Master’s Thesis

Master’s Thesis (Academic Year 2018)
Crop modelling of rain-fed maize response to minimum-tillage and cattle manure: A case for subsistence farmers in Swaziland.

Keio University
Graduate School of Media and Governance
MANYATSI, Lindelwa Lomaqhawe
81625531

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Declaration of honourI, MANYATSI LINDELWA LOMAQHAWE, declare that this thesis titled, ‘Crop modelling of rain-fed maize response to minimum-tillage and cattle manure: A case for subsistence farmers in Swaziland’ and the work presented in it are my own. I confirm that:
This work was done wholly while in candidature for a Masters degree in Media and Governance at Keio University.
Where any part of this thesis has previously been submitted for a degree or any other qualification at Keio University or any other institution, this has been clearly stated.
Where I have consulted the published work of others, this is always clearly attributed.
Where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work.

I have acknowledged all main sources of help.
Where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself.
Signed:
_____________________________
Date:
_____________________________

DedicationThis research is dedicated to my late father, Dr. David Manyatsi; the funniest man I had the opportunity of knowing, a source of inspiration for me and my siblings. He never ceased to show us the importance of education by leading by example.

A special dedication to my mother Dumisile Manyatsi for being my rock and allowing me to be the author of the story of my life.

Ngiyabonga boS’candza.

Acknowledgments
I would first and foremost like to thank the Almighty for the wisdom, protection, strength and courage to successfully complete this research. I owe a lot a gratitude to the Japan International Cooperation Agency (JICA) for the opportunity to be part of the ABE Initiative and being able to study in Japan.

My sincere gratitude goes to my main supervisor, Professor Wanglin Yan and my co-supervisors, Professor Lynn Thiesmeyer and Professor Tomohiro Ichinose, for their invaluable advice and guidance throughout this research study.

To the staff of Keio University, thank you for your efforts in making my studies pleasant and comfortable. I am also grateful to my colleagues, whom I have learnt a lot from, for their contributions towards this research.

A special appreciation to the main contributors in this research study, the subsistence farmers in Siphofaneni for their cooperation and willingness to be interviewed. To the Siphofaneni agricultural extension officers who made my fieldwork possible, I thank you.

I am indebted to Tomas Kapiye for his endless support throughout this research work.

And last but definitely not the least, I would like to express my sincere gratitude to my siblings for their emotional and moral support during my time in Japan; without them, I would not have been able to complete this research.

Abstract of Master’s Thesis Academic Year 2018
Crop modelling of rain-fed maize response to minimum-tillage and cattle manure: A case for subsistence farmers in Swaziland.

Maize is the main crop in Swaziland and is a vital food source that needs to be cultivated annually in order to maintain food security. Maize cultivation and the resultant yields have been negatively affected by many problems among which soil fertility is a major issue. To exacerbate this problem, maize is cultivated mainly by resource-poor subsistence farmers who invest very little in the application of inorganic fertilisers to their crop. This research proposes the use of cattle manure which is already readily available to the farmers and minimum tillage methods to reduce tractor hire costs and improve the soil’s biophysical properties. In order to be able to convince the farmers on the adoption of the proposed methods of maize cultivation, it is imperative to have concrete, scientific information to support the proposed benefits of the farming methods. These can be done through field experiments but these require a lot of labour, time and money. Crop simulation models provide an excellent alternative. It is for these reasons that the need for this research was established.

The objective of this research was investigate the potential of increasing maize yields for subsistence farmers through integrating minimum tillage and cattle manure as a soil amendment, as a measure to improve soil fertility. This was done by using the DSSAT-CERES Maize model to simulate maize crop responses to minimum-tillage and cattle manure. There were six treatments with 0, 5 and 10 tonnes of cattle manure with tractor and minimum tillage, replicated 15 times up to the year 2030. Inputs to the model included site information, daily weather data, soil properties, soil initial conditions, field management and crop phenotype data. The field management data was obtained through questionnaire interviews conducted in the study area, Siphofaneni consistuency in Swaziland with the respondents being 60 subsistence farmers.

The model was successful in simulating yield values with R2 =74%, RMSE = 43kg/ha and d = 0.82. The yield results showed that the yield for the year 2030 would be highest for minimum tillage with 10t/ha of cattle manure at 8.084t/ha while the lowest would be tractor tillage with no cattle manure at 1.480t/ha. The tillage method did not seem to have a significant impact on the yield amount.The results show that subsistence farmers in Swaziland could positively exploit the application of cattle manure to increase maize yield and net benefits and that the DSSAT CERES-Maize model may be applied with confidence to study effects of tillage practices and soil amendments on maize yield.

Keywords:
1. Swaziland, 2. maize, 3. cattle manure, 4. minimum tillage, 5. DSSAT-CERES Maize model
Keio University
Graduate School of Media and Governance
MANYATSI, Lindelwa Lomaqhawe
Table of Contents TOC o “1-3” h z u Declaration of honour PAGEREF _Toc517951023 h iiDedication PAGEREF _Toc517951024 h iiiAbstract PAGEREF _Toc517951025 h vList of tables and figures PAGEREF _Toc517951026 h viiiList of Appendices PAGEREF _Toc517951027 h viiiAbbreviations PAGEREF _Toc517951028 h viiiCHAPTER 1: INTRODUCTION PAGEREF _Toc517951029 h 11.1Background PAGEREF _Toc517951030 h 11.2Research problem PAGEREF _Toc517951031 h 21.3 Significance of the study PAGEREF _Toc517951032 h 31.3Nominal definitions PAGEREF _Toc517951033 h 41.4Research objective and question PAGEREF _Toc517951034 h 51.5Description of study area PAGEREF _Toc517951035 h 51.6Thesis layout PAGEREF _Toc517951036 h 8CHAPTER 2: LITERATURE REVIEW PAGEREF _Toc517951037 h 92.1Overview PAGEREF _Toc517951038 h 92.2Literature PAGEREF _Toc517951039 h 92.2.1Low maize yields in subsistence farming PAGEREF _Toc517951040 h 92.2.2Low maize yields and soil fertility PAGEREF _Toc517951041 h 102.2.3Minimum tillage farming PAGEREF _Toc517951042 h 112.2.4Effectiveness of cattle manure as soil amendment PAGEREF _Toc517951043 h 122.2.5Maize yield simulation using DSSAT PAGEREF _Toc517951044 h 132.3Hypothesis and Conceptual framework PAGEREF _Toc517951045 h 17CHAPTER 3: METHODOLOGY PAGEREF _Toc517951046 h 183.1Introduction PAGEREF _Toc517951047 h 183.2Data requirements and acquisition PAGEREF _Toc517951048 h 183.2.1Socio-economic data of the farmers PAGEREF _Toc517951049 h 193.2.2Daily weather data and pre-processing PAGEREF _Toc517951050 h 193.2.3Soil types PAGEREF _Toc517951051 h 203.2.5Maize crop genotype characteristics PAGEREF _Toc517951052 h 203.3Design of method PAGEREF _Toc517951053 h 203.3.1Establishing a baseline of actual yields PAGEREF _Toc517951054 h 213.3.2Model calibration PAGEREF _Toc517951055 h 213.3.3 Model validation PAGEREF _Toc517951056 h 223.3.4Projection of maize yields for the year 2030 PAGEREF _Toc517951057 h 223.3.5Feasibility and practicality analysis PAGEREF _Toc517951058 h 233.4Selection of the model PAGEREF _Toc517951059 h 264.1 Brief overview of fieldwork PAGEREF _Toc517951060 h 294.2 Fieldwork activities PAGEREF _Toc517951061 h 294.3. Problems encountered PAGEREF _Toc517951062 h 32CHAPTER 5: FINDINGS AND DISCUSSION PAGEREF _Toc517951063 h 345.1Brief overview PAGEREF _Toc517951064 h 345.2Socio-economic status of farmers in the study area PAGEREF _Toc517951065 h 345.3Fieldwork results PAGEREF _Toc517951066 h 375.4DSSAT CERES-Maize results PAGEREF _Toc517951067 h 385.4.1Model validation PAGEREF _Toc517951068 h 385.4.2Yield projection and nitrogen dynamics PAGEREF _Toc517951069 h 395.5Feasibility of implementation analysis results PAGEREF _Toc517951070 h 43CHAPTER 6: CONCLUSION AND RECOMMENDATIONS PAGEREF _Toc517951071 h 456.1Conclusions PAGEREF _Toc517951072 h 456.2Implications of study PAGEREF _Toc517951073 h 466.3Summary PAGEREF _Toc517951074 h 47APPENDICES PAGEREF _Toc517951075 h 47Questionnaire PAGEREF _Toc517951076 h 47Raw data PAGEREF _Toc517951077 h 47REFERENCES PAGEREF _Toc517951078 h 47

