Determinants of Resource Use and Use Efficiency in Urban and Peri-Urban Dairy, Beef Cattle and Pig Production Units in Ouagadougou, Burkina Faso (West Africa)

Home , Dairy farming

DETERMINANTS OF RESOURCE USE AND USE EFFICIENCY IN URBAN AND PERI-URBAN DAIRY, BEEF CATTLE AND PIG PRODUCTION UNITS IN OUAGADOUGOU, BURKINA FASO (WEST AFRICA)

Serge E. Mpouam Section Animal Husbandry in the Tropics and Subtropics Faculty of Organic Agricultural Sciences University of Kassel

DETERMINANTS OF RESOURCE USE AND USE EFFICIENCY IN URBAN AND PERI-URBAN DAIRY, BEEF CATTLE AND PIG PRODUCTION UNITS IN OUAGADOUGOU, BURKINA FASO (WEST AFRICA)

Serge Eugene Mpouam

Dissertation presented to the Faculty of Organic Agricultural Sciences University of Kassel, Witzenhausen 2019 Dissertation zur Erlangung des akademischen Grades eines Doktors der Agrarwissenschaften (Dr. agr.) vorgelegt durch Serge Eugene Mpouam, M.Sc.

Erste Gutachterin: Prof. Dr. Eva Schlecht Universitat Kassel und Georg August- Universitat Gottingen

Zweiter Gutachter: Assoc. Prof. Dr. Luc Hippolyte Dossa Universitat Abomey-Calavi

Pruter: Prof. Dr. Andreas Burkert Universitat Kassel

Pruter: Prof. Dr. Detlev Moller Universitat Kassel

Datum der Disputation: 13 September 2019

Dissertation to obtain the academic degree of Doctor of Agricultural Sciences (Dr. agr.) Presented by Serge Eugene Mpouam, M.Sc.

1 st Evaluator: Prof. Dr. Eva Schlecht University of Kassel and Georg-August University of Gottingen

nd 2 Evaluator: Assoc. Prof. Dr. Luc Hippolyte Dossa University of Abomey-Calavi, Benin

Examiner: Prof. Dr. Andreas Burkert University of Kassel

Examiner: Prof. Dr. Detlev Moller University of Kassel

Date of oral exam (disputation): 13 September 2019 Dedication

the memory of my late grandparents,

my beloved wife Nno Nkassa Claire Murielle our children Allongo Mpouam Cassidy Allisson,

Mandjom Mpouam Gamaliel Bryan, Mbang Mpouam Wendy Asaelle

Summa

Summary

The (peri-) urban demand for food of animal origin is increasing in sub-Saharan Africa as a result of the rapidly growing population and urbanisation. Livestock production in the vicinity of West African cities is well established but currently undergoes changes triggered by key transforming factors such as access to resources, production management related to tamers' knowledge or know-how, increasing demand for livestock and/or animal products, policies and the legal framework in which livestock production systems operate. Urbanisation and population growth have made (peri-) urban agriculture increasingly diversified such that several livestock production systems operate in parallel in West African cities, with different degree of specialisation, intensification, and integration with crop farming. Thus determinants of resource use and use efficiency vary within and across cities and are impacting on farm performances and profitability (Chapter 1 ). Therefore attempts to understand those key determinants in order to improve production practices and performances are highly relevant in the context of increasing resource scarcity.

For this study, 181 (peri-) urban commercial livestock keeping households in Ouagadougou, Burkina Faso, were interviewed using a semi-structured questionnaire (Chapter 2). The collected data enabled the identification and characterisation of four homogeneous (peri-) urban production systems (clusters) based on household socio economic characteristics, farm assets, crop production (yes/no), livestock ownership (species, numbers) and management as well as livestock intakes (purchases, births) and offtakes (sales of animals and products). The identified (peri-) urban livestock production systems were OUA-1 (pig production; 31.8%), OUA-2 (mixed crop-pig and non-dairy ruminant production; 10.8%), OUA-3 (mixed crop-dairy (cattle} production; 34.4%), and OUA-4 (mixed crop-dairy production (Fulani owners); 22.9%). Specialisation was mostly found in pig and dairy farms, whereas most farmers located in the peri-urban area integrated livestock keeping with crop farming. Intensification and market orientation also depended on the type of livestock production system.

In order to analyse resource use and use efficiency in those production systems exposed to the highest increase in consumer demand, namely dairy, beef and pig production, a 16- month quantification of input and output variables was conducted in the concerned livestock production systems. For dairy production (Chapter 3) the average daily milk offtake per animal in cluster OUA-3 was 5 to 7 times higher than in OUA-4. Dairy farmers in

vi Summa

OUA-3 were more interested in genetic improvement of their herd (mainly by using crossbred cattle) and production intensification (offering 5-20 fold more feed dry mater (OM) to their animals, depending on physiological status) compared to the farmers in OUA- 4. The animals of the latter were highly depending on year round grazing of natural pastures (supplying insufficient amounts and quality of feed), while animals on OUA-3 farms only grazed during part of the year. Important farm to farm and seasonal variation in feeding intensity was observed in both dairy production systems. Homestead feeding was dominated by the supply of protein roughages and protein-rich concentrate feeds, resulting in imbalanced diets with inadequate metabolisable energy supply. This resulted in an inefficient resource use, increased risk of metabolic disorders such as ketosis in both dairy production systems (general prevalence of 70%) and limited dairy cows' productivity and farm performances.

In pig (OUA-1 and OUA-2) and beef cattle (OUA-2 and OUA-3) production systems the use of non-local breeds and intense homestead feeding were also key features of intensifying segments of the respective (peri-) urban livestock sectors (Chapter 4 ). Beef cattle of households supplementing their animals had higher performances than non-supplemented beef cattle. In particular, supplemented adult male Sahelian breed beef cattle showed 1 highest average daily weight gain (AOG; 646±339 g d- ) as compared to local zebu (AOG 1 of 322±440 g d- ). Meanwhile, AOG of non-supplemented adult males was lowest (62±266 g d-1). Similar to dairy cattle, beef cattle were fed in an opportunistic manner with varying quantities and qualities of feeds according to availability. Feed OM offer was similar for supplemented local zebu and Sahelian zebu cattle. Even without grazing, fiber requirements were widely met in supplemented beef cattle. Utilized roughages were diverse straws from grain processing and hays, but also protein roughages, all with high lignin content and low nutritional value. Even though OM supply increased from rainy season to early dry and late dry season, total metabolisable energy and crude protein offered to the animals did not vary strongly as diet quality decreased over the seasons.

The overall productivity of pig rearing was low, with mortality rates of 61% and 14% in local and crossbred pigs, respectively, and average litter sizes of 5.4±2.3 versus 5.1±1.9 piglets in crossbred and local sows (Chapter 4 ). The inter-farrowing interval averaged 206±42 days in both breeds. The AOG of growing animals was 110±116 and 70±69 g d-1 in crossbred and local pigs, respectively. Suckling crossbred pigs had a 60% higher AOG

vii Summa than local pigs while their mother sows were losing weight about twice as fast during lactation than the local sows, even though they received about 40% more feed OM per kg metabolic weight and day. Most feeds offered to pigs were protein-rich industrial by products, and supply of commercial feeds was limited to crossbred pigs. There was a high variability in feeding intensities between farms and seasons. In most cases, the provision of metabolisable energy was by far more limiting pig growth than the supply of digestible protein, with severe energy deficits observed in 50% and 40% of local and crossbred animals, respectively.

The efficient use of resources and in consequence the improvement of production performances in West Africa's (peri-) urban livestock production systems can be realised through an optimised farm management including cancelation of animal scavenging, adoption of proper housing structures, year-round optimised allocation of feedstuffs, controlled breeding, appropriate provision of healthcare and improved farm biosecurity that might reduce the risk of diseases and reduce animal mortality, especially in pigs (Chapter 5).

viii Zusammenfassung

Zusammenfassung

Infolge von rasch wachsender Bevölkerung und Verstädterungsphänomenen steigt die (peri-) urbane Nachfrage nach Lebensmitteln tierischen Ursprungs in Sub-Sahara Afrika (SSA) sehr rasch an. Die Tierhaltung ist in westafrikanischen Städten und ihrem unmittelbaren Umfeld seit jeher etabliert, unterliegt derzeit jedoch raschen Veränderungen, zum Beispiel hinsichtlich der Richtlinien und rechtlichen Rahmenbedingungen für städtische Tierhaltung, des Zugangs zu Ressourcen, des Tiermanagements der Landwirte und der steigenden Nachfrage nach Lebendvieh und / oder tierischen Produkten. Unter anderem führen diese Faktoren zu einer zunehmenden Diversifizierung der städtischen Landwirtschaft, so dass heute in westafrikanischen Städten gleichzeitig verschiedene Tierproduktionssysteme mit unterschiedlichem Spezialisierungs-, Intensivierungs- und Integrationsgrad (mit Ackerbau) zu finden sind. Die Determinanten der Ressourcennutzung und der Nutzungseffizienz variieren daher innerhalb und zwischen den Städten und wirken sich auf die Leistung und Rentabilität der landwirtschaftlichen Betriebe aus (Kapitel 1 ). Daher ist es höchst relevant, Schlüsseldeterminanten zu identifizieren und ihre Wirkungsweise und -richtung zu verstehen, um im Kontext zunehmender Ressourcenknappheit Produktionspraktiken und -leistungen zu verbessern.

Für die vorliegende Untersuchung wurden 181 marktorientierte Tierhaltungshaushalte befragt, die im (peri-) urbanen Raum Ouagadougous lokalisiert waren, der Hauptstadt von Burkina Faso. Die Mithilfe eines semistrukturierten Fragebogens erhobenen Daten (Kapitel 2) erlaubten die Identifizierung und Charakterisierung von vier homogenen (peri-) urbanen Produktionssystemen (Clustern), auf der Grundlage von sozio-ökonomischen Haushaltsmerkmalen, dem landwirtschaftlichen Vermögen, des Engagements in der Pflanzenproduktion Ga/nein), dem Viehbesitz (Tierarten, Anzahl} und des Tiermanagements sowie der Zugänge (Ankauf und Geburten von Tieren) und Entnahmen (Verkauf von Tieren und Produkten). Folgende Tierhaltungssysteme wurden unterschieden: OUA-1 (Schweinehaltung; 31,8%), OUA-2 (Ackerbau, Schweine- und Rinderhaltung; 10,8%), OUA-3 (Ackerbau und Milchviehhaltung; 34,4%) und OUA-4 (Ackerbau und Milchviehhaltung durch Fulani; 22,9%). Eine Spezialisierung zeigte sich vor allem in den urbanen Schweinehaltungs- und Milchviehbetrieben; die meisten periurban lokalisierten Landwirte integrierten dagegen Viehhaltung und Ackerbau. Intensivierung und Marktorientierung hingen aber auch von der Art des Tierhaltungssystems ab.

ix Zusammenfassung

Um Ressourcennutzung und Nutzungseffizienz in den Milch, Rindfleisch und Schweinefleisch erzeugenden Betrieben zu analysieren, für deren Produkte die Verbrauchernachfrage momentan am stärksten zunimmt, wurden die wichtigsten Input und Output-Variablen in den entsprechenden Betrieben über einen Zeitraum von 16 Monaten quantifiziert. Hinsichtlich der Milcherzeugung (Kapitel 3) war die durchschnittlich ermolkene tägliche Milchmenge in Cluster OUA-3 fünf- bis siebenmal höher als in OUA-4. Die Milcherzeuger in OUA-3 waren im Vergleich zu denen in OUA-4 stärker an einer genetischen Verbesserung ihrer Herde (hauptsächlich durch die Verwendung von Kreuzungsrindern) und einer Produktionsintensivierung Ue nach physiologischem Status der Kühe 5- bis 20-fach höheres Futterangebot) interessiert. Tiere der OUA-3 Farmen waren stark von einer ganzjährigen Beweidung natürlicher Weiden abhängig (unzureichende Futtermenge und -qualität), während Tiere der OUA-4 Farmen meist nur während der frühen Trockenzeit weideten. 1 n beiden Milchproduktionssystemen wurden bedeutende betriebliche und saisonale Unterschiede in der Fütterungsintensität beobachtet. Die (Zu-)Fütterung war generell durch eine Versorgung mit eiweißhaltigem Rauhfutter und proteinreichen Kraftfuttermitteln gekennzeichnet, was zu einer unausgewogenen Ernährung und unzureichender Versorgung mit umsetzbarer Energie führte. Dies resultierte für beide Milchproduktionssystemen in einer ineffizienten Ressourcennutzung, einem erhöhten Risiko für Stoffwechselstörungen - vor allem Ketose (allgemeine Prävalenz von 70%) - und einer eingeschränkten Produktivität und Leistung der Milchkühe.

Für die Schweine- (OUA-1 und OUA-2) und Rindermastbetriebe (OUA-2 und OUA-3) waren die Verwendung exotischer Rassen und eine intensive Fütterung ebenfalls wichtige Merkmale der Intensivierung (Kapitel 4). Auf Betrieben die ihre Rinder zufütterten zeigten die Tiere eine höhere Leistung als auf Betrieben die nur Weidegang praktizierten. Insbesondere supplementierte männliche Zebus der Sahelrasse zeigten Tageszunahmen von 646±339 g; supplementierte männlichen Tiere der lokalen Zeburasse nahmen dagegen pro Tag nur 322±440 g zu und nicht supplementierte männliche Tiere sogar nur 62±266 g. Ähnlich wie beim Milchvieh wurden Fleischrinder mit sehr unterschiedlichen Futtermengen und -qualitäten gefüttert, abhängig von deren Verfügbarkeit. Dabei war die angebotene Futtermenge (Trockenmasse, TM} für supplementierte Rinder der lokalen Zeburasse und der Sahelrasse ähnlich. Auch ohne Weidegang war die Aufnahme an

X Zusammenfassung

Rauhfuttermitteln physiologisch ausreichend; angeboten wurden Getreidestroh, Grasheu und Leguminosenheu - alle mit hohem Ligningehalt und niedrigem Nährwert. Obwohl die täglich vorgelegte Futtermenge (TM} von der Regenzeit über die frühe Trockenzeit zur späten Trockenzeit hin zunahm variierte die den Tieren angebotene Menge an umsetzbarer Energie und Rohprotein nur geringfügig, da die Futterqualität im Verlauf der Trockenzeit abnahm.

Aufgrund hoher Sterblichkeitsraten war die Produktivität der Schweinehaltung (Kapitel 4) gering - sowohl für die Lokalrasse (Mortalität: 61% ) als auch für Kreuzungstiere (Mortalität: 14%); hinzu kam eine geringe Wurfgröße von 5,4±2,3 (Kreuzungstiere) beziehungsweise 5,1 ±1,9 Ferkeln (Lokalrasse). In beiden Fällen betrug die Zwischenwurfzeit durchschnittlich 206±42 Tage. Bei Masttieren betrug der tägliche Lebendmassezuwachs im Schnitt 110±116 g (Kreuzungstiere) beziehungsweise 70±69 g (Lokalrasse). Auch Saugferkel hatten im Falle von Kreuzungstieren eine um 60% höhere Tageszunahme als die Lokalrasse, jedoch nahmen die entsprechenden Muttersauen während der Laktation etwa doppelt so schnell ab wie die einheimischen Sauen, obwohl sie täglich etwa 40% mehr Futter (g TM pro kg metabolische Masse) erhielten. Den Schweinen wurden hauptsächlich proteinreiche Nebenprodukte aus der Nahrungsmittelerzeugung verfüttert, der Einsatz kommerzieller Schweinefuttermittel war auf Kreuzungstiere beschränkt. Die Fütterungsintensitäten variierten stark, sowohl zwischen Betrieben als auch zwischen Jahreszeiten. In den meisten Fällen wurde die tierische Leistung durch das Angebot an umsetzbarere Energie begrenzt - bei 50% der Schweine der Lokalrasse und bei 40% der Kreuzungstiere wurden erhebliche Energiedefizite beobachtet.

