0.4*4+0.4*5 = 4.2

Thirdly this average was multiplied by the percentage of farmers saying the resource is a
constraint. The outcome is then given a score following (Table 7), following a fuzzy logic.

Table 7: Conversion of scores to ‘score equivalents’

enabling the matching of interventions.
Outcome calculated Score equivalent
0-0.99 0
1-1.99 0.25
2-2.99 0.5
3-3.99 0.75
4-5 1

2.9.3.2 Expert opinion

A questionnaire was made, to ask experts about their opinion on requirement of cropping land,
grazing land, livestock, education, capital and technology for the implementation of the different
interventions (Appendix IV).

The received answers were compared and the most frequently given answer was taken into
account. If two scores ended up with the same frequency, the strongest score (e.g. 2 instead of 1)
was taken into account. The resulting scores were 0, 1 or 2 (see Table 8, row 1). A 0 (green)
answer meant there was no input needed of the specific element. A 1 (orange) meant there was
some input needed, while a 2 (red) meant a lot of input was required.

The cost of implementation given by the experts was multiplied by the score equivalent given by
the farmers (Table 8, row 2). When the final outcome was under 0.25, it was coloured green (Table
8, row 3), while if outcome was below 0.5 it was coloured orange and above that the outcome was
coloured red.

Green interventions faced no or little constraints for adoption, orange interventions faced medium
constraints for adoption, while red outcomes signalled a strong constraint for the adoption of
interventions.

19
Table 8: Scoring equivalents to match farmer constraints to expert constraints and come to the final
constraints.

Sources and Scoring equivalents

outcome for
constraint
strength

Cost of
implementation
0 1 2
(information from
experts)

Score equivalent
(information given 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1
by farmers)

0.5
Final outcome 0 0 0 0 0 0 0.13 0.25 0.38 0 0.25 0.5 0.75 1

20
21
3 Results
3.1 Literature study
3.1.1 Improvement of livestock forage
In this thesis the following interventions are assumed to have as a main aim the increase of the
quantity and/or quality of livestock feed and/or the timing of feed availability.

3.1.1.1 Pasture planted with native or exotic species

Depending on the species planted grazing intake differs. If the changed species composition leads
to increased biomass, increased digestibility or a better spreading of the feed throughout the year,
livestock is positively affected.

To implement this intervention there is a need to plant the species, requiring labour and knowledge
about the species to plant, as well as access and availability of the seeds or plantlets required.

3.1.1.2 Forage planted as live hedge

Hedges reduce wind and water erosion (Ayuk, 1997). Depending on the species used, a live hedge
may be able to access water from different soil layers. The shedding of leaves can lead to an
increase in soil organic matter around the hedge. In this project the hedges are considered as a
measure to provide extra feed for the livestock, improving livestock productivity.

For this intervention hedges need to be planted and maintained, requiring seeds or plantlets and
labour for planting and maintenance.

3.1.1.3 Planting of trees with multiple uses

Trees can improve soil fertility by increasing nutrient supply, through nitrogen fixation and/or the
retrieval of nutrients from lower layers in the soil. Trees can also increase nutrient availability
through the enhanced cycling of nutrients and the conversion of nutrients to more labile forms.
(Buresh & Tian, 1998). Trees can also improve water infiltration and storage. Depending on the
tree species they may also play a role as livestock feed, the role of main importance in this thesis.

There is a labour need for planting, protecting, maintaining and propagating the trees. Feed for
livestock has to be harvested. When the tree is grown close to the crops it can also lead to
competition with the crop.

3.1.1.4 Collection of herbaceous biomass

Herbaceous biomass may be collected and fed fresh to the livestock when corralled. Biomass may
also be collected and dried, to serve as hay fed during the dry season. This increases the amount
of feed available to the livestock around the year, improving livestock productivity.

In both cases labour is involved for the collection of biomass and the biomass should be available.

3.1.2 Dual-purpose varieties

Dual-purpose varieties yield both grain and stover in good quantities. The grain is used as food or
seed, while stover is used for animal feed or as organic amendment to the soil. The grain and
stover yields should be evaluated in terms of both quality and quantity (Singh et al., 2003;
Kristjanson et al., 2005). Depending on the crop, the improvement of crop residue quality may or
may not occur at the expense of yield (Zerbini & Thomas, 2003). The dual-purpose crops most
used in sub-Saharan Africa are cowpea and a special variety of sweet sorghum.

Dual-purpose varieties can alleviate the pressure on land for the production of grains and stover.
Income can be increased through the sale of extra grain and stover. The soil can be enriched
through the return of more residues to the soil or the production of more crop residues may lead to

22
the enhancement of animal performance (Descheemaeker et al., 2010), as 45 to 80% of the
livestock diet consists of crop residues (Rattunde, 1998).

For the implementation of dual-purpose crops seeds have to be available and accessible. There is
also a requirement for knowledge about the crop, so as to manage it well (Kristjanson et al.,
2005).

3.1.3 Drought tolerant crops

Drought tolerant crop are crops which give a higher yield than the normal crop under drought
conditions (CGIAR, n.d.). Either new varieties can be bred with a higher drought tolerance, or crops
or varieties which are inherently drought tolerant can be taken into production, such as cassava.
When breeding a new variety drought tolerance can be achieved through the improved efficiency of
water capture, or through the improved efficiency of water conversion (Schafleitner et al., 2007).

When crops are more drought tolerant farmers have more security about their yield, through the
reduced chance of a failed crop as crop production in the dry years is assured. To be able to
cultivate drought-tolerant crops, there is a need for seed access and availability (Cairns et al.,
2000). There is no change in labour requirement.

3.1.4 Fallow
Fallow is a traditional means of increasing soil fertility in West-Africa (Bilgo et al., 2007) by taking
land out of production. This can last from one season to many years depending on land use,
population pressure and soil qualities (Corbeels, Shiferaw & Haile, 2000). Due to the increasing
population and an increased need for fertile land, a decrease in fallows has occurred in sub-
Saharan Africa.

Fallows can break weed and pest and disease cycles. They can also improve soil fertility, reducing
erosion and increasing soil organic matter and biological activity through the input of organic
matter from primary production into the soil (Styger & Fernandes, 2006; Aguilera, 2013). Usually
there is less nutrient mining in areas where fallows are still common (Drechsel et al., 2001).
Fallows can also be used for the provision of medicinal plants and fibres, timber, etc., to improve
livelihoods, or be used as pasture (Styger and Fernandes, 2006).

However, if the fallow is not well managed (organic matter input smaller than organic matter
outputs from the system, including mineralization and erosion), SOC may decrease, reducing soil
fertility. Fallowing takes land out of production and thus reduces the potential food and feed
production of the farm. The impact of the fallow depends on the duration and production during the
fallow (Langyintuo & Dogbe 2005) and knowledge is needed on how to maximise and maintain soil
productivity (what crops to plant, for how long, etc.).

3.1.5 Mulching
Mulching is a technology whereby the soil surface is covered by organic matter (OM), often applied
in association with zero or reduced tillage. There are many different ways of mulching: the mulch
can be dead (e.g. straw) or alive (e.g. low growing intercrop), it can be produced on the field or it
can be acquired from outside (Erenstein, 2003).

The benefits obtained from mulching are multiple. Organic matter is added to the soil, improving
soil fertility. Mulch acts as soil cover, reducing erosion, run-off and water evaporation. Through
these functions it can play a role in increasing yields. Mulching increases moisture retention by
buffering the soil against extreme temperatures (Twomlow et al., 2008). It also can reduce weed
infestation through shading weeding (Erenstein, 2003; Twomlow et al., 2008).

Mulching can be a complicated technology. The user needs knowledge about the application of a
good amount of mulch and in the case of live mulch, the system should be managed so as not to
compete with the main crop. It is usually stated that mulch cover should be at least 30% to act as
effective soil protection (SSSA, 1986). This may be difficult in sub-Saharan Africa due to many

23
alternative uses for the mulch (livestock feed, construction materials, composting, etc.). When the
mulching material comes from outside the field or farm labour is required to transport and apply
the mulch. Tillage needs to be altered, so as to minimise the amount of residues being
incorporated into the soil (Erenstein, 2003).

3.1.6 Crop residue management

Once the crop residues are produced, there are often competing claims on them. There is a
pressure on the crop residue resource as low yields in addition to a growing population and a
reduction of communal resources have led to a dependence on the use of crop residues for
livestock feed, selling, fuel, construction and soil amendment (Valbuena et al., 2015). Two
potential uses explored in this project are the use of crop residues as animal feed and the use of
residues as soil amendment.

At this time the crop residues are mostly used as livestock feed (Valbuena et al., 2012), increasing
the availability of draught power, meat and manure and fulfilling a function as insurance and
savings. Livestock are very dependent on the crop residue feed source, especially in the dry season
and in case of droughts (Rattunde, 1998).

When crop residues are left on the soil they play a role in the improvement of soil fertility in the
long term, through the addition of organic matter to the soil (Omotayo & Chukwuka, 2009). They
also act as a cover, protecting the soil from erosion. Leaving crop residues on the field does
however not suffice to counter the loss of nutrients due to grain harvesting.

When the crop residues are left on the field no labour is needed. However, when the residues are
transported and handled labour is needed.

3.1.7 Manure management

Manure management contains three important steps influencing the quantity and quality of manure
returned to the soil: collection, storage and application. Manure is used to return organic matter
and nutrients to the soil, in the aim of increase crop yields. Manure is part of an internal flow of
nutrients within the farm. It can add nutrients to the farming system when livestock graze outside
the farm but manure is collected. However, as feed is scarce and the number of animals is limited,
this is often not the case (Bationo et al., 2001). The quantity and quality of manure is often low
and inadequate to meet the crop requirements (Vanlauwe & Giller, 2006; Bationo et al., 2001).

3.1.7.1 Manure collection

The more manure is collected, the more manure is available for the next handlings and the more
will be available as an organic amendment. Housing (e.g. corralling or not) also influences the
amount of manure collected (Rufino et al., 2007).

Manure collection is labour intensive, depending on the way the cattle is kept (cattle kept in a
corral or allowed to roam). Faster and more frequent manure collection leads to less C and N losses
and higher manure quality (Rufino et al., 2007).

3.1.7.2 Storage
Manure can be stored in many ways: in open or closed compartments, manure can be turned over
or left as it is, organic material can be added, etc. The conditions of storage affect decomposition
and nutrient losses, as aerobic conditions allow a faster decomposition of carbon than anaerobic
conditions (Rufino et al., 2007). Rufino et al. (2007) state that even the size of the manure heap
influences decomposition process, due to heat distribution within the heap.

Labour and knowledge are needed to store manure in a way to maintain quality.

3.1.7.3 Amount applied to the field

The more manure is applied to the fields and the higher the manure’s quality, the more positive the
effect on the soil fertility. Spreading manure is labour intensive and there is a need for farming
equipment (Vanlauwe & Giller, 2006).

24
3.1.8 Physical and chemical treatment of crop residues
Treatment of crop residues is done to improve the quality of crop residues as animal feed.

3.1.8.1 Physical treatment of CR

Physical treatments of crop residues include grinding, chopping, shredding and pelleting. These
lead to an increased area of the feed exposed to air, resulting in a greater activity of digesta,
followed by a greater voluntary intake of crop residues (Lawrence, 1993). This can potentially
increase livestock productivity. Digestibility of crop residues is usually not affected (Balch, 1977).

There is a need to collect the crop residues and to process them, requiring power and equipment.

3.1.8.2 Chemical treatment of crop residues

Treating the crop residues with chemicals such as hydroxide, ammonium hydroxide, anhydrous
ammonia or hydrogen peroxide enhances fibre digestion, resulting in a higher digestibility of the
crop residues and enhancing the voluntary crop residues intake (Lawrence, 1993; Owen, 1994;
Zerbini & Thomas, 2003).

Crop residues need to be collected, transported and processed, requiring labour and chemicals,
which are not prevalent in West Africa (Lawrence, 1993). There is also a need for knowledge for
the handling of chemicals.

3.2 Typology
Type 5 did not own livestock, type 6 had the most cattle and ruminants, type 2 had the most land
and the biggest number of household members, type 4 had some animals and a small piece of land
and type 1 had the smallest land area (Table 9). In FARMSIM, the maximum number of cattle
heads (20 instead of 33) was used for type 6.

Table 9: Characteristics of the farms used for the modelling exercise.

Type

Type 1 2 3 4 5 6
Household
8 13 8 9 6 15
members (#)
Area (ha) 1.5 10.6 4.8 2.3 3.0 17.9

Cattle (#) 2 5 1 5 0 33
Small ruminants
3 20 9 15 0 36
(#)

3.2.1 Crops grown

All farm types cultivated millet and groundnut. Except for type 5 all types grew sorghum and
except for type 6 cowpea was grown everywhere. Type 3 was the only type where maize was taken
up in the baseline scenario. Refer to Appendix V, Table 1, for a detailed account of percentages of
farmers growing particular crops within each farm type.

With the exception of type 3, millet was the crop which was grown over the largest area. The
legume areas were the smallest, except for farm type 3, where maize the maize area was even
smaller (Table 10).

25
Table 10: Area of crops (ha) per farm type and used in the baseline scenario
of the modelling exercise.
Type
Crop T1 T2 T3 T4 T5 T6
Millet 0.6 5.2 1.3 0.9 1.9 10.5
Sorghum 0.4 3.2 1.8 0.6 0.0 4.8
Maize 0.0 0.0 0.5 0.0 0.0 0.0
Cowpea 0.2 1.0 0.6 0.5 0.0 0.0
Groundnut 0.3 1.2 0.6 0.3 0.3 2.5
Total area 1.5 10.6 4.8 2.3 3.0 17.9

3.2.2 Manure and fertilizer use

From the CCAFS database it was inferred that less than half of the farmers from all farm types
used manure or fertilizer on their fields (Appendix V, Table 2). For this reason manure and fertilizer
application was assumed to be 0 in the baseline scenario

3.2.3 External livestock feed

With the exception of type 6, less than 50% of the farmers gave extra feed to their livestock during
one or more seasons. In farm type 6, 60% of the households fed concentrates to the livestock
during the early dry season (Appendix V, Table 3). It was therefore assumed that the livestock of
type 6 were fed with an extra daily ration of 0.5kg of concentrates during the early dry season
(November, December and January).

3.3 Modelling
3.3.1 Baseline scenario
In the baseline scenario there was an exponential decrease of crop yields over time, nearing a
plateau at the end of the simulation period around 500kg/ha (Figure 6: Evolution of crop yields
(kg/ha) over time.). A big jump in yields from the first to the second year was observed. Similarly
to the yields, SOC also decreased over time. The decrease in SOC decelerated over time and was
similar for all crop fields, going from approximately 0.75% to 0.6% (Figure 7).

3000

2500
Crop yield (kg/ha)

2000 Millet
1500 Sorghum

1000 Cowpea
Groundnut
500
Maize
0
1 2 3 4 5 6 7 8 9 10 11 12
Year

Figure 6: Evolution of crop yields (kg/ha) over time.

26
 0.8
0.7
0.6
0.5 Millet
SOC (%)

0.4 Sorghum
0.3 Cowpea
0.2 Groundnut
0.1 Maize
0
1 2 3 4 5 6 7 8 9 10 11 12
Year

Figure 7: Evolution of SOC (%) in the different fields over time.

Milk production was highest in the first years of simulation, after which it went to 0 (Table 11).
There was no milk production in farm types 1 and 3, as bulls were the only cattle present. Milk
production was highest for type 6, which had the highest number of animals. Similar to the trend in
milk production there was a decrease in livestock weight: cattle quickly lost weight and deceased
(Figure 8). As the total livestock weight of the herd was summed across the animals, farmers with
more livestock had a higher livestock weight. The seasonal variation in weight of the livestock is
illustrated by the wavy pattern in Figure 8. Livestock weight decreasing fastest in the dry seasons
but even in the wet season there was a loss of cattle weight.

Table 11: Amount of milk (kg/year) produced by the cattle for the different farm types.
Type
Year T1 T2 T3 T4 T6
1 0 395 0 355 1483
2 0 101 0 124 199
3 0 0 0 2 0
4 0 0 0 0 0
5 0 0 0 0 0
6 0 0 0 0 0
7 0 0 0 0 0
8 0 0 0 0 0
9 0 0 0 0 0
10 0 0 0 0 0
11 0 0 0 0 0
12 0 0 0 0 0

27
 4500
4000
3500
Cattle weight (kg)

3000
T1
2500
T2
2000
T3
1500
T4
1000
T6
500
0
1 13 25 37 49 61 73 85 97 109 121 133
Month

Figure 8: Monthly total cattle weight of the herd (kg, summed across the animals) for the farm
types.

The average yearly amount of crop calories produced over the 11 simulation years (first year of
simulation not taken into account) was highest for type 6, the farm type with the most cropping
land, while it was lowest for farm type 1, which had the least land (Figure 9). Except for farm type
1 all the farms produced on average enough food to feed their household. For types 2 and 6
respectively, the calorific production was approximately 4 and 6 times higher than the calorific
needs. This was related to the ratio of cropping land per household member.

70000000
Amount of calories produced averaged
over the 11 simulation years (kcal)

60000000

50000000

40000000
Calories produced
30000000 (kcal)

20000000
Calorific
10000000 requirement (kcal)

0
T1 T2 T3 T4 T5 T6
Farm type

Figure 9: Amount of crop calories (kcal) produced on farm, averaged over 11 years of the
simulation, compared to the amount of calories needed within the household.

Similarly to crop yields, the yearly calorific production of the farms decreased over the years
(Figure 10). Type 1 and 4 ended with a calorific production which was not sufficient to feed the
household, while type 5 was close to reaching the same situation. Type 6 was the farthest away
from insufficiently producing enough calories to feed the family, followed by types 2 and 3.

28
 18
Energy self-sufficiency ratio 16
T1
14
T2
12
10 T3

8 T4
6 T5
4 T6
2 Self-sufficiency
0
1 2 3 4 5 6 7 8 9 10 11 12
Year

Figure 10: Fraction of calorific food self-sufficiency for the farm types over the years.

3.3.2 Impact of interventions

The percentage change in calorific yield production, livestock productivity and SOC for the
interventions compared to the baseline scenario, are reported in Appendix VII, Tables 1 and 2. The
number of the interventions given in the following paragraphs can be used to navigate in the table
in annex (Appendix VII, Tables 1 and 2).

3.3.2.1 Improvement of livestock forage (interventions 1 to 7)

3.3.2.1.1 Increasing grass quantity available (interventions 1 and 2)

Increasing the quantity of grass fed led to higher livestock productivity as the overall quantity and
quality of the feed increased. In type 1 the increase in livestock weight reached 26.6%. This
intervention had no effect on crop yield or SOC, as there was no feedback mechanism (because in
baseline settings manure application was set at 0).

3.3.2.1.2 Improving grass quality (interventions 3 and 4)

Improving grass quality improved livestock productivity. The increase found depended on the ratio
of cropping land: number of cattle (the lower the ratio the stronger the effect) and on crops grown.

3.3.2.1.3 Increasing external forage in the form of legumes, live fences and/or trees (interventions 5 to 7)
Increasing the amount of legumes fed improved livestock weight. An increase in weight by 36%
compared to the baseline was reached for type 6 when 2kg of legumes were fed per day. This was
related to the improvement of overall feed quality. There was no effect on food production or SOC,
as there is no feedback mechanism (in the baseline settings manure application was set at 0).