List of tables and figures TOC c “Table” Table 1. The causes of low maize yields for subsistence farmers in Swaziland PAGEREF _Toc517949986 h 10
Table 2. Description of the parameters used by the DSSAT CERES-Maize model in simulating maize crop growth. PAGEREF _Toc517949987 h 15
Table 3. Table showing the various combinations of treatments that were applied in the study PAGEREF _Toc517949988 h 23
Table 4. The amount of labour requirements for the implementation of the proposed methods PAGEREF _Toc517949989 h 25
Table 5. Other crop simulation models that are available and their strengths and limitations. PAGEREF _Toc517949990 h 28
Table 6. The comparison of the sizes of the areas planted by subsistence farmers and the harvest area size PAGEREF _Toc517949991 h 30

TOC c “Figure” Figure 1. Distribution of crops grown by subsistence farmers in Swaziland (Source: Ministry of Agriculture 2016) PAGEREF _Toc517949563 h 2
Figure 2. The national annual maize, consumption requirements and rainfall (Source National Maize Corporation 2016, Meteorology Department 2016) PAGEREF _Toc517949564 h 3
Figure 3. Map of Swaziland’s position in Southern Africa and the location of the study area PAGEREF _Toc517949565 h 6
Figure 4. Average monthly minimum and maximum temperatures PAGEREF _Toc517949566 h 7
Figure 5. Average monthly rainfall in Siphofaneni consistuency PAGEREF _Toc517949567 h 8
Figure 6. A screenshot of the DSSAT model showing the various functions available. PAGEREF _Toc517949568 h 16
Figure 7. Conceptual framework of the variables in the study and how they fit together PAGEREF _Toc517949569 h 17
Figure 8. A summary of the methodology applied in the study. PAGEREF _Toc517949570 h 21
Figure 9. The distribution of the ages (left) and gender (right) of the farmers who were respondents to the questionnaire interviews PAGEREF _Toc517949571 h 31
Figure 10. The sizes of the farmers’ household. PAGEREF _Toc517949572 h 35
Figure 11. Levels of education of the farmers. PAGEREF _Toc517949573 h 36
Figure 12. Maize cultivation field sizes of the the farmers. PAGEREF _Toc517949574 h 36
Figure 13. The distribution of the herd sizes owned by the farmers. PAGEREF _Toc517949575 h 38
Figure 14. The accuracy of the model in simulating maize yields PAGEREF _Toc517949576 h 39
Figure 15. The maize grain development under the different treatments. PAGEREF _Toc517949577 h 41
Figure 16. The amount of nitrogen available in the soil at different times during the growing season. PAGEREF _Toc517949578 h 42
Figure 17. The amount of nitrogen absorbed by the maize crop at various stages of the growing period. PAGEREF _Toc517949579 h 42
Figure 18. The reasons for farmers to be sceptic about implementing the proposed methods (left) and reasons they would be willing to implement them (right). PAGEREF _Toc517949580 h 44

List of AppendicesAbbreviationsCHAPTER 1: INTRODUCTIONBackgroundGlobally, about 17% of agricultural lands are irrigated, and they produce 40% of the total cereal production. Irrigation is also associated with negative economic, environmental, social and political consequences as well as positive ones. For this reason, there is a great deal of interest in meeting future food needs through rain-fed agriculture which currently produces 60% of the world’s food, as a partial substitute for irrigationCITATION Dro01 m Coo08 l 2057 (Droogers 2001, Cooper 2008). This is especially true for sub-Saharan Africa where currently about 90% of the staple food production will continue to come from rain-fed agriculture under subsistence farming CITATION Ros2b l 2057 (Rosegrant 2002b).
Summer maize is the main crop in Swaziland (Figure 1) and is a vital food source that needs to be cultivated annually in order to maintain food security. Its supply for water is solely precipitation. Maize cultivation and the resultant yields have been negatively affected by many problems such as insufficient fertility, dry spells, lack of farming inputs etc. CITATION Gen17 l 2057 (Geng 2017). In 2016, the national maize yields dropped by over 60% and prompted the government to declare a national state of emergency.Ref. This research, however will focus on the problems of insufficient fertility and the lack of farming inputs to a lesser extent. Nitrogen is the most limiting soil nutrient in the production of maize in Swaziland Ref . To exacerbate this problem, maize is cultivated mainly by resource poor subsistence farmers who invest very little in the application of inorganic fertilizers.

Attempts to increase maize production for subsistence farmers should therefore seek methods to improve soil fertility coupled with practices that enhance in-situ soil moisture retention but at the same time using already available resources and inexpensive methods of farming. Identifying these methods usually requires years of field experiments in order to make meaningful deductions and this can be expensive and time consumingCITATION Bha l 2057 (Bhatia 2008). Crop simulation models offer an excellent alternative approach. The Decision Support System for Agrotechnology Transfer CITATION Jon03 l 2057 (Jones 2003) is one of the extensive application and highly praised crop model. It is an non-linear dynamic model that simulates daily crop growth and yield as a function of input parameters: weather, soil, crop characteristics and field management practices CITATION Hoo17 l 2057 (Hoogenboom 2017).

The CERES (Crop Estimation through Resource and Environment Synthesis) model simulates the growth and development of crops including maize. DSSAT-CERES has been used worldwide under a wide range of climatic and soil conditions and it has been proven as a useful tool for determining management practices due to its outstanding performance in simulating the growth, biomass accumulation, yield and water resource use efficiency in response to various environmental factors and management.

Figure SEQ Figure * ARABIC 1. Distribution of crops grown by subsistence farmers in Swaziland (Source: Ministry of Agriculture 2016)Research problemMaize, the staple food of Swaziland is the most important crop cereal crop and is grown by 80% of subsistence farmers which translates to a majority of the households in the country. Over the past few years, erratic weather and droughts have contributed to chronic food insecurity and gradually weakening livelihoods. Due to the heavy dependence on an annual maize crop under rain-fed conditions, households’ vulnerability to erratic weather is very high. According to MoFA (Ghana), low soil fertility is also a major reason for low productivity of maize. The farmers in the study area attributed their low yields to dry spells, low soil fertility and the high cost of inputs. To exacerbate this problem, maize is cultivated mainly by resource-poor subsistence farmers who invest very little in the application of inorganic fertilisers to their crop.

Up until the year 2000, Swaziland was routinely harvesting over 100,000 tons of maize but in 2016, the yield was estimated to be at a well below-average level of 33 000 tonnes, 60 percent down from the previous year’s output. The production season 2006/07 to 2010/11 maize production was steadily increasing positively, and in 2011/12 it dropped and slowly picked in 2012/13, and reached its peak in 2013/14 (bumper harvest). Even the 2013/14 bumper harvest could not meet the national requirement. The 2015/16 season was the worst season ever recorded in the history of maize production in the country CITATION Swa16 l 2057 (Swaziland Vulnerability Assessment Committee 2016).The sharply lower production is on account of the El Niñorelated drought that caused a contraction in the area planted and lower yields CITATION Foo16 l 2057 (Food and Agricultural Organisation 2016).

Figure SEQ Figure * ARABIC 2. The national annual maize, consumption requirements and rainfall (Source National Maize Corporation 2016, Meteorology Department 2016)The key is to improve the soil fertility status through a combination of soil amendments and field management practices that enhance in-situ moisture conservation under rain-fed maize farming despite low incomes. This prompts the usage of resources that are already available to the subsistence farmers.

1.3 Significance of the studyEvery developing economy has the agriculture sector as the irreplaceable pillar and the same applies in Swaziland. Rain-fed maize production by subsistence farmers in Swaziland is the maize source of food for most of the population. The livelihoods of the country’s population depend on the stable and sustainable ability of the farmers to continually produce high yields year after year.

Swaziland as a signee of the Sustainable Development Goals has a responsibility to deliver in terms of achieving these goals. This study will be placing emphasis on SDG 2: Zero hunger, which aims to “end all forms of hunger and malnutrition by 2030, making sure all people – especially children – have access to sufficient and nutritious food all year round”. This involves promoting sustainable agricultural practices: supporting small scale farmers and allowing equal access to land, technology and markets CITATION UND18 l 2057 (UNDP Swaziland 2018). The information that will be generated through this study will reveal the potential maize yield increase that can be achieved by subsistence farmers through integrating minimum tillage and cattle manure as a soil amendment. This information will then be used to inform the decision makers on whether it would be beneficial to encourage and promote these methods of farming. This study is even more important because it is being applied in Swaziland, where the farmers lack finances to hire tractors and Nguni cattle are traditionally bred by most farmers on Swazi Nation Land. Minimum tillage farming and cattle manure are free/low cost and with proper extension work, research outcome could help achieve higher maize yields and reduce input costs for subsistence farmers and also improve food security of rural families.

There is no previous study that has attempted to evaluate the use of a mechanistic crop simulation model (DSSAT) for estimating actual yields in Swaziland using the same set of internal parameters across all site-years, and based on measured weather, soil, and management data.