Der effiziente Einsatz von Ressourcen und damit die Verbesserung der Produktionsleistung in den (peri-) urbanen Tierhaltungssystemen Westafrikas setzt eine generelle Verbesserung des Tiermanagements voraus. Dazu gehören der grundsätzliche Verzicht auf das sogenannte Scavenging, die Bereitstellung tiergerechter Unterstände beziehungsweise Stallungen und eine ganzjährig bedarfsgerechte Fütterung. Außerdem strategische Züchtungsansätze, eine angemessene Gesundheitsvorsorge und eine insgesamt verbesserte Hygienesituation auf den landwirtschaftlichen Betrieben, die das Krankheitsrisiko und die Tiersterblichkeit, insbesondere bei Schweinen, verringern (Kapitel 5).

xi Zusammenfassung Résumé Résumé

La demande (péri-) urbaine en denrées alimentaires d'origine animale augmente en Afrique subsaharienne en raison de la croissance rapide de la population et de l'urbanisation. L'élevage dans les environs des villes d'Afrique de l'Ouest est bien établie, mais il subit actuellement des changements provoqués par des facteurs de transformation clés tels que l'accès aux ressources, la gestion de la production liée aux connaissances ou au savoir-faire des agriculteurs (éleveurs notamment), la demande croissante d'animaux et / ou de produits animaux, les politiques et le cadre règlementaire dans lequel les systèmes de production animale opèrent. L'urbanisation et la croissance de la population ont rendu l'agriculture (péri-) urbaine de plus en plus diversifiée, de sorte que plusieurs systèmes de production animale fonctionnent parallèlement dans les villes d'Afrique de l'Ouest, avec des degrés de spécialisation, d'intensification et d'intégration différents. Ainsi, les déterminants de l'utilisation des ressources et de l'efficacité d'utilisation varient d'une ville à l'autre et ont un impact sur les performances et la rentabilité des exploitations (Chapitre 1 ). Par conséquent, essayer de comprendre ces déterminants clés afin d'améliorer les pratiques de production et les performances s'avère très pertinent dans un contexte de ressources limitées. Pour cette étude, 181 ménages d'éleveurs commerciaux (péri-) urbains de Ouagadougou (Burkina Faso) ont été enquêtés à l'aide d'un questionnaire semi-structuré (Chapitre 2). Les données collectées ont permis d'identifier et de caractériser quatre systèmes de production (clusters) homogènes (péri-) urbains basés sur les caractéristiques socio économiques des ménages, les actifs agricoles, la production agricole (oui/non), la possession du bétail (espèces, nombres) et la gestion de l'exploitation ainsi que les flux de bétail en terme d'acquisitions (achats, naissances) et de prélèvements (vente d'animaux et de produits). Les systèmes de production animale (péri-) urbains identifiés étaient: OUA-1 (production porcine; 31 ,8% ), OUA-2 (production mixte de porcs et de ruminants non laitiers; 10,8%), OUA-3 (production mixte, bovins laitiers et ruminants non laitiers); 34,4%) et OUA-4 (production laitière mixte, propriétaires peuls}; 22,9%). La spécialisation concernait principalement les exploitations porcines et laitières, tandis que la plupart des ménages situés dans les zones périurbaines ont intégré l'élevage aux cultures agricoles. L'intensification et l'orientation vers le marché dépendaient également du type de système de production animale.

xiii Résumé Afin d'analyser l'utilisation des ressources et l'efficacité de leur utilisation dans les systèmes de production de plus en plus exposés à une forte augmentation de la demande des consommateurs, à savoir la production laitière, bovine et porcine, une quantification étalée sur 16 mois des flux entrant et sortant a été réalisée dans les systèmes de production animale concernés. En ce qui concerne la production laitière (Chapitre 3), la production journalière moyenne de lait par animal dans le cluster OUA-3 était 5 à 7 fois plus élevée que dans le cluster OUA-4. Les producteurs laitiers de OUA-3 appliquaient plus l'amélioration génétique de leur troupeau (principalement en utilisant des bovins croisés) et l'intensification de la production (offrant 5 à 20 fois plus de matière sèche (MS) à leurs animaux, en fonction de l'état physiologique) par rapport aux producteurs de OUA- 4. Les animaux de ces derniers dépendaient tout au long de l'année de pâturages naturels (apportant des quantités et une qualité d'aliments insuffisantes), tandis que les animaux des fermes OUA-3 ne pâturaient que pendant une partie de l'année. D'importantes variations saisonnières d'intensité alimentaire ont été observées entre les deux systèmes de production laitière et d'une ferme à l'autre dans chaque système. L'alimentation à la ferme a été dominée par l'apport d'aliments concentrés riches en protéines, résultant en un déséquilibre de la ration avec une alimentation en énergie métabolisable insuffisante. Cela a entrainé une utilisation inefficace des ressources, un risque accru de troubles métaboliques tels que la cétose dans les deux systèmes de production laitière (prévalence générale de 70%) et une productivité et des performances à la ferme des vaches laitières limitées. Dans les systèmes de production de porcs (OUA-1 et OUA-2) et de bovins à viande (OUA- 2 et OUA-3), l'utilisation de races exotiques (croisées ou non) et l'alimentation intensive à la ferme étaient également des caractéristiques essentielles des segments de plus en plus intensifs de l'élevage (péri) urbains (Chapitre 4 ). Les bovins à viande complémentés avaient de meilleures performances que ceux non complémentés. En particulier, les bovins mâles de race Sahélienne complémentés présentaient un gain moyen quotidien (GMQ) plus élevé ( 646±339 g f1 ) par rapport au zébu local ( 322±440 g f1 ). A l'opposé le GMQ des mâles adultes non supplémentés était le plus faible (62±266 g r1 ). Comme les bovins laitiers, les bovins de boucherie étaient nourris de manière opportuniste avec des quantités et des qualités d'aliments différentes selon les disponibilités. L'offre en MS de la ration était similaire pour les zébus locaux supplémentés et les zébus Sahéliens. Même sans tenir compte de l'apport du pâturage, les besoins en fibres étaient largement satisfaits chez les

xiv Résumé bovins de boucherie complémentés. Les fourrages utilisés étaient des types divers de pailles tels que du foin, mais également des fourrages riche en protéines, tous à forte teneur en lignine et à faible valeur nutritive. Même si les apports en MS ont augmenté entre la saison des pluies, le début de la saison sèche et la fin de la saison sèche, ceux de l'énergie totale métabolisable et des protéines brutes offertes aux animaux n'ont pas beaucoup varié, la qualité de l'alimentation diminuant au fil des saisons. La productivité globale de l'élevage porcin était faible, avec des taux de mortalité de 14% et 61% respectivement chez les porcs croisés et locaux, et une taille de portée moyenne de 5,4±2,3 par rapport à 5,1±1,9 chez les truies croisées et locales (Chapitre 4). L'intervalle entre les mises bas était en moyenne de 206±42 jours chez les deux races. Le GMQ des animaux en croissance était de 110±116 et 70±69 g f1 chez les porcs croisés et locaux, respectivement. Les porcs croisés avaient un GMQ supérieur de 60% à celui des porcs locaux, tandis que les truies (allaitantes) croisés maigrissaient environ deux fois plus vite que les truies locales, bien que recevant environ 40% de MS en plus par kg de poids métabolique et par jour. La plupart des aliments offerts aux porcs étaient des sous-produits industriels riches en protéines, et l'apport en aliments commerciaux était limité aux porcs croisés. Il y avait une grande variabilité dans les intensités d'alimentation entre les fermes et les saisons. Dans la plupart des cas, l'apport en énergie métabolisable était beaucoup plus restreint que l'apport en protéines digestibles et des déficits énergétiques sévères ont été observés chez 50% et 40% des animaux locaux et croisés, respectivement. L'utilisation efficace des ressources et, par conséquent, l'amélioration des performances de production dans les systèmes de production animale (péri-) urbaine d' Afrique de l'Ouest peuvent être réalisées grâce à une gestion optimisée de l'exploitation comprenant l'annulation de la divagation des animaux, l'adoption de logements appropriées, une allocation optimisée tout au long de l'année d'aliments du bétail, des croisements contrôlés, une fourniture de soins de santé appropriés et une amélioration de la biosécurité à la ferme, susceptibles de réduire les risques de maladies et de réduire la mortalité animale, en particulier chez les porcs (Chapitre 5).

XV Resume Preface

This PhD research was carried out within the UrbanFoodPlus Project (http://www.urbanfoodplus.org), funded by the German Federal Ministry of Education and Research (BMBF) and the German Federal Ministry for Economy, Cooperation and Development (BMZ) under the initiative "GlobE-Research for the Global Food Supply", Grant number 031 A242-A. The research focuses on resource use and use efficiency in (peri-) urban dairy, beef cattle and pig production units in Ouagadougou, Burkina Faso (West Africa). The first chapter introduces the thesis through a systematic summary providing insight into the socio-cultural, agro-ecological and economical embeddedness of (peri-) urban livestock production in and around West African cities with a particular emphasis on cattle (dairy and beef) and pig farming. Chapter 2 classifies and characterises livestock production systems in Ouagadougou (Burkina Faso), and analyses whether livestock farms develop towards specialisation or integration, and intensification or extensification. Chapter 3 analyses livestock and resources management practices in dairy enterprises of Ouagadougou (Burkina Faso) and links farm management to production performances, milk composition, metabolic diseases and resource-use efficiency. Chapter 4 analyses the performance of pig and beef cattle farms in Ouagadougou (Burkina Faso) including farm management, feeding management, and animal growth. Chapter 5 provides a general discussion and draws conclusions and recommendations from the various aspects of the thesis that may be useful in the future to different stakeholders interested in (peri-) urban livestock production such as researchers, government extension services, non-governmental organisations, animal feed processing and supplying companies, livestock farmers and their organisations.

xvi Table of contents

Table of contents

Summary ...... vi Zusammenfassung ...... ix Resume...... xiii Preface ...... xvi Table of contents...... xvii List of tables ...... xxi List of figures ...... xxvi List of abbreviations ...... xxviii 1. General introduction ...... 1 1.1 Livestock production in and around West African cities: demand, production, commercialisation and consumption ...... 2 1.2 Urban and peri-urban dairy production systems in West Africa ...... 4 1.2 .1 Demand for milk and dairy products ...... 4 1.2.2 Socio-economic characteristics of (peri-) urban dairy production systems ...... 5 1.2.3 Management features of (peri-) urban dairy production systems ...... 6 1.2.4 Cattle breeds used in (peri-) urban dairy production ...... 7 1.2.5 Feeding strategies, access to feed resources and use efficiency ...... 9 1.2.6 Marketing and value chain of milk and dairy products ...... 12 1.2.7 Constraints and opportunities ...... 12 1.3 Urban and peri-urban pig production systems in West Africa ...... 13 1.3.1 Socio-economic characteristics of (peri-) urban pig production systems...... 14 1.3.2 Management features of (peri-) urban pig production systems ...... 14 1.3.3 Pig breeds used in (peri-) urban production systems ...... 15 1.3.4 Feeding strategies and access to feed resources ...... 16 1.3.5 Pig marketing and value chain ...... 16 1.3.6 Urban and peri-urban pig production in West Africa: constraints and opportunities .. 17 1.4 Research problem, objectives and hypotheses ...... 19 1.5 Thesis structure ...... 20 2. Characterisation of urban and peri-urban livestock production systems and trends toward integration, specialisation and intensification ...... 21 2.1 Introduction ...... 22 2.2 Material and methods ...... 24 2 .2 .1 Study site ...... 24 2.2.2 Livestock keeping households' sampling approach and data collection...... 24 2.2.3 Statistical analyses ...... 26 2.3 Results ...... 27 2.3.1 Identification of livestock production system types...... 27

xvii Table of contents 2.3.2 Trends of livestock production in Ouagadougou ...... 29 2.3.2.1 Crop-livestock integration or specialisation in livestock production ...... 29 2.3.2.2 Intensification or extensification of livestock production ...... 34 2.3.2.3 Feed and manure use ...... 38 2.4 Discussion ...... 41 2.4.1 Specialisation of the (peri-) urban livestock (dairy and pig) production systems ...... 42 2.4.2 Trends towards intensification of (peri-) urban livestock production...... 42 2.4.3 Urban and peri-urban mixed crop-livestock farming ...... 44 2.4.4 Demand-offer of livestock products and farm performance ...... 45 2.5 Conclusions ...... 46 3. Urban and peri-urban dairy production systems on the cross road between intensive and extensive managements in Ouagadougou, Burkina Faso ...... 49 3.1 Introduction ...... 50 3.2 Material and methods ...... 51 3.2.1 Study area ...... 51 3.2.2 On-farm quantification of input and output ...... 53 3.2.2.1 Animal identification ...... 53 3.2.2.2 Animal inflow and outflow records ...... 53 3.2.2.3 Weight and body condition scoring ...... 53 3.2.2.4 Daily milk yield and milk composition ...... 53 3.2.2.5 Stall feeding ...... 54 3.2.3 Pasture use, grazing behavior and pastoral value ...... 55 3.2.4 Animal data calculation ...... 57 3.2.5 Statistical analyses ...... 59 3.3 Results ...... 60 3.3 .1 Herd structure and dynamic ...... 60 3.3.2 Feeding strategies on (peri-) urban dairy farms ...... 61 3.3.2.1 Homestead feeding ...... 62 3.3.2.2 Dairy cattle activities on pasture ...... 65 3.3.3 Pasture quality, biomass production and carrying capacity ...... 74 3.3.4 Body weight development ...... 74 3.3.5 Milk production and composition ...... 77 3.3.6 Feed use efficiency ...... 82 3.4 Discussion ...... 95 3.4.1 Herd management ...... 95 3.4.2 Feeding of dairy cattle ...... 96 3.4.2.1 Homestead feeding ...... 96 3.4.2.2 Activities on pasture of year round grazing (peri-) urban dairy cattle in Ouagadougou (OUA-4 cluster) ...... 97

xviii Table of contents 3.4.2.3 Pastoral value ...... 99 3.4.3 Weight development and body condition ...... 101 3.4.4 Milk offtake ...... 102 3.4.5 Milk composition and energy balance...... 103 3.4.6 Feeds and feeding efficiency ...... 105 3.5 Conclusions ...... 106 4. Performances and efficiency of (peri-) urban pig and beef cattle breeds under different production managements in Ouagadougou, Burkina Faso ...... 108 4.1 Introduction ...... 109 4.2 Material and methods ...... 11 O 4 .2 .1 Study site ...... 110 4.2.2 Households and animals ...... 111 4.2.3 Data collection ...... 111 4.2.4 Data calculation ...... 112 4.2.5 Statistical analyses ...... 113 4.3 Results ...... 113 4.3.1 Pig production ...... 113 4.3.1.1 Herd structure and dynamic...... 113 4.3.1 .2 Reproductionperformances ...... 114 4.3.1 .3 Feed quantity and quality...... 115 4.3.1.4 Weight development ...... 121 4.3.1.5 Feed use efficiency ...... 126 4.3.2 Beef cattle production ...... 134 4.3.2.1 Herd structure and dynamic...... 134 4.3.2.2 Homestead feeding...... 136 4.3.2.3 Weight development ...... 138 4.3.2.4 Feed use efficiency ...... 139 4.4 Discussion ...... 153 4.4.1 Pig production ...... 153 4.4.1 .1 Production management...... 153 4.4.1 .2 Feeding management and intensity ...... 155 4.4.1 .3 Weight development ...... 156 4.4.1.4 Feed use efficiency of pig farms ...... 157 4.4.2 Beef cattle production ...... 157 4.4.2.1 Production management...... 157 4.4.2.2 Feeding management and intensity ...... 158 4.4.2.3 Weight development ...... 159 4.4.2.4 Feed use efficiency in beef cattle production ...... 160 4.5 Conclusions ...... 160

xix Table of contents

5. General discussion ...... 162 5.1 General aspects...... 163 5.2 Trends in (peri-) urban livestock production systems and farm characteristics ...... 163 5.3 Trend towards intensification of (peri-) urban dairy production ...... 166 5.4 Efficiency of (peri-) urban fattening and traditional beef cattle production...... 168 5.5 Efficiency of (peri-) urban pig farming ...... 169 5.6 Testing of initial hypotheses ...... 171 5.7 General conclusions and recommendations ...... 172 References ...... 176 Appendix Related publications Acknowledgements Eidesstattliche Erklarung Affidavit