3.3.2.2 Dual-purpose varieties (Interventions 6 to 14)

3.3.2.2.1 Dual-purpose cowpea (interventions 6 to 10)

A trade-off was visible between the optimization of livestock productivity and calorie production
(Figure 1112). The increase of the area under cowpea cultivation increased livestock productivity
but reduced calorie production. Livestock weight and milk production were increased thanks to the
elevated fodder quality (higher ME, MP and DMD). The rise in livestock productivity was dependent
on the number of cattle (e.g. type 3 owns only one male cattle head which does not have the
possibility to reproduce while type 2 owns 5 cattle head with the possibility to reproduce) and on
the ratio land area : cattle head. When this ratio was higher the increase in livestock productivity
when enlarging cowpea area was larger. The livestock still only survives in the first couple of years.

29
Enlarging cowpea area lowered farm calorie production due to the low calorific content of cowpea
than millet, groundnut and maize and a lower yield per hectare than all the other crops. The
decrease in amount of calories produced was dependent on the percentage of farm area under
groundnut cultivation, as groundnut was the crop with the most calories per kilogram yield and the
crop with the second highest yield, after maize (Figure 11, b). Reducing the area of groundnut
therefore strongly affected farm calorie production.

No significant changes in SOC were observed, as the difference between the SOC in a cowpea field
or in any other field was of 0.01% at most.

a. b.
140 10

Change in calorie production

Change in livestock weight (%)

5
100
0 T1
60 (%) -5 T2
-10 T3
20
-15 T4

-20 -20 T6
0 25 50 75 100 0 25 50 75 100
Cowpea area (%) Cowpea area (%)

Figure 11: Change in livestock weight (a) and calorie production (b) relative to the baseline
scenario, depending on the percentage farm area dedicated to cowpea.

3.3.2.2.2 Implementation of dual-purpose sorghum

3.3.2.2.2.1 Increasing sorghum crop residue availability (interventions 11 and 12)

Increasing sorghum crop residue production was negative for livestock productivity, especially for
type 3 (up to 13.6% loss of cattle weight), as this was the type where the ratio of arable land:
livestock is the greatest, followed by type 2. Type 3 was also the type with the highest proportion
of land under sorghum cultivation. For the other types the decrease was negligible. This is due to
the fact that sorghum residues are of low quality and therefore increasing the amount of sorghum
crop residues decreases overall fodder quality. Farm calorific production was slightly reduced.

3.3.2.2.2.2 Improving sorghum crop residue quality (interventions 13 and 14)

Improving the quality of sorghum crop residues was positive for livestock productivity as the
overall quality of feed was increased. In types 2 and 6 increases of more than 15% of livestock
weight were reached. The effect on livestock productivity was dependent on the area of sorghum
on the farm, on the ratio of farm area: livestock head and on the possibility of the livestock to
reproduce. The greater the ratio of land under sorghum cultivation and the greater the ratio farm
area: livestock head, the greater the increase. When cows are present on farm they can produce
offspring, explaining a greater increase in livestock weight.

The improvement of the crop residue quality led to a small decrease in yields but had no significant
effect on SOC.

3.3.2.3 Drought-tolerant crops (interventions 15 to 18)

Improvement of water conversion efficiency (WCE) was associated with a trade-off between food
and livestock productivity. Improving WCE raised yields slightly (a maximum increase of 2.9% for
type 6 at a WCE increased by 20%). This intervention was negative for livestock, as the amount of

30
millet or sorghum crop residues increased, decreasing overall forage quality in the dry seasons.
The magnitude of the effects depended on the proportion of the farm under cultivation of millet and
sorghum. Therefore the farm calorific production increased most in farm types 5 and 6 for millet
and types 2 and 3 for sorghum. SOC was not affected as there was no change in the amount of
crop residues going to the soil.

3.3.2.4 Fallow (interventions 19 to 22)

Increasing the amount of fallow reduced the size of arable land, decreasing farm calorific
production and the amount of crop residues produced. The percentage decrease in calories
produced was equal to the percentage of land taken out of production.

The fallow interventions did not impact SOC in the fields; whether sorghum was fallowed or not the
SOC at year 12 was of 0.61 and whether cowpea was fallowed or not the SOC at year 12 was of
0.60. This meant that increasing the area of sorghum (crop + fallow) slightly increased SOC while
increasing the amount of cowpea (crop + fallow) slightly decreased average farm SOC.

Increasing the amount of fallow decreased the amount of crop residues available for feed and
livestock productivity.

3.3.2.5 Mulching / crop residue management (interventions 23 to 27)

More crop residues on the field led to negligible higher yields and SOC but lower cow productivity.
The increases in yield and SOC were of no more than 0.1% and 0.4% respectively compared to the
baseline. Animal body weight decreased by 13% in some of the farms (Figure 12), while milk yield
was reduced by more than 50% in farm type 4, due to the reduction of feed available. Cattle
weight changes were similar for all farm types, with the exception of type 6. Livestock weight loss
when no crop residues were fed was less for this type because livestock receive a daily ration of
0.5kg of concentrates.

14
Livestock weight loss (%)

12
10
T1
8
T2
6
T3
4
T4
2
0 T6
0 25 50 75 100
CR left on the field (%)

Figure 12: Percentage cattle weight loss relative to the baseline scenario depending on the
percentage of crop residues left on the field as soil amendment.

3.3.2.6 Manure management (interventions 28 to 37)

The more manure was applied to the fields, the higher the SOC, increasing by 2% compared to the
baseline for type 1. The higher SOC became, the more calories were produced. The increase in
calorie production is related to the ratio of cattle head per land area, as more manure applied per
field led to a higher increase in SOC and calorie production. For this reason the greatest increase
was found for type 4, followed by the types 1, 6, 2 and 3.

Using a plastic sheet to cover the manure heap further improved these results slightly, as carbon
and nutrient losses were reduced. There was no a substantial impact on livestock productivity.

31
 8
produced on farm (%)
Increase in calories

6
T1
4 T2
T3
2
T4
0
T6
0 25 50 75 100
Percentage of produced manure applied to fields

Figure 13: Relative increase of calories (%) produced on farm compared to the baseline scenario,
depending on the application of manure to the fields. Manure stored in an open heap.

3.3.2.7 Physical and chemical improvement of crop residues (interventions 38 and 39):
Improving the quality of crop residues improved livestock weight gain and milk production up to
respectively 520% (in type2) and 642% (in type3), through the increased availability of MP and
ME. More cows live on all farms and the duration of life of the cows is prolonged.

There were no negative effects on food production or SOC, as this intervention had no effect on the
soil (because in the baseline scenario it was defined that manure was not collected and applied to
the fields).

3.4 Interview
3.4.1 Basic farm typology
There was a temporal space between the collection of the data collected for the WLE-Volta project
(June 2015) and the data collected for this thesis (June 2016). During this time changes occurred
leading to households shifting from type (Appendix VIII, Table 1). There were no more farms in
type 6, therefore this type was no longer included in the analysis of the interview answers. The
number of farmers in types 5, 1 and 2 increased. The main cause for changes in farm types was
the loss of livestock (14 out of 39 households experienced livestock losses, moving the household
into other farm types). In four cases the change was due to a gain of cattle and in four cases it was
due to a gain of ruminants. Only in two cases was the change of type due to a shift in land
holding. For full information about household characteristics, refer to Appendix VII, Tables 2 and 3.

3.4.1.1 Primary farm activity

Most farmers considered themselves both crop cultivators and livestock holders (Table 12). Only
one famer interviewed was only livestock holder. This farmer belonged to type 4 and possessed
many small ruminants. The respondent saying the primary farm activity was ‘other’ was involved in
small business. Some farmers in type 1 and 5 (without cattle) consider themselves solely
cultivators (Table 12).

Table 12: Number of farmers considering themselves cultivator, livestock holder, both cultivator
and livestock holder and other.
Farm type
Primary farm activity 1 2 3 4 5
Crop cultivation 3 0 0 0 5
Livestock holding 0 0 0 1 0
Agriculture and livestock 3 8 11 5 2
Other 0 0 0 1 0

32
3.4.1.2 Livestock holding
In almost all cases cattle is owned by the men in the household. Small ruminants are more often
under shared ownership or women ownership (Appendix VIII, Table 4).

3.4.2 Constraints
Capital was a constraint for all respondents. It was on average also the constraint which got the
highest score (Table 13). Livestock was the second most important constraint, getting for all types
a median score of 4 (Table 13). However, for type 5 less than 80% of the respondents considered
livestock as constraining (Figure 14). Technology and education were most often seen as a
constraint by farmers in types 1 and 5, while respondents from type 2 considered technology and
education as constraints the least often. The constraints of education and technology got a score of
4 for types 1, 5 and 3, while types 2 and 4 gave them a score of 3. Cropping land was for all types
a medium to weak constraint and it was present for 60% to 80% of farmers in all types. Grazing
land was less often a constraint and it had the lowest constraint scores for all types. For farm type
5a median constraint strength of 0 was found.

For grazing land, education and technology a significant difference in the proportion of respondents
listing elements as “constraint” or “no constraint” was found (Figure 14). The outcomes of the
statistical tests can be found in Appendix VIII, Table 5.

Cropping land Education

100% 100%
80% 80%
60% 60%
40% 40%
20% 20%
a a a a a ab a ab ab b
0% 0%
T1 T2 T3 T4 T5 T1 T2 T3 T4 T5

Grazing land Technology

100% 100%
80% 80%
60% 60%
40% 40%
20% 20%
a a ab a b ab a ab ab b
0% 0%
T1 T2 T3 T4 T5 T1 T2 T3 T4 T5

Livestock
100%
80%
60%
no
40%
yes
20%
a a a a a
0%
T1 T2 T3 T4 T5
Figure 14: Proportions of ‘yes’ and ‘no’ answers given when asked if cropping land, grazing land,
livestock, education and technology were constraints. Proportions are significantly different if no
letter is in common.
33
 Table 13: Average importance given to the constraints by farmers in the different farm types,
when element is given as a constraint. Reddish and greenish colours indicate higher and lower
percentages of respondents mentioning that the element is a constraint.
Farm type
T1 T2 T3 T4 T5
Constraints
(n=12) (n=16) (n=22) (n=14) (n=14)
Capital 4.3 4.3 4.5 3.4 4.6
Livestock 4.3 3.4 3.8 3.2 3.3
Technology 3.7 2.4 3.6 2.9 4
Education 3.8 2.1 3 2.6 3.8
Cropping land 2.9 1.9 2.8 3.1 2.9
Grazing land 0 0.9 0.7 0.9 0

No significant difference was found between men and women when comparing the proportions of
respondents listing elements as constraints (Appendix VIII, Table 6).

3.4.3 Goals
For all farm types social cohesion was the most important goal (Table 20). This was the case
because respondents worked on communal projects, such as the building of roads. Social cohesion
was also important as it was a premise for living in a community. Next to this the village
inhabitants lived closer to each other than they lived to family and mutual aid was important.
Especially in case of problems it was important to know there was someone who could help.

Sustainable management of natural resources was the next most important goal, explained by the
will to preserve the environment for the following generations and thus to ensure the survival of
the family. To this end, many different management interventions were applied, including tree
planting, zaï practice and the building of stone bunds to reduce erosion.

Food self-sufficiency was very important in all types except type 2 because it enabled feeding the
family and decreased dependence on the market, remittances or other inhabitants of the
community. However, it was also stated that food self-sufficiency was dependent on rainfall and
was thus variable. For most farmers self-sufficiency was not achieved and food had to be bought
for a varying number of months per year (from two to ten months). For type 2 self-sufficiency was
a less important goal as 50% of the farmers produced enough food to survive throughout the year.
This percentage was respectively of 0%, 4%, 7% and 0% for types 1, 3, 4 and 5. Once again type
2, which was considered the wealthiest type was significantly different from the other farm types
for this goal (Appendix VIII, Table 7).

Maximising whole farm production and maximising yield were the next most important goals,
except for type 5. Maximising whole farm production meant that not only yields were maximised,
but also the crop residue production for livestock feed and manure production for amending the soil
were maximised. This goal was relatively important because famers said more crop residues
increased feed availability, increasing livestock productivity and manure production. Some farmers
also sold crop residues for extra income. Sometimes it was however added that food production
was the main aim and overall farm production should not reduce food production. This goal was
less important for type 5 as respondents did not own livestock (64% of the respondents from this
type did not put emphasis on maximising whole farm production).

Maximising yields was important. However, similarly as for maximising whole farm production, the
score given by type 5 was significantly lower than the score given by other types (Appendix VIII,
Table 7). This was due to the fact that 57% of the farmers said not to have enough means to be
able to maximise yields. For types 1, 2, 3 and 4 the percentage of farmers who made the same
statement was respectively of 8%, 13%, 9% and 0%. Maximising yields was important to produce
more food and to be able to sell produce.

34
The goal labour minimization was not interpreted as total labour (hired + household labour). Only
hired labour was taken into account by the respondents. For all farm types the average score for
minimising labour was approximately equal (between 2.6 and 3.1). Farmers of type 1 explained
that land area was relatively small and therefore there was no need for external labour (42%), or
that they did not have enough capital to hire external labour (33%). In type 2 the majority of the
respondents (75%) said labour was hired because of the big amount of work that needed to be
done (timeliness of achieving activities). Type 2 farmers were the most resource endowed because
they owned a lot of livestock and land. In type 3 only 45% of the respondents hired labour. In this
type, 32% of the farmers stated they would like to hire labour but this was impossible due to
capital constraints. In type 4 43% of the respondents said they hire labour, while 36% stated not
to have enough capital to hire labour. Only 14% of the respondents in type 4 declared they did not
need external labour force. For type 5 the scenario is different: 71% of the respondents told they
needed more labour but they did not have enough capital to hire labour. The rest of the
respondents of type 5 said family was big enough to meet the labour requirement.

Table 14: Average score given by the different farm types for the 11 goals.

Type

Average score for goals T1 T2 T3 T4 T5

Ensuring social cohesion 4.9 4.1 4.6 4.4 4.4

Maximising sustainable management

4.2 3.8 3.8 4.0 3.7
of natural resources

Maximising self-sufficiency 4.2 2.3 3.9 3.9 4.1

Maximising whole farm production 4.3 3.5 3.8 4.1 2.1

Maximising yield 4.2 3.3 3.3 3.9 2.4

Minimising labour input 2.7 2.9 3.1 3.1 2.6

Achieving activities in a timely

2.3 2.8 3.0 2.9 3.7
manner

Maximising off-farm income 3.9 3.1 2.6 2.4 2.1

Ensuring land ownership 3.0 1.8 2.7 3.1 2.8

Maximising income 3.4 2.8 2.5 2.4 1.9

Maximising market orientation 3.2 2.7 1.6 2.1 1.6

Achieving activities in a timely manner received a medium score, except for types 3 and 5, who
scored this objective higher. For type 5 this was because family is the main workforce. Type 5 had
the smallest number of household members and was seen as the least endowed farmer type as no
livestock was kept. For these respondents rural exodus was also an important factor limiting labour
availability and thus the timely achievement of activities. Fifty percent of the interviewees of type 3
also said they depend on household labour. Thirty-six percent of the respondents of both type 3
and 5 said they had difficulties finding sufficient labour to perform the required tasks, while for the
types 1, 2 and 4 this proportion was respectively 16%, 25% and 21%. This was due to the fact
that type 1, 2 and 4 had enough labour available within the household and because the area to be
cultivated was small enough to take care of.

Maximising off-farm income was most important for types 1 and 2 and least important for type 5.
Type 1 was significantly different for this goal compared to the other types (Appendix VIII, Table

35
7). For the respondents having a vegetable garden, the selling of products from this vegetable
garden was seen as generating off-farm income. In farm type 5 more than half of the farmers did
not have another job, while as much as 36% of these respondents performed vegetable gardening.
In types 4 and 1 the amount of farmers who did not earn income with off-farm origin was above
49%. In type 1 an often recurring reason for giving a high score was the absence of other sources
of income. In types 2 and 3 less farmers expressed that there was a lack of off-farm opportunities
for gaining money.

Land ownership got a score below 3 for all farm types except 1 and 4. In most farm types the
percentage of land owners was around 55%, the other respondents rented land. This meant that
more than half of the respondents were not afraid their land would be taken from them, as they
were the legitimate owners. For types 1 and 4 the amount of land owners was lower, which might
explain the importance of land ownership for those types. These two farm types were also the
types with the smallest area of land.

Income maximisation was not the most important objective. However, in all types there were
farmers with income generating activities. Type 1 farmers gave an average score of 3.4 to income
maximization. In that type 33% of the farmers generated income either through the growing of
crops for selling or through off-farm activities. In type 2 81% of the respondents declared crops
were cultivated to generate income (such as sesame and groundnut). Thirty-six percent of farmers
in type 3 had activities in the aim of increasing income (livestock raising, vegetable growing). In
type 4 only one farm (7%) was busy with milk production in the dry season with the aim of
increasing income. In type 5 36% of the farmers had activities to increase income: production of
cowpea and groundnut, as well as vegetable production.

Market orientation was on average the least important goal for the respondents, except for type 1,
for whom the average score was of 3.2 (this farm type was significantly different from the other
farm types in this response). In this type 66% of the respondents said they sold products, when
possible at the moment prices were highest. In type 2 80% of the farmers sold products. However,
only 50% of the farmers sold products at the moment when the prices were best. In type 4, 5 and
3 respectively 40%, 14% and 12% of the respondents sold products, when possible. The farmers
who did not sell their produce consumed it, as they were food insecure. Here again a difference
was observed between type 2, the type considered most resource endowed and the other farm
types.

Appendix VIII, Table 7 presents the Fisher exact test outcomes, signalling the presence (or not) of
significant differences in the proportion of answers given by the different farm types.

For neither of the goals there was an association between the gender of the respondent and the
scores given to constraints (Appendix VIII, Table 8).

3.4.4 Attitudes

3.4.4.1 Analysis of attitude

The most striking answers were those to the statements number 1, 3, 8, 11 and 13 where almost
all respondents gave the same answer (Table 15). Of the 77 households only 4 households stated
they were not proud of being farmer, because they felt they have no other choice than being
farmer (statement 1). The feeling of proudness was sometimes said to depend on crop yield. Most
farmers were proud to be farming as it enabled to produce their own food and support the family.
This was also the reason why most farmers perceive farming as a fulfilling activity (statement 3).
All the farmers thought there were problems in the current farming system and there was a need
for new and improved technologies to improve the farming system (statements 8 and 11). Except
for 3 % of the farmers all of them were also open towards new practices and technologies
(statement 13). The justification for not being open towards new interventions is related to the
absence of means to implement new interventions.

36
On the statement that it is a destiny to be farming (statement 2) the answers were mostly in
agreement, because the respondents grew up in agriculture and farming was performed by their
parents. For some participants it was also because they did not go to school. When farmers did not
think it farming is a destiny, they said it was because there always is a choice.

Table 15: Percentage of farmers agreeing, disagreeing and giving a neutral answer to the
statements.
Percentage of farmers giving the answer

Statement Disagreement Neutral Agreement

1. I am proud to be a farmer 4 1 95
2. It is my destiny to be a farmer 13 4 83
3. Farming is fulfilling 4 9 87
4. Farming is a good way to meet my needs and those of my 18 14 68
family
5. There is no better investment than an investment in 26 10 64
farming
6. If I had a choice I would be a (full-time) farmer 43 8 49

7. I would prefer if my children would not end up farmers 22 25 53

8. There is/are no problems in the current farming system 100 0 0
9. The current way of farming functions well enough to cover 48 5 47
my family’s needs now and in the future
10. There is no hope for farmers like us to improve our 66 8 26
standard of living
11. There is a need for new interventions and technologies to 0 0 100
improve the current farming system
12. I seek information on good farming practices from 18 8 74
extension officers / farm groups / other farms / external
sources
13. I am open towards or practice new methods and 3 1 96
interventions
14. I am prepared to give labour and capital to do new 8 3 90
interventions and technologies
15. I want to avoid risk as much as possible 82 8 10

A little more than the majority of the farmers said there is no better investment than farming
(statement 5). This had several different reasons: there are no opportunities except agriculture,
respondents are proud to be farming, investments in agriculture enable maximising production and
because agriculture is the activity the respondents know and master. When it was said that there
are better investments than agriculture this was due to respondents saying they would invest in
something else if there was a choice and stating agriculture does not meet the needs.