Nominal definitionsSubsistence farmers – self-sufficiency farmers who focus on growing enough food to feed themselves and their families. The output is mostly for local requirements with little or no surplus trade. CITATION Per13 l 2057 (Perera 2013) CITATION Lin11 l 2057 (Lininger 2011).

Minimum tillage farming – describes a form of cropping system which does not use mechanical tillage of the soil for crop establishment. This system avoids disturbing the soil with tools like chisel plows, field cultivators, disks, and plows CITATION Fre12 l 2057 (Freidrich 2012).

Soil amendments – inorganic and organic substances mixed into the soil for achieving a better soil constitution i.e. improve its physical properties, regarding plant productivity. Organic soil amendments usually are derived from plants or plant products that occur naturally or are the by-products of processing plants or mills or waste disposal plants CITATION Sta10 l 2057 (Stauffer 2010). For the purposes of this research the cattle manure will be considered as the chosen soil amendment.

Crop simulation model – a simulation model to predict growth, development, and yield of crops that uses various inputs such as the local environmental conditions including weather and soil physical and chemical characteristics, crop management, and genetic information CITATION Kri11 l 2057 (Krishnan 2011). For the purposes of this study, the CERES-Maize model from the Decision Support System for Agrotechnology Transfer (DSSAT) package was used.

Research objective and questionGeneral objective
To investigate the potential of increasing maize yields for subsistence farmers through integrating minimum tillage and cattle manure as a soil amendment by using the DSSAT-CERES Maize model to simulate maize crop responses to minimum-tillage and cattle manure.
Specific objectives
Determine the current annual maize yields of the farmers in the study area.

Determine the field management practices of the farmers i.e. tillage methods and soil amendments application.

Simulate the maize yield that can be achieved by integrating no-tillage and cattle manure as a soil amendment
Research question
What are the current annual maize yield amounts in Siphofaneni?
What is the maize yield response to a combination of minimum tillage and cattle manure as a soil amendment?
Description of study areaThis section details the description of the study area where the research was carried out.
Siphofaneni consistuency is home to 23 448 people out of the approximately 1.3 million people in Swaziland. The dominant activity is susistence farming.

Figure SEQ Figure * ARABIC 3. Map of Swaziland’s position in Southern Africa and the location of the study areaGeographic location
Swaziland is a landlocked nation bordered by South Africa to the North, West and South and Mozambique to the East with an area of 17,364km2, of which 15-20% is arable. There are four distinct geographical regions or ecological zones: Highveld – the westernmost, mountainous and forested belt; Middleveld – where most of Swaziland’s agricultural activities take place; Lowveld – where much of the nation’s cattle farming and cultivation of export crops occurs and Lubombo – where mixed farming occurs.

The study area, Siphofaneni consistuency is located in the low-veld of Swaziland, in the Lubombo region. The population is approximately 23 448 on a land area of 539 km². Geographically, it is located between latitudes 2630S and 2641S and longitutes 3130E and 3130E.

Climate
Siphofaneni has a tropical climate. It is very hot during the summer season and cold in winter. Summer average temperature range from 20? to 30? and winter temperatures range from 7? to 25?. Siphofaneni is a low lying area with and elevation of between 190 – 280 metres above sea level.

According to the Swaziland Meteorology Services the average annual rainfall for the years 1996 to 2017 was 548mm in the study area. The area is characterized by long dry spells. Summer rainfall usually falls as thunderstorms with high intensity over a short period of time. The rainy season falls between the months October and March.

Figure SEQ Figure * ARABIC 4. Average monthly minimum and maximum temperatures
Figure SEQ Figure * ARABIC 5. Average monthly rainfall in Siphofaneni consistuencyThesis layoutThe thesis of this study is presented in six chapters. The first chapter gives an introduction of the study by highlighting the background of the low maize productivity and food security issue in Swaziland as well as previous interventions and government policies that have tried to tackle the issue. It continues to present the research problem, objectives and the significance of the study.

Chapter 2 covers the literature review on minimum tillage farming, soil amendments and describes theoretical framework of the research topic. The chapter also describes the hypothesis and the scope (assumptions and limitations) of the study.

The methodology is presented in chapter 3 which also includes the design of the methods adopted to achieve the specific objectives and the analysis procedures. The fifth chapter presents the findings while chapter 6 is a discussion of these results and their implications. Lastly, is the conclusion which is a summary of the thesis and suggestions for future research.

CHAPTER 2: LITERATURE REVIEWOverviewIn light of climate change and the ever increasing population, low maize remain a huge issue globally but more specifically in Sub-Saharan Africa where maize is the staple food. This literature review will cover the general concept of the problem, its causes and impacts. As this study aims to investigate the potential maize yield increase that can be achieved by integrating minimum tillage agriculture and cattle manure as a soil amendment, this section will also give a critical review of previous cases where these methods were used.

LiteratureLow maize yields in subsistence farmingMaize remains the dominant crop in subsistence farming in Swaziland due to its cultural importance. Even in the face of decreasing maize yields, farmers favour cultivating maize compared to other persistent crops like sorghum, beans etc. According to the South African Department of Agriculture, Successful maize production depends on the correct application of production inputs that will sustain the environment as well as agricultural production. These inputs are, inter alia, adapted cultivars, plant population, soil tillage, fertilisation, weed, insect and disease control, harvesting, marketing and financial resources.

There are quite a number of causes of low yields in subsistence farming of maize in Swaziland, some of which are illustrated in Table 1 below.

Main problem Causes from farmers Mid-level causes Low level causes
 
 
 
 
 
 
 
Low maize yields Low rainfall Location
Geography
Deforestation (upwind) Firewood, building material, more farmland
Poor soil fertility Monoculture
Soil erosion
Overgrazing (field ; pasture)
Topography Ease of weeding
Steep slopes
Too much livestock
Lack of inputs Unemployment
Land tenure system Outdated farming methods Lack of extension services
Low level of education No motivation, low salary
Poverty
Culture Sense of security in growing maize
Fear of land being taken away Staple food
Land tenure system
Table SEQ Table * ARABIC 1. The causes of low maize yields for subsistence farmers in SwazilandLow maize yields and soil fertilityThe International Journal of Science mentions that although drought may be the best known barrier to successful crops in Africa, the poor soils are a huge part of the equation. Farmland in Africa has been robbed of chemicals such as nitrogen, phosphorus and potassium, which are vital for plant growth. And these have not been replaced with organic and chemical fertilizers, as they are in most other countries, because of the expense CITATION Pea06 l 2057 (Pearson 2006)Researchers Julio Henao and Carlos Baanante of the International Centre for Soil Fertility and Agricultural Development (IFDC) did a study where they totted up the nutrients that feed the soil, and subtracted losses from leaching, volatilization, erosion and crop uptake. They plugged these numbers, which differ from region to region, into a computer model to come up with a picture for the entire continent over time. The study showed that around three-quarters of the continent’s land is severely degraded of nutrients, with many regions losing as much as 60 kilograms of nutrients per hectare each year. As a result, the production of cereals in sub-Saharan Africa has stagnated at one tonne per hectare compared with around three tonnes in the rest of the world.

A hectare of maize crops absorbs 191kg of nitrogen, 89kg of phosphorus and 235kg of potassium to produce 9.5 tonnes of grain yield on average CITATION Rai17 l 2057 (Raimi 2017)It is because of these nutrient deficient soils that subsistence farmers in Swaziland are unable to produce enough maize to feed their families. Poor soil fertility is one of the fundamental causes of low agricultural production in hunger-endemic areas like Swaziland CITATION Har06 l 2057 (Hartemink 2006). It is important now more than ever, to boost the fertility of the soil if the country is to improve its peoples’ food security.

Minimum tillage farmingThe practice of intensive tillage can be effective for weed management. However, it increases the risk of soil erosion. Frequent tillage can increase land degradation, soil erosion and soil compaction, and these are major challenges that smallholder farmers are facing in crop production CITATION Ham05 l 2057 (Hamza 2005). Minimum tillage practices have significant potential to reduce expenses and the potential negative environmental effects caused by intensive tillage operations CITATION Dar14 l 2057 (Darby 2014).

Bescanca (2006) noted that minimum tillage practices simultaneously conserve soil and water resources, reduce farm energy and increase or stabilise crop production which often leads to positive changes in the physical, chemical and biological properties of the soil. Soil physical properties that are influenced by conservation tillage include bulk density, infiltration and water retention. Improved infiltration of rainwater into the soil increases water availability to plants, reduces surface runoff and improves groundwater recharge. Reduced soil cultivation reduces farm energy requirements and overall farming costs as less area has area has to be tilled CITATION Mup l 2057 (Mupangwa n.d.).