XX List of tables

List of tables

Table 1 .1: Management features of dairy farms in West Africa cities ...... 8

Table 1.2: Age at first calving (AFC), calving interval (Ci), lactation length (LI) and total milk yield (TMY)/ daily milk yield (DMY) of main dairy cattle breed types used in (peri-) urban dairy farms of West Africa cities...... 1 O

Table 1.3: Age at first calving (AFC), calving interval (Ci), lactation length (LI) and total milk yield (TMY)/ daily milk yield (DMY) of main crossbreds used in (peri-) urban dairy farms of West Africa cities...... 11

Table 2.1: Main occupation (self-assessed) of the heads of four different (peri-) urban farm types in Ouagadougou (n = 157) ...... 30

Table 2.2: Number of livestock species per household (HH) and livestock species kept in four different (peri-) urban farm types in Ouagadougou (n = 157)...... 31

Table 2.3: Logistic regression parameters for variables predicting crop and livestock integration or specialisation in dairy enterprises across households in Ouagadougou (n = 157) ...... 32

Table 2.4: Livestock ownership in four different (peri-) urban farm types in Ouagadougou (n = 157)...... 33

Table 2.5: Milk, eggs and livestock sales by four different (peri-) urban farm types in Ouagadougou (n = 157) ...... 33

Table 2.6: Cattle types owned by four (peri-) urban farm types in Ouagadougou (n = 157) ...... 34

Table 2.7: Use of controlled mating in different livestock species by (peri-) urban households (HH) in Ouagadougou (OUA; n = 157) ...... 35

Table 2.8: Logistic regression parameters for variables predicting intensification of livestock management in relation to breeding, housing and healthcare across (peri-) urban farm households (n = 157) in Ouagadougou...... 37

Table 2.9: Labour use for livestock activities in four different (peri-) urban livestock farm types in Ouagadougou (n = 157) ...... 38

Table 2.10: Feeding management of livestock owned of four different (peri-) urban farm types in Ouagadougou (n = 157) ...... 39

Table 2.11: Homestead feeding of grain residues and forage to ruminants in four different (peri-) urban farm types in Ouagadougou (n = 157) ...... 39

Table 2.12: Homestead feeding of energy and protein rich feeds to poultry and pigs in four different (peri-) urban farm types in Ouagadougou (n = 157) ...... 39

Table 3.1: Animal inflow and outflow in dairy farms of clusters OUA-3 and OUA-4 during the rainy season (RS), the early dry season (EDS), and the late dry season (LDS) (October 2014 - February 2016) ...... 61

xxi List of tables Table 3.2: General overview of feeding strategies and feeding related activities used by (peri-) urban dairy farmers in Ouagadougou Burkina Faso ...... 64

Table 3.3: Feed categories and percentages of feedstuffs (on dry matter basis) offered to (peri-) urban dairy cattle in Ouagadougou during rainy season (RS), early dry season (EDS), and late dry season (LDS) (October 2014 - February 2016)...... 66

Table 3.4: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to pregnant (a) and lactating (b) dairy cows in the (peri-) urban area of Ouagadougou (October 2014 - February 2016) ...... 67

Table 3.5: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry (a) and adult male (b) dairy cattle in the peri- urban area of Ouagadougou (October 2014 - February 2016) ...... 68

Table 3.6: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to suckling (a) and young (b) dairy cattle in the peri- urban area of Ouagadougou (October 2014 - February 2016) ...... 69

Table 3.7: Characteristic activities of dairy cattle from cluster OUA-4 on pastures in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) (RS: July 2015 - October 2015, EDS: November 2015 - February 2016, LDS: March 2015 - June 2015). Italic values indicate averages across sites ...... 70

Table 3.8: Relative contribution (%) of functional woody plant to daily browsing time of dairy cattle in cluster OUA-4. Data are shown for the rainy season (RS: July 2015 - October 2015), the early dry season (EDS: November 2015 - February 2016), and the late dry season (LDS: March 2015 - June 2015) ...... 73

Table 3.9: Number of species per species group and specific contribution of each species group in pastoral lands used by dairy cattle of cluster OUA-4 in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) during the rainy season (September 2015) ...... 75

Table 3.10: Number of species per specific quality index, pastoral value of pastoral lands used by dairy cattle of cluster OUA-4 in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) during the rainy season (September 2015) ...... 75

Table 3.11: Average biomass yield, relative contribution of each species group to biomass production and carrying capacity of pastoral lands used by dairy cattle of cluster OUA-4 in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) during the rainy season (September 2015) ...... 76

Table 3.12: Seasonal variation in live weight (LW) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016)...... 78

Table 3.13: Seasonal variation in average daily weight gain (kg) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016)...... 79

Table 3.14: Seasonal variation in body condition score (BCS) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016)...... 80

xxii List of tables Table 3.15: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016)...... 81

Table 3.16: Milk yield and milk composition of cows kept in two different production systems in the (peri-) urban area of Ouagadougou (October 2014 - February 2016)...... 83

Table 3.17: Effects of season on milk yield and milk composition in two different (peri-) urban dairy production systems of Ouagadougou (October 2014 - February 2016)...... 84

Table 3.18: Intake at the homestead (mean±S .D. for the 16 month study period) of feed dry matter, crude protein, phosphorus and metabolisable energy by cows, males and young tropical livestock units (TLU) in two dairy production systems in Ouagadougou (October 2014 - February 2016) ..... 89

Table 3.19: Energy and protein requirements for maintenance, milk synthesis and growth of different groups of cattle in two dairy production systems in Ouagadougou (October 2014 - February 2016)...... 90

Table 3.20: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in dairy cattle of two different dairy productionsystems in Ouagadougou (October 2014 - February 2016)...... 92

Table 3.21: Adequacy of coverage of CP requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou, (October 2014 - February 2016)...... 94

Table 4.1: Number of pigs per group and breed at the start of the study ...... 112

Table 4.2: Number of cattle per group, feeding intensity (non-/supplemented) and breed at the start of the study ...... 112

Table 4.3: Reproduction performance ...... 114

Table 4.4: Animal intakes and offtakes in two different pig breed types in Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LDS) (October 2014 - February 2016)...... 115

Table 4.5: Feed categories and percentages of feedstuffs (on dry mater basis) offered to pigs in Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LDS) (October 2014 - February 2016) ...... 117

Table 4.6: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to pregnant (a) and lactating (b) sows of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 118

Table 4.7: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry sows (a) and adult males (b) of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 119

Table 4.8: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to young pigs (a) and suckling piglets (b) of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 120

xxiii List of tables Table 4.9: Seasonal variation in live weight (kg) of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 122

Table 4.10: Seasonal variation in average daily weight gain (kg/day) of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 123

Table 4.11: Seasonal variation in body condition score of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 124

Table 4.12: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) in pig groups of different breeds in Ouagadougou (October 2014 - February 2016). 125

Table 4.13: Intake at the homestead (mean±S.D. for the 16 month study period) of feed dry matter, digestible protein, phosphorus and metabolisable energy per tropical livestock unit (TLU) of different groups of two pig breed types in Ouagadougou (October 2014 - February 2016) ...... 129

Table 4.14: Energy and digestible protein requirements (mean±S.D.) for maintenance and growth in groups of two pig breed types in Ouagadougou (October 2014 - February 2016) ...... 130

Table 4.15: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 131

Table 4.16: Adequacy of coverage of DP requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 133

Table 4.17: Animal intake and offtake in supplementing and non-supplementing beef cattle farms during the rainy season (RS), the early dry season (EDS), and the late dry season (LDS) (October 2014 - February 2016) ...... 135

Table 4.18: Feed categories and percentages of feedstuffs offered to beef cattle (on dry mater basis) in the (peri-) urban area of Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LDS) (October 2014 - February 2016)...... 137

Table 4.19: Seasonal variation in live weight (kg) of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 140

Table 4.20: Seasonal variation in average daily weight gain (kg) of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016)...... 141

Table 4.21: Seasonal variation in body condition score of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 142

Table 4.22: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) of different groups of supplemented beef cattle kept in Ouagadougou (October 2014 - February 2016)...... 143

Table 4.23: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) of different groups of non-supplemented local beef cattle kept in Ouagadougou (October 2014 - February 2016) ...... 144

xxiv List of tables Table 4.24: Intake at the homestead (mean±S.D. for the 16 month study period) of feed dry matter, crude protein, phosphorus and metabolisable energy by cows, males and young tropical livestock units (TLU) of two different beef zebu cattle breed types in Ouagadougou (October 2014 - February 2016) ...... 146

Table 4.25: Energy and protein requirements (mean±S.D.) for maintenance and growth of different groups of beef cattle of two different breed types receiving homestead supplementation in Ouagadougou (October 2014 - February 2016) ...... 147

Table 4.26: Energy and protein requirements (mean±S.D.) for maintenance and growth of different groups of non-supplemented local zebu beef cattle in Ouagadougou (October 2014 - February 2016) ...... 148

Table 4.27: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1): requirements(MJ ME d-1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso(October 2014 - February 2016) ...... 150

Table 4.28: Adequacy of coverage of CP requirements(intake (g d-1): requirements (g d-1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016) ...... 152

Table 5.1: Study hypotheses (Chapter 1) as verified by the study results(Chapter 2, 3, 4 )...... 171

Table A.1: Potential classification variables for two-step cluster analysis(% for nominal, mean±S.D. for continuous variables) ...... 191

Table A.2: Proximate composition of feedstuffs offered to dairy, beef cattle and pigs in the (peri-) urban area of Ouagadougou (October 2014 - February 2016)...... 194

Table A.3: Applied metabolisable energy (ME) values and nitrogen digestibility for feeding analysis ...... 195

Table A.4: Requirements (Req) of metabolisable energy (ME) and crude protein (CP) or digestible protein (DP), respectively, for maintenance (m), protein and fat accretion (pf; i.e. growth) and lactation (I) of the different animal types distinguished in the study, as well as their physiological limit for dry matter intake (DMI) ...... 196

Table A.5: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to pregnant(a) and lactating(b) cows in two different beef cattle breed types in Ouagadougou, Burkina Faso(October 2014 - February 2016) ...... 198

Table A.6: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry cows (a) and adult males (b) in two different beef cattle breed types in Ouagadougou, Burkina Faso(October 2014 - February 2016) ...... 199

Table A.7: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to young cattle (a) and suckling calves (b) in two different beef cattle breed types in Ouagadougou, Burkina Faso(October 2014 - February 2016) ...... 200

XXV List of figures

List of figures

Figure 2.1: Location of the 181 interviewed (peri-) urban livestock keeping households in Ouagadougou, Burkina Faso, from January to April 2014...... 26

Figure 2.2: Main characteristics of the identified four livestock farm types in Ouagadougou (OUA-n, n = 1, 2, 3, 4) using spider web diagrams (n = 157)...... 29

Figure 2.3: Breeding, health care provision and housing management of livestock in four different (peri-) urban farm types in Ouagadougou (n = 157)...... 36

Figure 2.4: Manure use across four different (peri-) urban farm types in Ouagadougou (n = 157). 41

Figure 3.1: Overview of herd structure (mean±S.E. for each group) in dairy farms of clusters OUA- 3 and OUA-4 during the rainy season (RS), the early dry season (EDS) and the late dry season (LOS) (October 2014- February 2016)...... 61

Figure 3.2: Proportion of daily a) feeding, b) walking and c) resting (resting without and with rumination) time (%) spent by dairy cattle of cluster OUA-4 in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) on different land cover classes. Data are shown for rainy season (RS: July 2015 - October 2015), early dry season (EDS: November 2015 - February 2016 and late dry season (LOS: March 2015 - June 2015 )...... 72

Figure 3.3: Milk fat-to-protein ratio of different lactation stages (1: 1-30 days, 2: 31-120 days, 3: >120 days) in two different (peri-) urban dairy production systems in Ouagadougou (OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners)) (October 2014 - February 2016)...... 85

Figure 3.4: Percentage of cows depicting a negative energy balance (NEB), a risk for ketosis (RK), and a risk for acidosis (RA) at least once in different lactation stages (1: 1-30 days, 2: 31-120 days, 3: >120 days) in two different (peri-) urban dairy production systems in Ouagadougou (OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners)) (October 2014 - February 2016)...... 86

Figure 3.5: Coverage of metabolisable energy (ME) requirements (intake (MJ ME d-1):requirements (MJ ME d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou, Burkina Faso (October 2014 - February 2016), 3: OUA-3 cluster, mixed crop-(dairy) cattle production, 4: OUA-4 cluster, mixed crop-dairy production (Fulani owners), Preg: pregnant cow, Lact: lactating cow, Dry: dry cow, Male: adult male cattle, you: young cattle...... 88

Figure 3.6: Coverage of crude protein (CP) requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou (October 2014 - February 2016), 3: OUA-3 cluster, mixed crop-(dairy) cattle production, 4: OUA-4 cluster, mixed crop-dairy production (Fulani owners), Preg: pregnant cow, Lact: lactating cow, Dry: dry cow, Male: adult male cattle, You: young cattle...... 93

Figure 4.1: Weight-age diagram for crossbred (CB) and local breed (LB) pigs in Ouagadougou, Burkina Faso...... 126

Figure 4.2: Coverage of metabolisable energy (ME) requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LB: local breed pig, CB: crossbred pig, Preg: pregnant sow, Lact: lactating sow, Dry: dry sow, Male: adult male pig, Young: young pig...... 128

xxvi List of figures Figure 4.3: Coverage of digestible protein (DP) requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LB: local breed pig, CB: crossbred pig, Preg: pregnant sow, Lact: lactating sow, Dry: dry sow, Male: adult male pig, Young: young pig ...... 132

Figure 4.4: Weight-age diagram for non-supplemented local zebu breed (LBN), supplemented local zebu breed (LBS) and supplemented Sahelian zebu breed (SBS), in Ouagadougou, Burkina Faso ...... 145

Figure 4.5: Coverage of metabolisable energy (ME) requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LZ: Local zebu, SZ: Sahelian Zebu, Preg: pregnant cow, Lact: Lactating cow, Dry: Dry cow, Male: Adult male cattle, Young: Young cattle...... 149

1 1 Figure 4.6: Coverage of crude protein (CP) requirements (intake (g d- ):requirements (g d- )) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LZ: Local zebu, SZ: Sahelian zebu, Preg: pregnant cow, Lact: Lactating cow, Dry: Dry cow, Male: Adult male cattle, Young: Young cattle ...... 151

xxvii List of abbreviations

List of abbreviations

% Percent oc Degree celcius ADF Acid detergent fiber

ADG Average daily weight gain ADMc Air dry matter content

AFC Age at first calving

Apv Average pastoral value

AU-IBAR African Union Inter-African Bureau for Animal Resources

BCS Body condition score

BMBF German Federal Ministry of Education and Research

BMZ German Federal Ministry of Economic Cooperation and Development CA Crude ash

Ci Calving interval

CP Crude protein

OM Dry matter

DMD Dry matter digestibility DMI Dry matter intake

DMY Daily milk yield

DP Digestible protein

DWG Daily weight gain

ECOWAS Economic Community of West African States

EDS Early dry season FAO Food and Agricultural Organization

FAOSTAT Statistics/ Food and Agricultural Organization

FCM Fat content of milk

FPR Milk fat to protein ration

F/P Fat to protein ratio g Gram

xxviii List of abbreviations

GPS Global Positioning System

Gpv Good pastoral value ha Hectare

IBM International Business Machines Corporation

ID Identification number (for feed, milk samples and animals)

ILRI International Livestock Research Institute

Is Specific quality index K Utilisation coefficient kJ Kilojoule km2 Square kilometer kg Kilogram

LOS Late dry season

LI Lactation length

Lpv Low pastoral value LW Live weight

ME Metabolisable energy

MJ Megajoule

MW Metabolic body weight (of animal) m Meter mm Millimeter m2 Square meter n Sample size

N Nitrogen n.s. Not significant

NDF Neutral detergent fiber

NEB Negative energy balance NIRS Near lnfrared Spectroscopy OECD Organisation for Economic Co-operation and Development

OM Organic matter OMO Organic matter digestibility xxix List of abbreviations

P Phosphorus

P P-value; probability pt Protein and fat accretion

PPS Probability proportional to size

PV Pastoral value

RA Risk for acidosis

Req Requirement RK Risk for ketosis

RS Rainy season

S.D. Standard deviation

SEM Standard error of the mean

SFi Specific frequency of species i

SPSS Statistical Package for the Social Sciences

SSA Sub Saharan Africa SWAC Sahel and West Africa Club

TLU Tropical livestock Unit

TMY Total milk yield

USA United States of America

XOF West African CFA Franc

WA West Africa WCA Wet chemical analysis

WECARD West and Central African Council for Agricultural Research and development (CORAF/WECARD)

Wpv Without pastoral value

XXX Cha ter 1

1. General introduction

1 Cha ter 1 1.1 Livestock production in and around West African cities: demand, production, commercialisation and consumption

The fast increase in the global and urban population associated with a rapid expansion of cities and a much slower increase in food production triggers the risk of food insecurity in developing countries in general and in West African countries in particular (SWAC OECD/ECOWAS, 2008). Urban consumers in developing countries in particular have been responsible for the growing demand for milk and meat in recent decades, and this increasing demand is projected to double by 2050 for the baseline year 2000 (Rosegrant et al., 2009). Diverse livestock production systems occur in sub-Saharan countries ranging from pastoral/grassland systems, through mixed crop-livestock systems, to intensive, sometimes landless (industrial) systems. In West Africa, (peri-) urban agriculture is a popular practice (Thys et al., 2005; Amadou et al., 2012). Rearing animals in and around cities is important throughout African countries and contributes substantially to eradicate extreme poverty and hunger (Schiere et al., 2006).