Half of the farmers would be farmer if they had the choice (statement 6). The other half would not
be farmer, for several reasons: either they would prefer a job as civil servant, either they would
prefer to raise livestock.

The answers concerning the hope for children to end up farming were variable (statement 7).
Either the respondents said they are farming and therefore it is also good enough for their children.
Some respondents also gave a neutral answer, mentioning that children should choose for
themselves. Lastly some respondent preferred if their children did not end up farming, in the hope
of building up a better life. For this reason some of the children were sent to school. The answers
as to whether farming was a good way to fulfil all needs were also variable (statement 4). Half of
the respondents answered this was the case, as their own food is produced and there is no
dependency on other people. Other respondents stated yields are not high enough and farming is a
tiring activity, which therefore does not fulfil all the needs.

37
The opinions are divided about the ability of the current farming system to cover current and future
needs (statement 9). Positive answers are given thanks to the use of technologies (e.g. stone
bunds) and due to the ownership of land. Negative answers are supported by stating that there is a
need for new interventions and technologies, that climate change and the reduction in soil fertility
will limit agricultural production in the future.

Mostly it was hoped to increase the quality of life through farming (statement 10). This was related
to the fact that respondents relied on new technologies and information to improve yields and soil
quality. The farmers who did not believe farming can increase the quality of life stated yields were
too low and they did not have enough money to progress. The majority of the respondents stated
they seek information about good agricultural practices (statement 12). The negative answers on
information were given by women, who answered it is their husband who seeks information.

The biggest portion of the farmers was prepared to give labour and capital for the implementation
of new technologies and techniques and they are open to taking (moderate) risks if the gains are
high enough (statements 14 and 15).

3.4.4.2 Attitude cluster formation

Clusters of the respondents with the most similar answers given on the attitude statements were
made. Some statements were not taken into account for the making of the cluster, for the
following reasons:

- Statement 12 was omitted because the negative answers came from women, who said only
their husband seeks for information. This meant that there was a search for information
within the household.
- Statement 6 was not taken into account as livestock was seen as an activity different from
farming.
- Statements 13, 14 and 15 were not included as some farmers did not answer the questions
in the way the questions were intended.

The dendogram for the construction of clusters can be seen in Appendix VIII, Figure 1. The chosen
cut-off point was located at a distance of 17.5, leading to the formation of 4 clusters (Table 16).

Farmers from all the clusters were proud to be farming, they thought it was their destiny to be
farming, they felt farming was fulfilling and a good way to fulfil needs and that there was a need
for new interventions to improve the current situation (Table 16). They also all stated there are
problems in the current farming system. Differences between clusters are as follows:

- Cluster 4 had a neutral answer on the statement whether farming is a good investment.
Respondents from cluster 1, 2 and 3 considered farming to be a good investment
- Cluster 2 was neutral as of the future of the children, whether they end-up farming or not,
while clusters 1, 3 and 4 hoped their children do not end up farming.
- The opinions about farming being a good way to suffice needs were divided. Cluster 1 and
2 did think so, while cluster 3 and 4 were more hesitant.
- Respondents from clusters 1 and 4 were neutral about the statement that there is hope to
improve the quality of life. Respondents from clusters 2 and 3 thought there is hope to
improve the quality of life

A significant association between the clusters and the interviewers who conducted the interviews
was found (Appendix VIII, Table 9). There was no significant difference between the interviewers 1
and 2 but both were significantly different from interviewer 3, meaning either the interviewers
themselves influenced the answers through the way of conducting the interviews, or there was a
location effect, as each interviewer conducted interviews in a different village.

38
Table 16: Average score given to the statement accompanied by standard deviation, for the four
clusters (1=disagreement, 2=neutral, 3=agreement).
Cluster
Statement 1 2 3 4
1. Proud to be farmer 3.0 (±0.0) 2.9 (±0.4) 2.9 (±0.5) 2.9 (±0.5)

2. Destiny to be a farmer 2.7 (±0.7) 2.9 (±0.4) 2.7 (±0.7) 2.4 (±0.9)

3. Farming is fulfilling 3.0 (±0.0) 2.9 (±0.3) 2.9 (±0.5) 2.6 (±0.7)

4. Good way to fulfil needs 2.8 (±0.6) 2.5 (±0.8) 2.4 (±0.9) 2.4 (±0.8)

5. Good investment 2.9 (±0.3) 2.4 (±0.9) 2.4 (±0.9) 1.9 (±0.9)

7. I would prefer if my children would 2.6 (±0.6) 1.9 (±0.9) 2.4 (±0.8) 2.5 (±0.8)
not end up farming

8. There are no problems in the 1.0 (±0.0) 1.0 (±0.0) 1.0 (±0.0) 1.0 (±0.0)
current farming system

9. The current way of farming is good 2.7 (±0.7) 2.4 (±0.9) 1.2 (±0.5) 1.6 (±0.9)
enough to suffice my needs

10. There is no hope for farmers like 2.4 (±0.9) 1.4 (±0.9) 1.1 (±0.5) 1.6 (±0.9)
us to improve our quality of life

11.There is a need for new 3.0 (±0.0) 3.0 (±0.0) 3.0 (±0.0) 3.0 (±0.0)
interventions and technologies to
improve the current farming system

3.5 Expert opinion on constraint strengths

Only the interventions ‘leaving crop residues on the field’ had no constraints at all (Table 17).
Cropping land was a constraint only for the intervention fallowing, for which cropping land is taken
out of production. The only interventions for which livestock was constraining was when manure
was applied to the fields. For the planting of native and exotic species, pasture area was necessary
and therefore considered a constraint. Education, technology, capital and labour were constraining
for the following interventions:

- Pasture planted with exotic species,

- Planting of drought tolerant crops,
- Planting of dual-purpose crops,
- Collection and storage of crop residues as animal feed,
- Application of manure on the fields,
- Physical treatment of crop residues,
- Chemical treatment of crop residues

Next to this, education, technology and labour were constraining for the planting of the pasture
with native species. Labour was the major constraint for the collection of crop residues livestock
feed.

39
Table 17: Expert opinion on the strength (0=weak, 1=medium and 2=strong) of constraints for the
implementation of interventions.
Constraining element
Intervention Cropping land Pasture Livestock Education Technology Capital Labour
Pasture planted with native
0 2 0 1 1 0 1
species
Pasture planted with exotic
0 2 0 2 2 2 1
species
Planting of improved dual-
0 0 0 1 1 1 1
purpose cowpea and sorghum
Planting of drought-tolerant
0 0 0 1 1 1 1
millet and sorghum
Leaving 10-20% of the land
2 0 0 0 0 0 0
fallow for 5 years
Collection of herbaceous
0 0 0 0 0 0 1
biomass for animal feed
Collection and storage crop
0 0 0 1 1 1 2
residues for animal feed
Leaving of crop residues on the
0 0 0 0 0 0 0
field
Application of manure 0 0 1 2 1 1 2
Physical treatment of crop
residues for improved 0 0 0 1 2 1 2
palatability
Chemical treatment of crop
residues for improved 0 0 0 2 2 1 1
palatability

3.6 Matching expert opinion to farm constraints

When matching expert opinion (Table 17) to farmer-reported constraints (Table 13), similarities
and differences were seen between types (Tables 18 to 22).

The collection of herbaceous biomass for animal feed and leaving crop residues on the field did not
pose a constraint for any farm type, as there was no constraint to the implementation of these
interventions following the expert opinion (labour was not taken into account).

For the planting of pastures with native species, education and technology were constraints in all
the farm types. All constraints were of medium strength, except for type 5 which was strongly
constrained in technology. When exotic species were planted the same constraints played a role,
with the addition of capital and all constraints, for all types, were strong. The same was true for
the chemical treatment of crop residues, except for the medium constraint of capital in type 4. The
results for the physical treatment of crop residue were similar to the results for the chemical
treatment of crop residues, minus the reduced strength for the education constraint, which was
medium for all types.

For the planting of dual-purpose crops, drought tolerant crops and for the collection and storage of
crop residues as animal feed the same three constraints were present. For most types education
and technology were medium constraints, while capital was strongly constraining. This was not the
case for type 4 for which capital was only mildly constraining. For type 5 technology was also
strongly constraining.

The only constraint for all types for the implementation of fallowing was cropping land. Depending
on the strength of the constraint in the different types the final constraint was either medium or
strong (medium for type 2, which has a lot of land and strong for all the other types, which own
less land).

40
Table 18: Resulting constraint scores from the matching of the expert evaluation with the
constraints of type 1, for the different interventions. Green boxes show there is no constraint,
orange boxes show a medium constraint and red boxes show a strong constraint. The higher the
score the more important the constraint.
Constraining element
Intervention Cropping
Pasture Livestock Education Technology Capital
land
Pasture planted with native
0 0 0 0.38 0.38 0
species
Pasture planted with exotic
0 0 0 0.75 0.75 1
species
Planting of improved dual-
0 0 0 0.38 0.38 0.5
purpose cowpea and sorghum
Planting of drought-tolerant
0 0 0 0.38 0.38 0.5
millet and sorghum
Leaving 10-20% of the land
0.5 0 0 0 0 0
fallow for 5 years
Collection of herbaceous
0 0 0 0 0 0
biomass for animal feed
Collection and storage crop
0 0 0 0.38 0.38 0.5
residues for animal feed
Leaving of crop residues on the
0 0 0 0 0 0
field
Spreading of manure on fields 0 0 0.5 0.75 0.38 0.5
Physical treatment of crop
residues for improved 0 0 0 0.38 0.75 0.5
palatability
Chemical treatment of crop
residues for improved 0 0 0 0.75 0.75 0.5
palatability

The application of manure on the field required many elements: livestock, education, technology
and capital. The education constraint was always strong, as this was defined as a strong constraint
for the intervention by the experts and all the farm types were in greater or lesser extent limited
by education. Type 1 was the only type strongly constrained by livestock. For manure application
type 5 was strongly constrained by technology while for the other types this was a medium
constraint. All types were strongly constrained by capital, except for type 4, for which it was a
medium constraint.

41
Table 19: Resulting constraint scores from the matching of the expert evaluation with the constraints
of type 2, for the different interventions. Green boxes show there is no constraint, orange boxes show
a medium constraint and red boxes show a strong constraint. The higher the score the more
important the constraint.
Constraining element

Intervention Cropping
Pasture Livestock Education Technology Capital
land

Pasture planted with native species 0 0 0 0.5 0.25 0

Pasture planted with exotic species 0 0 0 0.5 0.5 1
Planting of improved dual-purpose
cowpea and sorghum 0 0 0 0.25 0.25 0.5

Planting of drought-tolerant millet and

0 0 0 0.25 0.25 0.5
sorghum
Leaving 10-20% of the land fallow for
0.25 0 0 0 0 0
5 years
Collection of herbaceous biomass for
0 0 0 0 0 0
animal feed
Collection herbaceous biomass for
0 0 0 0 0 0
selling
Collection and storage crop residues
0 0 0 0.25 0.25 0.5
for animal feed
Spreading of manure on fields 0 0 0.38 0.5 0.25 0.5
Physical treatment of crop residues for
0 0 0 0.25 0.5 0.5
improved palatability
Chemical treatment of crop residues
0 0 0 0.5 0.5 0.5
for improved palatability

Table 20: Resulting constraint scores from the matching of the expert evaluation with the constraints
of type 3, for the different interventions. Green boxes show there is no constraint, orange boxes show
a medium constraint and red boxes show a strong constraint. The higher the score the more important
the constraint.
Constraining element

Intervention Cropping
Pasture Livestock Education Technology Capital
land
Pasture planted with native species 0 0 0 0.38 0.38 0
Pasture planted with exotic species 0 0 0 0.75 0.75 1
Planting of improved dual-purpose
0 0 0 0.38 0.38 0.5
cowpea and sorghum
Planting of drought-tolerant millet and
0 0 0 0.38 0.38 0.5
sorghum
Leaving 10-20% of the land fallow for
0.5 0 0 0 0 0
5 years
Collection of herbaceous biomass for
0 0 0 0 0 0
animal feed
Collection and storage crop residues
0 0 0 0.38 0.38 0.5
for animal feed
Leaving of crop residues on the field 0 0 0 0 0 0
Spreading of manure on fields 0 0 0.38 0.75 0.38 0.5
Physical treatment of crop residues
0 0 0 0.38 0.75 0.5
for improved palatability
Chemical treatment of crop residues
0 0 0 0.75 0.75 0.5
for improved palatability

42
Table 21: Resulting constraint scores from the matching of the expert evaluation with the constraints
of type 4, for the different interventions. Green boxes show there is no constraint, orange boxes show
a medium constraint and red boxes show a strong constraint. The higher the score the more
important the constraint.
Constraining element

Intervention
Cropping
Pasture Livestock Education Technology Capital
land

Pasture planted with native species 0 0 0 0.25 0.25 0

Pasture planted with exotic species 0 0 0 0.5 0.5 0.75
Planting of improved dual-purpose
0 0 0 0.25 0.25 0.38
cowpea and sorghum
Planting of drought-tolerant millet and
0 0 0 0.25 0.25 0.38
sorghum
Leaving 10-20% of the land fallow for
5 years 0.75 0 0 0 0 0
Collection of herbaceous biomass for
0 0 0 0 0 0
animal feed
Collection and storage crop residues
0 0 0 0.25 0.25 0.38
for animal feed
Leaving of crop residues on the field 0 0 0 0 0 0
Spreading of manure on fields 0 0 0.38 0.5 0.25 0.38
Physical treatment of crop residues for
0 0 0 0.25 0.5 0.38
improved palatability
Chemical treatment of crop residues
0 0 0 0.5 0.5 0.38
for improved palatability

Table 22: Resulting constraint scores from the matching of the expert evaluation with the constraints
of type 5, for the different interventions. Green boxes show there is no constraint, orange boxes show
a medium constraint and red boxes show a strong constraint. The higher the score the more important
the constraint.
Constraining element

Intervention Cropping
Pasture Livestock Education Technology Capital
land
Pasture planted with native species 0 0 0 0.38 0.5 0
Pasture planted with exotic species 0 0 0 0.75 1 1
Planting of improved dual-purpose
0 0 0 0.38 0.5 0.5
cowpea and sorghum
Planting of drought-tolerant millet and
0 0 0 0.38 0.5 0.5
sorghum
Leaving 10-20% of the land fallow for
0.5 0 0 0 0 0
5 years
Collection of herbaceous biomass for
0 0 0 0 0 0
animal feed
Collection and storage crop residues
0 0 0 0.38 0.5 0.5
for animal feed
Leaving of crop residues on the field 0 0 0 0 0 0
Spreading of manure on fields 0 0 0.38 0.75 0.5 0.5
Physical treatment of crop residues for
0 0 0 0.38 1 0.5
improved palatability
Chemical treatment of crop residues
0 0 0 0.75 1 0.5
for improved palatability

43
3.7 Decision tool
For the selection of the appropriate intervention per farm type, information of different parts of this
thesis should be assembled. Take the following steps (Figure 15):

1. Determine the aim of the project

2. Determine the farm type for a particular household (Figure 5)
3. Use Tables 18 to 22 (depending on the farm type) to determine which interventions are
possible without encountering any constraint, which interventions are possible when
alleviating medium constraints and which interventions are possible when alleviating one or
several strong constraints.
4. Assess the potential impacts of the intervention for the specific farm type, using Appendix
VII, tables 1 and 2.
5. Determine which interventions are applicable, weighing the potential impact on various
indicators, against the constraint limiting their adoption.
6. Check the intervention supports the goals of the farmer (Table 14). If this is not the cases
repeat the steps 5 and 6

2.
Classify farm in
resource-based
farm type
(THESIS, Figure
5)

3.
1. Identify
Determine aim of constraints
the project present
(USER) (THESIS, Tables
18-22)

6.
Consider if farm 4.
goals fit with Assess potential
intervention impact of
outcomes. If not, interventions for
return to step 5 the specific farm
and find another type (THESIS,
possibility. Appendix VII,
(THESIS, Table tables 1 and 2)
14)
5.
Choose the best
intervention
(USER,
considering steps
1, 3 and 5)

Figure 15: Flow chart of the steps in the decision process. User mean the step should be
taken by the user, thesis means the information is available in the thesis.

44
3.8 Example
In the following paragraph an example is given on how the decision tool could be used. The steps
are taken in the order given in Figure 15.

1. The project aim is to increase calorific production.

2. One household is considered, with 5ha but no cattle. Using Figure 5 it is determined that
this household belongs to Type 1.
3. Refer to table 18 (as the household belongs to type 1) to see which interventions bring
along which constraints.
4. Using Appendix VII and Table 1 it was found that interventions the most beneficial for the
improvement of calorific production are applying manure followed by decreasing cowpea
area (reducing the area of dual-purpose crop) and by increasing water conversion
efficiency of crops (taking drought-tolerant crops into use).
5. Spreading manure is strongly constrained by livestock, education and capital and the
medium constraint of technology is also present.
Using drought-tolerant crops is slightly negative for livestock production. It is strongly
constrained by capital and medium constrained by education and technology.
Reducing cowpea area is negative for livestock production and it has the same constraints
as the implementation of drought-tolerant crops.
With this information the choice is made to go for using drought tolerant crops: there are
fewer constraints than for the implementation of manure spreading and the impact on
calorie production is greater than for the reduction of cowpea area, it also has less negative
impact on livestock production.
6. This intervention is mostly in line with farmer goals: it will increase food self-sufficiency
and maximise yield. Crop residue production is also increased.

45
4 Discussion
4.1 Setting of baseline scenario
4.1.1 Crop
For this exercise only one farm type grew maize. This has two main causes. Firstly maize is not a
major crop in the Ouahigouya region (Funk et al., 2012). Secondly the cut-off points chosen for
including a crop in a representative farm played a role: a crop was taken up in the simulation if
more than 50% of farmers grew the crop. If a lower cut-off point would have been chosen maize
would have been taken up in the simulation exercise more frequently. The area over which maize is
grown was on average quite small (over all farms growing maize the average area was of 0.6ha) so
that including maize would have only a small impact on the overall farm outcomes.

The area allocated to cereals on the farms is larger than the area allocated to legumes. This is
among others due to the fact that women are largely responsible for the cultivation of legumes
(Nhamo et al., 2003), traditionally women’s crops. Women are most often only allowed to perform
agriculture on land which is owned by the husband but not used in that particular season, as they
have to get access to productive resources in accordance with their husband (Kevane & Gray,
1999). Next to this the calorie yield of legumes is also smaller for legumes as for cereals.

4.1.2 Livestock
When only a small number of cattle are owned, the herd is predominantly male, while the fraction
of females increases with herd size. This is due to the fact that the first need is for ploughing and
draft power. Once this need is satisfied farmers will invest in cows for milk production and the
provision of calves. In Kenya a similar relation was found between the function of cattle and the
structure of the herd: when cattle was important for providing draft power keeping male animals
was more important, while for the production of milk the fraction of females in the herd was
increased (Rege et al., 2001).

4.1.3 Fertilizer
Fertilizer and manure use was low due to the lack of capital at the start of the growing season (S.
Coulibaly, personal communication). The limited use of manure is partly due to the labour needs
for collection, transport, storage and application. Transport is limited as manure is bulky and
transportation equipment is required. Manure use is also limited because quantity and quality are
often low (Bayu et al., 2005), primarily due to feed limitation. Manure might also have different
uses within the household, such as fuel (Bayu et al., 2005). From the interview it is however clear
those farmers recognise the importance of manure in their farming system and they voice their will
to use it. Thus far it is mainly restricted due to transport and handling costs.