A study was conducted by the University of Vermont where the objective was to evaluate the impacts of minimum tillage methods, namely no-till, vertical-till, and strip-till on maize yield and quality. The average yield for the minimum tillage trial was 48.9 tonnes/ha, which is very good when compared to yields of similar relative maturity maize planted by means of conventional tillage. The average yield of this same variety in conventional tillage trials was 39.1 tonnes/ha. This was attributed to the fact that the study area soil had been in reduced tillage for four years and it is likely that the soil had improved to a point where higher yields could be supported.

A research was done in Romania where the effects of the three tillage systems, no-tillage, minimum tillage and tractor tillage on the yield of wheat, maize and soybean, obtained on an argic Faeoziom from the Somes Plateau where investigated. The experiment was done in split-plot design and replicated over three years for each crop. It was found that wheat had equal yields between 3745-3856 kg/ha, with no significant differences between tillage systems. Maize responded better to the soil loosening, at the mobilization of soil fertility and nutrient mineralization of tractor tillage, providing a yield of 6310 kg/ha. The yield was between 5890-6145 kg/ha for minimum tillage and the lowest yield for no-tillage at 5774 kg/ha, being distinct significant negative. The soybean crop productions were between 2112-2341 kg/ ha, being significant positive for minimum tillage and no-tillage. It was concluded that the positive benefits of a tillage method on the yield depends of the crop CITATION Rus11 l 2057 (Rusu 2011).

A similar study was conducted by researchers at the University of Agriculture in Pakistan where a field experiment was conducted to evaluate the effect of different tillage systems (minimum, deep and shallow tractor tillage) and mulch levels on soil physical properties and growth of maize. The results showed with regards tillage, the maximum mean value of grain production was observed for observed deep tillage at 5.57 followed by 5.38 tonnes/ha for shallow tillage and in the case of minimum tillage, the minimum mean value was 5.37 tonnes/ha CITATION Khu06 l 2057 (Khurshid 2006).
Effectiveness of cattle manure as soil amendmentThe use of manure is an old technology that is appropriate for small-scale farmers in Swaziland, as most farmers practice mixed livestock and crop farming. Regardless of the fact that the use of manure dates back many years, small-scale farmers in Swaziland are not fully exploiting the available manure for replenishing the fertility of their soils CITATION Mkh06 l 2057 (Mkhabela 2006).

Cattle manure has been proven by various researchers to be beneficial as a soil amendment because it can can increase and maintain soil fertility by providing nitrogen, Phosphorus, Potassium, Selenium, Calcium, Magnesium, Sodium and other trace elements such as Iron, Manganese, Copper and Zinc. It also improves the pH of acid soils and calcareous soils, increases soil organic matter content and cation exchange capacity, improves soil aggregate stability, soil macro-structure, infiltration, water holding capacity and erosion resistance CITATION Bay05 l 2057 (Bayu 2005). It contains nitrogen in the form of ammonia (NH4) and nitrates (NO3), phosphorus in the form of P2O5 and potassium as K20.

Inorganic fertiliser has been suggested to be the short-term solution for increasing crop yield in Africa, and the present fertiliser consumption needs to be increased by 10–18% annually to reach this target. A very small percentage of sub-Saharan African farmers are also concerned about the adverse environmental impacts of inorganic fertiliser while a larger percentage of farmers consider the benefits and access to other efficient nutrient management system such as organic fertilisers as a good alternative CITATION Rai17 l 2057 (Raimi 2017).

Sakala et al (2003) conducted a study exploring the potential of green manures to increase soil fertility and maize yields in Malawi. Maize was cultivated for two successive cropping seasons at five locations in Malawi from 1996 to 1999. It was found that over the two seasons and across the five sites the application of inorganic fertilizer (35 or 69 kg N ha?1) to maize significantly increased maize yields at all the sites. Maize yields following green manures without inorganic fertilizer additions were much higher than yields from continuous maize with no fertilizer addition.

Researchers, Nyamangara et al (2005) investigated the effectiveness of cattle manure and nitrogen fertilizer application on the agronomic and economic performance of maize. A field experiment was carried out in Harare, Zimbabwe using manure that had been aerobically composted with a carbon content of 8.4%, a nitrogen content of 0.93% and a soil content of 83%. It was found that in the first year, increase in grain yield was much higher when manure (12.5 t ha-1 and 37.5 t ha-1) was combined with the 60 kg N ha-1 mineral nitrogen rate (40% and 25.1 %, respectively), and a relatively smaller further increase of 17.5% was recorded for the 37.5 t ha-1 rate while there was a decrease of 3.7% for the 12.5 t ha-1 rate, when mineral N rate was doubled to 120 kg N ha-1. In the third season increase in grain yield was also much higher when manure (12.5 t ha-1 and 37.5 t ha-1) was combined with the 60 kg N ha-1 mineral N rate (66.2% and 16%, respectively)
and relatively smaller further increases were recorded when the mineral N rate was doubled to 120 kg N ha-1 (21.4% and 15.1 %, respectively). It was concluded the residual effects of cattle manure can last for at least three seasons and thus farmers could apply up to 40 t ha-1 in the first season and benefit from its residual fertility in subsequent seasons. It was also a recommendation of the study that smallholder farmers in Zimbabwe and other countries of Sub-Saharan Africa could positively exploit the combined application of manure and nitrogen fertilizer to increase maize yield and net benefits.
Maize yield simulation using DSSATCrop simulation models are increasingly being used for describing and predicting soil-crop variables under a wide variety of conditions for providing guidelines at farming level. There are various crop simulation models available around the world used to simulate the processes of crop growth, development, yield formation, and its response to environment. Such models include Decision Support System for Agrotechnology Transfer (DSSAT), Agricultural Production System Simulator (APSIM), Environmental Policy Integrated Climate (EPIC), Agricultural Land Management Alternatives with Numerical Assessment (ALMANAC), and Cropping Systems Simulation Model (CROPSYST) among others.
The Decision Support System for Agrotechnology Transfer (DSSAT) CERES-Maize model is the most commonly used crop simulation model. It is a non-linear dynamic model that simulates daily maize crop growth and yield as a function of input parameters namely; daily weather, soil characteristics, crop phenotype characteristics and field management practices CITATION Hoo17 l 2057 (Hoogenboom 2017). This model has been proven by various researchers as outstanding in terms of crop simulation performance compared to other models CITATION Sal16 l 2057 (Salo 2016). This, DSSAT version 4.7, was the model that was chosen to be used for this research.

A study was done in China where the objective was to forecast maize yield with DSSAT-CERES-Maize model driven by historical meteorological data. The results showed that the model was suitable for the yield simulation since the absolute relative error was smaller than 15%; and that the data distribution of predicted yields began to converge and the uncertainty decreased rapidly after the tasseling stage in the growth of the maize crop. This study showed that there was a big potential for yield prediction using the model CITATION Che17 l 2057 (Chen 2017).

Another study was done in China to assess the performance and application of the DSSAT CERES-Maize model on simulating maize yield under water-stress conditions. The study recognised that in order to maximize maize yield and improve water use efficiency, the rule of yield loss caused by drought stress should be clarified. The results showed that comparing the simulated yield with measured yield under different drought stresses at different growth stages, it was found that simulated losses under mild drought were higher than that of measured in all growing stages, on the contrary, simulated data under moderate and severe drought were lower than measured ones in each stage. , the performance of simulated model was considered to be perfect. Results indicated that the calibrated model simulated grain yields and measured grain yields fitted each other quite well in response to different drought scenarios CITATION Gen17 l 2057 (Geng 2017).

The DSSAT CERES-Maize model was also used to model conservation agriculture maize response to climate change in Malawi. The importance of this study lies the fact that the adoption of conservation agriculture is increasingly being promoted as a way of adapting agricultural systems to increasing climate variability, especially for areas such as southern Africa where rainfall is projected to decrease.
The results of the long term simulations showed that maize–cowpea rotation gave 451kg/ha more maize grain yield and 1.62 kg/mm more rain and rain water productivity compared to conventional tillage. This study suggested that conservation agriculture, especially when all three principles i.e. no-tillage, crop residue retention and crop rotation, are practiced by smallholder farmers in the medium altitude of Lilongwe and similar areas, has the potential to adapt the maize based systems to the effects of climate change. The use of DSSAT simulation of the effects of CA was successful for no-till and crop residue retention, but poor for crop rotation CITATION Ngw14 l 2057 (Ngwira 2014).

Parameters of DSSAT-CERES-Maize model
Module Parameter Function
Weather Reads or generates daily weather parameters used by the model. Adjusts daily values if required, and computes hourly values.

Soil Soil dynamics Computes soil structure characteristics by layer.