Livestock production units contribute substantially towards filling the large demand-supply gap in fresh livestock products such as broiler chicken, milk and eggs in and around African cities, where their consumption is noticeably high (Marichatou et al., 2005; Moustier and Danso, 2006). Urban and peri-urban agriculture improves daily life of West African urban dwellers who are engaging in this activity, by supplying food, household income and employment, but it also has negative and positive environmental and health impacts (Cohen and Garrett, 2009; De Zeeuw et al., 2011 ). The role of urban livestock production is furthermore important as a part of (peri-) urban agriculture. But animals can be both a nuisance and an asset, due to their various direct and indirect functions in urban ecosystems, each with different priorities at household, city, and national level (Schiere et al., 2006). Urban and peri-urban livestock production have evolved dynamically and in various ways within and across cities of the same country and between countries (Dossa et al., 2011 ). In a comparative study, Amadou et al. (2012) showed that sheep and goats are dominating in Kano, Nigeria (kept by 76% and 75% of farming households) compared to Bobo-Dioulasso, Burkina Faso (48% and 40%) and Sikasso, Mali (28% and 40%). Cattle and poultry are more frequent in Bobo-Dioulasso (kept by 82% and 69% of farming households) as well as Sikasso (65% and 79%) but not in Kano (29% and 20%). In Ouagadougou, almost 27% of urban households keep livestock including cattle, small

2 Cha ter 1 ruminants, pigs and poultry (Thys et al., 2005). For sub-Saharan Africa however, despite the growing interest in (peri-) urban livestock production and the increasing recognition of its importance, limited data about the productivity, the profitability and the sustainability of the different urban livestock production systems are available (FAO, 2005).

In West Africa, (peri-) urban livestock farmers face several constraints such as animal diseases, low production performance of their breeds, high costs of feed and labour, as well as policy and institutional constraints ( excessive regulations, lack of policy formulation and planning, marketing and processing; Schiere et al. 2006). Different strategies such as production intensification with feed supplementation, artificial insemination, crossbreeding, and better health care are used by farmers to improve their livestock production, especially on market-oriented farms (Sidibe-Anago et al., 2006; Duku et al., 2010; Amadou et al., 2012). Yet, resource use efficiency also varies throughout seasons, for instance in cattle and small ruminant farms, where pasture-based feeding systems are still widely encountered (Diogo et al., 2010; Amadou et al., 2012). Urban and peri-urban livestock farming in West African countries is market-oriented with sales being the main reason for animal offtake as shown by previous studies (Amadou et al., 2012).

According to Herrero et al. (2014), milk consumption is likely to triple in Africa by 2050 (with East Africa leading the growth in milk consumption) and the consumption of meat and eggs from poultry and pork have the highest growth projection rates. The region of West Africa is projected to have a six- to sevenfold increase in the consumption of monogastric products until 2050, followed by Southern and Eastern Africa (four fold increases; Herrero et al., 2014 ). The foreseen increase in consumption of dairy and monogastric products is likely to trigger the development of intensive segments of the related production systems (Thornton, 2010; Roessler et al., 2016) which therefore, in the context of resource scarcity and climate change underlines the relevance of studying their efficiency (Herrero et al., 2014) as well as production systems such as beef cattle production that might have a moderate increase of consumption but intensifies simultaneously with dairy production intensification. Thus within this context the present study carried out within the BMBF-funded GlobE-Project UrbanFoodPlus (http://www.urbanfoodplus.org; Grant number 031A242-A} tried to contribute, by zooming into (peri-) urban dairy, beef and pig production in West Africa, to the understanding of the determinants of resource use and use efficiency of these

3 Cha ter 1 important production systems and identify site-specific resource use strategies that are economically and environmentally sustainable and socially accepted.

1.2 Urban and peri-urban dairy production systems in West Africa

Dairy cattle farms operate at different scales and supply a certain proportion of the growing cities' demand for fresh milk and meat in many West African cities. Yet, the local milk production is low and does not keep track with the fast growing demand in milk and dairy products in the expanding cities, which therefore rely on imported dairy products (Metzger et al. 1995; Marichatou et al. 2005; Oudet 2005). Production intensification (genetic improvement of domestic breeds, feed supplementation, and health care) and milk value chain development are important factors that contribute to an increase in (peri-) urban dairy production (Millogo et al. 2008; Amadou et al. 2012).

Production, processing and trading of dairy products in West Africa cities have been described in various studies carried out in countries such as Cameroon (Bayemi et al., 2005), Cote D'Ivoire, Mali and Nigeria (Agyemang et al., 2006), Niger (Chaibou et al., 2011), Gambia, Guinea and Guinea Bissau (Somda et al., 2004), Mali (Bonfoh et al., 2007), Burkina Faso (Millogo et al., 2008), Ghana (Okantah et al. 1998; Gidiglo 2014), Chad (Duteurtre and Koussou, 2001 ), and Senegal (Dieye et al., 2003). Even though most of those studies differ in terms of the methods used, results obtained and therefore conclusions, there exist similarities and differences across cities and countries that deserve to be looked at from a regional perspective to better understand the development of the urban milk production sector across West Africa. In chapter 1.2, we attempt to extract, from published and unpublished documents, relevant information and data related to the structural, organisational, functional, geographic and economic characteristics of (peri-) urban dairy production in West Africa cities.

1.2.1 Demand for milk and dairy products

There is a growing demand for milk and dairy products in West Africa cities as a result of population expansion. Yet, there is a disparity concerning the demand and consumption of milk and dairy products with a decreasing gradient from Sahelian to costal countries, mainly due to dietary habits of the (former) pastoralist and agro-pastoralist groups in these locations. With globalization and southward migration of (former) pastoralist groups to the

4 Cha ter 1 coastal West African countries (especially in the wake of droughts and political crises as well as recently emerging terrorist attacks in the Sahelian region) the consumption of local milk is increasing in the whole region (FAOSTAT 2015), except for areas of out-migration. The data from 1961 to 2011 shows that the average annual per capita milk consumption in the sub-region is 40-70 liters of milk equivalents in the Sahelian countries (Mali, Niger, and Senegal), and 10-20 liters of milk equivalents in the costal countries (Benin, Togo, Nigeria, Ghana, Cote d'Ivoire, Guinea). In comparison to the nutritional recommendation of 100 liters of milk equivalents per person and year, milk consumption in West Africa is low (FAOSTAT, 2015). However countries national dietary guidelines are based on local food availability, cost, nutritional status, consumption patterns and food habits (FAO, 2019).

1.2.2 Socio-economic characteristics of (peri-) urban dairy production systems

The social background of those involved in dairy farming in the cities of West Africa has been described in several studies (Okantah et al., 1998; Somda et al., 2004; Agyemang et al., 2006; Bonfoh et al., 2007; Mingoas et al., 2014). In general, across countries, traditional cattle keeping ethnic groups such as Fulani pastoralists grant animal ownership to other urban ethnic groups who then take advantage of the high demand of dairy products in the cities and policies promoting dairy production (Bonfoh et al., 2007; Millogo et al., 2008; Toure et al., 2015). Dairy cattle are acquired by different social and economic channels such as heritage and trade (Bonfoh et al., 2007). Dairy farming is often carried out as a secondary activity in association with other activities such as crop farming and off farm activities (Somda et al., 2004; Agyemang et al., 2006; Mingoas et al., 2014). Reasons for keeping dairy animals vary from completely commercial reasons to non-commercial or cultural reasons. The involvement of tamers or investors from ethnic groups with different socio-cultural and economic background as compared to the pastoralists triggers changes in the way that dairy enterprises are being run (Bonfoh et al., 2007). Usually they tend to intensify milk production by improving the genetic potential of domestic cattle breeds through cross breeding and supplement feeding (Millogo et al., 2008; Toure et al., 2015). This then leads to diverse ways of resource use and management, stretching from purely cultural and traditional production systems of pastoralists or agro-pastoralists to "modernized" (peri-) urban dairy production systems with different levels of intensification, diversification and market orientation, characterised by higher cash flow and changes in labour type with hired labour replacing family labour (Somda et al., 2004).

5 Cha ter 1

1.2.3 Management features of (peri-) urban dairy production systems

In West African cities, the size of dairy farms varies from a few to hundreds of animals. This wide range has been described by several authors (Table 1.1 ). Due to anthropogenic pressure and shrinkage of agricultural areas around cities, small herds are predominant in the vicinity of city centers and large herds are mostly found further away from city centers where land is available for farming activities (Herrero et al., 2014). It is still unclear whether the distance to a city's commercial center is related to the level of intensification. Animals are managed in different ways ranging from grazing only to zero grazing, partial or seasonal grazing that combine different degrees of homestead or on-farm supplementation. Transhumance of dairy cattle away from cities has been described by a few authors but is declining in importance (Marichatou et al., 2005; Millogo et al., 2008). In the traditional cattle farming systems, of which the peri-urban dairy enterprises originate, a low emphasis is put on milk production which often is solely for home consumption. Despite the emerging urban demand for fresh milk and dairy products and the development of a local market, home consumption still prevails in urban dairy farming. Most farms use dual purpose breeds (milk and meat) and a considerable number of non-reproductive mature males (bulls or castrated} are usually reared on-farm for long periods, although one would expect the majority of adult animals to be productive cows (Table 1.1). Natural mating is predominantly used but artificial insemination is promoted by the government and non-governmental organisations (Bonfoh et al., 2007; Millogo et al., 2008; Toure et al., 2015) in order to improve the performance of indigenous breeds. Animal are mostly hand milked once or twice a day depending on milk yield and farmers' interest in calf growth. Different hand milking techniques are used ("full hand grip", "thumb in" and "pull down") but do not differ in their effects on teat treatment, milk yield or milk composition (Millogo, 2010). Milk and dairy products are either marketed or consumed by the dairy farming households at different degree across production systems and countries, emphasizing the traditional heritage of (peri-) urban dairy farms and the role of the urban demand in milk and dairy products as a driving force towards marketing (Table 1.1). At farm level, crop-livestock integration has been described but is less common as compared to rural settings. At city level, however, manure is used to fertilize crop fields and crop residues represent an important source of feed for dairy cattle which therefore enhance crop-livestock integration (Agyemang et al., 2006; Amadou et al., 2012; Roessler et al., 2016).

6 Cha ter 1

1.2.4 Cattle breeds used in (peri-) urban dairy production

The performance of breeds that are used for dairy production in West Africa has been widely studied in different settings (on-farm or on station; Table 1.2; Table 1.3). Yet, there have been decades of informal (not regulated, or implemented by farmers) and formal (organised by state authorities) pure breeding and cross breeding of local breeds, cross breeding of local breeds with exotic breeds, and pure breeding of exotic (European) breeds. This resulted in a wide range of genetic variability of dairy cattle used in (peri-) urban dairy production systems in West Africa (Marichatou et al., 2005; Millogo et al., 2008; Toure et al., 2015). Toure et al. (2005) identified 14 genetic types among 1428 pure breed and cross breed animals in the peri-urban dairy farms of Bamako, Mali, without further details on the genetic makeup of the crossbreds. Kouriba et al. (2002), however, identified 140 genetic types out of 3000 crossbreds in the same area. The situation is likely to be the same in other cities of West Africa, where dairy farming is expanding and animal breeding is poorly regulated and mainly carried out by farmers or farmer organisations who may or may not adopt government and non-governmental initiatives. It is well established that phenotypic traits or performances are determined by both genetic makeup and environmental influences. Natural selection over centuries has made local breeds well adapted to local conditions such as heat, long distance walking, poor quality fodder and tropical animal diseases, but with low reproduction and production performances. In contrast, exotic breeds have high performances (under specific management conditions) but a low tolerance of tropical environmental and management conditions (McDowell et al., 1985). By using cross breeds farmers take advantage of positive traits (disease resistance or tolerance and high performance) of local and exotic breeds while diluting the effect of negative traits (disease susceptibility and low performance; McDowell et al. 1985). However, critical analysis and comparison of production performances of dairy breeds should be done under the same environmental and management conditions in order to take into consideration on-farm conditions as there are multiple non-genetic factors of variation such as feeding intensity, disease pressure, production and reproduction management that might influence animal and farm performances (Messine et al., 2007; Coffie et al., 2015). Urban and peri-urban dairy systems in West Africa have a high potential for a profitable use of exotic germplasm and local breeds provided good management conditions (Roessler et al., 2019).

7 Table 1.1: Management features of dairy farms in West Africa cities.

Mainly Herd System H. stru. Crop Milk use Country City Authors dairy(%) size type(%) (% cow) farm (%) n P. gra. Z. gra. Con. Sol. Ouagadougou, Millage et al. (2008), Burkina Bobo- 76- Gnanda et al. (2016), Faso 54.5 76±22 57-100 2 95 Dioulasso 81 Roessler et al. (2016) Ngaoundere, 65 Fogwe (2015), Cameroon 5-240 89-100 0-11 49 35 Bamenda M ingoas et al. (2014) Khorogo, Cote Agyemang et al. (2006), Abidjan, 70 12-35 0 100 d'Ivoire N'goran et al. (2008) Bouake Western, Central River Gambia and lower 36 57±41 79 33 97 23 77 Somda et al.(2004) River Divisions Ghana Tamale 100 80 Roessler et al. (2016) Guinea Haute Guinea 34 10±5 59 52 80 53 47 Somda et al. (2004) CXl Bafata and G. Bissau 93 92±89 99 32 93 38 62 Somda et al. (2004) Gabu Regions Agyemang et al. 2006), Mali Bamako 44-47 10-120 66 42 38 67 Toure et al. (2015) Niger Niamey 17 56 32 68 Marichatou et al. (2005) Ayanwuyi et al. (2012), Nigeria Kufuna, Zaria 24-31 14 86 12-76 24-88 Agyemang et al. (2006) Fatick, Sow Dia et al. (2007), Senegal Kaolack, 26 100 38 60 22-38 62-75 Dieye et al. (2003), Dakar Ba Diao et al. (2006) H: herd, P. gra: partial grazing, Z. gra: zero grazing, H. stru.: herd structure, Con.: consumed, Sol.: Sold. Cha ter 1

Yet, there still exists a knowledge gap concerning suitable cross breeds or types that allow optimal milk production across the (peri-) urban dairy production systems in West Africa. In West Africa dairy calves usually suckle milk from their dams (Agyemang et al., 2006; Millogo et al., 2008) making it difficult to estimate the true milk yield of lactating cows of a given breed. Table 1.2 presents production performances of local breeds and exotic breeds, and Table 1.3 focusses on the production performance of crossbreds in West Africa.