4.1.4 Crop residues

For the simulation purposes it was assumed that 100% of the crop residues were eaten by
livestock. This is however an overstatement as not all residues are edible (low feeding values of
certain parts) (McDowell, 1988). It is therefore possible that a small amount of crop residues stay
on the fields, impacting soil fertility. Crop residues also have many alternative functions (fuel,
construction material, etc.), leading to a pressure on this biomass resource (Powell et al., 2004;
Valbuena et al., 2015). The farmer’s choice of crop residue use is dependent on his or her
preference, total crop residue production, the availability of other biomass sources and on the
demand for crop residues (Valbuena et al., 2015). This implies that in reality less than 100% of the
produced crop residues are available for the livestock. In the simulation exercise this would lead to
a reduction in livestock productivity (taking away 10% of crop residues is similar to the 10% fallow
situation, which led to a reduction in livestock productivity).

46
4.2 Outcomes baseline scenario
Calorie production, SOC and livestock productivity decrease over time. SOC decreases due to the
fact that there is no organic matter input into the soil (crops and residues are taken away, no
manure is applied), thus there is more organic matter decomposition than input. This decrease in
SOC is followed by a decrease in yields. This is in general the case in sub-Saharan Africa, where
low soil fertility is a major factor in the declining per-capita food production (Bayu et al., 2004).
The final value of SOC (year 12 of simulation) is still in the range of values found in Burkina Faso
(Beal et al., 2015). Livestock productivity decreases due to the low quality of feed available.

4.2.1 Yields
At the start of the simulation exercise the yields are in the higher part of yields currently found in
the area. It is possible that the simulation outcomes overestimate yields, as for example diseases
and pests are not taken into account (Marinus et al., 2015). In the last simulation year the yields
are in the lower range of the yields currently found in sub-Saharan Africa. A big jump in yields is
observed between the first and second year. This was due to the fact that there was strong water
shortage in years 2 and 3 (years of drought) (Nagothu, 2016).

4.2.2 Food self-sufficiency

It was expected some households would not be self-sufficient, as there is a great number of people
undernourished in Burkina Faso (around 4 million in 2016, an amount that has increased since
1990) (FAO, 2015). It is assumed all the crops produced are eaten, while this is not always the
case in reality, as part of the harvest may be sold (e.g. to cover expenses for health). This means
that the amount of food eaten might be lower, decreasing the amount of calories available and
reducing food self-sufficiency. At the same time households might buy external food, increasing
calorie intake.

Livestock products were not included in the calculation of food self-sufficiency. Including these
products would not have had a great effect, as livestock productivity was low. On average the
consumption of livestock products in Burkina Faso is low and meat and milk are only rarely eaten
(Lykke et al., 2002). Even in case of shocks, such as crop failure, livestock products are not
consumed (Kazianga & Udry, 2006), rather livestock is sold to buy staple foods (Binswanger &
McIntire, 1987; McDermott et al., 2010).

4.2.3 Livestock production

Livestock weight gain is related to quantity and quality of available feed (Powell et al., 1996). In
sub-Saharan Africa feed shortage is one of the most important constraints limiting milk and meat
production (Sanogo, 2011; Bayu et al., 2005) as traditionally the pastures comprise the main
source of animal feed (Sanogo, 2011; Sanon et al., 2007; Bayu et al., 2005). Crop residues are
also important, especially in the dry season when the nutritional value of rangelands decreases
(Rattunde, 1998). Weight is thus gained during the wet season and lost during the dry season.

From the model outcomes it was found that livestock lose weight during all the seasons. As in all
seasons more biomass is offered than is eaten, meaning that feed quality and not quantity was
limiting. Livestock weight loss was lowers during the wet season. In this season only the basic
amount of feed is fed (grass + roughage) which is then of the highest quality (Table 23). In the
early dry season grass quality decreases, reducing the quality of the basic amount of feed given
(Table 23). During the late dry season this decrease is even stronger. During these dry seasons
also crop residues are fed. The quality of these residues is also below the quality of the basic
amount of feed given during the wet season (Table 24). This means that the overall feed quality
(basic feed + crop residues) is lower than feed given during the wet season. This explains the
stronger decrease in livestock productivity in the dry seasons compared to the wet season.

47
 Table 23: Quality during the different seasons of the mixture
of grass and roughage given throughout the year.
Season CP (g/kg DM) ME (MJ/kg DM)
Wet season 92.20 8.14
Early dry
78.60 7.50
season
Late dry
54.58 6.62
season

Table 24: Average over 12 years of crop residue quality

(crude protein and metabolisable energy).
Type CP (g/kg DM) ME (MJ/kg DM)

T1 92.89 7.83
T2 82.06 7.52
T3 85.32 7.58
T4 93.91 7.86
T5 76.02 7.34
T6 76.51 7.36

As in general livestock gain weight during the wet season in Burkina Faso, the quality of the feed
given during this season should have been increased to get more realistic results. This could have
been done through increasing the quality of the fodder or through increasing the ratio of grass fed
compared to the ratio of roughage given.

4.3 Impact of interventions

Cattle weight and milk production can be greatly increased through the interventions but none of
the interventions assessed had a big effect on SOC or crop yields. This is because of the sensitivity
of livestock to feed changes. The soil is less sensitive to small changes because it is a more static
component. Yields are also less sensitive for changes than livestock, mainly due to the fact that
yield depends on soil quality (Marinus et al., 2015).

The trends of the results were similar for all farm types, due to the fact that the soil and rainfall
conditions are the same. The magnitude of the impacts of the interventions is different between
the farm types, depending on crops grown, herd size, the area of land and the ratio between the
land area and the number of cattle head. A smaller area of land leads to more cows per land area
and the fodder quality will depend less on the crop residues and more on the basic amount of feed
given. More cows per land area also means that more manure is produced per area land and thus
the impact of manure collection is bigger.

Of the list of interventions assessed there are no interventions which promote all objectives (food
production, SOC and livestock productivity) at the same time. Sometimes the interventions
promoted only one or two of the three goals (e.g. increase of forage quality) and sometimes a
trade-off was found between the optimization of different objectives.

The increase of feed quality, the physical or chemical treatment of crop residues, the improvement
of livestock forage (quality and quantity) and the implementation of dual-purpose sorghum through
increased residue quality all only affect livestock production. This is due to the fact that in the
baseline scenario manure application was set to 0. This means that there is no return of nutrients
from the crop residues back to the fields in the form of manure. Allowing the collection and
application of manure to the fields increased yields. For future research it would be interesting to
see what the effect is on SOC and calorie production of the previously mentioned interventions
when manure collection and application happens. Livestock production is increased for all these
interventions as fodder quality is improved.

48
Manure application was positive not only for calorie production, it also increased SOC. However, it
had no significant impact on livestock as fodder quality is only slightly impacted (the increase in
the amount of crop residues produced on per hectare is of 180kg for sorghum and 200kg for
cowpea, negligible quantities compared to the total amount of crop residues fed and the small
amount of basic feed given).

A trade-off is found for the establishment of dual-purpose crops and the establishment of drought-
tolerant crops. The implementation of dual-purpose cowpea reduced calorific yield as cowpea yield
is lower than the other crops and calorific value was lower than groundnut, millet and maize. It
however improved livestock productivity as cowpea residues are of a better quality than residues of
sorghum and millet, and grass in the dry seasons. Increasing the ratio of cowpea residues in
livestock feed therefore increase fodder quality, the limiting factor for livestock productivity. The
uptake of drought tolerant sorghum and millet of the uptake of dual-purpose sorghum with an
increased crop residue production both led to a greater availability of sorghum (and millet) crop
residues. This residue is of lower quality of residue from cowpea or groundnut, therefore
decreasing overall fodder quality in the dry seasons, and decreasing livestock productivity. These
interventions however improved calorie production, as grain yields were increased. When
implementing these interventions the farmer has to make a choice in improving livestock
productivity or yield.

Fallow is the only intervention which was not advantageous for any objective. Calorie production
decreased proportionally to the amount of land not taken into production. Fallowing also did not
improve the SOC, due to the low qantity of residues applied and due to the high decomposition
rate. Livestock productivity is also decreased, as less crop residues are available during the dry
seasons, reducing forage quality in these seasons (refer to table 23 and 24 to compare the quality
of the basic feed given and the quality of crop residues produced on farm).

4.4 Modelling considerations

4.4.1 Weather and soil data
The weather data used was for the years 1981-1992, which was the most complete data-set
available. This is however already 24 years ago and the overall rainfall pattern may have slightly
changed. For example in the years 1931-1960 and 1968-1990 there was a decline in rainfall
(Nagothu, 2016). The data also encompasses the years 1983-1984 and 1991-1992, which were
years of widespread droughts (Nagothu, 2016) and may therefore not be representative for
average years. The effect of the drought in the years 1983-1984 can be seen in the yields of the
baseline scenario (Figure 6).

The available soil data for the region was variable, with for example SOC reaching from 0.5% (A.
Ayantunde, personal communication, soil in the proximity of Ouahigouya, in the villages Sabouna
and Ziga) to 2% (Leenaars et al., 2014). In this thesis a SOC of 0.8% was chosen as starting
situation. Higher SOC leads to higher yields, as yields are strongly related to soil quality. For
example taking an SOC of 1% as a starting condition for the simulation exercise would have led to
yields higher during the whole simulation period (Figure 17).

Modelling has many advantages in this research, as the use of the model has enabled the mixing of
‘generalised science based knowledge’ to ‘local specific data’ (van Paassen et al., 2011), to help in
the choice of more locally-adapted interventions. Exploratory modelling is a good tool for assessing
the effect of different interventions on different farm types. However, the model does not take into
account socio-economic circumstances and traditions. Therefore the possibility for implementing
the different interventions still has to be checked (labour shortage, input availability, etc.). It also
has to be reviewed whether the interventions fit the knowledge of farmers and their values (van
Paassen et al., 2011). The use of a model also has limitations, in the sense that the model cannot
contain all the details of the farming system and that approximations are sometimes needed (e.g.
fallow modelled through growing a crop not producing grain that is entirely left on the field). The
starting conditions used for the model are also approximations, as they are not well known. The

49
accuracy of the model is also reduced through several set rules, such as the obligatory selling of a
bull after 6 years of work and the inability to buy another bull before that time.

Sorghum Cowpea
3000 3000
2500 2500
Yield (kg/ha)

Yield (kg/ha)
2000 2000
1500 1500
1.0% SOC
1000 1000
500 500 0.8% SOC
0 0
1 3 5 7 9 11 1 3 5 7 9 11
a Year b Year

Figure 14: Yields of sorghum (a) and cowpea (b) when the initial SOC is set at 1.0% and 0.8%.

As the model was used to perform an exploratory study, it is interesting to consider how realistic
the interventions are. It is possible for the forage quality of crop residues to be improved with
20%? Would a farmer convert 100% of the farm area to cowpea if this increases livestock
productivity? Is it possible for farmers to collect and apply 100% of the produced manure to the
fields? The answer to these questions would also impact the adoptability of the interventions for
farmers. Another interesting question is how well the interventions were simulated. There is for
example a difference between a real fallow and growing a (grain less) crop that which is left on the
field.

4.4.2 Modelling limitations

The way the livestock module was configured was restrictive for the exploratory modelling, as it
was impossible for the farms to buy extra bulls. In reality there is a chance that livestock would be
bought, as it is a saving which can be sold in time of need (Herrero et al., 2003). For one farm type
the number of cattle had to be reduced considerably (from 33 to 20 head) to enable the running of
the model. This greatly influences the results, as livestock is one of the outcomes monitored.
Increasing the number of cattle is important for the production of manure but also reduces the
amount of crop residues available per animal. Next to this livestock have many other functions,
such as insurance in case of crop failure (Herrero et al., 2003). These aspects are not (yet)
captured by the model.

Sesame, one of the crops which accordingly to the set rules should have been taken into account in
the simulation exercise could not be included because sesame cannot yet be simulated within
FARMSIM. This crop could be developed to enable the inclusion in the model, to make the
simulation more realistic.

It would have been interesting to test the interventions with the farmers or to simulate the
interventions in the presence of farmers, to see what their opinions and ideas are. The knowledge
of the field agents could also have been further exploited, as they have a more concrete knowledge
about the (im)possibilities for farmers in the region. It was however not done due to time limitation
and because it was beyond the scope of this thesis. This would be an interesting field of research
for the future.

50
4.5 Interview
4.5.1 Basic farm characteristics
Trading and death of livestock led to households changing types. This shows that a typology which
is static on paper is dynamic in real life. The change from one farm type to another occurs rapidly
(e.g. disease causing death of livestock) and can have big impacts on the farm strategies and
management. This might be an indication that livestock holding was not the best indicator for
making a typology, although it is an important farm asset and a source of farm heterogeneity.

The result about how a farmer sees himself (e.g. livestock holder or crop cultivator) gives an
insight in where the farmer puts the focus in the farming system and which are the most important
key elements in the farming system. This may be important in determining which interventions
interest the farmer.

Livestock are mostly owned by men. This should be taken into account when promoting
interventions. If this tools is to be used for different regions, it should be checked whether livestock
ownership is similar, to promote the best interventions (Oladele & Mankhei, 2008).

4.5.2 Constraints related to type

Resource endowment plays a role in the setting of many constraints. For this reason farm types
should be taken into account in the formulation of interventions. Labour should also be considered
as a potential constraint, although it was not part of this thesis. The livestock constraint was
interpreted differently between the farmers, leading to the need to clarifying the notion of livestock
constraint.

The capital constraint is the strongest constraint, for all farm types. As stated before this
corresponds to the lack of capital to buy farming inputs at the start of the growing season and the
low rate of fertilizer and pesticide use in Burkina Faso. It would be interesting to see whether the
answer to the question if capital is constraining would be different if asked in another period, for
example just after harvest.

The grazing land constraint is low as there is at this moment still enough space for cattle to graze.
Similarly, cropping land can also still be expanded. However, population increase and urbanization
will limit the options for expanding cropping and grazing land in the future (Asadu et al., 2008) and
land will become a more pressing constraint. This should be taken into account in future studies.

The similarity in results for technology and education can be related to the fact that technology and
education (in the form of on-farm management) are often brought through the same channels (the
same project will educate and bring the resources for field trials, etc.). It is also possible that
higher education stimulates farmers to try new technologies, as education is found to have a
positive effect on the adoption of interventions (Strauss et al., 1991). At the same time higher
resource endowment was found to reduce education and technology constraints. This might be due
to the fact that resource endowment enables the purchase of required technological inputs
(Vanlauwe et al., 2010).

4.5.3 Goals related to type

Resource endowment plays a role in the determination of most goals. The farm types are therefore
important in choosing the right interventions.

When promoting an intervention the social traditions and culture should be considered as social
cohesion is important. An example of this is the construction of a vegetable garden, an activity
which is traditionally conducted by women. It would require a big change for males to grow a
vegetable garden, even if the sale of vegetables could potentially be a good way to gain capital (S.
Coulibaly, personal communication).

51
As sustainable management is important, interventions promoting this have a greater chance of
being adopted. However, there are often constraints for the implementation of these interventions.
For example many farmers state they want to use manure to improve soil fertility. However, due to
the labour need for the collection and spreading of manure this is not often done (S. Coulibaly,
personal communication). Even when a compost heap is used the quality and the quantity of the
compost is often not enough to improve soil quality (Bayu et al., 2005).

The importance of maximising yields, whole farm production and land ownership was dependent on
farm type. Resource endowment should therefore be taken into account when promoting and
implementing interventions with those aims.

Minimising labour and achieving activities in a timely manner was of medium important for all
types, emphasizing the need to take these goals into account when deciding of interventions.
Maximising off-farm income is in general less important for all types, signalling that most effort
goes into the farm. This is a positive signal for the chance of adoption of interventions, as it means
there is a greater chance farmers will put effort in the farming system, instead of putting energy in
maximising off-farm opportunities. One farm type is different as it puts more emphasis on off-farm
income (type 1). This might be the case because this farm type is searching for alternative sources
for income. This can either make the respondents less likely to take-up new interventions (as they
are less dependent on agriculture for their income) or this could stimulate respondents from this
type to take-up new interventions to be able to rely more on farming.

Maximising income and market orientation are the least important goals for less resource endowed
farms. Interventions aiming to increase income and market orientation will probably less easily be
adopted by these farms. This means that once again farm resource endowment will play a role in
the setting of household goals, influencing interventions adoption.

4.5.4 Constraints and goals related to gender

For both constraints and goals no association was found between gender and the constraints and
goals mentioned. This could mean that the effect of resource endowment (type) is stronger than
the effect of gender. This result could also emanate from the way the interviews were conducted.
For example is the interview is held at the same time for both household members or one of the
household members is present while the other gives the answers the answers given could have
been influenced.

It was expected that constraints would have been different between men and women, as for
example the only land to which women have access is the land, owned by the husband, which is
not cultivated (Kevane & Gray, 1999). To improve the view on the constraints present it is also
possible to focus the constraint questions more on the respondents rather than on the farm.

4.5.5 Attitude
The results on attitude indicate the need for clearer statements, with a more straightforward goal
as it was not possible to create typical farm attitude groups. It is positive that respondents are
proud to be farming and that they are open towards new interventions as this might facilitate the
spreading of new techniques. However, most farmers hope their children will not end up farming.
Before promoting interventions, information is needed on why farmers do not want their children to
end up farming, as this might shape the future of farming in the region.

Especially in this part of the interview, where personal questions were asked, it is possible that
respondents tried to give the ‘appropriate’ answer. The field agent conducting the interview also
has a great effect on results through the way the questions are asked and the answers are
reported. This could have led to the association between field agent and attitude cluster. It is
however also possible that farming attitude cluster was dependent on the village (e.g. differences
in education, tradition and customs).

52
For future research the attitude statements should be made clearer, to reduce ambiguity and
collect more reliable data (e.g. define what is meant with ‘investment’ when asking if farming is a
good investment). It is also possible to develop the ‘attitude’ part of the thesis in another way. An
example is the measuring of the attitude towards risk as the aversion to risk leads to diversification
and risk can be either restricting or favouring new interventions (Bidogeza et al., 2009; Kebede et
al., 1990). For example, Kebede et al. (1990) found that risk aversion is usually negative for the
adoption of new interventions.

4.6 Considerations interviews

4.6.1 Overall considerations
The use of a Likert scale posed advantages and constraints. Farmers easily understood the
answering in terms of “agreement” and “disagreement”, which allowed the surveys to go relatively
quickly. However, the answers were probably influenced by the way the questions were asked, as it
is more obvious to answer in agreement. Farmers had more difficulties in discerning the gradient
between “disagreement” and “strong disagreement”. For this reason those answers were combined
in the final data analysis. It is also possible that respondents avoid choosing the “extreme”
answers, distorting the results. Using a Likert scale was a good way to obtain quantifiable data on
qualitative subjects, directly from the farmer, also enabling statistical analysis. Many different ideas
however exist on the right methods for the statistical analyses (Clason, 1994). In this thesis the
average was calculated for constraint and goals, while this is not advised as the use of a Likert-
type scale gives rise to ‘greater than’ relationships, with the amount of ‘greater than’ not being
expressed. This makes the data ordinal and leads to the requirement of the calculation of medians
and not averages (Ott & Longnecker, 2015; Boone & Boone, 2012).

When performing interviews many flaws about this way of collecting data were discovered. There
were difficulties in translation (between English, French and the local language) and the
understanding of the interview questions was different between the enumerators. The notation and
interpretation of results was also different per enumerator, making the quality of the data
dependent on the ability of the enumerator. These problems probably also occurred during the data
collection for the CCAFS and WLE projects. Using interviews for data collection was however
interesting as it was possible to capture verbal and non-verbal communication and it was also
possible to go into depth into certain answers to get all the information required.