Soil temperature
Computes soil temperature by layer
Soil water
Computes soil water processes including snow accumulation and melt, runoff, infiltration, saturated flow and water table depth. Volumetric soil water content is updated daily for all soil layers. Tipping bucket approach is used
Soil nitrogen and carbon
Computes soil nitrogen and carbon processes, including organic and inorganic fertilizer and residue placement, decomposition rates, nutrient fluxes between various pools and soil layers. Soil nitrate and ammonium concentrations are updated on a daily basis for each layer
Crop CERES-Maize Each is a separate module that simulates phenology, daily growth and partitioning, plant nitrogen and carbon demands, senescence of plant material, etc.

Field Management Planting Determines planting date based on read-in value or simulated using an input planting window and soil, weather conditions
Fertiliser Determines fertilizer additions, based on read-in values or automatic conditions
Residue Application of residues and other organic material (plant, animal) as read-in values or simulated in crop rotations
Irrigation Determines daily irrigation, based on read-in values or automatic applications based on soil water depletion
Harvesting Determines harvest date, based on maturity, read-in value or on a harvesting window along with soil, weather conditions
Table SEQ Table * ARABIC 2. Description of the parameters used by the DSSAT CERES-Maize model in simulating maize crop growth.

Figure SEQ Figure * ARABIC 6. A screenshot of the DSSAT model showing the various functions available.Position of this research:
Attempts to increase maize production for subsistence farmers should seek to identify methods to increase soil fertility, coupled with practices that enhance in-situ soil moisture retention. These methods should use already available resources and inexpensive methods of farming. Identifying these methods usually require many years of field experiments in order to make meaningful decisions. These can be expensive, require a lot of labour and also time consuming. Crop simulation models like DSSAT CERES-Maize model provide an alternative approach to tackling this issue.

Many researchers have investigated the individual beneficial effects of minimum tillage and cattle manure on the maize crop yield. It has already been established that they hold the promise of improving yields, especially for subsistence farmers. There has not been any research done on investigating their combined effects on the maize yields.
There is also a vast amount of literature available on the usage of DSSAT CERES-Maize model to estimate and project maize yields. It has been repeatedly proven to be a very effective model available. The gap lies in the fact that there has not been a study conducted using DSSAT CERES-Maize model to investigate the combined effects of minimum tillage and cattle manure of the maize crop yield. This research attempted to do exactly that using Swaziland as a case study.

Hypothesis and Conceptual frameworkIntegrating minimum-tillage farming and cattle manure may improve soil fertility and in-situ soil moisture retention at critical dry spell periods during the growing season, thereby also increase rain-fed maize yields for subsistence farmers.

Figure SEQ Figure * ARABIC 7. Conceptual framework of the variables in the study and how they fit together
CHAPTER 3: METHODOLOGYIntroductionThe research questions which needed to be interrogated in this study were the following:
What are the current annual maize yield amounts in Siphofaneni?
What are the prospects and challenges the adoption of a combination of minimum tillage and cattle manure as a soil amendment by subsistence farmers?
What is the maize yield response to a combination of minimum tillage and cattle manure as a soil amendment?
In order to be able to answer these questions, both qualitative and quantitative methods were employed in this research. The quantitative method provided answers to the following aspects:
The socio-economic background of the subsistence farmers: In order to propose an optimal farming system for the farmers, a good understanding of their socio-economic background was required.

The current status quo of the farming practices and maize yield amount: An understanding of the current system of farming was important to be able to see how it is performing.

The extent to which the maize yields can be improved using the proposed methods: In order to present an incentive to the farmers to adopt the methods, there has to be substantial evidence to show that it will be indeed beneficial in increasing their yields.

The willingness and prospects to adopting the proposed farming method to increase their yields: Since the system that would be proposed will have to be implemented by the farmers. It is crucial to find out if they are willing to make the switch.

Data requirements and acquisitionData for this study included socioeconomic data of the farmers which was important to lay a foundation for the study, daily weather data (minimum and maximum temperature, precipitation, solar radiation, dew point, relative humidity) for the years 2011 to 2016, soil types, field management data, maize crop genotype characteristics. This data was required to perform the maize yield simulation on the DSSAT Ceres-Maize model.

Socio-economic data of the farmersIn order to determine the necessity for a field survey, secondary data from past literature had to be examined to also find out there was any existing data to provide answers to the research questions. Due the tendency of developing countries to be behind of research and also the lack of available data records, and Swaziland being a developing country as well, it was difficult to find this information. Another concern with some of the data that was available was that it was heavily generalised and could be useful for this research.

This prompted an actual field survey to collect primary data. Field observations, experiments and a questionnaire survey was used to collect data
The questionnaire respondent sample size which was 60 rural small-scale farmers was determined from approximately 300 maize farming homesteads and systematic random sampling was used to select the interviewees from a list of names of households obtained from the local agricultural extension officer. Due to the fact that homesteads in the study area are quite a considerable distance away from each other, only 10 soil samples were taken and used. The 10 homesteads where the soil samples were taken were chosen by systematic random sampling from the list of 60 interviewees selected for the the questionnaire interview.

A questionnaire survey was done to learn about the socio-economic background of the rural small-scale subsistence farmers in question. The socio-economic questionnaire survey covered the following aspects.

Demographic information: gender, age, family size
Education: highest level of education
Economic: family income
A sample of the questionnaire is included in the appendix.

Daily weather data and pre-processingThe daily weather data for the years 2011 t0 2016 were obtained from Swaziland Meteorological Services for the Siphofaneni weather station. This data included the geographic coordinates and altitude of the weather station, the minimum and maximum temperature, precipitation, solar radiation, dew point, relative humidity for the months of October to March which is the maize growing season in Swaziland. The data was received as an Excel file (*.xlsx) therefore it had to be converted to DSSAT weather file (*.wth) in order to be able to import it in the DSSAT CERES- Maize model.

Soil typesThe data about the geology and soil type of the study area were acquired from the Geology Department. Since the study area was quite small, there was no variability in the soil type and geology. The data was received in shape file format (*.shp) and therefore had to be converted to DSSAT soil file (*.sol) to enable incorporation into the model.

Field management data and willingness to switch to other farming methods
This was the most important part of the research and fieldwork was conducted in the study area, Siphofaneni consistuency. A questionnaire survey was done to learn about the current field management practices of the farmers in question. The questionnaire survey covered the following aspects:
Tillage method
Planting date
Fertilizer application
Weeding method
Irrigation method
Annual maize yields
The farming conditions including labour time, equipment used and problems encountered were also asked.

A sample of the questionnaire is included in the appendix.

The field management data to be used in the study (simulation) only focused on the tillage method and fertiliser i.e. soil amendment application.

The data for this objective did not require advanced technical analysis but Microsoft Excel was used to present the descriptive statistics such as frequency graphs, percentages etc.

Maize crop genotype characteristics
Design of method
In order to test the hypothesis of this study, a specific methodology was followed as shown in the diagram below:

Figure SEQ Figure * ARABIC 8. A summary of the methodology applied in the study.Establishing a baseline of actual yieldsThe first step was to establish a baseline of maize yields using the years 2010-2015 and field management methods from a field survey where questionnaire interviews where conducted. Statistical analysis was then performed to get variables that represents the group of farmers in the study area. From the field survey, farming practices which would be used to perform the initial simulation to validate the model.

Model calibrationThe DSSAT CERES-Maize model was calibrated using weather for the years 2010-2015, soil, and crop phenotype characteristics which were obtained from various government agencies as discussed in the data collection section. The model was also calibrated with the following variables which were obtained from the farmers through a questionnaire:
Planting dates: 25 Sept – 15 Oct
Tillage method: Tractor mouldboard plough
Basal fertilizer application method and amount: 100kg/ha, banded beneath surface, 5cm
Seeding type and variety: Dry seed, PIO 3147
Planting method, spacing and planting depth: Hand planting, 10 seeds/m2, 5cm
Plant row direction: North-South
Weeding: hand hoeing at 6 weeks
Top dressing fertilizer application and amount: 50kg/ha, banded on surface
3.3.3 Model validationTo validate the model, a simulation of maize crop development and yield was done for the years 2010-2015 using the values and the farmers’ farming practice variables from the results of the field survey. The simulated values of maize yields were then compared to the baseline yields collected from the field survey using statistical values. The indices that were chosen to analyse model fitting performance were Root Mean Square Error (RMSE), Normalised Root Mean Square Error (nRMSE), coefficient of determination (R2) and the index of agreement (d) to investigate the accuracy of the model in its simulation capability.

RMSE = i=1n(Si-Mi)2n …………………………………………………(1)
nRMSE = i=1n(Si-Mi)n ×100M …………………………………………..(2)
R2 = n?SM-(?S)(?M)n?S2-(?S)2n?M2-(?M)22 × 100 ……………………………….(3)
d = 1-M-S2S-Mmean+M-Mmean2 ………………………………………(4)
where:
M, actual yields collected from fieldwork
S, simulated yields
Mmean, mean of actual yields
Projection of maize yields for the year 2030After the validation and adjustment of the model, it was then used to project yields up to the year 2030. The same variables as those used to validate the model with the only exception being the daily weather values. These were projected up to the year 2030 using the DSSAT’s Weatherman module. There were six treatments with 0, 5 and 10 tonnes of cattle manure with tractor and minimum tillage as shown in the table below.