1.2.5 Feeding strategies, access to feed resources and use efficiency

Feeding occupies a central place in farm management and different feeding strategies have been described for dairy enterprises of West African cities. Being at the crossroads between pastoral, agro-pastoral and intensive systems, farmers exploit freely available and accessible pastures or grazing lands as the exclusive feeding source or in association with on-farm feed supplementation (Ba Diao et al., 2006; Millogo et al., 2008; Mingoas et al., 2014; Fogwe, 2015). Therefore, their dependence on natural feed resources and vulnerability vary widely across production systems in the region. According to Diogo et al. (2010), (peri-) urban dairy farmers in Niger were using resources inefficiently. Likewise, dairy farmers in Sikasso were less efficient, highly extensive and therefore, underexploited their feed resources (Amadou et al., 2015). Limited community pasture areas and poor forage quality have been described by several authors in West Africa (Boudet, 1975a; Gaston and Lamarque, 1994; Coulibaly, 2006). Factors affecting composition and quality of the vegetation include rainfall, soils, plant species and pasture management. Feeds that are used on-farm can be classified as industrial by-products, crop residues, green forages produced on-farm, grasses, concentrates. Their chemical composition has been described in several studies, and when available the extent of their use varies across production systems (Somda et al., 2004; Ba Diao et al., 2006; Mingoas et al., 2014). Though the composition of a specific feed type changes under different conditions (the development stage of the plants at harvest, processing, conservation), a stronger determinant of whether fodder can fulfill the animals' nutrient requirement and insure full production performance is diet formulation. Farmers have access to feeds through cash (feed markets; Ayantunde et al. 2014) and non-cash ways (on-farm production, roadside/ bush cut and carry, grazing open access grasslands within and around cities; Hamadou et al. 2005).

9 Cha ter 1 Table 1.2: Age at first calving (AFC), calving interval (Ci), lactation length (LI} and total milk yield (TMY)/ daily milk yield (DMY) of main dairy cattle breed types used in (peri-) urban dairy farms of West Africa cities. Breed Country-city AFC Ci or LI TMY (kg) or Author {month} (daz:.SL * DMY (ll * Cameroon- 33±4 425±60 4284±1626 Njwe et al. (2001) Bamenda Osei et al. (1991), 480±68 2499±148 Ghana - Accra 34±1 Kabuga and LI: 332±15 4225±161 Holstein** Agyemang, (1984) Nigeria - Vom, Ngodigha & Etokeren 5252±365 Plateau State (2009) Senegal- 4541±1730 Ba Diao (2005), 36±7 511±149 Dakar DMY: 7 Ba Diao et al. (2006) Cameroon - IEMVT (1975) 31 419 2681 Bamenda Adebayo & Oseni Jersey** Nigeria 390 2160±35 (2016) Senegal - Ba Diao (2005) 36±7 511±149 3 096±784 Dakar Mont- Senegal - 3 605±1 356 Ba Diao (2005), 36±7 511±149 beliarde** Dakar DMY: 7 Ba Diao et al. (2006) Burkina Faso DMY· 1-2 Millogo et al. (2008) Cameroon - 513-295-536- IEMVT (1975) LI: 114-175 Fulani/ Bamenda 465 Peul*** Niger LI: 160-200 DMY· 2-3 Marichatou et al. (2005) Nigeria - Mrode (1988) 36-72 1018±18 lbadan Azawak*** Niger DMY: 7-8 Marichatou et al. (2005) Cameroon - IEMVT (1975) 48 511 373 Gudali*** Bamenda Niger LI: 230 1000-1100 Marichatou et al. (2005) DMY: 1.1±0.5 Ndama*** Gambia 1.4±0.6 Nouala et al. (2003) Ghana 162±12 Darfour-Oduro et al. Ghana LI: 164±9 Sanga*** DMY: 1 {201 O} *Precision is given when LI and DMY are presented, **European breeds, ***West African breed.

10 Cha ter 1

Table 1.3: Age at first calving (AFC), calving interval (Ci), lactation length (LI} and total milk yield (TMY)/ daily milk yield (DMY) of main crossbreds used in (peri-) urban dairy farms of West Africa cities. Crossbred Country-city AFC Ci orLI TMY (kg) or Author (month) (days)* DMY (lj * Cameroon- Mbah et al. (1991), 399 1575 Ngaoundere Mbah et al. (1987) Holstein X Ahunu and Acquaah Gudali Ghana LI: 268- 3047 (1987), 387 Ahunu et al. (1994) Holstein X 266±10 Darfour-Oduro et al. Ghana LI: 201±7 Ghana Sanga DMY: 1.4 (2010) Holstein X Cameroon- Mbah et al. (1991), 403 1551 Fulani Ngaoundere Mbah et al. (1987) Holstein X Gambia DMY:4±1.8 Nouala et al. (2003) Ndama 5±2.1 Jersey X Gudali Ghana LI: 238- 1563±90 Rege et al. (1994) 11.7 Jersey X Ghana Ghana 1459±47 Rege et al. (1994) Shorthorn LI: 254±6 Cameroon- Mbah et al. (1991), Jersey X Fulani 382 1011 Ngaoundere Mbah et al. (1987) Ahunu and Acquaah Ghana 1514±40 LI: 298 (1987) Cote d'Ivoire LI: 257 1277 Charray et al. (1977) - Bouake Jersey X Ndama DMY: 3±1.3 Gambia Nouala et al. (2003) 4±1.3

Mali LI: 281 1130 Tamboura et al. (1982) Montbeliarde X Cameroon- Mbah et al. (1991), 38 399 1380 Gudali Ngaoundere Mbah et al. (1987) Cote d'Ivoire -Abidjan, LI: 258±17 DMY: 5±1.4 N'goran et al. (2008) Montbeliarde X Bouake Ndama Mali LI: 326 1268 Tamboura et al. (1982) Senegal- Montbeliarde Fatick, Sow Dia et al. (2007), (Holstein) X Kaolack, DMY: 2.4-5.6 Ba Diao et al. (2006) Gobra Dakar *Precision is given when LI and DMY are presented.

11 Cha ter 1

1.2.6 Marketing and value chain of milk and dairy products

Historically, dairy farming in West Africa cities has gone through different political and structural changes that have affected not only the management and production level of dairy enterprises but have also shaped the value chain and the marketing of milk and dairy products (Bonfoh et al., 2007). Raw milk and dairy products are marketed through two different channels in West Africa. The first is the direct and informal channel where the farmer sells raw milk and self-processed dairy products directly to final consumers. The other is the indirect and formal channel where milk is collected from the farmers by collectors who then supply dairy plants (processing units) that will sell dairy products to final consumers directly or through specific shops or markets (Omore et al., 2004; Corniaux et al., 2005; Gidiglo, 2014). A well-organised dairy value chain is a prerequisite to ensure an efficient, permanent and sustainable supply of local milk and dairy products to consumers West Africa cities, but this reliable value chain is lacking so far. Until today the supply of milk and dairy products is strongly influenced by forage availability: milk production is usually high in the rainy season, but the share of marketed milk is relatively low, whereas in the dry season milk production is constrained by forage availability and quality but the marketing rate is higher, due to a better price for the milk (Dieye et al., 2003). Milk and dairy products are highly perishable and their safety, that is, hygienic quality throughout the milk value chain is of high concern for public health. To be able to supply quality milk and dairy products, all stakeholders throughout the value chain have to be involved and must collaborate. Problems are poor milking hygiene, poor hygiene of milk containers and milk handling personnel, and limited or no cooling - often milk contaminations increases as one moves down the chain to retailers (Omore et al., 2004; Millogo et al., 2008).

1.2.7 Constraints and opportunities

The lack of pro (peri-)urban dairy policies in most West African countries, and the general land tenure, land use and urbanisation policies do normally not promote the development of livestock keeping in and around cities (Schiere et al., 2006) despite government commitments towards assisting farmers (Grace et al., 2007). In addition, almost all West Africa countries import dairy products from Europe and America to meet the growing urban demand for milk and dairy products (Duteurtre and Corniaux, 2013). These imports are a

12 Cha ter 1 constraint to the domestic production since subsidized imported dairy products are dumped on local markets at less than local milk production costs, leading to very limited investments in local milk production and value chain development (Oxfam, 2002). In addition, city expansion and shrinkage of agricultural lands in their vicinities, climatic constraints inducing shortage of natural resources, namely water and feedstuffs, constitute additional problems for (peri-) urban dairy farming in West Africa. Animal diseases and their control still represent an important constraint to the development of dairy enterprises in many West Africa cities; both infectious endemic diseases such as foot and mouth disease, tuberculosis and brucellosis (Traore et al., 2004; Boussini et al., 2012; Boukary et al., 2013; Tialla et al., 2014) and parasitic diseases (Wymann, 2005) have been frequently reported and their impact on dairy production evaluated.

However, the availability of an expanding market for milk and dairy products, the development of local milk and dairy value chains with dairy plants collecting and processing fresh milk, supported by the willingness and the commitment of these processing units, non-governmental organisations and governments to support (peri-) urban dairy production in West Africa can boost the dairy value chain development (Duteurtre & Corniaux 2013; Corniaux 2015). Furthermore, in and around cities, veterinary inputs are more accessible and affordable as compared to rural areas, and health extension services or private veterinarians provide necessary technical assistances to improve the management of dairy enterprises as well as other livestock farming systems (Turkson, 2008). Given the poor feeding management already evidenced by other authors (Diogo et al., 2010; Amadou et al., 2015) there is still room to make West Africa (peri-) urban dairy enterprises more competitive and profitable by enhancing the efficiency of resource use. Therefore key features affecting resource use and use efficiency can be identified and site specific strategies for the improvement of West Africa (peri-) urban dairy production developed.

1.3 Urban and peri-urban pig production systems in West Africa

Pig farming has a considerable potential to supply the market with meat where pork consumption is not limited by cultural or religious factors, both because of high prolificacy and fast growth of this livestock species when adequately managed. In West Africa, pig farming has expanded at an annual rate of 2% to 6% (between 2004 and 2013) and the

13 Cha ter 1 annual pork production has approximated 800,000 tones representing 1 % of the global pig meat production (FAOSTAT, 2019). The consumption of monogastric products (from pig and poultry) in West Africa is projected to have a six to sevenfold increase until 2050 (Herrero et al., 2014). In this context, pig farming in and around cities plays an important role to ensure food security and adequate protein supply to urban dwellers. Yet, the opportunities provided to pig farmers by the important demand of the urban market are limited by several challenges inherent to urbanisation. In sub-Sahara Africa like in most developing countries, the potential of pig farming has not been fully exploited because of several reasons such as poor sanitation, poor disease control practices and poor nutrition (Kiendrebeogo et al., 2008; Mopate Logtene et al., 2014).

1.3.1 Socio-economic characteristics of (peri-) urban pig production systems

Pig farming is closely associated to the cultural or religious background of farmers. In West Africa mainly non-muslim people are involved in this agricultural activity (Kiendrebeogo et al., 2008; Fualefac et al., 2014; Mopate Logtene et al., 2014). As cities are conglomerates of diverse ethnics groups, ethnicity can be hardly considered as a determinant of engaging in pig production. Pig farming in West African cities is mainly carried out by small-holders owning a few pigs who generate income through selling of livestock and livestock products, which are also the main purposes for rearing pigs (Rekwot et al., 2005; Machebe et al., 2007; Kiendrebeogo et al., 2008; Fualefac et al., 2014). Owned by the household head who in most cases is male, small-size pig farms use mostly, if not exclusively, family labour (Kiendrebeogo et al., 2008; Fualefac et al., 2014). In some West African cities such as Dschang and Douala in Cameroon (Fualefac et al., 2014), Bobo-Doulasso, Gaoua, Kaya and Dori in Burkina Faso (Kiendrebeogo et al., 2014) there is a growing trend towards important private investments from richer households in pig farming. This translates into production intensification, employment of hired labour and a strong commercial orientation of pig farms.

1.3.2 Management features of (peri-) urban pig production systems

Both traditional and modern intensive pig farming systems have been described for West Africa cities (Mopate Logtene and Kabore-Zoungrana, 2013). In the traditional pig farming systems, pigs are either scavenging or kept indoors (year round or part of the year). The stable is usually made of local material requiring low investment and is of poor hygienic

14 Cha ter 1 standard (Kiendrebeogo et al., 2008). In order to avoid conflicts within the neighborhood, farmers in most cases enclose their animals, leading to an increased need for feed and labour. However, this not always translates to an increase in outputs because of additional inputs charges. Mating is hardly controlled in the traditional system as animals are free ranged. In the modern pig production systems evolving in the vicinity of West Africa cities, however, there is an important investment in housing, which is constructed of durable, easy to clean industrial material (Kiendrebeogo et al., 2008). Pig farming in West Africa cities is a profitable enterprise (Adetunji and Adeyemo, 2012). Kadurumba et al. (2014) showed that labour costs, costs of feedstuffs and supplements, farm size and to some extent the cost of veterinary care and medication determine the profit level in pig enterprises. Due to this, the traditional pig farming system (when existing) is likely to be more profitable still than the modern system because of low fixed costs (housing) as well as lower variable costs and input prices (labour, feed, medication) than the modern system. In addition, traditional pig farmers are less dependent on pork prices and in the end achieve a positive gross margin (Kadurumba et al., 2014). Furthermore, their animals have also social functions besides earning money (Porphyre, 2009).

1.3.3 Pig breeds used in (peri-) urban production systems

Several pig breeds are used for pig farming in and around West African cities. Reproduction and production performances of the so-called African local pig named Somo in Mali, Bakossi in Cameroun, Ashanti Black Pig in Ghana, Nigerian native in Nigeria, and Houehanou Gounouhanou egouhan in southern Benin have been described in several studies under specific management conditions, and reviewed by Agbokounou et al. (2016). Though the local African pig shows poor production and reproduction performance, it is well adapted to the local climatic conditions characterised by high ambient temperatures, diseases, and insufficient feed in the traditional pig farming system. Yet production/reproduction performances of the local pig types are subject to an important variability, which is not only due to genetic variability but also - and mainly - to the variability of management systems. Therefore, pig performance can be improved by improving the husbandry practices (Koutinhouin et al., 2009).

The semi-intensive and intensive pig farming systems are dominated by exotic breeds such as Large White, Landrace, and Duroc that have higher production performance when

15 Cha ter 1 properly managed (Adesehinwa et al., 2010; Kiendrebeogo et al., 2012b; Kouamo et al., 2015). The local types have been continuously and arbitrarily crossed with exotic breeds in West Africa, with the aim of improving the performance through genetic improvement. Yet, cross breeding of local pigs with exotic breeds is a threat for the local porcine genetic resources and there is an increasing number of voices calling for preserving the well adapted local breeds, which are especially valuable for the resource-poor rural farming communities (Amills et al., 2013).

1.3.4 Feeding strategies and access to feed resources

For pig feeding in and around West African cities, farmers mainly rely on locally available by-products from industrial and local cereal processing units (Machebe et al., 2007; Fualefac et al., 2014; Mopate Logtene et al., 2014). However, in the traditional extensive and semi intensive pig farming systems, organic household and food wastes and grasses contribute substantially to fulfilling the nutrient requirements of pigs (Machebe et al., 2007; Fualefac et al., 2014). Lekule and Kyvsgaard (2003) have reviewed the nutritional situation of scavenging pigs. An unbalanced diet based on cereals such as maize or sorghum, which are common feeds in poor pig raising households, supplies around 30% of the pig's requirements of the most limiting amino-acids (lysine and methionine). If combined with 20% scavenged feed, this amino acid supply can increase to about 80% of the requirement. In semi- intensive and intensive pig farming systems all nutrient requirements have to be provided through stall-fed balanced diets. The intensive pig farming thus only relies on purchased commercial feeds and is therefore highly sensitive to seasonal availability of feed, feed prices and price fluctuations - in West Africa (Porphyre, 2009) as well as in East Africa (Okello et al., 2015). There is still a knowledge gap to fill with respect to identifying and characterising locally available feedstuffs, in order to formulate diets that match the genetic material and to improve pig husbandry in West Africa (Lekule and Kyvsgaard, 2003).