The sample size of the farmers was sufficient to get an idea of the situation, although it is unknown
which fraction of the population has been interviewed in each village.

Both paradigms for intervention adoption are used in this thesis (economic and innovation
diffusion, refer to paragraph 1.5). The importance of the innovation diffusion paradigm was often
stressed during the interviews, when farmers provided an explanation when asked if and how they
sought for information (when asked if education was a constraint). This ranged from: “I am always
looking for information and new technologies” over “I only implement interventions when promoted
by field agents” to “I only adopt an intervention when it has proven to be worthwhile for my
neighbors”. In the future it might be an interesting option to ask where respondents get their
information from, as this might influence the adoption and spread of interventions.

4.6.2 Considerations constraints

In this thesis the labour availability constraint was not taken into account when conducting the
interventions. This is an important gap as the shortage of labour can be an important factor in the
(non)adoption of interventions: it can be determinant in the non-adoption of labour requiring
interventions but it may also promote the adoption of labour-saving technologies (Bidogeza, 2009).
Labour is also one of the three main farming assets, together with land and capital (van Vliet et al.,
2015). Labour was also considered an important constraint by experts.

53
In this thesis the constraints are seen as ‘fixed’ and reducing interventions adoption. It is not
considered how easily constraints can be alleviated. If the ‘strong constraint’ can be easily
removed, the implementation of the intervention becomes easier.

4.6.3 Considerations goals

Farming goals are not fixed and can evolve over time. This could be taken into account in further
studies. For example, several cases have been found where an increase in population pressure led
to the increase in soil maintenance practices (Asadu et al., 2008). This could mean that in the
future there could be more demand for implementing soil maintenance interventions. Next to this,
the trend in farm development in sub-Saharan Africa is expected to be different from many in
developed regions, where the number of farms decreased and the area expanded. It is expected
that smallholders will remain dominant, with a larger number of farmers and a decrease in farm
area (McDermott et al., 2010; van Vliet et al., 2015). Therefore interventions promoting
intensification have a chance of becoming more important for farmers in sub-Saharan Africa in the
future.

4.7 Decision tool

The final decision tool is a very broad tool, which can lead to many outcomes depending on the aim
of the project and on the emphasis put by the user on farming constraints, intervention outcomes
and farmer goals. This makes the decision tool easy for use. However the tool is now made
specifically for the Yatenga region in Burkina Faso and adaptation might need to be made if the tool
is to be used in any other region as constraints and goals might be completely different.

To make good use of the decision tool, the user should know with certainty to which type the farm
belongs to. This might be tricky as farms (as seen in this thesis) can rapidly change from type.

The presence of constraints through the matching off farming constraints and expert opinions led
to relatively similar results for all the types. Mostly the same constraints were present for the same
interventions, even if the strength of the constraints was different. This is related to the method
used: scores were categorised and the use of the key matrix to match interventions to types
globalised scores and smoothed out differences.

In the decision tool only farming constraints and goals are taken into account. It is however known
that the farmer’s perception of the interventions, as well as traditions and habits may also play an
important role in the adoption of interventions (Adesina & Baidu-Forson, 1995). To enhance the
adoption rate of the interventions it is therefore interesting to study these aspects.

For the modelling part of this thesis only calorie production, livestock productivity and SOC
outcomes were analysed. If the aim of the project is different the modelling exercise should be
done anew. Social outcomes, such as social cohesion are not taken into account in the FARMSIM
model. Other models also including social outcomes could be created.

54
55
5 Conclusion
In this thesis it was shown how modelling, household survey data and qualitative data on goals,
constraints and attitudes can be combined to create a decision tool which may help in increasing
adoption potential of promising interventions.

In the modelling baseline scenario yield, livestock productivity and SOC decreased over time. Yields
declined due to the decline in soil fertility. SOC decreased because the input of organic matter into
the soil did not off-set the output of organic matter into the soil. Food self-sufficiency decreased
over the years due to the reduction in yields but was also dependent on farm typology, defining the
ratio of land cultivated per household member. Feed quality was low and led to the reduction of
livestock productivity, especially in the dry season, when grass quality was lowest.

When modelling the interventions, the impacts on food production, livestock productivity and SOC
showed the same trends for all farm types. The magnitude of the impacts however differed
depending on farm type, due to the ratio of farm area: livestock and the area of crops grown.
Therefore the interventions potentially increasing SOC, livestock productivity of calorie production
were the same for all types but the quantitative impact depended on farm type. As manure
collection was set at 0, there was no return of nutrients to the field in the form of manure and
there was no feedback mechanism between the improvement of livestock diet and crop yield or
soil. The shortage in feed quality led to the livestock being sensitive to changes in feed quality,
while farm calorific production and SOC were more static. Not one of the interventions improved all
the objectives (livestock productivity, SOC and crop production). Either 1 or 2 objectives were
improved or a trade-off was found between different objectives when implementing the
interventions.

Farming constraints and goals were dependent on resource endowment and therefore the use of a
resource-based typology is a good way to classify farmers to find the most suitable interventions to
promote. Farming goals and constraints, as well as farm typology are not constant and need to be
measured and reviewed frequently. The labour constraint should be taken into account in the
future as labour is an important farm asset. For the good implementation of the decision tool the
difficulty of alleviating the constraints should be analysed, as this impacts the selection of suitable
interventions for the farms. From this thesis, no conclusions can be made based on farmer attitude
and the willingness or not to implement interventions. A different way for quantifying farming
attitude should be found.

Gender was not found to play a role in the defining of farming goals or constraints. This can be due
to the fact that resource endowment is more important in setting goals and constraints. It could
however also be due to the way in which the interviews were conducted. More research is needed
to confirm these hypotheses.

The final decision tool is broad and leaves space for the user to fill in the goals of the project. When
having been through all the steps and having chosen an intervention, cultural values, practices and
tradition should be taken into account as they can play a role in intervention adoption. The decision
tool is constructed for farms in the Yatenga region. It is possible the impacts of interventions and
farming goals and constraints are different in other regions. These should then be measured again.

56
57
 6 References
Adesina, A. A., & Baidu-Forson, J. (1995). Farmers' perceptions and adoption of new agricultural
technology: evidence from analysis in Burkina Faso and Guinea, West Africa. Agricultural
economics, 13(1), 1-9.
Aguilera, J., Motavalli, P., Valdivia, C., & Gonzales, M. A. (2013). Impacts of Cultivation and Fallow Length
on Soil Carbon and Nitrogen Availability in the Bolivian Andean Highland Region. Mountain
Research and Development, 33(4), 391-403.
Alvarez, S., Paas, W., Descheemaeker, K., Tittonell, P., & Groot, J. Typology construction, a way of dealing
with farm diversity.
Asadu, C. L. A., Nweke, F. I., & Enete, A. A. (2008). Soil properties and intensification of traditional
farming systems in Sub Saharan Africa (SSA).Journal of Tropical Agriculture, Food, Environment
and Extension, 7, 186-192.
Ayuk, E. T. (1997). Adoption of agroforestry technology: the case of live hedges in the Central Plateau of
Burkina Faso. Agricultural systems, 54(2), 189-206.
Balch, C. C. (1977). The potential of poor-quality agricultural roughages for animal feeding. FAO Animal
Production and Health Papers (FAO).
Bambara, A. S. (2010, June 6). Accompagnement touristique au Burkina Faso [blog post].
http://sbambara.canalblog.com/
Barbier, B., Yacouba, H., Karambiri, H., Zoromé, M., & Somé, B. (2009). Human vulnerability to climate
variability in the Sahel: farmers’ adaptation strategies in northern Burkina Faso. Environmental
management, 43(5), 790-803.
Bationo, A., & Mokwunye, A. U. (1991). Alleviating soil fertility constraints to increased crop production in
West Africa: The experience in the Sahel. In: Alleviating soil fertility constraints to increased crop
production in West Africa(pp. 195-215). Springer Netherlands.
Bationo, A., Nandwa, S.M., Kimetu, J.M., Kinyangi, J.M., Bado, B.V., Lompo, F., Kimani, S., Kihanda, F., &
Koala, S. (2001). In T.O. Williams, S. Tarawali, P. Hiernaux & S. Fernandez-Rivera (Eds),
Sustainable crop-livestock productivity for improved livelihoods and natural resource management
in West Africa (pp. 173-198). Nairobi, Kenya.
Bationo, A., & Buerkert, A. (2001). Soil organic carbon management for sustainable land use in Sudano-
Sahelian West Africa. Nutrient cycling in Agroecosystems, 61(1-2), 131-142.
Bationo, A., Kihara, J., Vanlauwe, B., Waswa, B., & Kimetu, J. (2007). Soil organic carbon dynamics,
functions and management in West African agro-ecosystems. Agricultural systems, 94(1), 13-25.
Bayu, W., Rethman, N. F. G., & Hammes, P. S. (2005). The role of animal manure in sustainable soil
fertility management in sub-Saharan Africa: a review. Journal of Sustainable Agriculture, 25(2),
113-136.
Beal, T., Belden, C., Hijmans, R., Mandel, A., Norton, M. & Riggio, J. (2015). Country profiles: Burkina
Faso. Sustainable Intensification Innovation Lab. http://gfc.ucdavis.edu/profiles/rst/bfa.html
Bidogeza, J. C., Berentsen, P. B. M., De Graaff, J., & Lansink, A. O. (2009). A typology of farm households
for the Umutara Province in Rwanda. Food Security, 1(3), 321-335.
Bilgo, A., Masse, D., Sall, S., Serpantié, G., Chotte, J. L., & Hien, V. (2007). Chemical and microbial
properties of semiarid tropical soils of short-term fallows in Burkina Faso, West Africa. Biology and
Fertility of Soils, 43(3), 313-320.
Binswanger, H. P., & McIntire, J. (1987). Behavioural and material determinants of production relations in
land-abundant tropical agriculture. Economic Development and Cultural Change, 36(1), 73-99.
Boone, H. N., & Boone, D. A. (2012). Analyzing likert data. Journal of extension, 50(2), 1-5.
Bossuet, J. (March 2010). Fodder Innovations to Help Indian Dairy Farmers. New Agricuturist.
http://www.new-ag.info/en/
Breman, H. & de Ridder, N. (1991). Manuel sur les pâturages des pays sahéliens. KARTHALA Editions.
Buresh, R. J., & Tian, G. (1998). Soil improvement by trees in sub-Saharan Africa. In Directions in Tropical
Agroforestry Research (pp. 51-76). Springer Netherlands.

58
Cairns, J. E., Hellin, J., Sonder, K., Araus, J. L., MacRobert, J. F., Thierfelder, C., & Prasanna, B. M. (2013).
Adapting maize production to climate change in sub-Saharan Africa. Food Security, 5(3), 345-360.
CGIAR. (n.d.). Drought-Tolerant Crops for Drylands. http://www.cgiar.org/web-archives/www-cgiar-org-
impact-global-des_fact2-html/
Chan, K. Y. (2010, July). The important role of soil organic carbon in future mixed farming systems.
In Proceedings of 25th Annual Conference of the Grassland Society of NSW Inc. on ‘Adapting Mixed
Farms to Future Environments (pp. 28-29).
Clason, D. L., & Dormody, T. J. (1994). Analyzing data measured by individual Likert-type items. Journal of
Agricultural Education, 35, 4.
Corbeels, M., Shiferaw, A., & Haile, M. (2000). Farmers' knowledge of soil fertility and local management
strategies in Tigray, Ethiopia. IIED-Drylands Programme.
Critchley, W. (1991) Case Studies, Burkina Faso. In: Graham O. (ed) Looking after our Land: Soil and
Water Conservation in Dryland Africa. Oxford, UK
Defoer T., Budelman A., Toiulmin C., Carter S.E. (2000). Building Common Knowledge. Participatory
Learning and Action Research (Part 1). In: Defoer T, Budelman A. (eds). Managing Soil Fertility in
the Tropics. A Resource Guide for participatory learning and action research. Amsterdam, The
Netherlands: Royal Tropical Institute.
Descheemaeker, K., Amede, T., & Haileslassie, A. (2010). Improving water productivity in mixed crop–
livestock farming systems of sub-Saharan Africa. Agricultural water management, 97(5), 579-586.
Douxchamps, S., Ayantunde, A., & Barron, J. (2014). Taking stock of forty years of agricultural water
management interventions in smallholder systems of Burkina Faso. Water resources and rural
development, 3, 1-13.
Drechsel, P., Gyiele, L., Kunze, D., & Cofie, O. (2001). Population density, soil nutrient depletion, and
economic growth in sub-Saharan Africa. Ecological economics, 38(2), 251-258.
Drechsel, P., Olaleye, A., Adeoti, A., Thiombiano, L., Barry, B., & Vohland, K. (2005). Adoption driver and
constraints of resource conservation technologies in sub-Saharan Africa. Berlin: FAO, IWMI,
Humbold Universitaet.
Elmqvist, T., Tuvendal, M., Krishnaswamy, J., & Hylander, K. (2013). Managing trade-offs in ecosystem
services. Values, Payments and Institutions for Ecosystem Management. Ed. Kumar P and Thiaw I,
70-89.
Erenstein, O. (2003). Smallholder conservation farming in the tropics and sub-tropics: a guide to the
development and dissemination of mulching with crop residues and cover crops. Agriculture,
Ecosystems & Environment, 100(1), 17-37.
Falconnier, G. N., Descheemaeker, K., Van Mourik, T. A., & Giller, K. E. (2016). Unravelling the causes of
variability in crop yields and treatment responses for better tailoring of options for sustainable
intensification in southern Mali. Field Crops Research, 187, 113-126.
FAO. (2014). Socio-economic context and role of agriculture. FAPDA country fact sheet on food and
agriculture policy trends, Burkina Faso.
FAOSTAT (2013). Agro-Environmental Indicators. Fertilizers. Nitrogen + Phosphate Fertilizers (N+P2O5
total nutrients) + (Total). http://faostat.fao.org/site/677/DesktopDefault.aspx?PageID=677%20-
%20ancor#ancor
FAOSTAT. (2014). Production, Crops, Burkina Faso, 2013. Retrieved from http://faostat3.fao.org/home/E
FAPDA (2014). Socio-economic context and role of agriculture. In: Country Fact Sheet on Food and
Agriculture Policy Trends, Burkina Faso (pp 1). Food and Agriculture Policy Decision Analysis team
of FAO.
FAO (2015) FAO Statistical Pocketbook. World Food and Agriculture. Food and Agriculture Organization of
the United Nations, Rome, 2015.
Feedipedia. (2012). Animal Feed Resources Information System. http://www.feedipedia.org/node/12361
Funk, C. C., Rowland, J., Eilerts, G., Adoum, A., & White, L. (2012). A Climate Trend Analysis of Burkina
Faso (No. 2012-3084). US Geological Survey.
Gobin, A., Campling, P., Janssen, L., Desmet, N., van Delden, H., Hurkens, J., Lavelle, P. & Berman, S.
(2011). Approach to Assess Regional Soil Organic Matter Balances and Scenario Analyses. In: Soil

59
 organic matter management across the EU – best practices, constraints and trade-offs. Final
Report for the European Commission’s DG Environment, September 2011.
Haub C. & Kaneda T. (2014). 2014 World Population Data Sheet. Population Reference Bureau, 2014,
Washington DC.
Herrero, M., Grace, D., Njuki, J., Johnson, N., Enahoro, D., Silvestri, S., & Rufino, M. C. (2013). The roles
of livestock in developing countries. Animal,7(s1), 3-18.
HHS/USAD (2010). Dietary guidelines for Americans.
History & maps. (2015). Country map, Google map Burkina Faso.
http://www.lahistoriaconmapas.com/atlas/country-map03/google-map-burkina-faso.htm
Kazianga, H., & Udry, C. (2006). Consumption smoothing? Livestock, insurance and drought in rural
Burkina Faso. Journal of Development Economics, 79(2), 413-446.
Kebede, Y., Gunjal, K., & Coffin, G. (1990). Adoption of new technologies in Ethiopian agriculture: The case
of Tegulet-Bulga district Shoa province. Agricultural economics, 4(1), 27-43.
Kevane, M., & Gray, L. C. (1999). A woman's field is made at night: Gendered land rights and norms in
Burkina Faso. Feminist Economics, 5(3), 1-26.
Kingori, A. M., Wachira, A. M., & Tuitoek, J. K. (2010). Indigenous chicken production in Kenya: a
review. International Journal of Poultry Science, 9(4), 309-316.
Köhn, H. F., & Hubert, L. J. (2006). Hierarchical cluster analysis. Wiley StatsRef: Statistics Reference
Online.
Körschens, M., Weigel, A., & Schulz, E. (1998). Turnover of soil organic matter (SOM) and long‐term
balances—tools for evaluating sustainable productivity of soils. Zeitschrift für Pflanzenernährung
und Bodenkunde, 161(4), 409-424.
Kristjanson, P., Okike, I., Tarawali, S., Singh, B. B., & Manyong, V. M. (2005). Farmers' perceptions of
benefits and factors affecting the adoption of improved dual‐purpose cowpea in the dry savannas of
Nigeria. Agricultural Economics,32(2), 195-210.
Langyintuo, A. S., & Dogbe, W. (2005). Characterizing the constraints for the adoption of a Callopogonium
mucunoides improved fallow in rice production systems in northern Ghana. Agriculture, ecosystems
& environment, 110(1), 78-90.
Lawrence, P. R. (Ed.). (1993).Research on the nutrition of working animals : needs, experiences and
methods. In: Recherche Pour Le Développement de la Traction Animale en Afrique de L'Ouest (pp
163-174). ILRI (aka ILCA and ILRAD).
Leenaars J.G.B., van Oostrum, A.J.M. & Ruiperez Gonzalez, M. (2014) Africa Soil Profiles Database,
version 1.2. A compilation of geo-referenced and standardised legacy soil profile data for Sub-
Saharan Africa (with dataset). ISRIC report 2014/01. Africa Soil Information Service (AfSIS)
project. ISRIC – World Soil Information, Wageningen, the Netherlands.
Lykke, A. M., Mertz, O. L. E., & Ganaba, S. (2002). Food consumption in rural Burkina Faso. Ecology of
Food and Nutrition, 41(2), 119-153.
Mando, A., & Stroosnijder, L. (1999). The biological and physical role of mulch in the rehabilitation of
crusted soil in the Sahel. Soil Use and Management,15(2), 123-127.
Marinus, W., van de Ven, G., Descheemaeker, K., Zijlstra, M., Koot, T. (2015). NUANCES-FARMSIM –
Analysing Smallholder crop-livestock systems in sub-Saharan Africa. Course document,
Wageningen University, The Netherlands, Wageningen.
McCown, R. L., Hammer, G. L., Hargreaves, J. N. G., Holzworth, D. P., & Freebairn, D. M. (1996). APSIM: a
novel software system for model development, model testing and simulation in agricultural
systems research. Agricultural systems, 50(3), 255-271.
McDermott, J. J., Staal, S. J., Freeman, H. A., Herrero, M., & Van de Steeg, J. A. (2010). Sustaining
intensification of smallholder livestock systems in the tropics. Livestock science, 130(1), 95-109.
McDowell, R. E. (1988). Importance of crop residues for feeding livestock in smallholder farming systems.
In Plant breeding and the nutritive value of crop residues. Proceedings of a workshop. ILCA, Addis
Ababa (Ethiopia) (pp. 3-27).
MEA (Millennium Ecosystem Assessment) 2003. Ecosystems and their services, In: Hassam R., Scholes R.
& Ash N. (eds) Ecosystems and Human Well-being. Washington DC, USA.