For the potential maize yield simulation, all variables were assumed to be static except for tillage method and application of cattle manure.

Tillage method Cattle manure amount (tonnes)
0 5 10
Tractor xxx xxx xxx
Minimum xxx xxx xxx
Table SEQ Table * ARABIC 3. Table showing the various combinations of treatments that were applied in the studyThe DSSAT CERES-Maize model uses the following series of formulas to calculate the amount of maize yields:
Rm = LEAF 0.3 + STEM 0.015 + ROOT 0.01 + SO 0.01……………….(5)
Pg = Kp PGMAX(PAR) fL fO fN fT …………………………………………(6)
W = E (Pg – Rm ) ……………………………………………………………………(7)
dWdt=W-SL-Ss- SR……………………………………………….………….(8)
where:
t – time
SL , SS , SR – aging pare of leaf, stem, root
E – energy transfer
Pg – gross photosynthesis
Rm – maintenance respiration
Kp – adaptability of soil fertility
PGMAX(PAR) – response on photosynthetic active radiation
fL , fO , fN , fT – response on LAI, water supply, nitrogen, daily temperature
LEAF, STEM, ROOT, SO – protein activity of leaf, stem, root, storage organ
Feasibility and practicality analysisThis study was about implementing a new method of farming subsistence farmers; it was necessary to investigate the feasibility of implementation. This was done by studying the following aspects:
Cattle manure availability
According to….the average cow weighing at 450kg produces about 25kg of manure each day. This includes both the urine and faeces. To calculate the amount of cattle manure, it was taken into consideration that the cattle breed kept by farmers in Swaziland is Nguni cattle and these weigh an average of 500kg depending on the gender.
To customize the average values in order for them to be applicable to the case study, the daily manure production was calculated as:
DMP = (DM/CW) * NC……………………………………………………………..(9)
With:
DMP, daily manure production
CW, Average cow weight
DM, Average daily manure produced
NC, Average weight of Nguni cow
During fieldwork it was gathered that the farmers kept their herd of cattle in an enclosed area (kraal) during the night and the cattle spend an average of 12 hours (half a day) in one place meaning that the daily manure production had to be divided by 2.

NP = DMP / 2…………………………………………………………………………(10)
With:
NP, Manure produced at night time
Then to get the amount of manure produced by one cow in a year, the cattle manure produce at night time was multiplied by 365 days. This value was then multiplied by the average number of cattle (calculated form fieldwork data) owned by each farmer.

AM = NP * 365………………………………………………………………………(11)
AH = AM * HS……………………………………………………………………….(12)
With:
AM, Annual manure produced by each cow
AH, Annual manure produced by herd
HS, Size of the cattle herd
Labour requirements
Without a knowledge of labour availability and allocation there is a danger that one may become preoccupied with technologies which may subsequently be found to be unacceptable to the target group (Grandin 1983). It is for this reason that this part of the research was important to find out whether the demand for labour resulting from the use of the technology can be matched by the supply of household labour.

In order to reduce the amount of human labour required, the research proposes the use animal drawn implements and equipment winch could be used by smallholders with reasonably low financial investments to achieve increased agricultural production.

It was gathered that the farmers spend about 5 hours each day working in the fields during the maize cultivation season therefore the man-day was taken to be 5 hours. Conversion factors were applied on the man-day for females and children. For female adults the factor of 0.75 was used and for children under 14 years old, a factor of 0.5 was used. The calculations were done for one hectare of land. The today amount of labour required to complete each task is shown in the table below. It was found that implementing these methods would require an additional 25 man-days.

Task Man-days required
Tilling the soil 2
Digging manure 20
Application of manure 3
Total 25
Table SEQ Table * ARABIC 4. The amount of labour requirements for the implementation of the proposed methodsThe following is the formula that was used to calculate the number of man-days (labour) available to each farmer’s family.

L = (M*0.75 + W*0.75 + C*0.5) * 5…………………………………………………………..(13)
Where:
L, labour available
M, average number of men
W, average number of women
C, average number of children
Willingness to implement
Since the proposed farming methods will be for the benefit of the farmers and its success will depend on their extent of implementation, it was imperative that to investigate if they are willing to consider trying out these methods. This was done through questionnaire interviews. The extension officers were also interviewed about their interest in learning more about these proposed methods as they will be the ones responsible for disseminating this new information to the farmers.

Selection of the modelThe DSSAT CERES-Maize model was selected among many models that exist for maize crop daily growth and subsequently maize yield simulation. This model was because it has been successfully used by many researchers (ref.) to estimate maize yield around the world. There was therefore enough information available to provide guidance on how to simulate maize yields using minimum tillage and cattle manure as soil amendment using the DSSAT CERES-Maize model in Swaziland where there are limited studies on maize yields with which an already locally approved model could be selected from. Table 5 shows some of the most widely used models for estimating maize yields.

Model Description Strengths Limitations Sources
Decision Support System for Agrotechnology Transfer (DSSAT) CERES-Maize Dynamic simulation model
Most widely used model -Allows simulation options for crop management over a number of years to assess the risks associated with each option
-Has been used in advising farmers and decision-making
-Provides an accurate description of the development stages of a specific cultivar e.g. growth, development and yield for maize using a daily time step
-Does not include factors that may limit growth and yield of crops grown under field conditions, such as different pests or other soil constraints.

CITATION Boa17 l 2057 (Boa 2017) CITATION Jon03 l 2057 (Jones 2003)Agricultural Production System Simulator (APSIM) Modular modelling framework -User friendly interface and does not require additional coding to set up simulations
-Simulates various levels of production situations including potential, attainable and actual situations.

-Combines accurate yield estimation in response to management with prediction of the long-term consequences of farming practice on the soil resource
-Presents a challenge in specifying a new manure type CITATION Kea03 l 2057 (Keating 2003) CITATION Arc14 l 2057 (Archontoulis 2014)Environmental Policy Integrated Climate (EPIC) Type 2 model -Includes hydrology, erosion-sedimentation, nutrient cycling, crop growth, tillage, soil temperature, economics, and plant environment control functionalities
-Useful in simulating long-term crop yield as well as complex rotations and fallow-cropping systems
– Does not use species specific crop genotypes
-Does not provide outputs for crop developments stages
CITATION Boa17 l 2057 (Boa 2017)Agricultural Land Management Alternatives with Numerical Assessment (ALMANAC) Mechanistic simulation model
-Can generate weather when data is not available or to fill in data gaps
-Able to simulate the growth of up to ten plants in a single simulation run and, thus, it is possible to reflect both competitions by weeds and mixed cropping
-Fails to predict the positive fertilizer response of maize in low fertility soils
CITATION Hil00 l 2057 (Hilger 2000)Cropping Systems Simulation Model (CROPSYST) Generic crop simulation model -Has a link to GIS software and a weather generator
-Takes water, heat, freezing, and nutrient stresses into account to calculate the actual yield.

-Calculates crop growth and development as a function of thermal accumulation, affecting actual harvest date and other growth stages.

-Factors other than those included in the crop growth module may tend to limit the ability of the model to simulate crop growth in response to nitrogen.

CITATION Usi03 l 2057 (Bellocchi 2003) CITATION Sto03 l 2057 (Stockle 2003)Table SEQ Table * ARABIC 5. Other crop simulation models that are available and their strengths and limitations.

CHAPTER 4: FIELDWORK
4.1 Brief overview of fieldworkTo recapture the research questions of this research which were;
What are the current annual maize yield amounts in Siphofaneni?
What are the prospects and challenges the adoption of a combination of minimum tillage and cattle manure as a soil amendment by subsistence farmers?
What is the maize yield response to a combination of minimum tillage and cattle manure as a soil amendment?
To answer these questions, it was considered necessary to conduct a fieldwork study to satisfy the needs of this study. The fieldwork study was conducted for three weeks in August/September 2017. It was an opportunity to collect concrete information about the socio-economic situation and maize farming activities of the subsistence farmers in Siphofaneni consistuency, Swaziland.
The principal part of this research which was to use DSSAT CERES-Maize model to answer the question: ‘What is the maize yield response to a combination of minimum tillage and cattle manure as a soil amendment?’ required questionnaire interviews to be conducted in the study area. One of the inputs of the crop simulation model was the field management activities of the farmers when cultivating maize. The questionnaire interviews included questions about their tillage, planting, weeding, fertilizer application and their ownership of cattle.

The research fieldwork was done in collaboration with the extension officers under the Ministry of Agriculture in the study area. The actual specifics of the fieldwork will be discussed in the next section.