1.3.5 Pig marketing and value chain

The West African pig and pork marketing system is characterised by formal and informal marketing channels and various segments moving mainly pigs and pork to producers, rural assemblers, wholesalers, commission agents and retailers (Ndebi and Ongla, 2006; Ajala and Adesehinwa, 2008; Kiendrebeogo et al., 2012a). The governments are usually

16 Cha ter 1 involved to different extents in the formal channel through meat quality control, disease surveillance, data collection, and provision of public market infrastructures (slaughtering houses) in main West African cities, with no direct participation or regulatory measures. Major participants in the pig market chain are the pig farmers who play a dual role as pig producers and buyers, the traders of live pigs and pork, and finally the pork consumers (Ajala and Adesehinwa, 2007; Ajala and Adesehinwa, 2008). Middlemen at different levels such as assemblers, commissioned agents, wholesalers and retailers play a major role for collection and purchasing pigs at farm gates, processing carcasses and pork at the slaughtering house or processing plant, and selling red meat at the market, or roasted pork at roadside snack bars (Ndebi and Ongla, 2006; Ajala and Adesehinwa, 2008; Kiendrebeogo et al., 2012a; Mopate Logtene and Kabore-Zoungrana, 2013). In some West African countries, market participants are organised in socio-professional groups; these associations advocate and lobby on the behalf of their members (Mopate Logtene and Kabore-Zoungrana, 2013). Using a market margin and efficiency analysis, Ajala and Adesehinwa (2008) showed that pig marketing was profitable but inefficient in various Nigerian cities. The same conclusion was drawn by Ndebi and Ongla (2006) in Cameroon as well as by Mopate Logtene and Kabore-Zoungrana (2013) in the capital cities of Cameroon (Yaounde), Chad (N'Djamena), Cote D'Ivoire (Abidjan) and Burkina Faso (Ouagadougou).

1.3.6 Urban and peri-urban pig production in West Africa: constraints and opportunities

There are several constraints to pig production in sub-Saharan Africa in general and in and around West Africa cities specifically. These are mainly related to urbanisation which implies a lack of space, and poor compatibility with formal or informal urban living standards (of the traditional as well as the modern pig systems) due to numerous nuisances (Porphyre, 2009). Marketing constraints are high transportation costs for (peri-) urban farms, absence of pricing according to standard weights and measures, lack of or inadequate slaughterhouses, absence of refrigerators, and lack of access to formal credit sources (Ajala and Adesehinwa, 2007). In addition the lack of technical knowledge, sufficient feeding base and management objectives of the farmer, added to the absence of appropriate supervision are important constraints to pig production in West Africa (Ndebi et al., 2009). Of much concern is the high mortality rate in many (peri-) urban farms (Mopate

17 Cha ter 1

Logtene et al., 2014) that is mainly due to highly contagious diseases such as African Swine fever. The observed high mortalities are favored by free-range production systems, local pig trade and a lack of biosecurity measures (Muhangi 2014; Awosanya et al. 2015). Other health problems relate to parasitic diseases such as cysticercosis that occur when pigs scavenge in areas with poor human sanitation systems (Zoli et al., 2003), but they decrease in importance as pigs are reared in intensive (housed) production systems (Lekule and Kyvsgaard, 2003).

Despite the production and marketing constraints highlighted above pig production is one of the most potential livestock sub-sectors that can contribute to the food security and income generation at smallholder farmer level in sub-Saharan Africa (Abdu and Gashaw, 2010; Moreki and Mphinyane, 2011; Muhanguzi et al., 2012; Atherstone et al., 2018). Pig production has several opportunities such as (1) An increasing number of people getting interested in consuming pork leading to the expansion of the market base, (2) In several African countries, the price of other meats is rising compared with pork, and recent campaigns encouraged people to consume more poultry, rabbit and pork products, (3) An increase in the number of outlets, such as local butcheries and small-scale processors, gives farmers a wider choice of markets for their product, (4) With the shrinkage of agricultural lands due to cities' expansion, human population increase and fragmentation of land, land-intensive livestock species such as pigs gain advantage over those that require more land such as cattle, (5) Farmers when properly trained can increased their profit margins by producing their own feeds (or balanced diet) ; the potential to reduce production costs makes pig-rearing even more viable, (6) Intra African countries export markets for pigs and pork products already exist, enhancing the market base, (7) The short production cycle of pigs gives them an advantage over other livestock such as cattle and, (8) The training of pig farmers in several African countries has been progressively affecting pig farmers' interest to adequate farm management (FAO, 2012). Furthermore, in recent years, the international community, national authorities, pork industry stakeholders and researchers have made a commitment to sustainably resolve the constraints affecting swine production leading to the development of an Africa-wide strategy for the prevention and control of African swine fever (FAO/AU-IBAR/ILRI, 2017).

18 Cha ter 1

1.4 Research problem, objectives and hypotheses

Rapid population growth in parallel with a considerable increase in the demand for livestock products in West African cities constitute a driving force that is likely to shape size, breed selection and livestock management in (peri-) urban livestock production units. In addition, livestock farming households are of different socio-economic backgrounds (Thys et al., 2005) that affect not only the production level of the farm but also resource use and use efficiency.

Access to resources and their use by livestock farmers vary within and across livestock production systems (Diogo et al., 2010). Thus, determinants of resource use and use efficiency are specific to farm types and are still to be understood in West Africa cities both for dairy, beef and pig farm types.

Urban livestock farmers are highly market oriented and trends towards intensification have been frequently reported. Yet the impact of intensification and market orientation on resource use and use efficiency still shows knowledge gaps (Dossa et al., 2011; Amadou et al., 2012).

Against the above, the main objective of this study was to identify and analyse factors that influence resource use and use efficiency in (peri-) urban dairy, beef cattle and pig production systems in the West African city of Ouagadougou, Burkina Faso. The specific objectives were:

• to characterise the different livestock production systems that are operating in that city,

• to identify the determinants of resource use and use efficiency of the different dairy, beef cattle and pig production systems,

• to identify bottlenecks of optimised livestock production and to make appropriate recommendations for the improvement of resource use efficiency and farm production performances in the study area.

Access to biophysical and socio-economic resources and their combined use as inputs into livestock production by farmers vary across different (peri-) urban livestock production systems and determine farm production level, resource use efficiency and consequently

19 Cha ter 1 sustainability of these systems. In addition, urban dwellers keep livestock for different purposes that affect production decisions and therefore resource use directly or indirectly, such as generation of tangible benefits ( cash revenue, products for household consumption, manure) and intangible values (cultural aspects, savings, insurance).

It is therefore hypothesised that:

1. Intensification and specialisation lead to high animal performances, due to continuously high levels of feed input through purchase and storage in intensive dairy, beef cattle and pig farms, and to a reduced variability of feed offer throughout the year or production cycles.

2. Market orientation is positively correlated to resources use efficiency. The more market oriented a farm is, the higher is its production and resource use efficiency.

3. Resource use efficiency in (peri-) urban livestock production systems is related to the breeds used in the different production systems, in other words: the use of improved breeds enhances resource use efficiency.

4. In cattle, resource use efficiency is similar for dairy and beef cattle production systems, with comparable level of resource use efficiency in both production systems according to the level of intensification.

1.5 Thesis structure

This PhD thesis is structured in five chapters. Chapter 1 presents the general introduction with a systematic summary insight into the socio-cultural, agro-ecological and economical embeddedness of (peri-) urban livestock production in and around West African cities and a particular emphasis on cattle (dairy and beef) and pig farming. Chapter 2 classifies and characterises livestock production systems in Ouagadougou (Burkina Faso), and analyses whether livestock farms develop towards specialisation or integration, and intensification or extensification. Chapter 3 analyses livestock and resources management practices in dairy enterprises of Ouagadougou and links farm management to production performances, milk composition and metabolic diseases and resource-use efficiency. Chapter 4 analyses the performance of pig and beef cattle farms in Ouagadougou including farm management, feeding management, and animal growth. Chapter 5 provides a general discussion and draws conclusions and recommendations from the various aspects of the thesis.

20 Cha ter 2

2. Characterisation of urban and peri-urban livestock production systems and trends toward integration, specialisation and intensification

21 Cha ter 2

2.1 Introduction

Between 2000 and 2030, the continuous population growth in West Africa cities will increase the demand for proteins of animal origin by more than 200%, and by 300% for commodities such as milk and pork (FAO, 2011 ). This change is three times the projected increase in the rural demand for the same products (FAO, 2011 ). Livestock production systems have evolved within the main and secondary West African cities since decades and supply the city dwellers with fresh livestock products (milk, meat and eggs). Yet, this supply is insufficient (FAO, 2011) and improving the efficiency of livestock enterprises is highly relevant in the context of limited resources (Diogo et al., 2010; Amadou et al., 2015). Various types of livestock enterprises operate in different ways in and around West African cities (Dossa et al., 2011; Abdulkadir et al., 2012). Livestock farming in those cities has evolved in various production systems characterised by different levels of intensification/resource use efficiency driven by changing agro-ecological, socio-political and economic forces which have shaped their features within and across countries throughout the years (Tiffen, 2004; Bonfoh et al., 2007). Population growth and urbanisation have had several impacts on livestock rearing in West Africa such as diminishing land available for agricultural activities, triggering intensification of livestock production with labour intensive cut-and carry feeding methods, and use of industrial by products in feeding strategies, amongst others (Tiffen, 2004). In a poorly regulated political environment, shrinkage of agricultural land related to urbanisation and cities' expansion leads to a fierce competition for the few remaining open spaces (Dossa et al. 2015). Thus, the socio-economic background of livestock keeping households, the geographical location of the farm, the species and numbers of livestock kept, their management (e.g., feeding and breeding strategies, health care) and product processing, marketing and consumption are important determinants of (peri-) urban livestock production systems. Characterising livestock production systems allows for a better understanding of the underlying key production features that have to be taken into consideration in order to improve the systems' efficiency and productivity, along with food security and farmers' livelihood (Dossa et al., 2011 ). Thus by identifying homogenous livestock production systems, site specific bottlenecks for the optimisation of production performances are characterised and improved resource use and management strategies for problem mitigation can be meaningfully developed and efficiently implemented (Notenbaert et al., 2009) in the ever

22 Cha ter 2 changing (peri-) urban environment (Tiffen, 2004). In West Africa, (peri-) urban livestock production systems have been studied and characterised by several authors (Okantah et al., 1998; Sidibe et al., 2004; Somda et al., 2004; Rekwot et al., 2005; Agyemang et al., 2006; Ajala et al., 2007; Fualefac et al., 2014; Kiendrebeogo et al., 2014; Mingoas et al., 2014; Tindano et al., 2015). However most of those authors have been mainly focusing on one livestock species, missing out the full picture of multispecies (or mixed farming) systems as studied by Amadou et al. (2012). Thus, there is still more insight to be gained about the development trends and trajectories that (peri-) urban livestock systems may undergo under the changing and challenging urban environment of West African cities. The aim of this chapter is to identify and characterise livestock production systems in and around Ouagadougou, Burkina Faso a typical densely populated and growing West African capital city (Commune de Ouagadougou, 2011; INSD, 2014) characterised by the keeping of several livestock species by more than a quarter of households (Thys et al., 2005; Bellwood-Howard et al., 2015). In accordance with Baltenweck et al. (2003), livestock production intensification is characterised by increased inputs in terms of labour and capital flowing into the livestock production unit through feeding of grains, improved housing, controlled breeding and better health management and higher outputs (animals and animal products). Unlike intensification, extensification requires less labour and capital inputs and delivers low marketed outputs. Rearing of one livestock species or producing one main product (i.e. milk or egg) is the key feature of specialised livestock farms and livestock production is their main activity. Crop farms or cultivated lands when present are expected to be smaller in specialised livestock farms than in integrated multiple livestock species crop-livestock farms. Thus, through livestock keeping households' surveys, classification, cross-tabulation and logistic regression, trends towards specialisation or integration, and towards intensification or extensification of (peri-) urban livestock production systems were analysed. Furthermore determinants of trends were identified and their relation to resource use (feed, manure) of livestock tamers assessed. We hypothesised that (i) market orientation is the key determinant of intensification, that (ii) specialisation and intensification are mostly found on dairy cattle and pig farms, and that (iii) a great majority of (peri-) urban livestock farmers still rely on freely available feed (open grassland for pasture, and food waste).

23 Cha ter 2

2.2 Material and methods

2.2.1 Study site

The study site was Ouagadougou the densely populated capital city of Burkina Faso, located in the hot semi-arid Sudan-Sahelian zone. The average annual rainfall is about 815 mm. The rainy season stretches from May to October. The cool dry season runs from December to January and the hot dry season runs from March to May (WorldClimate, 2016). Since 2009, Ouagadougou is divided in 12 subdivisions and 55 sectors, forming together the (peri-) urban area and covering a surface of about 318 km2 (Commune de Ouagadougou, 2011 ). Ouagadougou is the most populous city of Burkina Faso. During the last three decades, the population of Ouagadougou has been increasing significantly from year to year. Indeed, from 465,969 inhabitants in 1985, the population reached 1,915,102 inhabitants in 2012 corresponding to 4 to 5 folds increase over a period of three decades (INSD, 2017). Ouagadougou city dwellers are involved in several occupations including agriculture, industry, arts and trade at different scales. In fact, across and around the city there are small crop farms of millet, maize, sorghum, and market gardening or off-season crops (Bellwood-Howard et al., 2015). In 2010 urban agriculture was estimated to be carried out on an area of 151 ha with 1244 producers around the main dams of the city and dams located on the outskirts of the city (Commune de Ouagadougou, 2011 ). Urban and peri-urban livestock keeping is a popular activity in Ouagadougou such that one and a half decades ago already more than a quarter of the households in Ouagadougou owned livestock including all major livestock species (Thys et al., 2005). The main livestock species that are reared are cattle, sheep, goats, pigs, poultry, and horses (Thys et al., 2005; Bellwood-Howard et al., 2015). Livestock production is dominated by the extensive production type and is sedentary. Nevertheless, modern and semi-modern livestock production is developing in the peripheral zone, with the establishment of farms specialising in milk production and fattening (Gnanda et al., 2016).

2.2.2 Livestock keeping households' sampling approach and data collection

The list of livestock farmers provided by the extension and advisory services of the Ministry of Animal Resources of the Kadiogo Region, Burkina Faso, was updated by livestock farmers' associations and used to randomly select 13 out of 55 sectors of the city of Ouagadougou by probability proportional to size (PPS) sampling. In each selected sector

24 Cha ter 2 tamers keeping livestock for commercial purpose were stratified according to the main species kept, i.e. ruminants and non-ruminants. The main species kept was the livestock species they declared keeping at the extension services of the Ministry of Animal Resources. Ten livestock farmers each (5 ruminant farmers and 5 non-ruminant farmers) were then selected by simple random sampling in 12 selected sectors, and 20 farmers (10 of each type) were selected in one sector because of its high number of livestock keeping households as compared to the other sectors. This resulted in 140 livestock-keeping households. 41 additional households from a previous survey on (peri-) urban agriculture carried out by the UrbanFoodPlus project (Bellwood-Howard et al., 2015) were included in the sample; these owned at least 5 tropical livestock units (TLU; 1 cattle represents 0.8 TLU, one pig 0.2 TLU, one sheep or goat 0.1 TLU and one head of poultry 0.01 TLU). In total 181 livestock keeping households were surveyed from January to Apri I 2014 using a semi-structured questionnaire. The questionnaire was pre-tested and later on necessary corrections were taken into consideration for the improvement of the final version. Prior to the interview informed consent was obtained from the interviewee. The data collected was related to household socio-economic characteristics (education level, location, migration status, education level, ethnicity, labour type), farm assets (ownership of farming equipment such as tractor, milking machine, water pump, stall, motorbike, wheelbarrow, plough ... ), crop production (yes/no), livestock ownership (species, numbers) and management (feeding methods, number of livestock units per species, mating, practice of cattle transhumance ( type of pastoralism) characterised by a seasonal movement of livestock in search of better pastures), as well as livestock intakes (purchases) and offtakes (sales of animals, and of products). Labour force was calculated according to the definition of the International Labour Organisation of the working age group which excludes household members younger than 16 years, and by applying the following conversion factors: 1.0 for males aged between 16 and 55 years, 0. 75 for females between 16 and 55 years, 0. 75 for males above 55 years and 0.5 for females above 55 years.