60
Mertz, O., Mbow, C., Reenberg, A., & Diouf, A. (2009). Farmers’ perceptions of climate change and
agricultural adaptation strategies in rural Sahel. Environmental management, 43(5), 804-816.
Mortimore, M. J., & Adams, W. M. (2001). Farmer adaptation, change and ‘crisis’ in the Sahel. Global
environmental change, 11(1), 49-57.
Nagothu, U. S. (Ed.). (2016). Climate variability and extremes, relevance of agro-ecological-based climate
smart farming systems in the sub-Saharan Africa. In: Climate Change and Agricultural
Development: Improving Resilience Through Climate Smart Agriculture, Agroecology and
Conservation. Routledge.
Nhamo, N., Mupangwa, W., Siziba, S., Gatsi, T., & Chikazunga, D. (2003). The role of cowpea (Vigna
unguiculata) and other grain legumes in the management of soil fertility in the smallholder farming
sector of Zimbabwe. In: Grain legumes and green manures for soil fertility in southern Africa:
taking stock of progress (pp 119-127).
Nijhof, K., (1987 a). The concentration of macro-elements in economic products and residues of
(sub)tropical field crops. Staff working paper SOW-87-08-. Stichting onderzoek
wereldvoedselvoorziening / Centre for world food studies, Wageningen, The Netherlands
Nijhof, K., (1987 b). The concentration of macro-nutrients in (sub)tropical crops of minor importance. Staff
working paper SOW-87-17-. Stichting onderzoek wereldvoedselvoorziening / Centre for world food
studies, Wageningen, The Netherlands
Oladele, O., & Monkhei, M. (2008). Gender ownership patterns of livestock in Botswana. Livestock
Research for Rural Development, 20(10).
Omotayo, O.E. & Chukwuka, K. S. (2009). Soil fertility restoration techniques in sub-Saharan Africa using
organic resources. African Journal of Agricultural Research, 4(3), 144-150.
Ott, R. L., & Longnecker, M. T. (2015). Categorical data. Taylor M., Siebert, D. & Walsh, S. (Eds). In: An
introduction to statistical methods and data analysis. Nelson Education.
Ouedraogo, M., Kabore, V., Riej, C., Scoones, I., & Toulmin, C. (1996). The zaï: a traditional technique for
the rehabilitation of degraded land in the Yatenga, Burkina Faso. Sustaining the soil: indigenous
soil and water conservation in Africa., 80-84.
Owen, E. (1994). Cereal crop residues as feed for goats and sheep. Livestock Research for Rural
Development, 6(1), 13.
van Paassen, A., de Ridder, N., & Stroosnijder, L. (2011). Role of an explorative model for learning about
sustainable agricultural development in Burkina Faso. International journal of agricultural
sustainability, 9(2), 310-321.
Panthou, G., Vischel, T., & Lebel, T. (2014). Recent trends in the regime of extreme rainfall in the Central
Sahel. International Journal of Climatology,34(15), 3998-4006.
Powell, J. M., Fernandez-Rivera, S., Hiernaux, P., & Turner, M. D. (1996). Nutrient cycling in integrated
rangeland/cropland systems of the Sahel.Agricultural Systems, 52(2), 143-170).
Powell, J. M., Pearson, R. A., & Hiernaux, P. H. (2004). Crop–livestock interactions in the West African
drylands. Agronomy journal, 96(2), 469-483.
Power, A. G. (2010). Ecosystem services and agriculture: trade-offs and synergies. Philosophical
transactions of the royal society B: biological sciences, 365(1554), 2959-2971.
Pretty, J. N., Noble, A. D., Bossio, D., Dixon, J., Hine, R. E., Penning de Vries, F. W. T., & Morison, J. I. L.
(2006) Resources-Conserving Agriculture Increases Yields in Developing Countries. Environmental
science & technology, 40(4), 1114-1119.
Professional Nutrient Management Group. (2013). Feed planning for cattle and sheep.
www.nutrientmanagement.com
Quiros, C. (2013) IMPACT Lite Data Dictionary; East and West Africa Datasets. From the CCAFS Detail
Household Characterisation Survey.
Rattunde, H. F. W. (1998). Early‐maturing dual‐purpose sorghums: Agronomic trait variation and
covariation among landraces. Plant Breeding, 117(1), 33-36.
Rege, J. E. O. (2001). On-farm phenotypic characterisation of Kenyan zebu cattle. In: Zebu cattle of
Kenya: Uses, performance, farmer preferences, measures of genetic diversity and options for
improved use (Vol. 1). ILRI (aka ILCA and ILRAD).

61
de Ridder, N., Stroosnijder, L., Cisse, A.M. & van Keulen, H. (1982). The Importance of Animal Husbandry
Systems in the Sahel. In: Productivity of Sahelian Rangelands: A Study of the Soils, the Vegetation
and the Exploitation of that Natural Resource. PPS Course Book, Volume 1, Theory. Wageningen
Agricultural University, Department of Soil Science and Plant Nutrition, Wageningen, The
Netherlands.
de Ridder, N. (1991). La production des pâturages. Manuel sur les pâturages des pays sahéliens.
KARTHALA (Eds).
de Ridder, N., Breman, H., van Keulen, H., & Stomph, T. J. (2004). Revisiting a ‘cure against land hunger’:
soil fertility management and farming systems dynamics in the West African Sahel. Agricultural
systems, 80(2), 109-131.
De Ridder, N., Sanogo, O. M., Rufino, M. C., van Keulen, H., & Giller, K. E. (2015). Milk: the new white
gold? Milk production options for smallholder farmers in Southern Mali. animal, 9(07), 1221-1229
Rufino, M. C., Tittonell, P., Van Wijk, M. T., Castellanos-Navarrete, A., Delve, R. J., De Ridder, N., & Giller,
K. E. (2007). Manure as a key resource within smallholder farming systems: analysing farm-scale
nutrient cycling efficiencies with the NUANCES framework. Livestock Science, 112(3), 273-287.
Sanogo, O. M. (2011). Le lait, de l’or blanc? (Doctoral dissertation, Wageningen University).
Sanon, H. O., Kaboré-Zoungrana, C., & Ledin, I. (2007). Behaviour of goats, sheep and cattle and their
selection of browse species on natural pasture in a Sahelian area. Small Ruminant Research, 67(1),
64-74.
Schafleitner, R., Rosales, R. O. G., Gaudin, A., Aliaga, C. A. A., Martinez, G. N., Marca, L. R. T. &
Bonierbale, M. (2007). Capturing candidate drought tolerance traits in two native Andean potato
clones by transcription profiling of field grown plants under water stress. Plant Physiology and
Biochemistry,45(9), 673-690.
Singh, B. B., Ajeigbe, H. A., Tarawali, S. A., Fernandez-Rivera, S., & Abubakar, M. (2003). Improving the
production and utilization of cowpea as food and fodder. Field Crops Research, 84(1), 169-177.
SSSA (Soil Science Society of America). (2008). Glossary of soil science terms 2008. ASA-CSSA-SSSA.
Strauss, J., Barbosa, M., Teixeira, S., Thomas, D., & Junior, R. G. (1991). Role of education and extension
in the adoption of technology: a study of upland rice and soybean farmers in Central-West
Brazil. Agricultural Economics, 5(4), 341-359.
Styger, E., & Fernandes, E. C. (2006). Contributions of managed fallows to soil fertility recovery. Biological
approaches to sustainable soil systems. New York: Taylor & Francis. p, 425-437.
Twomlow, S. J., Urolov, J. C., Jenrich, M., & Oldrieve, B. (2008). Lessons from the field–Zimbabwe’s
conservation agriculture task force. Journal of SAT Agricultural Research, 6(1), 1-11.
USDA, Natural Resources Conservation Services Soils. (n.d. a). Soil Texture Calculator.
http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/?
USDA, Natural Resources Conservation Services Soils. (n.d. b).Estimating Moist Bulk Density by Texture.
http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/
USDA (2016). USDA National Nutrient Database for Standard Reference. Release 28.
https://ndb.nal.usda.gov/ndb/search
Valbuena, D., Erenstein, O., Tui, S. H. K., Abdoulaye, T., Claessens, L., Duncan, A. J. & van Wijk, M. T.
(2012). Conservation Agriculture in mixed crop–livestock systems: Scoping crop residue trade-offs
in Sub-Saharan Africa and South Asia. Field crops research, 132, 175-184.
Valbuena, D., Tui, S. H. K., Erenstein, O., Teufel, N., Duncan, A., Abdoulaye, T. & Gérard, B. (2015).
Identifying determinants, pressures and trade-offs of crop residue use in mixed smallholder farms
in Sub-Saharan Africa and South Asia. Agricultural Systems, 134, 107-118.
Vanlauwe, B., & Giller, K. E. (2006). Popular myths around soil fertility management in sub-Saharan
Africa. Agriculture, ecosystems & environment,116(1), 34-46.
Vanlauwe, B., Bationo, A., Chianu, J., Giller, K. E., Merckx, R., Mokwunye, U. & Smaling, E. M. A. (2010).
Integrated soil fertility management operational definition and consequences for implementation
and dissemination. Outlook on agriculture, 39(1), 17-24.
van Vliet, J. A., Schut, A. G., Reidsma, P., Descheemaeker, K., Slingerland, M., van de Ven, G. W., & Giller,
K. E. (2015). De-mystifying family farming: features, diversity and trends across the globe. Global
Food Security, 5, 11-18.

62
Wang, G., LUO, Z., Wang, E. & Huang, Y. (2013). Contrasting effects of agricultural management on soil
organic carbon balance in different agricultural regions of China. Pedosphere, 23(6), 717-728.
Whitbread, A. M., Robertson, M. J., Carberry, P. S., & Dimes, J. P. (2010). How farming systems simulation
can aid the development of more sustainable smallholder farming systems in southern
Africa. European Journal of Agronomy, 32(1), 51-58.
van Wijk, M. T., Tittonell, P., Rufino, M. C., Herrero, M., Pacini, C., de Ridder, N., & Giller, K. E. (2009).
Identifying key entry-points for strategic management of smallholder farming systems in sub-
Saharan Africa using the dynamic farm-scale simulation model NUANCES-FARMSIM. Agricultural
Systems, 102(1), 89-101.
WLE (Water, Land and Ecosystems) (2014). Project: Realising the full potential of mixed crop-livestock
systems in rapidly changing Sahelian agro-ecological landscapes. Project Proposal, November
2014.
Woolverton A., Okello, J., Binci, M. & Never, D. (2014). Attitudes and decision-making in smallholder
commercialization. In Woolverton, A., & Neven, D. (eds) Understanding Smallholder Farmer
Attitudes to Commercialization, the Case of Maize in Kenya. FAO (Food and Agriculture
Organization of the United Nations), Rome, Italy.
the World Bank (2013). Transforming Agriculture in the Sahel: What would it take? Background paper for
summits on pastoralism and irrigation in the Sahel, in Nouakchott and Dakar, on 29 and 31
October 2013 respectively
Zerbini, E., & Thomas, D. (2003). Opportunities for improvement of nutritive value in sorghum and pearl
millet residues in South Asia through genetic enhancement. Field Crops Research, 84(1), 3-15.
Zhang, W., Ricketts, T. H., Kremen, C., Carney, K., & Swinton, S. M. (2007). Ecosystem services and dis-
services to agriculture. Ecological economics,64(2), 253-260.
Zorom, M., Barbier, B., Mertz, O., & Servat, E. (2013). Diversification and adaptation strategies to climate
variability: A farm typology for the Sahel. Agricultural Systems, 116, 7-15.

63
7 Appendices
7.1 Appendix I: Nutritional values for food

Table 1: Nutritional values of food elements (source: USDA, 2016).

Product kcal / 100 g
Sorghum grain 329
Millet raw 378
Cowpea, common, mature seeds, raw 336
Peanut, all type, raw 567
Corn grain yellow 365

64
7.2 Appendix II: Parametrising FARMSIM

Table 1: Monthly and yearly rainfall (mm) in Ouahigouya, for the years 1981 to 1995.

Rainfall (mm)

Year Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec Yearly Total

1981 0 0 3 11 222 82 182 175 106 55 0 0 836

1982 0 4 3 3 43 68 67 129 33 15 0 0 365

1983 0 0 0 6 8 60 110 122 51 1 0 0 358

1984 0 0 0 1 33 25 133 82 98 19 2 0 393

1985 0 0 0 4 8 62 148 133 64 2 0 0 421

1986 0 0 0 1 40 101 128 172 141 8 0 0 591

1987 0 0 0 0 0 124 106 89 131 6 0 0 456

1988 0 0 0 43 0 9 207 294 154 1 0 0 708

1989 0 0 0 0 34 25 147 311 61 34 0 0 612

1990 0 0 0 3 20 24 154 97 76 29 0 16 419

1991 0 0 0 1 65 159 83 229 106 22 0 0 665

1992 0 0 0 0 28 75 184 255 135 6 30 0 713

65
Table 2: FARMSIM input parameters for Méré cattle.
Cattle
category Parameter Unit Value
All cattle
Birth weight of calves kg 22
Distance walked by a cow per day km 6
Energy requirement to walk J/m/kg 1.5
Maturity groups (early, medium, late) - 3
Maximum fraction of concentrates in diet as part of
total dry matter intake - 0.4
Fraction of the produced milk given to the calf - 0.6
Milk substitute supplied to the calves depending on
age kg/year 800 800 0
Age steps of the calf for milk allowance year 0 0.1667 1.5
Crude protein content of milk fed to calves g/kg DM 3.7
Dry matter content of milk fed to calves g/kg DM 142
Metabolizable energy in milk fed to calves MJ/kg DM 23.3
Age at which the calf stops receiving milk year 1.4166
Minimum weight curve depending on age. If below
this value the cow dies kg 18 40 65 78 95 130 140 140
Age steps for a specific minimum weight year 0 1.4 3.1 4.1 6.1 12 20 25
Live weight gain value MJ/kg DM 19
Live weight loss while lactating g/kg 138
Bulls
The effect of age on mortality rate - 0 0 0 0 0 0 0 1 1
Age steps for mortality rate year 0 1 2 3 4 5 12 20 25
Potential growth curve, maximum weight
depending on age kg 30 159 280 345 388 400 400 400
Age steps for potential weight years 0 2.1 4.1 6.7 9.1 12 20 25
Activity allowance for beef cattle MJ/day 0.00696
Correction factor for fasting metabolism
requirements - 1.15
Correction factor for the emery value for weight
gain - 1
Correction factor for nutrition level 1

66
 Sex-specific correlation factor for energy retention - 1.15
Correction factor for net protein in weight gain,
depending on maturity type - 1
Compensatory growth rate depending on
metabolisability of the feed kg/year 182.5 182.5 365 547.5 730 730 730
Metabolisability of the feed steps for compensatory
growth rate MJ/MJ 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Cows
Effect of age on the calving rate of a dam - 0.75 1 1 0.85 0 0
Age steps for calving rate year 3.5 5 9 10 20 25
Maximum calving rate per year /year 0.95
Fraction affecting the effect of age on lactation 0.8 0.8 1 1 0.6 0.3 0.3
Age steps for effect of age on lactation year 3.5 4.5 5.5 8 15 20 25
Reduction factor for lactation based on he condition
index - 0 1 1
Condition index steps for the effect on lactation - 0 0.3 1
Max annual body weight loss depending on
lactation phase kg/year 128 128 10 10
Lactation phase steps for max annual body weight
loss year 0 0.0625 0.625 1.5
Max length of lactation year 18
Fat content of the milk depending on the lactation
phase g/kg 40.5 48 55.5 60 63
Lactation phase steps for fat content of the milk year 0 0.25 0.5 0.833333 1.5
Modifier for feed intake of lactating cows depending
on lactation phase - 1.1 1.35 1.286 1.2 1
Potential lactation curve depending on lactation
phase kg/year 2200 2200 1250 0 0
Lactation phase steps for potential lactation curve year 0 0.166667 0.75 1.5 2
The effect of age on mortality rate - 0 0 0 0 0 0 0 1 1
Age steps for mortality rate year 0 1 2 3 4 5 12 20 25
Length of gestation period days 282
Feasible age set for reproductive age depending on
weight year 3.5 4.23 5 6.1 12 13 20 25
Weight steps for the reproductive age kg 230 169 115 110 120 140 145 145
Potential growth curve, maximum weight
depending on age kg 30 139 214 255 280 300 300 300

67
Age steps for potential weight years 0 1.4 3.1 4.1 6.1 12 20 25
Activity allowance for beef cattle MJ/day 0.00917
Correction factor for fasting metabolism
requirements - 1
Correction factor for the energy value for weight
gain - 1.3
Correction factor for nutrition level - 1
Sex-specific correlation factor for energy retention - 1.1
Correction factor for net protein in weight gain,
depending on maturity type - 0.8
Compensatory growth rate depending on
metabolisability of the feed kg/year 120 182.5 182.5 365 547.5 547.5 547.5
Metabolisability of the feed steps for compensatory
growth rate MJ/MJ 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Efficiency for growth of concepta - 0.133

68
Table 3: Crop characteristics used as input in FARMSIM (source: Nijhoff 1987 a, b).

Crop characteristics Unit Maize Cowpea Millet Sorghum Groundnut

Number to identify the crop - 1 2 3 4 5

Legume (0 = no, 1 = yes) - 0 1 0 0 1

Harvest index - 0.4 0.26 0.26 0.27 0.33

Shoot : root ratio - 6 11 6 6 6

Light determined yield kg/ha 24400 19200 19200 18500 19200

Water conversion efficiency Kg/ha/mm 89 72 72 70 72

Minimum and maximum values for N, P and K in grains and stover

Maximum grain N (%) kg N/kg DM 0.032 0.047 0.0385 0.032 0.056

Minimum grain N (%) kg N/kg DM 0.0095 0.029 0.008 0.01 0.035

Maximum stover N (%) kg N/kg DM 0.01 0.027 0.01 0.012 0.022

Minimum stover N (%) kg N/kg DM 0.004 0.0075 0.0032 0.0035 0.007

Maximum grain P (%) kg P/kg DM 0.009 0.006 0.009 0.0065 0.008

Minimum grain P (%) kg P/kg DM 0.0015 0.002 0.0012 0.0013 0.0027

Maximum stover P (%) kg P/kg DM 0.0035 0.003 0.007 0.003 0.0052

Minimum stover P (%) kg P/kg DM 0.0005 0.0008 0.0004 0.0005 0.0011

Maximum grain K (%) kg K/kg DM 0.008 0.018 0.008 0.007 0.026

Minimum grain K (%) kg K/kg DM 0.003 0.009 0.002 0.0025 0.012

Maximum stover K (%) kg K/kg DM 0.024 0.037 0.036 0.028 0.024

Minimum stover K (%) kg K/kg DM 0.01 0.0085 0.009 0.008 0.007

Dry matter content (fraction) of total - 1 1 1 1 1

biomass

Carbon content of stover kg C/kg DM 0.42 0.45 0.42 0.45 0.45

Root C content kg C/kg DM 0.32 0.32 0.32 0.32 0.32

Root N content kg N/kg DM 0.005 0.005 0.005 0.005 0.005

Root P content kg P/kg DM 0.001 0.001 0.001 0.001 0.001

Ration of nictrogen (N) from N2 fixation to - 0 0.6 0 0 0.6

the total above ground plant N (only for
legumes)

Parameters for reduction factor relationship for water capture efficiency

Parameter a - 0.9889 0.9889 0.9889 0.9889 0.9889

Parameter b - -0.4706 -0.4706 -0.4706 -0.4706 -0.4706

Parameter r - 0.9028 0.9028 0.9028 0.9028 0.9028

69
Table 4: Pasture quality over the seasons, data input for FARMSIM.

Parameter Units Rainy season Early dry Late dry

(June, July, season season
August, (November, (February,
September and December and march, April
October) January) and may)

Dry Matter (g/kg) 200 500 800

Metabolisable Energy MJ/kg DM 10.3 8.7 6.5

Gross Energy MJ/kg DM 22 18 16

Fermentable Metabolisable Energy MJ/kg DM 12 12 12

Crude Protein g/kg DM 134 100 40

Acid Detergent Insoluble N g/kg DM 1.235 1.116 0.906

Dry Matter Digestibility g/kg DM 0.65 0.48 0.41

A: proportion of water soluble N in Kg N/kg N 0.4 0.56 0.56

the total N in feed

B: proportion of potentially Kg N/kg N 0.6 0.38 0.38

degradable (slowly degradable) N
other than water soluble N in total
N in feed

C: Fractional rumen degradation Kg N/kg N/h 0.2 0.04 0.04

rate per hour of the b fraction of
the feed N with time

Neutral Detergent Fibre g/kg DM 670 670 670

P concentration Kg P/kg DM 0.002 0.002 0.002

K concentration Kg K/kg DM 0.016 0.016 0.016

70
7.3 Appendix III: Interviewing protocol

Protocol for Questionnaire ‘Assessing biomass enhancing

options for the Ouahigouya region, Burkina Faso’
7.3.1 Introduction
In this document we present the protocol for a farmer survey that aims to collect information about
the farm production constraints and the goals and attitudes of farmers. This survey is conducted
within the project ‘Realising the full potential of mixed crop-livestock systems in rapidly changing
Sahelian agro-ecological landscapes’, WLE Volta and Niger Focal Region (WLE-Volta project). This
project, led by the International Livestock Research Institute (ILRI), puts emphasis on the impact
of biomass enhancing interventions on agricultural productivity, natural resource sustainability,
food security and livelihoods.