4.2 Fieldwork activitiesAs it has been already mentioned before, the study area was Siphofaneni consistuency in Swaziland. The study area was chosen because it is one of the areas most affected by food scarcity due to low yields. The farmers in the area routinely harvest only a fraction, 58% of what they had planted as shown by the Lowveld AEZ in Table ??

Table SEQ Table * ARABIC 6. The comparison of the sizes of the areas planted by subsistence farmers and the harvest area sizeThe questionnaire interviews were conducted over a duration of 10 days spent in the field. There were 6 interviews conducted per day and each interview took about 30 minutes to complete and another 30 minutes was spent observing the fields’ size, soil type etc. Due to the distance between each farmer’s homestead, a lot of time was spent walking to the next homestead.

The respondents were 60 local small-scale subsistence farmers between the ages of 30 to over 60 years old as shown in Figure ??.. There was an unequal distribution between the genders with female respondents taking a bigger portion as shown in Figure ??. The reason for this might be that the interviews were conducted during the day when males are usually out of the homestead and only the females are left to perform daily chores.

Figure SEQ Figure * ARABIC 9. The distribution of the ages (left) and gender (right) of the farmers who were respondents to the questionnaire interviewsThe DSSAT CERES-Maize model requires field management practices as one of the inputs. The questionnaire interviews included questions about their tillage, planting, weeding, fertilizer application. The details are listed below:
Tillage – date, method, tillage depth, direction
Planting – date, seed variety, depth, method, density,
Weeding – method, date
Fertilizer application – date, type, method, cattle manure application
Irrigation – method
To perform the feasibility analysis, the farmers were interviewed about their cattle ownership, number of family members (labour), and what they think would be a good incentive for them to implement the proposed methods of farming.

Picture SEQ Picture * ARABIC 1. Picture taken during fieldwork while conducting interviews.

Picture SEQ Picture * ARABIC 2. Picture take during fieldwork showing cattle grazing in a maize field
4.3. Problems encounteredThe fieldwork was conducted outside of the maize growing period because the growing period did not coincide with the university holidays. Due to this reason it was not possible to observe the actual situation of the farmers’ maize cultivation practices and the only option was to depend entirely on their questionnaire answers (Pictures 1 ; 2 above). The farmers do not keep diary records of their farming activities, therefore some of their answers were estimations especially the dates when the activities were performed.

Another difficulty that was encountered was convincing the farmers that I was doing the study own my own capacity as a student and not affiliated with the Ministry of Agriculture or any non-governmental organisation. This was exacerbated by the fact that I was being accompanied by extension officers from the Ministry of Agriculture. The farmers expect the government to be doing something to assist them in their situation, therefore for them to see a study being done presents them with a glimmer of hope that they will be receiving something soon. They, therefore had to be convinced that the study was about testing the applicability of a new farming method.

CHAPTER 5: FINDINGS AND DISCUSSIONBrief overviewThis chapter presents the results that were obtained in the research study in order to achieve the objectives. The main objective of the study was to investigate the potential of increasing maize yields for subsistence farmers through integrating minimum tillage and cattle manure as a soil amendment by using the DSSAT-CERES Maize model to simulate maize crop responses to minimum-tillage and cattle manure; hence the yield improvement to determine the applicability of using minimum tillage and cattle manure in maize cultivation. The results presented in this study include the findings from fieldwork, the results of the evaluation of the DSSAT CERES-Maize model’s ability to predict maize yield, the results of the yields projections up to the the year 2030 under the different cultivation methods using the DSSAT CERES-Maize model, the soil nitrogen variation between the different methods and the results of the implementation feasibility analysis.

Socio-economic status of farmers in the study areaAge and gender distribution
Among the 60 farmers that were interviewed female farmers accounted for 90% while 10% were males. The inequality of the gender distribution towards females may be due to a number of reasons. Though the reasons were never asked during the survey but it is probable that some male heads of households were away in pursuit of wage employment outside of Siphofaneni or it might be that the interviews were conducted during the day when males are usually out of the homestead and only the females are left to perform daily chores. Another reason could probably be some may have passed on as it is a well known fact that males have a shorter life expectancy compared to females. The majority (57 %) of the farmers were aged between 50-59 years and above while very few farmers (7%) were between the ages 30-39 years old (Table 7).
Gender Age Male Female Total Percentage
30-39 0 4 4 7
40-49 0 14 14 23
50-59 4 30 34 57
60+ 2 6 8 13
Total 6 54 60 100
Percentage 10 90 100 Table 7. Distribution of the age and gender of the farmers.

Family size
The size of the farmers’ families i.e. number of family members varied between 1 and more than 12 people per family. Of the farmers interviewed, 30% had 4-6 members in their families. This is about the same as the national average which is 4.6 members per family. There was 3% of the farmers who have 13 or more family members (Figure 10). The members of the family usually include up to three generations living together in one homesteads.

Figure SEQ Figure * ARABIC 10. The sizes of the farmers’ household.Levels of education
A majority of the farmers (43%) had up to secondary level of formal education and about 27% acquired primary school education. Only 13% of the farmers acquired high school education and none had tertiary education (Figure 11). The results indicate a grim literacy picture among the farmers in the study area with even 17% with no formal education at all. The literacy level among the farmers in Siphofaneni is the same as the national rate of about 70% of the adult (15 years and above) population.

All the farmers did not have any formal employment; therefore, they could not give their income amounts.

Figure SEQ Figure * ARABIC 11. Levels of education of the farmers.Size of maize field per farmer
Most of the farmers cultivate small land holdings where 30% of the farmers cultivate less than 0.75 hectares of land. A greater portion of the farmers (40%) had a field that was between 0.76-1 hectare each. Only 3% of the farmers cultivate more than 2 hectare parcels of land (Figure 12). The average field size was 1.13 hectares per farmer.

There are no standards pertaining to size in the allocation of land to farmers under the Swazi Nation Land tenure system; hence the diversity in the sizes of land owned by the farmers. Absentee land lords are very rare in Swaziland and there is no landownership that is temporary.

Figure SEQ Figure * ARABIC 12. Maize cultivation field sizes of the the farmers.Fieldwork resultsMaize field management practices
The farmers had varying information about how they conduct their maize cultivation during the growing season. In order to be able to use this information in the DSSAT CERES-Maize model, statistical analysis was done to obtain values that represent all the farmers.

The summarised information that was obtained from the farmers about their field management practices is as follows:
Planting dates: 25 Sept – 15 Oct
Tillage method: Tractor mouldboard plough
Basal fertilizer application method and amount: 100kg/ha, banded beneath surface, 5cm
Seeding type and variety: Dry seed, PIO 3147
Planting method, spacing and planting depth: Hand planting, 10 seeds/m2, 5cm
Plant row direction: North-South
Weeding: hand hoeing at 6 weeks
Top dressing fertilizer application and amount: 50kg/ha, banded on surface
Actual maize yields
The maize yield values obtained from the farmers through questionnaire interviews were averaged to get one since value representing all the farmers’ yields. The year 2013 had the highest amount of harvest at 5052kg/ha (Table 8). This corresponds to the national maize production statistics which also reported the year 2013 as a high yielding year.

It is important to note that the values obtained from the farmers were not the exact values but estimations because the farmers did not keep records.

Year Actual (kg/ha)
2010 4243
2011 4150
2012 4097
2013 5052
2014 4081
2015 4173
Table 8. The average amount of maize yield harvested by each farmer by the year.

Cattle ownership
In Swaziland, cattle are held as a store of wealth. Farmers are known to have a tendency to retain or increase cattle numbers even under adverse environmental conditions CITATION Dor16 l 2057 (Doran 2016). Cattle are usually herded to communal grazing areas during the day and are kept in an enclosed kraal during the night.

The results showed that 87% the farmers that were interviewed owned a herd of cattle. The size of the ranges from 1 to 25 cattle per farmer with the average being 11 cattle per farmer. A large portion, 25% of the farmers had a herd with between 11 and 15 cattle and the smallest portion, 10% had a herd with between 20 and 25 cattle (Figure 13).
The farmers do not use the cattle manure that is collected in the cattle kraal except for occasional events when it is used to smear the floors of traditional houses.

Figure SEQ Figure * ARABIC 13. The distribution of the herd sizes owned by the farmers.DSSAT CERES-Maize resultsModel validationThe model was evaluated by comparing the simulated maize yield values to the actual maize yield amounts obtained from the farmers during fieldwork. In order to validate the model’s performance in estimating the maize yield, statistical parameters RMSE, nRMSE, R2 and d were calculated. The RMSE was 438kg/ha which translated to an nRMSE value of 5%. This indicated that the simulated grain yields and the actual grain yields fitted well with each other. The regression of simulated and actual grain yields indicated a good fit with R2 being 74%. In addition, the index of agreement, d, showed that there was an agreement of 0.82 between the simulated and actual grain yields (Table 9 and Figure 14).