25 Cha ter 2

...... -. ..., . .. - ...... �···\

0 30km

Figure 2.1: Location of the 181 interviewed (peri-) urban livestock keeping households in Ouagadougou, Burkina Faso, from January to April 2014.

2.2.3 Statistical analyses

Prior to analysis the data collected from the 181 households was screened for consistency and completeness, and only 157 households remained for further statistical analyses. By using the software SPSS version 20.0 (IBM Corp., 2011) , categorical principal component analysis (CatPCA) and two-step cluster analysis were used to classify livestock production types operating in Ouagadougou into homogenous clusters. For CatPCA, 29 potential classification variables (Table A.1) related to socio-economic variables, farm characteristics and livestock management were pre-selected. Multicollinearity for continuous variables was tested by Spearman correlation test (correlation coefficient of �0.7 as benchmark). Effect sizes between nominal (independent) and continuous variables (dependent) (benchmark of 0.25 as suggested by Cohen, 1988) was estimated using rj2 statistics. The strength of the association between nominal variable was measured by Cramer's V using a benchmark of �0.7. The significance level for correlations was set at p s 0.01. Finally, 24 variables were used for CatPCA after exclusion of strongly correlated variables. The

26 Cha ter 2 number of dimensions was set at 2. The variables identified as the most discriminating variables to be used for the subsequent two-step cluster analysis had a loading score �0.5 on one of the two components. Thus, the following variables remained for two-step cluster analysis: Location, migration status, education level, ethnicity, labour type, sale of milk, number of pigs, practice of cattle transhumance, feeding method of goats and cattle, respectively, feeding method of pigs, controlled mating and number of TLU sold. Out of several cluster solutions explored, the final cluster solution was retained because of its sound equilibrium between the number of variables and the cluster quality ( measured by silhouette measure of cohesion and separation; IBM Corp., 2011). Variables with the lowest impact on cluster creation were discarded such that the best cluster solution was obtained with 6 variables for 4 clusters, and a silhouette measure of cohesion and separation of 0.5 (goodness-of-fit statistics BIC=1042.69). Figure 2.2 presents the cluster determining variables.

Cross tabulation and binary logistic regression were carried out to analyse trends of livestock production, to characterise factors influencing the identified trends and to assess the impacts of those factors on feed and manure use. Binary logistic regression with stepwise backward elimination of predictors was used to assess the odds of livestock keepers' involvement in crop farming (yes/no), production of milk (yes/no), provision of health care (yes/no), keeping of animals in barns (yes/no) and practice of controlled mating (yes/no) from a set of independent predictor variables such as location in or around the city, the education level of the household head, available household labour force, land ownership and number of animals owned, as well as migration status of the household head. An iterative elimination of predictor variables not useful in predicting the dependent variables was applied.

2.3 Results

2.3.1 Identification of livestock production system types

In total 4 clusters are identified. The identified clusters or livestock production systems (OUA-n, n = 1, 2, 3, 4) are described below:

27 Cha ter 2 Cluster 1 (OUA-1): Pig production; n = 50 (31.8%)

Households belonging to this cluster are characterised by the absence of cattle and related management variables which differentiate them from the farms of the three other clusters (Figure 2.2; Table 2.4). Three quarters of the households in OUA-1 keep pigs. This differentiates them from OUA-3 and OUA-4 where there are no pigs. There is no difference in terms of pig herd sizes and piggery management between OUA-1 and OUA-2 (Figure 2.2; Table 2.4; Table 2.7).

Cluster 2 (OUA-2): Mixed crop-pig and non-dairyruminant production; n = 17 (10.8%)

All households in cluster OUA-2 keep pigs, do not send cattle to transhumance and are not involved in milk production unlike households in OUA-3 and OUA-4 (Figure 2.2). All households of this cluster own cattle, and three quarter own small ruminants. Grazing the ruminants in the city periphery is common in households OUA-2 and also in the other 3 clusters (Table 2.10).

Cluster 3 (OUA-3): Mixed crop-(dairy) cattle production; n = 54 (34.4%)

All households of OUA-3 and OUA-4 own cattle and are not involved in pig production. One quarter of the households of OUA-3 are engaged in milk production (in comparison to 100% households in OUA 4, p � 0.001) (Figure 2.2; Table 2.5) and the average daily milk production per farm is higher in OUA-3 than in OUA-4. In addition the amount of milk sold is lower in OUA-4 than in OUA-3. Households in OUA-3 less frequently carry out cattle transhumance and grazing of cattle as compare to OUA-4 households (Figure 2.2).

Cluster 4 (OUA-4): Mixed crop-dairyproduction (Fulani owners), n = 36 (22.9%)

In contrast to the other 3 clusters household heads of OUA-4 belong to the Fulani ethnic group (Figure 2.2). All household are involved in milk production (Table 2.4; Table 2.5). Transhumance is practiced by 20% of the households in OUA-4 unlike OUA-2 and OUA-3 where it is not reported. Of the households practicing transhumance, 57.1% do so in the dry season, 28.6% in the rainy season and 14.3% year round. Transhumance duration is 3.1±1.35 months on average and the average covered distance is 59.3±66.74 km (one way).

28 Cha ter 2

Mossi ethnicity Mossi ethnicity 1.0 1.0 ·

Stall-feeding Cattle Stall-feeding Cattle pigs transhumance pigs 'transhumance Ij l l Pig herd size Sale of milk Pig herd siz Sale of milk

Cattle grazing OUA-1 (n=50) Cattle grazing OUA-2 (n=17)

Mossi0.8 ethnicity Mossi0.8 ethnicity 1.0 1.0 0.6 Stall-feeding Cattle Stall-feeding o.6 Cattle pigs , transhumance pigs transhumance I ] I r· to�� :0 0.�, j l . 1] I Pig herd size� jSale of milk Pig herd size

Cattle grazing Cattle grazing OUA-3 (n=54) OUA-4 (n=36) Figure 2.2: Main characteristics of the identified four livestock farm types in Ouagadougou (OUA-n, n = 1, 2, 3, 4) using spider web diagrams (n = 157).

2.3.2 Trends of livestock production in Ouagadougou

2.3.2.1 Crop-livestock integration or specialisation in livestock production

More than two thirds (70.7%) of the livestock keepers in Ouagadougou also carry out cropping activities (Table A. 1 of "Appendix"). Of the 157 household heads, 42.0% are crop-livestock farmers, 13.4% livestock farmers, 20.4% are self-employed, 12.1 % employed off-farm, 4.5% are retired and 7.6% are involved in other professional activities (Table 2.1). Concerning crop production, differences between clusters are significant (p :5 0.001) with fewer households in OUA-1 engaged in cropping (50%) than in OUA-2 (94.1%), OUA-3 (75.9%) and OUA-4 (80.6%).

Keeping multiple livestock species is common across households (89.8% of households) (Table 2.2). Local chicken and guinea fowls are the most frequently encountered species

29 Cha ter 2 (77.7% of households, with the highest and the lowest proportions in OUA-4 and OUA-1, respectively; p � 0.001 ), followed by cattle (68.2%, with all households in OUA-2, OUA-3 and OUA-4 keeping cattle unlike in OUA-1 with no cattle keeping household), sheep (65%, with the highest and the lowest proportions in OUA-4 and OUA-1, respectively; p � 0.01) and goats (58%, with the highest and the lowest proportions in OUA-2 and OUA-1, respectively; p � 0.05). Pig keeping households are less important (34.4%) and only found in OUA-1 and OUA-2 with a proportion higher in OUA-2 (100%) than OUA-1 (74%); p � 0.01 (Table 2.4).

Only 10 .2% of households keep one single livestock species (with the highest and the lowest proportions in OUA-1 and OUA-2, respectively; p > 0.05) and pig is the most frequently encountered animal in households keeping only one species (81.2%, p � 0.001) (Table 2.2). Purely pig keeping households are all clustered in OUA-1. The 18.8% remaining households keeping only one livestock species reared cattle. In OUA-1, 78.4% of the households are located in the peri-urban area and keep higher number of pigs than other households in OUA-1 located in the intra-urban areas of Ouagadougou. Milk producers are grouped in OUA-3 and OUA-4 with a significantly higher frequency in OUA-4 than in OUA-3 (100% in OUA-4 vs. 26% in OUA-3; p � 0.001 ). Average cattle herd size and total number of TLU are significantly higher (p � 0.001) in OUA-4 than in the other three clusters (Table 2.4). Across clusters average goat and sheep herd sizes are not significantly different (Table 2.4).

Table 2.1: Main occupation (self-assessed) of the heads of four different (peri-) urban farm types in Ouagadougou (n = 157).

Livestock Crop-livestock Self- Other HH Employed Retired keet!er farmer emt!loted occu t!ation Cluster n % % % % % % OUA-1 50 14.0 30.0 26.0 10.0 4.0 16.0 OUA-2 17 5.9 58.8 17.6 5.9 11.8 0.0 OUA-3 54 11.1 37.0 22.2 16.7 5.6 7.4 OUA-4 36 19.4 58.3 11.1 11.1 0.0 0.0 OUA 157 13.4 42.0 20.4 12.1 4.5 7.6 p 0.574 <0.05 0.398 0.669 0.229 <0.05 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n= number of HH per cluster.

30 Cha ter 2

Table 2.2: Number of livestock species per household (HH) and livestock species kept in four different (peri-) urban farm types in Ouagadougou (n = 157). One livestock Type of single livestock Two livestock Three or more HH s�ecies s�ecies s�ecies livestock s�ecies Cluster n % % % % OUA-1 50 26.0 Pigs (100.0) 48.0 26.0 OUA-2 17 0.0 0.0 100.0 OUA-3 54 1.9 Cattle (100.0) 11.1 87.0 OUA-4 36 5.6 Cattle (100.0) 5.6 88.9 OUA 157 10.2 Pigs (81.2), Cattle (18.8) 20.4 69.4 <0.001 <0.001 <0.001 e.p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number of HH per cluster.

The results of the logistic regression (Table 2.3) indicate that households in the peri-urban areas are more likely to cultivate crops (e13 = 7.5; p � 0.001) than households in the urban area. Also, as household heads get more educated, engagement in cropping activities decrease significantly. Households with heads that have only acquired primary education are more likely to cultivate crops (e13 = 8.0; p � 0.01) than households where the heads have secondary education and tertiary education. The type of land ownership (own, allocated, borrowed or leased} and total number of TLU does not significantly affect a household's decision to engage in crop farming. Households with heads that have migrated to Ouagadougou from another location are less likely to cultivate crops (e13 = 0.9; p � 0.05) than autochthonous households. Engagement in milk production is significantly affected by the education level of the household head (p � 0.05). Milk production is less likely adopted by households with heads that reach secondary (e13 = 0.3, p = 0.3) and tertiary education level (e13 = 0.9, p = 0.9). Furthermore, the total number of TLU owned influences household's decision to engage in dairy production ( e13 = 1.2; p � 0.001).

31 Cha ter 2

Table 2.3: Logistic regression parameters for variables predicting crop and livestock integration or specialisation in dairy enterprises across households in Ouagadougou (n = 157). Wald's Odds Predictors SE df x2 p Ratio Integration of livestock and crops Constant -2.437 1.646 2.191 1 0.139 NA Peri-urban location 2.013 0.566 12.627 1 0.000 7.486 Education level 8.793 3 0.032 Level 1 (primary education) 2.079 0.785 7.011 1 0.008 7.999 Level 2 (secondary education) 1.357 0.830 2.672 1 0.102 3.884 Level 3 (tertiary education) 0.643 0.823 0.611 1 0.434 1.903 Land ownership 1,409 2 0.494 Land ownership 1 (allocated land ) -1.413 1.192 1.405 1 0.236 0.243 Land ownership 2 (borrowed, leased) -1.208 1.168 1.070 1 0.301 0.299 Total TLU owned (n) 0.025 0.018 1.832 1 0.176 1.025 Available labour force (n) -0.058 0.076 0.583 1 0.445 0.944 Farmer migrated from another location -0.059 0.025 5.531 1 0.019 0.943 Goodness-of-fit test (model x2) 53.932 9 0.000 Milk production Constant -5.204 2.529 4.234 1 0.040 NA Peri-urban location -0.123 0.706 0.030 1 0.862 0.884 Education level 8.008 3 0.046 Level 1 (primary education) 1.246 0.915 1.854 1 0.173 3.476 Level 2 (secondary education) -1.094 1.123 0.949 1 0.330 0.335 Level 3 (tertiary education) -0.156 1.085 0.021 1 0.886 0.856 Land ownership 1.863 2 0.394 Land ownership 1 (allocated land) 1.657 2.076 0.637 1 0.425 5.241 Land ownership 2 (borrowed, leased) 2.245 2.009 1.248 1 0.264 9.442 Total TLU owned (n) 0.161 0.031 26.496 1 0.000 1.175 Available labour force (n) -0.078 0.101 0.593 1 0.441 0.925 Farmer migrated from another location -0.022 0.040 0.297 1 0.586 0.978 Goodness-of-fit test (model x2) 93.955 9 0.000 p: p-value, NA: Not applicable.

32 Table 2.4: Livestock ownership in four different (peri-) urban farm types in Ouagadougou (n = 157). Cattle �er HH Goats �er HH Shee� �er HH Pigs �er HH Poultry �er HH Cluster n %of Mean±S.D. % of Mean±S.D. % of Mean±S.D. % of Mean±S.D. %of Mean±S.D. HH HH HH HH HH OUA-1 50 0.0 40.0 13.2±7.95 36.0 18.2±10.40 74.0 26.2±17.32 58.0 29.7±17.18 OUA-2 17 100.0 8.3±5.38 76.5 9.8±9.85 76.5 11.9±6.99 100.0 28.8±34.17 94.1 50.7±52.79 OUA-3 54 100.0 15.9±17.82 63.0 12.2±10.46 77.8 18.7±13.85 0.0 33.0±23.14 OUA-4 36 100.0 35.0±19.29 66.7 15.5±14.78 80.6 20.3±18.55 0.0 91.7 26.3±18.16 Full 157 68.2 21.1±19.78 58.0 13.0±11.11 65.0 18.2±14.28 34.4 27.0±23.62 77.7 32.7±27.34 sample E. <0.01 ::;0.001 <0.05 0.476 <0.01 0.358 <0.01 0.711 <0.01 <0.05 p: p-value for differences between clusters, Fisher's exact test for proportions. Mean calculated only for those households with the respective livestock species, Small superscript letters for significant differences between clusters (Kruskal-Wallis test), HH: Household, n: number of HH per cluster, S.D.: Standard deviation.

Table 2.5: Milk, eggs and livestock sales by four different (peri-) urban farm types in Ouagadougou (n = 157).