Biomass serves many purposes in farming systems. It is important as a source of food, feed, fuel
and fibre, but it also provides organic matter and a cover for the soil, improving soil quality.
Improved soil quality can lead to improved food and feed availability. Burkina Faso is a country
with yearly food deficits and therefore increasing biomass production could decrease the hunger
period.

There are many ways for enhancing biomass production, such as the planting of dual-purpose
improved varieties or composting. However, not all interventions are suitable for all farmers. To
make a decision tool to help in choosing the right interventions it is important to have knowledge
about the environment of the farming system and about what is of interest and concern for
farmers. Consequently, when suggesting a new technology or intervention to a farmer, certain
points have to be kept into mind.

Firstly, the intervention has to be beneficial for a farmer and to the farming system. To assess this,
the impact of the intervention on the farming system has to be known. Secondly, it is of interest to
know if it is possible for a farmer to implement the intervention. This will depend on the production
constraints present in the farming system. The number of production constraints and the degree of
effort to overcome these production constraints will define which interventions can be
implemented. For example, if land is severely constraining, leaving land fallow is not a suited
intervention as it takes land out of production. Thirdly, the farmer’s goals and his or her attitude
towards farming have to be taken into account, as this will influence the decision to implement the
biomass enhancing intervention. Goals are the causes for which a farmer will strive and they define
what is important to farmers. The farmers’ attitude towards farming defines the farmer’s way of
looking at farming. The attitude towards farming characterizes the way farmers perceive
interventions (e.g. is he/she willing to try new options, is he/she a risk taker, etc.). Together, goals
and attitude influence the farmer’s perception of the technique’s advantages and therefore play a
role in defining how inclined a farmer is to adopting the intervention.

7.3.2 Aim
The main aim of this interview is to collect information to understand what are constraints in a
farming system, and what are different farmer goals and farmer attitude towards farming. This
information will be used to see how these concepts play a role in the decision of adopting or not
adopting a new intervention. From this information a decision support tool will be constructed,
which couples types of farmers to the interventions that are most promising for their situation.

As farmers have different assets and possessions, the most easily adoptable interventions will be
different for different farmers. Therefore farmers from different types (based on household size,
land and livestock possession) will be interviewed, as well as male and female household members.
The generated data will give information on the best suited interventions for different farm types
from the perspective of both men and women within these types.

71
To get the data for the construction of the decision tool, interviews have to be conducted, with
questions about farm constraints and farmer goals and farmer attitude. These interviews are
spread over the three different villages. Farmers from all the different farm types are interviewed,
and men and women are interviewed separately. Firstly the farmers to be interviewed are selected.
In a second step the survey is held. A deeper explanation on how to select the farmers to interview
can be found in Chapter 3, “Farmer selection”. More information about the way the interview
should be conducted can be found in Chapter 4, “Interview”. At the end of the protocol a checklist
is present, to make sure all the steps for farmer selection and setting up an interview are taken in
the right order.

7.3.3 Farmer Selection

Four villages are selected for the WLE-Volta project: Ziga, Thiou, Ninigui and Pogoro-silmimossin.
Select three of these villages to conduct the interview. In these three villages households
corresponding to every farm type should be interviewed. Both a male and female member of the
household should be interviewed within one household. The typology is based on household size,
total land area, cereal land area, number of small ruminants and number of cattle. Six farm types
have been made, which are explained below and can be found in Table 1.

Type 1 farmers have a small amount of land and animals. Type 2 farmers have the biggest number
of household members and have a big land area, compared to the other farm types. Farm type 3
are average, farmers having some land and some animals. In type 4 there are quite some animals,
but only a little land. Type 5 does not possess any livestock. According to farmers it is very
important to possess livestock, therefore this type is seen as being the less endowed farm type.
These are also the farmers which are hired and work on farms of the types 1, 2, 3, 4 and 6, as
farm type 5 is constrained by capital. The type 6 has the largest number of animals, both cattle
and small ruminants, and is seen as being the most endowed farm type. The average values for
the number of household members, total land area, number of small ruminants and number of
cattle can be seen in the following table.

Table 1: Average values for household size, cropping area and livestock possession for the six
different farm types.

Type 1 2 3 4 5 6
Household members 8 13 8 9 6 15
Area 1.5 10.6 4.8 2.3 2.9 17.9
Cattle 2 5 1 5 0 33
Ruminants 3 20 9 15 0 36

As it is impossible to interview all farmers, a selection of farmer is made for the interviews. Based
on the total number of farmers present in specific types, a varying number of farmers is selected
per specific type. In Table 2 the minimum number of households to interview in each type per
village can be found. The minimum amount of household interviews to be conducted for farm types
1, 2 and 3 is thrice per village. Farm type 4 is interviewed twice per village and farm types 5, 6 and
7 are interviewed once per village. During those household interviews both men and women have
to be questioned. If possible the woman interviewed within the household should be the main wife,
and the man interviewed should be the household head.

From information which was already collected within the WLE-Volta Project , it was possible to
select farmers to interview. The compiled list can be found at the end of the protocol ( “List of
farmers to be interviewed”). Use this list to know which farmers to interview. The list also presents
extra household head names, which can be interviewed in case an interview is cancelled.

72
Table 2: Number of households to be interviewed per village.

Farm Type V1 V2 V3 V4 V5 V6 V7 Total

Number of farms in type across all 30 44 73 19 3 7 7 183
villages
Number of households interviewed 3 3 3 2 1 1 1 14
per village
Number of interviews conducted 6 6 6 4 2 2 2 28
(men + women) per village
Total number of interviews 18 18 18 12 6 6 6 84
conducted (men + women)

7.3.4 Interview
The interview is for the main part a structured interview, but in the final section it allows space for
discussion. Within a household the interview is done twice; once with the male household head and
once with the main wife. Number each interview.

The interview is sub-divided in three parts. Parts one and three have to be asked to both household
members (man and wife separately). Part two of the interview has to be conducted only once
within a household as the answers are expected to be similar for the two household members.
Conducting the full interview (parts 1, 2 and 3) should take about 1h30. The second interview
within the household (which only includes parts 1 and 3) should last about one hour.

When starting the interview, a short introduction should be given about the project and the aim of
the survey should be explained, i.e. that information is collected about farming constraints, farmer
goals and farmer attitude towards farming in order to match interventions to farm types.

When no answer is given to some of the questions, fill in N.A. If the answer is zero, fill in 0.

7.3.4.1 Part One

7.3.4.1.1 To fill in by the interviewer

Date dd/mm/yyyy

Time of interview start hh:mm

Time of interview end hh:mm

Village Code A

Farm type Code B

Household ID

Interview number Number

Name of the interviewer Name

A) Village number: 1 = Ziga, 2 = Thiou, 3 = Ninigui, 4 = Pogoro-silmimossin

B) Type number: 1, 2, 3, 4, 5, 6 or 7, as described in Figure 1.

7.3.4.1.2 Basic Information

Ask for the following information and fill in the table.

73
Name of household head Name

Name of respondent Name

Gender of respondent Code A

Primary activity of the farm Code B

Marital status Code C

A) Gender: 1 = Male, 2 = Female

B) Primary activity of the farm: 1 = crop production, 2 = animal husbandry, 3 = crop production
and animal husbandry, 4 = other (specify)

C) Marital status: 1 = Married, 2 = Single, 3 = Divorced, 4 = Other (specify)

7.3.4.2 Part 2 Only for first interviewee

7.3.4.2.1 Household information

Please fill in the number of male and female household members above and below 15 years of age.
Household members are the people present for more than half of the year on the family property.

Household members Number

Male adult (above 16 years of age)

Female adult (above 16 years of age)

Boys (up to 16 years of age)

Girls (up to 16 years of age)

7.3.4.2.2 Land information

Total area of land

Unit of land area Code A

Area of land under cereal

Unit of land area Code A

A) Unit of land area: 1 = ha, 2 = m², 3 = other (specify)

7.3.4.2.3 Livestock information

Does the household possess livestock? Yes or no. If no, mark 0 in the column total. If yes, indicate
the number of animals for every species. Also indicate if the animals are owned by men, women or
if they are owned jointly. Finally ask if there are any other animals that have not been asked
about.

74
 Species Total number Owner (1=men, 2=women, 3=jointly)

Cattle

Small ruminants

7.3.4.3 Part Three

7.3.4.3.1 Perceived Constraints

Production constraints are the restrictions which limit or reduce the agricultural production. It is
important to assess the farming system constraints and their severity, as this information will help
identify the interventions which can most readily be implemented. These have a greater chance of
adoption. To understand how important certain constraints are for a farmer, fill in the following
table.

Begin by clarify the meaning of the constraint to the farmer, as explained under the table (E.g.:
grazing land is constraining when there is not enough biomass available to keep the livestock alive
throughout the year, and there is not enough biomass to be able to perform the wished amount of
cut and carry). Then ask if the farmer does or does not feel constrained. Note the answer in the
table, under the column ‘Constraint present’.

If the farmer identifies the item as a constraint, ask how the constraint plays a role, and when it is
most important (E.g. “Why is grazing land constraining? When is it most constraining?”. Typical
answers could be “There is more and more livestock on the pastures, there is not enough feed for
all the animals. This problem is worst at the end of the dry season, when the least feed is
vegetation is present”). This information should then be noted under the column ‘Explanation’.
After this, ask the farmer how important the constraint is to him/her. Note this in under the column
‘Importance’.

Once the mentioned constraints have all been filled in, ask the farmer if there are other production
constraints present in the farming system which have not been mentioned. For these constraints
follow the same routine as for the other constraints. This information should be filled in under the
row “Other”.

The last step, once all the production constraints have been filled in, is to ask the farmer what he
or she thinks the top three most important constraints are. For this it might be handy to repeat all
the constraints mentioned before. This information should be filled in the rightmost column.

Production constraint Constraint present (0 = no, Importance (code A) Explanation

1 = yes)

Area cropping land

Area grazing land

Number of livestock

Capital

Education

Technology

Other:

A) 1 = very low, 2= low, 3 = average, 4 = strong, 5 = very strong

B) Top three constraints: 1 = most important constraint, 2 = second most important constraint, 3

75
 = third most important constraint, 0 = constraint which is not in the top three

* Definition of the production constraints:

- Cropping land is the land used with the purpose of growing crops. It is constraining when there is
not enough land to feed the household and produce crop residues for the livestock. Or if there is
not enough land to generate products for the market, if desired.
- Grazing land is the area of common pastures. It is constraining if there is not enough area of
grassland or if the biomass production is too low for the livestock to eat from and survive
throughout the year. Or if biomass production is not sufficient to allow sufficient biomass for cut
and carry purposes.
- Education is the knowledge and skill which is gained through classes and contact with other
farmers, extension agents, etc.. Education is a constraint when the desired knowledge is not
available or accessible, or when access to school, the existence of (farming)groups and external
sources of information, the presence of extension agents, contact with fellow farmers is limited.
- Technology is seen as being improved agricultural practices, such as crop varieties, inputs, tools
and machinery. Technology is constraining when availability and accessibility is limited.
- Livestock are the cattle, small ruminants and other animals present on the farm. Livestock is
constraining when not enough manure is produced, not enough draught power is available, and the
livestock cannot serve as savings for difficult times.
- Capital are the financial assets. Capital is constraining when there is not enough money to buy the
necessary inputs for the farming system.

7.3.4.3.2 Farmer Goals

Farmer goals are the purposes for farming, the aims which farmers want to reach by farming. To
understand the goals of farmers, fill the following table.

Ask the question in bold, for every category. For supporting questions and to give help in giving an
answer, ask the questions which are not in bold. The answer to the question in bold should be of
the following range:

- not important,
- slightly important,
- important,
- very important
- or extremely important.

Write the answer in the column ‘Rank’. An explanation can be given in the column ‘Explanation’.
When the table is filled, ask the farmer if he/she has a goal that has not been mentioned before. If
this is the case write it down in the additional rows.

Category Statement Rank (code A) Explanation

Self- - How important is it for you to be self-subsistent?

Subsistence
-Do you provide all you own food? Do you depend on the
market to feed your family?

- Do you depend on others for subsistence (food, income,

etc.)?

76
Keeping land -How important is it for you to make sure you stay the
ownership owner of the land?

-Is the land yours or do you rent it?

Do you have security that you can keep farming on the land?

- Would you be ready to make sacrifices to keep the land

(money, time,...)

Maximizing - How important is it for you to increase the crop yield

yields or animal productivity as much as possible (kg/ha or
kg/animal respectively)?

- Do you choose crops for their ability to give maximum

yields?

- Do you apply external inputs to maximize yields?

- Do you feed animals for higher daily milk yield?

Maximizing - How important is it for you to increase the whole farm

whole farm production, and not only yields?
productivity
- De you try to maximize not only the crop yield and animal
production, but also the production of the by-products such as
crop residues or manure?

Maximizing - How important is it to you to maximize income?

income
- Do you select crops that will give you the most cash?

- Do you perform extra paid activities with the aim of

increasing income?

Maximize - How important is it to you to sell products on the

market market?
orientation
- Do you sell part of your harvest?

- Do you sell your products at the moment the prices are

best?

- Are the species you grow selected in the aim of selling them,
and does this influence your choice of crop to grow?

Minimizing - How important is it to minimize labour?

labour
- Is it worth paying for labour in order to save time?

Is it worth paying for labour to reduce the workload?

Achieving - How important is it to have enough labour when

activities in a needed?
timely
manner - Is there labour available when needed?

- Is it worth paying for labour in order to save time and

achieve activities in a timely manner?

77
Maximize - How important is the maximization of off-farm income
off-farm to you?
income
- Do you have other paid activities next to farming?

- Do other household members have other paid activities next

to farming?

- Is the off-farm income more important than the on-farm

income?

Sustainable - How important is it to you to think about the

management sustainable management of natural resources?
natural
resources - Do you think about the repercussion of farming actions on
the environment (soil, water,...)?

- Have you taken measures against declining yields or to keep

soil fertility high?

- Do you intend to keep the farm in a good state for the

coming generations?

Keeping - How important is it for you to keep a positive

social relationship with the inhabitants of the community and
cohesion how important a role do they play in your decision
making?

- Do you have good contact with other inhabitants and

farmers around you?

- Do you manage certain resources together with other

inhabitants (wells, manage your herds together,...)?

- Do you keep in mind the opinion of the community when you

start a new technology/intervention?

- Do your actions depend on the wills of other people?

Additional:

A) Rank: 1 =not important, 2 =slightly important, 3 =important, 4 = very important, 5 = extremely

important.

7.3.4.3.3 Attitudes
The farmer’s attitude towards farming is his view on farming. Fill in the table by taking the
following steps.

For every statement in the table, ask if the farmer agrees or disagrees, and how strong the feeling
is (strongly disagree, disagree, neutral, agree, strongly agree). Note the answer in the column
‘Rank’. Then ask for an explanation of the answer given, which can be noted in the column
‘Explanation’.

78
Number Statement Rank (code A) Explanation

1 I am proud to be a farmer

2 It is my destiny to be a farmer

3 Farming is fulfilling

4 Farming is a good way to meet my needs and those of my

family

5 There is no better investment than an investment in farming

6 If I had a choice I would be a (full-time) farmer

7 I would prefer if my children would not end up farmers

8 There is/are no problems in the current farming system

9 The current way of farming functions well enough to cover

my family’s needs now and in the future

10 There is no hope for farmers like us to improve our standard

of living

11 There is a need for new interventions and technologies to

improve the current farming system
12 I seek information on good farming practices from extension
officers / farm groups / other farms / external sources

13 I am open towards or practice new methods and

interventions

14 I am prepared to give labour and capital to do new

interventions and technologies

15 I want to avoid risk as much as possible

A) B) Rank: 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, 5 = strongly agree

7.3.5 Checklist for procedures.

7.3.5.1 Farmer selection

Set up a meeting with members who have knowledge about the farmers in the village you are
going to visit. Together organise the visits to the different households.

7.3.5.2 Interview set-up

- When at the household, conduct the first interview (with the household head or with the
main wife). This interview first interview within the household should include all three parts
of the questionnaire. It should take about 1h30.
- Then conduct the second interview within the household (with the household head or with
the main wife), this time only ask parts 1 and 3 of the questionnaire. This should take
about 1h.

79
7.4 Appendix IV: Determination of expert opinion
7.4.1 Questionnaire

7.4.2 Assessment of biomass enhancing interventions in Burkina Faso

I would like to ask for your help with a component of my MSc thesis, for which I rely on expert
knowledge. I would like you to fill out a short table, about the inputs needed for the
implementation of different interventions. This will only take a couple minutes, and I would be very
grateful if you would consider helping me out!

In my thesis I am assessing the adoptability of several interventions for different farm types based
on resource endowment. This is done through the evaluation of farmers’ constraints, goals and
attitude.

For the different farm types, I already know which constraints limit agricultural production. The
general idea is that the implementation of interventions should not further stress these constraints.
The constraints considered are:

- Cropping land (quality and quantity)

- Pasture (quantity)
- Education
- Technology
- Capital
- Labour

For the next step in my project I would like to assess the amount of resources needed (the cost)
for the implementation of different interventions. I will complement quantitative information from
the literature, with the opinion of experts like you.

Imagine a basic farm situated in Burkina Faso, with both cropping land and some cattle. To
implement a proposed intervention, is there a need for any extra cropping land, pasture, livestock,
education, technology, capital or labour, compared to the basic situation where the intervention is
not implemented?

The answer to this question, for several interventions, is what I am looking for. The table below
shows the interventions in the first column. Therefore the above stated question is asked for every
row of the table. Please give your answer on a scale from 0 to 2. A ‘0’ means there is no need for
extra resources, ‘1’ means there is a little need for extra resources, while ‘2’ means there is a big
need for extra cropping land, pasture, etc. In this way every cell in the table can be filled in.

Here is an example: for the intervention of fallow, there is an extra need for land, as land will be
taken out of production. Therefore in the cell ‘cropping land’*’fallow’ I would judge the answer as
being a ‘2’. For the implementation of dual-purpose crop the normal crop is replaced by a dual-
purpose variant, so I assume there is no need for extra cropping land, and my answer to this
question is of ‘0’.

The answers given do not need to be supported, I am asking for your personal opinion on the
situation. Please return the completed table to me by 11 July.