Year Actual yield (kg/ha) Simulated yield(kg/ha) RMSE
(kg/ha) nRMSE R2 d
2010 4243 5434 438 10.2% 74% 0.82
2011 4150 5237 2012 4097 5059 2013 5052 5469 2014 4081 5094 2015 4173 5113 Table 9. The results of the statistical analysis to validate the accuracy of the model.

Figure SEQ Figure * ARABIC 14. The accuracy of the model in simulating maize yieldsYield projection and nitrogen dynamicsMaize grain yield
The validated DSSAT CERES-Maize model was applied to simulate annual maize yields for 15 years, up to the year 2030 under both the current methods of cultivation and the minimum tillage and cattle manure method.

The maize yield calculation for the year 2030, showed that the annual yield would be 8065kg/ha in the treatment with minimum tillage and the highest amount of cattle manure in the treatment,10 t/ha while the treatment with tractor tillage and no cattle manure application resulted in an annual maize yield of 1480kg/ha (Table 10). When compared to the maize yields from 2015, the results showed that a 94% increase was possible.

This would be attributed to the fact that the soils receiving cattle manure would be continuously being nourished every year until the year 2030 and while the soil that is not receiving cattle manure would be losing fertility every year until until 2030. The results also showed that the tillage method did not have a significant impact on the maize yield according to the model simulation. The treatments with tractor tillage had less than 10kg/ha difference from the minimum tillage treatments.

The effect of different amounts of cattle manure on the maize grain yield of the crop was observed and there is a significant difference in the rate of growth and development in the plant grain weight as a result (Figure 15).
Tillage method Cattle manure amount (tonnes)
0 5 10
Tractor 1480 4487 8065
Minimum 1486 4501 8084
Table 10. The yields for the year 2030 as predicted by the model.

Figure SEQ Figure * ARABIC 15. The maize grain development under the different treatments.Nitrogen dynamics
From Figure ?? it can be seen that the application of cattle manure increased the amount of nitrogen in the soil tremendously

Figure SEQ Figure * ARABIC 16. The amount of nitrogen available in the soil at different times during the growing season.

Figure SEQ Figure * ARABIC 17. The amount of nitrogen absorbed by the maize crop at various stages of the growing period.Feasibility of implementation analysis resultsThe results showed that each Nguni cow is capable of producing 27.8kg of manure per day of which only half of it can be collected due to half of the day being spent in the grazing lands. In a year this translates to 5.069t of manure produced by each animal. With the average herd size being 11 cattle per farmer, this produces 55.8t per year. According to this result, each farmer only requires 2 cows to provide the highest amount of manure used in this study which was 10t. Based of the data on cattle ownership that was collected in the field, where the herd sizes ranged from 1-25 cows, there is more than enough cattle manure for the farmers to use in their fields to enjoy the benefits of increased maize yields.

The average farmer’s household consisted of 8 people. This was not necessarily the farmer’s immediate family but it includes people who help with the labour in the maize fields. There was an average of 2 adult women, 3 adult men and 3 children. It is important to not that an adult was defined as a person between the ages of 15-65 years old in this study. People over 65 years old were excluded from the household labour calculations. The result showed that the average household had a potential to contribute 30 man-days to working in the maize fields. The proposed methods will require 25 man-days to implement meaning that the average house can afford to use it as they have enough labour available.

When the farmers were asked about their willingness to implement the proposed methods, 60% of their had their reservations about it. At the same time, 53% of them would be motivated to implement them. 23% of the farmers thought that they were sceptical because these methods can increase the amount of weeds and also harden the soil while another 14% thought that it increased the amount of labour required in maize cultivation. These reasons to be sceptic are not scientifically accurate as the benefits of these methods have been explained in previous sections. Of the farmers that willing to try these methods, 30% said their motivation would be the prospects of higher yields and 23% said they would be motivated by the reduced input costs (Figure 18).

Figure SEQ Figure * ARABIC 18. The reasons for farmers to be sceptic about implementing the proposed methods (left) and reasons they would be willing to implement them (right).

CHAPTER 6: CONCLUSION AND RECOMMENDATIONSConclusionsThe main objective of the study was to determine, using the DSSAT CERES-Maize model, the maize yield changes that would results as a consequence of using minimum tillage and cattle manure as a soil amendment. This was to investigate the practicality of implementing of these farming methods by subsistence farmers by providing concrete scientific evidence of the benefits of these methods. Several sub-objectives were employed to meet the main objective of the study and the following conclusions were drawn:
It is possible to increase maize yields for subsistence farmers. The study revealed that the the yields can be increased by up to 94% by the year 2030 when compared to the yields obtained in 2015.

The model was successful in simulating yield values with R2 =74%, RMSE = 43kg/ha and d = 0.82. The yield results showed that the yield for the year 2030 would be highest for minimum tillage with 10t/ha of cattle manure at 8.084t/ha while the lowest would be tractor tillage with no cattle manure at 1.480t/ha. The tillage method did not seem to have a significant impact on the yield amount.

Minimum tillage does not have any significant contribution in determining the maize grain yield which was an unexpected result. Minimum is a system that is relevant to the resource poor subsistence farmers in Swaziland because of the significant reduction of cultivation costs.

The average farmer does have the minimum resources to implement the usage of cattle manure as a soil amendment which are a herd of cattle with at least two cows to produce about 10tonnes of manure annually and 25 man-days for the application of the manure in a hectare of land.

Achieving higher maize yields for subsistence does not necessarily mean that there should be a higher financial investment into maize farming. This research showed, with numerical evidence, that higher maize yields can be achieved while reducing the input costs.

The results of crop simulation models can never be 100% accurate because crop growth dynamics depend on many other environmental and physiological variables that current crop simulation models, including the DSSAT CERES-Maize model, cannot capture.

Implications of studyThis research could, hopefully, inform decision makers with concrete arguments for policy formulation and increasing awareness of farmers on the positive contribution of cattle manure in a sustained increase of maize yields.

Successful implementation of using cattle manure as a soil amendment might lead to:
A modification in soil water and nitrogen dynamics and an increased soil nitrogen content. These may have a positive impacts on crop yields by increasing them thereby increasing rural household food security and reducing poverty.

A reduction in input costs because cattle manure is already readily available at no cost to the farmers.

A reduction in the environmental footprint of maize farmers by reducing carbon emissions from tractors and the use of fossil fuels to power the tractors.

An avoidance and/or reduction in the dependency of farmers on inorganic fertilisers. A reduction in the usage of inorganic fertilisers will result in the reduction of pollution from chemical fertilisers.

It is important to emphasise that the government of Swaziland has a crucial role to play in ensuring that cattle manure is fully adopted as one of the alternatives in the quest to address soil fertility challenges. Their support can be in the form of information dissemination and training the subsistence farmers through the agricultural extension services.

Limitations to implementation
Labour challenges – moving the cattle manure from the cattle to the maize field requires added labour. Some farmers have their fields some distance from their homestead and this may present a challenge to farmers. This is especially true because it is a common phenomenon to have the young and abled people leaving the rural areas to go to the city in search of better education and employment opportunities. This leaves the farmers with less manpower to work in the fields.

In as much as cattle manure is a promising alternative strategy for resource-poor farmers in Swaziland, it cannot meet crop nutrient demand over large areas, because of the limited quantities available and the relatively low nutrient content of the materials. This may prompt the combined use of cattle manure and inorganic fertilizers.

SummaryThis research study highlighted the fertility issues and the benefits associated with the use of organic fertilisers (cattle manure) and minimum tillage in improving crop productivity among smallholder farmers in sub-Saharan Africa. Unfortunately, minimum tillage did not have any significant contribution in determining the maize grain yield which was an unexpected result based of previous literature. From the literature that was reviewed, the use of cattle manure was established to increase plant growth and yield, as well as improve soil quality. In addition, cattle manure could also protect the natural environment and soil biodiversity as it promotes the growth of microorganisms in the soil. The environmentally friendly property of cattle manure, as well as its great potential in sustainable agriculture, accentuates the need to reduce, if not replace, the use of agrochemical inputs with organic ones. The resource-poor farmers in Swaziland who cultivate on nutrient-poor soil need a cost effective and efficient technology to increase yield. It is, therefore, important to encourage farmers to to use cattle manure in their fields through extension education and agronomic training such as on-site training.
The production process of cattle manure is simple, natural and occurs organically. It requires less capital, no technology and workforce, unlike inorganic fertiliser production that requires huge energy, high capital base and a significant amount of manpower. Substituting inorganic fertilisers with cattle manure will not only increase productivity subsistence farmers, it will ultimately increase food security in Swaziland without causing any known environmental challenges. In conclusion, there is an urgent need to improve the awareness and use of cattle manure among subsistence farmers in Swaziland.

APPENDICESQuestionnaireRaw dataREFERENCES