Thank you for your cooperation!

80
Intervention Cropping Pastur Livestock Education Technolog Capital Labour
land e y
Pasture planted with
native species

Pasture planted with

exotic species

Planting of improved,
dual-purpose cowpea and
sorghum varieties

Planting of drought
tolerant sorghum

Leaving 10-20% of the

land fallow for 5 years.

Collection herbaceous
biomass for animal feed

Collection herbaceous
biomass for selling

Collection and storage

crop residues for animal
feed

Collection of crop residues

for selling

Leaving of crop residues

on the field

Spreading of manure on
fields

Use of crop residues as

building material

Physical treatment of crop

residues for improved
palatability

Chemical treatment of
crop residues for
improved palatability

Rotational grazing in the

rangelands

Answer:

- 0 = no extra need compared to a basic situation

- 1 = a little need compared to the basic situation
- 2 = a big need compared to the basic situation

81
7.5 Appendix V: Typology

Table 1: Fraction of farmers within types growing specific crops. Green boxes represent crops taken
up in the simulation exercise, red boxes represent crops which cannot be taken up in the simulation
exercise
Crops
Type Millet Sorghum Maize Rice Fonio Cowpea Groundnut Sesame
T1 1 0.6 0.20 0.20 0.1 0.57 0.63 0.37
T2 1 0.89 0.48 0.23 0.02 0.73 0.75 0.55
T3 0.90 0.73 0.60 0.20 0.00 0.75 0.68 0.18
T4 1 0.63 0.42 0.16 0.11 0.74 0.58 0.26
T5 1 0.40 0.10 0.30 0.10 0.50 0.50 0.20
T6 0.86 0.86 0.43 0.14 0.00 0.43 1.00 0.57

Table 2: Fraction of farmers within types using manure, urea and fertilizer

Fraction of farmers using fertilizer

Type Compost and manure Inorganic fertilizer

T1 0.24 0.24
T2 0.37 0.39

T3 0.22 0.41
T4 0.11 0.28

T5 0.00 0.00
T6 0.25 0.00

Table 3: Percentage of farmers feeding concentrates to the livestock during the different seasons.
Boxes are green when more than 50% of the farmers feed concentrates to the livestock, and thus
concentrates are taken up tin the simulation exercise.
Percentage of farmers feeding the livestock with cakes

Type Rainy season Early Dry season Late dry season

Jun-Oct Nov-Jan Feb-May
T1 3 07 10
T2 20 07 27
T3 19 07 27
T4 21 11 47
T5 00 00 00
T6 00 71 43

82
7.6 Appendix VII: Impact of interventions
Table 1: Change (%) in crop yield based calorie production, SOC, livestock milk and weight production when implementing the interventions compared to the
baseline scenario, for types 1, 2 and 3. Green boxes represent positive changes, while red boxes represent negative changes. The abbreviation ‘n.a.’ stands for
not applicable.
Type
T1 T2 T3
Intervention Yield SOC Milk Weight Yield SOC Milk Weight Yield SOC Milk Weight
1 Grass quantity +1kg 0 0 n.a. 14.4 0 0 21.2 11.2 0 0 n.a. 10.8
2 Grass quantity +2kg 0 0 n.a. 26.6 0 0 36.9 18.6 0 0 n.a. 18.4
3 Grass quality +10% 0 0 n.a. 20.6 0 0 35.2 17.2 0 0 n.a. 7
4 Grass quality +20% 0 0 n.a. 74.5 0 0 73.3 51.2 0 0 n.a. 15
6 0.5kg Legume addition 0 0 n.a. 4.2 0 0 14.1 5.1 0 0 n.a. 1.2
7 1kg legume addition 0 0 n.a. 7 0 0 26.3 13 0 0 n.a. 2.8
5 2kg legume additon 0 0 n.a. 13.3 0 0 46.8 34.9 0 0 n.a. 6.5
6 0% Cowpea are 3.3 0.2 n.a. -21.7 1.5 0.1 -17.9 -7.5 2.5 0.4 n.a. -17.5
7 25% cowpea area -2.1 -0.1 n.a. 6 -2.6 -0.2 76.2 18 -2 -0.2 n.a. 7.2
8 50% cowpea area -7.5 -0.4 n.a. 17.6 -6.6 -0.6 177.1 59.1 -6.5 -0.3 n.a. 6
9 75% cowpea area -12.9 -0.8 n.a. 24.6 -10.6 -0.9 286 101.3 -11 -0.6 n.a. 21.3
10 100% cowpea area -18.3 -1.1 n.a. 83.4 -14.7 -1.3 549.6 132.5 -15.5 -0.9 n.a. 44.4
11 10% quantity CR sorghum -0.1 0 n.a. -0.5 -0.2 0 -0.4 -1.1 -0.2 0 n.a. -13.5
12 20% quantity CR sorghum -1.2 0 n.a. -0.5 -1.4 0 -1.2 -2.7 -1.7 0 n.a. -13.6
13 10% quality sorghum CR 0 0 n.a. 4.5 0 0 16.8 9.4 0 0 n.a. 6.6
14 20% quality sorghum CR -0.5 0 n.a. 7.6 -0.7 0 78.6 20.1 -0.8 0 n.a. 13.5
15 10% quantity CR sorghum -0.1 0 n.a. -0.5 -0.2 0 -0.4 -1.1 -0.2 0 n.a. -13.5
16 20% quantity CR sorghum -1.2 0 n.a. -0.5 -1.4 0 -1.2 -2.7 -1.7 0 n.a. -13.6
17 10% quality sorghum CR 0 0 n.a. 4.5 0 0 16.8 9.4 0 0 n.a. 6.6
18 20% quality sorghum CR -0.5 0 n.a. 7.6 -0.7 0 78.6 20.1 -0.8 0 n.a. 13.5
19 10% sorghum fallow -10.0 0.1 n.a. -0.5 -10.0 0.0 -1.2 -1.7 -10.0 -0.9 n.a. -1.7
20 20% sorghum fallow -20.0 0.1 n.a. -1.0 -20.0 0.1 -2.1 -6.3 -20.0 -0.9 n.a. -1.9
21 10% cowpea fallow -10.0 -0.1 n.a. -0.5 -10.0 -0.2 -1.2 -1.7 -10.0 -1.1 n.a. -1.7

83
22 20% cowpea fallow -20.0 -0.2 n.a. -1.0 -20.0 -0.4 -2.1 -6.3 -20.0 -1.2 n.a. -1.9
23 0% CR on field Base Base Base Base Base Base Base Base Base Base Base Base
24 25% CR on field 0 0.3 n.a. -0.9 0 0.2 -2.7 -1.7 0 0.2 n.a. -0.1
25 50% CR on field 0 0.3 n.a. -3 0.1 0.2 -6.7 -2.3 0.1 0.2 n.a. -0.4
26 75% CR ton field 0.1 0.3 n.a. -6.5 0.1 0.2 -14.1 -4.3 0.1 0.4 n.a. -0.8
27 100% CR on field 0.1 0.3 n.a. -12 0.1 0.2 -42.4 -13.1 0.1 0.4 n.a. -11.5
28 Manure 0% open Base Base Base Base Base Base Base Base Base Base Base Base
29 Manure 25% open 1.4 0.3 n.a. 0 0.4 0.2 0 0 0.2 0.2 n.a. 0
30 Manure 50% open 2.7 0.3 n.a. 0 0.8 0.2 0 0 0.5 0.4 n.a. 0
31 Manure 75% open 4.1 1.6 n.a. 0 1.2 0.2 0 0 0.7 0.4 n.a. 0
32 Manure 100% open 5.5 2 n.a. 0 1.6 0.2 0 0 0.9 0.4 n.a. 0
33 Manure 0%plastic Base Base Base Base Base Base Base Base Base Base Base Base
34 Manure 25% plastic 1.5 0.3 n.a. 0 0.4 0.2 0 0 0.3 0.2 n.a. 0
35 Manure 50% plastic 3 0.3 n.a. 0 0.9 0.2 0 0 0.5 0.4 n.a. 0
36 Manure 75% plastic 4.5 1.6 n.a. 0 1.3 0.2 0 0 0.8 0.4 n.a. 0
37 Manure 100% plastic 6 2 n.a. 0 1.8 0.2 0 0 1 0.4 n.a. 0
38 Feed quality +10% 0 0 n.a. 44.6 0 0 252.3 93.6 0 0 n.a. 42
39 Feed quality +20% 0 0 n.a. 409.6 0 0 520.6 168.2 0 0 n.a. 642

84
Table 2: Change (%) in crop yield based calorie production, SOC, livestock milk and weight production when implementing the interventions compared to the
baseline scenario, for types 4, 5 and 6. Green boxes represent positive changes, while red boxes represent negative changes. The abbreviation ‘n.a.’ stands
for not applicable.
Type
T4 T5 T6
Intervention Yield SOC Milk Weight Yield SOC Milk Weight Yield SOC Milk Weight
31 Grass quantity +1kg 0 0 32.3 9 0 0 n.a. n.a. 0 0 14.4 14.8
32 Grass quantity +2kg 0 0 58.8 13.9 0 0 n.a. n.a. 0 0 30.8 22.8
33 Grass quality +10% 0 0 65.8 17 0 0 n.a. n.a. 0 0 43.7 25.4
34 Grass quality +20% 0 0 144.2 54.5 0 0 n.a. n.a. 0 0 110 61.6
35 0.5kg Legume addition 0 0 29.7 7.3 0 0 n.a. n.a. 0 0 19 8.4
36 1kg legume addition 0 0 56.8 14.7 0 0 n.a. n.a. 0 0 37.4 22.4
37 2kg legume additon 0 0 103.1 28.7 0 0 n.a. n.a. 0 0 78.6 36.9
10 0% Cowpea are 4.2 0.3 -37.7 -9.5 4.2 0.4 n.a. n.a. 0 0 -0.5 0
11 25% cowpea area -0.4 0 7.7 3.1 -0.4 -0.4 n.a. n.a. -4.3 -0.4 101.9 29.9
12 50% cowpea area -5.1 -0.4 62.2 14.2 -5.1 -0.7 n.a. n.a. -8.5 -0.7 214.6 68.7
13 75% cowpea area -9.7 -0.7 107.3 26.6 -9.7 -1.1 n.a. n.a. -12.8 -1.1 330.5 93
14 100% cowpea area -14.3 -1.1 168.9 58.1 -14.3 -1.4 n.a. n.a. -17 -1.4 375 125.4
10% quantity CR
15 -0.1 0 0 -0.1 0 0 n.a. n.a. -0.1 0 -1.3 -0.1
sorghum
20% quantity CR
16 -1.3 0 -1.8 -0.2 0 0 n.a. n.a. -1.2 0 -0.7 -0.1
sorghum
17 10% quality sorghum CR 0 0 16.4 3.9 0 0 n.a. n.a. 0 0 19.2 8.1
18 20% quality sorghum CR -0.6 0 33.6 8.6 0 0 n.a. n.a. -0.6 0 18.8 18.2
10% quantity CR
15 -0.1 0 0 -0.1 0 0 n.a. n.a. -0.1 0 -1.3 -0.1
sorghum
20% quantity CR
16 -1.3 0 -1.8 -0.2 0 0 n.a. n.a. -1.2 0 -0.7 -0.1
sorghum
17 10% quality sorghum CR 0 0 16.4 3.9 0 0 n.a. n.a. 0 0 19.2 8.1
18 20% quality sorghum CR -0.6 0 33.6 8.6 0 0 n.a. n.a. -0.6 0 18.8 18.2
6 10% sorghum fallow -10.0 0.1 -2.7 -0.3 -9.9 0.0 n.a. n.a. -10.1 0.0 -8.8 -2.9
7 20% sorghum fallow -20.0 0.1 -6.2 -0.6 -19.9 0.0 n.a. n.a. -20.1 0.0 -10.5 -3.8
8 10% cowpea fallow -10.0 -0.1 -2.7 -0.3 -9.9 -0.1 n.a. n.a. -10.1 -0.1 -8.8 -2.9
9 20% cowpea fallow -20.0 -0.2 -6.2 -0.6 -19.9 -0.3 n.a. n.a. -20.1 -0.3 -10.5 -3.8

85
1 0% CR on field Base Base Base Base Base Base Base Base Base Base Base Base
2 25% CR on field 0 0.2 -6.8 -1.1 0 0.2 n.a. n.a. 0 0.2 -10 -1.9
3 50% CR on field 0 0.2 -13.7 -2.2 0 0.2 n.a. n.a. 0.1 0.2 -13.3 -2.6
4 75% CR ton field 0.1 0.2 -35.1 -7.7 0.1 0.2 n.a. n.a. 0.1 0.2 -24.9 -5.4
5 100% CR on field 0.1 0.2 -53.6 -12.6 0.1 0.2 n.a. n.a. 0.1 0.2 -32.1 -8
19 Manure 0% open Base Base Base Base Base Base Base Base Base Base Base Base
20 Manure 25% open 1.9 0.2 0 0 n.a. n.a. n.a. n.a. 0.7 0.2 0 0
21 Manure 50% open 3.8 0.2 0 0 n.a. n.a. n.a. n.a. 1.5 0.2 0 0
22 Manure 75% open 5.7 1.6 0 0 n.a. n.a. n.a. n.a. 2.2 0.2 0 0
23 Manure 100% open 7.5 1.6 0 0 n.a. n.a. n.a. n.a. 2.9 0.2 0 0
24 Manure 0%plastic Base Base Base Base Base Base Base Base Base Base Base Base
25 Manure 25% plastic 2.1 0.2 0 0 n.a. n.a. n.a. n.a. 0.8 0.2 0 0
26 Manure 50% plastic 4.1 0.2 0 0 n.a. n.a. n.a. n.a. 1.6 0.2 0 0
27 Manure 75% plastic 6.2 1.6 0 0 n.a. n.a. n.a. n.a. 2.4 0.2 0 0
28 Manure 100% plastic 8.2 1.6 0 0 n.a. n.a. n.a. n.a. 3.2 0.2 0 0
29 Feed quality +10% 0 0 194.4 68 0 0 n.a. n.a. 0 0 286.5 80.4
30 Feed quality +20% 0 0 508.5 124.4 0 0 n.a. n.a. 0 0 571 149

86
7.7 Appendix VIII: Interviews

Table 1: Number of farmer within each type, based on WLE dataset and on the data
collected through interviews
Number of households in Farm type based on WLE Farm Type based on interviews
type dataset
T1 9 6
T2 9 8
T3 9 11
T4 6 7
T5 4 7
T6 2 0

Table 2: Average number of household members in the different types

Average number of household members T1 T2 T3 T4 T5

Total 13 20 15 16 11
Male adults 3 5 3 3 2
Female adults 3 5 4 4 2
Boys 3 6 4 5 4
Girls 4 5 4 4 4

Table 3: Average household characteristics for the farm types. number of cattle and small
ruminants in possession of the different farm types.
Farm type

Household characteristics T1 T2 T3 T4 T5

Cattle 0 7 3 3 0

Small ruminants 3 20 4 12 0

Total land area 2.2 9.7 5.5 2.9 5.5

Cereal land area 1.8 6.9 4.3 2.5 4.6

Table 4: Livestock possession by men and women in the farm types

Farm type
Livestock category Owner T1 T2 T3 T4
Cattle Male 100 75 90 86
Female 0 13 0 0
Both 0 13 10 14
Small ruminants Male 67 25 50 29
Female 17 13 10 0
Both 17 63 40 71

87
Table 5: P-value outcomes of fisher test for the
assessment of significant difference in the
proportion of respondents classifying elements as a
constraint, between different types. Green boxes
represent a significant difference. The abbreviation
‘n.a.’ stands for not applicable, used when the
proportions are in all cases the same.

Cropping land
T1 T2 T3 T4 T5
T1
T2 0.705
T3 0.410 0.147
T4 0.665 0.260 1.000
T5 0.665 0.260 1.000 1.000
Grazing land
T1 T2 T3 T4 T5
T1
T2 0.059
T3 0.075 0.088
T4 1.000 1.000 0.039
T5 0.001 0.001 0.039 0.000
Livestock
T1 T2 T3 T4 T5
T1
T2 n.a.
T3 n.a. n.a.
T4 1.000 0.467 0.389
T5 0.225 0.090 0.051 0.596
Technology
T1 T2 T3 T4 T5
T1
T2 0.053
T3 1.000 0.065
T4 0.483 0.399 0.547
T5 n.a. 0.045 0.611 0.481
Education
T1 T2 T3 T4 T5
T1
T2 0.088
T3 0.635 0.147
T4 1.000 0.118 1
T5 0.462 0.007 0.141 0.481

88
Table 6: Fisher test for significant association between gender and the proportion of 'yes'
and 'no' answers. The test is not run for capital as 100% of the respondents say capital is
constraining.
Constraint
Test Cropping Grazing Livestock Capital Education Technology
variables land land
P-value 0.209 0.493 1.000 n.a. 0.075 1.000

Significant No No No n.a. No No
difference

Table 7: P-value of fisher test for association between the

farm types and the score given to the goals. Green boxes
represent a significant difference.
Self- subsistence
T1 T2 T3 T4 T5
T1
0.003
T2
1.000 0.002
T3
1.000 0.020 0.885
T4
1.000 0.001 1.000 0.855
T5
Land ownership
T1 T2 T3 T4 T5

T1
0.098
T2
0.184 0.306
T3
0.681 0.147 0.360
T4
0.049 0.158 0.953 0.187
T5
Maximising yields
1 2 3 4 5

1
0.035
2
0.133 0.880
3
0.397 0.530 0.383
4
0.003 0.081 0.221 0.007
5
Maximising whole farm production
1 2 3 4 5

1
0.084
2
0.159 0.937
3
0.577 0.545 0.854
4
0.001 0.008 0.000 0.001
5
Maximising income
1 2 3 4 5
1
0.003
2
0.051 0.031
3
0.108 0.122 0.787
4

89
 0.010 0.000 0.172 0.077
5
Maximising market orientation
1 2 3 4 5

1
0.280
2
0.000 0.001
3
0.101 0.342 0.016
4
0.001 0.000 0.706 0.003
5
Minimising labour
1 2 3 4 5

1
0.085
2
0.099 0.110
3
0.510 0.159 0.700
4
0.077 0.443 0.635 0.345
5
Achieving activities in a timely manner
1 2 3 4 5

1
0.549
2
0.193 0.462
3
0.408 0.738 0.614
4
0.003 0.099 0.363 0.147
5
Maximising off-farm income
1 2 3 4 5
1
0.013
2
0.026 0.141
3
0.004 0.080 0.573
4
0.001 0.054 0.589 0.383
5
Sustainable management natural resources
1 2 3 4 5

1
0.313
2
0.497 1
3
0.577 0.038 0.088
4
0.234 0.840 0.917 0.026
5
Maximising Social cohesion
1 2 3 4 5

1
0.133
2
0.050 0.337
3
0.132 0.130 0.009
4
0.020 0.805 0.297 0.044
5

90
Table 8: Fisher test for association between the gender of the
respondents and the proportion of answers given for the
scoring of different goals
Goal P-value Significant
difference?
Self-subsistence 0.879 No
Land ownership 0.760 No
Yield 0.843 No
Whole farm production 0.603 No
Income 0.417 No
Market orientation 0.419 No
Labour 0.224 No
Timely manner 0.434 No
Off-farm income 0.208 No
Sustainable management natural 0.820 No
resources
Social cohesion 0.135 No

91
Cut-off

Figure 15: Dendogram with the classification of the different farm types based on the answering of the attitude statements

92
Table 9: Fisher test for association between interviewer and
the clusters in which the farms are classified
Interviewers between whom P-value Significant
difference is tested difference?
1 and 2 0.756 No
1 and 3 0.000 Yes
2 and 3 0.000 Yes

93