(,.) (,.) HH Milking Daily milk Sale of Milk sold Sale of TLU sold in Poultry Selling eggs with production milk daily animals 2013 keepers cattle (kg/farm) (kg/farm) Cluster n % of HH Mean±S.D. % of HH Mean±S.D. % of Mean±S.D. n % of poultry with with tota I keeping HH cattle cattle HH OUA-1 0 NA NA NA NA 78.0 3.7±4.95 29 3.4 OUA-2 17 0.0 NA 0.0 NA 88.2 4.3±3.01 16 25.0 OUA-3 54 25.9 46.7±74.81 24.1 48.7±77.59 77.8 3.8±4.04 44 6.8 OUA-4 36 100.0 16.4±10.30 97.2 12.0±9.90 97.2 3.8±2.71 33 6.1 Full samgle 107 46.7 24.9±41.83 44.9 21.9±43.35 83.4 3.8 ±3.90 122 8.2 p ::;0.001 0.761 ::;0.001 0.289 0.056 <0.05 0.066 p: p-value for differences between clusters, Fisher's exact test for proportions, Mean calculated only for those households with the respective livestock species, Small superscript letters for significant differences between clusters (Kruskal-Wallis test), Total number of households: OUA-1=50, OUA-2=17, OUA-3=54, OUA-4=36, HH: Household, n= number of HH with cattle for each cluster, S.D.: Standard deviation, NA: Not applicable. Cha ter 2

2.3.2.2 Intensification or extensification of livestock production Several cattle types are used by the different livestock keeping households. The majority of cattle keeping households own local breed (zebu Peul) (84.7% of Households) followed by exotic cross breeds between local zebu and European breeds (29% of households) then Sahelian zebu that is Gudali zebu (24.2% of households) and Azawak zebu (20.2% of households). All cattle keeping households in OUA-2 use the local zebu with higher proportion in OUA-2 than the other two cattle keeping clusters (p � 0.05), else there is no significant difference between clusters (p > 0.05) with regards to the percentage of households keeping a specific cattle type (Table 2.6). For small ruminants, two main breeds are used namely the Djallonke breed (84.7% for sheep, 94.2% for goats) and the Sahelian breed (26.5% for sheep, 7. 7% for goats). Crosses between Djallonke and Sahelian breeds are also used in 15.7% and 4.8% of the households with sheep and goats, respectively. Controlled mating is practiced by half of the livestock owners. Cattle are more subject to controlled mating (48.6% vs. 46.3% for pigs, 29.4% for sheep, 16. 7% for goats and 1.6% for local poultry) (table 2.7). About half of households keep animals in barns and Households in OUA-1 and OUA-2 more frequently confine their animals than those in OUA-3 and OUA-4 (p � 0.001 ). The majority of livestock keepers (90.4%) provide health care to their animals, such as prophylactic vaccination, curative medical treatments or consultation of veterinary services in case of health problems. The proportion of households keeping a specific livestock species and providing healthcare to that species varies between species such that local poultry (4.9% of local poultry keepers) receive less health care than goats (26.4 % goat keepers), pigs (61 .1% of pig keepers), sheep (66. 7% of sheep keepers) and cattle (93.5% of cattle keepers). In contrast a significantly lower proportion of livestock keepers in OUA-1 provide health care to their animals as compared to other three clusters (p � 0.001) (Figure 2.3).

Table 2.6: Cattle types owned by four (peri-)urban farm types in Ouagadougou (n=157). HH with Fulani Gudali Azawak European Local cattle Zebu Zebu Zebu crosses crosses Cluster n % ofHH % ofHH % ofHH % ofHH % ofHH OUA-1 0 NA NA NA NA NA OUA-2 18 100.0 11.1 0.0 11.1 0.0 OUA-3 63 77.8 31.7 23.8 31.7 6.3 OUA-4 43 88.4 18.6 23.8 32.6 2.3 Full sample 124 84.7 24.2 20.2 29.0 4.0 p <0.05 0.112 0.070 0.193 0.376 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number ofHH per cluster, NA: Not applicable.

34 Table 2.7: Use of controlled mating in different livestock species by (peri-) urban households (HH) in Ouagadougou (OUA; n = 157).

HH with Cattle* HH with Sheep HH with Goats HH with Pigs HH with local Local cattle shee� goats �igs �oultry �oultry Cluster n %of n %ofHH n %ofHH n % ofHH n %ofHH HH OUA-1 0 NA 18 27.8 20 10.0 38 52.6 30 3.3 OUA-2 17 41.2 13 30.8 12 0.0 16 31.2 16 0.0 OUA-3 54 40.7 42 26.2 34 20.6 0 NA 48 2.1 OUA-4 36 63.9 29 34.5 24 25.0 0 NA 33 0.0 Full sample 107 48.6 102 29.4 90 16.7 54 46.3 127 1.6 p 0.089 0.890 0.203 0.232 0.802 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number of HH per cluster, NA: Not applicable * Artificial insemination is used by 12.7% of cattle keepers, of which 55%in OUA-3 and 35% in OUA-4.

(,,) 01 Cha ter 2

100

Cl) 60 0

Cl) 0 40

0 Mating controlled Health care provided Animals housed in barns •OUA-1, n = 50 oOUA-2, n = 17 li!'.IOUA-3, n = 54 riOUA-4, n = 36

Figure 2.3: Breeding, health care provision and housing management of livestock in four different (peri-) urban farm types in Ouagadougou (n = 157).

The decision to control mating is significantly influenced by the education level of household heads with those who have only received primary education being less likely to control mating (ef-l = 0.18 for primary education level, p � 0.05) than those who have reached secondary and tertiary education level. Likewise, the practice of keeping animals in barns is significantly influenced by the education level of the household head (p � 0.001 ), with household heads having attended only primary or secondary school being less likely to keep animals in barns than those who reached tertiary education level (ef-l =0.2 and ef-l =0.5 with p � 0.001 and p � 0.05, respectively, for primary and secondary education level) (Table 2.8). Households with large herd sizes are less likely to keep animals in barns (ef-l = 0.96; p � 0.01) than households with small herd size. Likewise, households with heads who migrated from another location are less likely to keep animals in barns (ef-l = 0.95; p � 0.05) (Table 2.8).

36 Cha ter 2 Table 2.8: Logistic regression parameters for variables predicting intensification of livestock management in relation to breeding, housing and healthcare across (peri-) urban farm households (n = 157) in Ouagadougou.

Wald's Odds Predictors /J SE df p x2 ratio Controlled mating Constant 0.653 1.548 0.178 1 0.673 1.921 Peri-urban location 0.470 0.532 0.780 1 0.377 1.599 Education level 7.521 3 0.057 Level 1 (primary education) -1.714 0.841 4.155 1 0.042 0.180 Level 2 (secondary education) -1.153 0.887 1.690 1 0.194 0.316 Level 3 (tertiary education) -0.469 0.937 0.251 1 0.617 0.626 Land ownership 4.386 2 0.112 Land ownership 1 (allocated land) -1.138 0.993 1.313 1 0.252 0.321 Land ownership 2 (borrowed, leased) -0.127 0.953 0.018 1 0.894 0.881 Total TLU owned (n) 0.028 0.014 4.250 1 0.039 1.028 Available labour force (n) -0.037 0.064 0.337 1 0.562 0.963 Farmer migrated from another -0.011 0.025 0.181 1 0.670 0.989 location Goodness-of-fit test (model x2} 35.924 9 0.000 Animals kept in barns Constant 5.181 1.787 8.408 1 0.004 NA Peri-urban location -0.732 0.540 1.841 1 0.175 0.481 Education level 17.602 3 0.001 Level 1 (primary education) -3.983 1.219 10.684 1 0.001 0.019 Level 2 (secondary education) -2.974 1.236 5.793 1 0.016 0.051 Level 3 (tertiary education) -2.036 1.233 2.728 1 0.099 0.130 Land ownership 3.586 2 0.166 Land ownership 1 (allocated land) 0.042 0.973 0.002 1 0.965 1.043 Land ownership 2 (borrowed, leased) 0.951 0.973 0.955 1 0.328 2.587 Total TLU owned (n) -0.043 0.015 8.516 1 0.004 0.958 Available labour force (n) -0.031 0.065 0.222 1 0.638 0.970 Farmer migrated from another -0.049 0.024 4.227 1 0.040 0.952 location Goodness-of-fit test (model x2} 42.210 9 0.000 Animal health care Constant 21.888 12852.156 0.000 1 0.999 NA Peri-urban location -1.040 0.936 1.233 1 0.267 0.354 Education level 1.070 3 0.784 Level 1 (primary education) 0.393 1.276 0.095 1 0.758 1.481 Level 2 (secondary education) 0.827 1.361 0.369 1 0.543 2.287 Level 3 (tertiary education) 1.303 1.635 0.636 1 0.425 3.682 Land ownership 0.088 2 0.957 Land ownership 1 (allocated land) 19.378 12852.156 0.000 1 0.999 0.000 Land ownership 2 (borrowed, leased) 19.151 12852.156 0.000 1 0.999 0.000 Total TLU owned (n) 0.085 0.046 3.431 1 0.064 1.089 Available labour force (n) -0.003 0.104 0.001 1 0.974 0.997 Farmer migrated from another -0.023 0.031 0.556 1 0.456 0.977 location 2 Goodness-of-fit test (model x ) 10.679 9 0.298 p: p-value, NA: Not applicable.

37 Cha ter 2 Only 8.3% households are exclusively using hired labour for livestock activities. Also, 61.8% and 29.9% of the livestock keepers use exclusively family labour and both labour types, respectively. Compared to OUA-1 and OUA-3, households in OUA-2 and OUA-4 more frequently use hired labour for farming activities (p � 0.05) (Table 2.9). Yet households in OUA-1 have significantly less family labour force than households of other clusters with an average of 3.8±2.39 in OUA-1, 5.34±2.42 in OUA-2, 6.1 ±3.52 in OUA-3 and 5.7±2.54 in OUA-4 (p� 0.001).

Table 2.9: Labour use for livestock activities in four different (peri-) urban livestock farm types in Ouagadougou (n = 157). HH Family labour Hired labour Cluster n % ofHH % ofHH OUA-1 50 98.0 24.0 OUA-2 17 88.2 52.9 OUA-3 54 88.9 37.0 OUA-4 36 88.9 52.8 Full sample 157 91.7 38.2 p 0.282 <0.05 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number ofHH per cluster.

The number of TLU sold is significantly higher in OUA-2 than in other clusters (p � 0.05). As a consequence of high variation (high standard deviation) there is no significant difference in average amount of milk sold per day by the milk producing households in OUA-3 and OUA-4, although average amount of milk sold daily by households in OUA-3 is four times higher than the amount sold by households in OUA-4 (p > 0.05). Likewise, the sale of eggs is four and seven times more common among households in OUA-2 than in OUA-3/0UA-4 and in OUA-1, respectively, although differences were not significant (p > 0.05) (Table 2.5).

2.3.2.3 Feed and manure use For ruminants grazing is practiced by almost all households (except for 3.7% households, Table 2.10) and is complemented at the same extent by homestead or stall feeding with forages and/or grains (Table 2.11 ), but with a significantly higher frequency in OUA-3 and OUA-4 than in OUA-1 and OUA-2 (p � 0.001 ). The same trends are observed for pigs where only 1.9% of the households keeping pigs do not provide homestead feeding. Scavenging of pigs is reported by 29.9% of pig owners (Table 2.10).

38 Cha ter 2 Table 2.10: Feeding management of livestock owned of four different (peri-) urban farm types in Ouagadougou (n = 157). Cluster HH with Grazing Stall fe eding HH with Scavenging Stall fe eding ruminants {ruminants) {ruminants) 12igs (12igs) (12igs) n % ofHH % ofHH n % ofHH % ofHH OUA-1 27 100.0 74.1 37 24.3 100.0 OUA-2 17 94.1 88.2 17 41.2 94.1 OUA-3 54 92.6 100.0 OUA-4 36 100.0 100.0 Full sample 134 96.3 93.3 54 29.6 98.1 p 0.196 :5:0.001 0.221 0.315 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number ofHH per cluster.

Table 2.11: Homestead feeding of grain residues and forage to ruminants in four different (peri-) urban farm types in Ouagadougou (n = 157). Cattle Shee12 Goats Cluster % ofHH- % ofHH- % ofHH- % ofHH- % ofHH- %ofHH Grain Forage Grain Forage Grain Forage residues {n} {n} residues {n} {n} residues {n} {n} OUA-1 68.8(11) 87.5(14) 47.4( 9) 63.2(12) OUA-2 88.2( 15) 94.1 ( 16) 46.2( 6) 76.9 (10) 16.7( 2) 50.0 ( 6) OUA-3 94.4( 51) 100.0 ( 54) 71.4(30) 78.6 (33) 55.9(19) 64.7 (22) OUA-4 97.2( 35) 91.7 ( 33) 34.5(10) 31.0( 9) 24.0( 6) 24.0 ( 6) Full sample 94.4(101) 95.4 (103) 57.0(57) 66.0(66) 40.0(36) 51.1 (46) e. 0.414 :5:0.001 <0.05 :5:0.001 <0.05 <0.05 p: p-value for differences between clusters, Fisher's exact test for proportions. HH: Household, n: number ofHH per cluster.

Table 2.12: Homestead feeding of energy and protein rich feeds to poultry and pigs in four different (peri-) urban farm types in Ouagadougou (n = 157). Pigs Poultry Cluster % ofHH % ofHH % ofHH % ofHH Energy rich(n) Protein rich (n) Energy rich(n) Protein rich(n) OUA-1 100(37) 27.0(10) 22.2( 6) 3.7(1) OUA-2 100(17) 17.6( 3) 25.0( 3) 0.0(0) OUA-3 100 ( 1) 0.0( 0) 8.9 ( 4) 0.0 (0) OUA-4 0.0 ( 0) 0.0 (0) Full sample 100 (55) 23.6(13) 11.7 (13) 0.9 (1) p 0.643 <0.05 0.371 p: p-value for differences between clusters, Fisher's exact test for proportions, HH: Household, n: number ofHH per cluster.

All cattle keeping households in OUA-3 provide homestead forage to their cattle, a proportion which is significantly higher than for the other clusters (p � 0.001 ). Feeding grain residues to cattle is a common practice amongst cattle keeping households.

39 Cha ter 2 Similarly, in OUA-3, the share of farmers feeding homestead forage to sheep and goats is significantly higher than in other clusters (p � 0.001, for sheep, p � 0.05 for goats), and the same trend is noted for homestead provision of grain residues (p � 0.05 for both species). All pig owners provide energy rich feeds to their animals and less than a quarter provides protein rich feeds. The frequency of homestead feeding of poultry is remarkably low for both energy and protein rich feedstuffs (Table 2.12).

Maize bran, cowpea and groundnut hay, cereal straws, fresh grass, and brewer's spent grain are the most frequently used feed types for homestead feeding. The source of brans is mostly mills (58% of households using this feed), but also fodder markets (39%). Cereal straws (14%), groundnut hay (25%), grass hay (21%) and cotton seed cake (88%) also originate from fodder markets. The households' own crop fields also supply an important share of feeds especially crop residues (56% for cowpea hay, 32% for groundnut hay, 27% for cereal straws) compared to other (peri-)urban crop fields (26.5% for cowpea hay, 19.4% for grass hay, 14% for cereal straw). Cut and carry of forages such as bush hay (43.5% of households) and fresh grasses (86.9%) also constitute another source of fodder in (peri-) urban livestock keeping households of Ouagadougou. The share of households purchasing feed is the highest in OUA-1: maize bran from grain mills is purchased by 80.4% of households in OUA-1, 52.9% in OUA-2, 45.0% in OUA-3, and 42.9% in OUA-4 (p � 0.05), and groundnut hay from urban fodder markets is purchased by 81.4% of households in OUA-1, 17.5% in OUA-3 and 0% in OUA-2 and OUA-4 (p � 0.001 ). Peri-urban crop land is the reported source of groundnut hay for 56% 18%, 25% and 11 % households in OUA-4, OUA-1, OUA-2 and OUA-3 respectively (p � 0.001 ), while other sources are less important.

Manure is used to fertilize crops on own land, but also sold, given away, thrown away or used in multiple ways by 58%, 18%, 10%, 7% and 8% of all surveyed households, respectively. The frequency of households using manure on their own land is significantly higher in OUA-2 than in the other three clusters (p � 0.05) (Figure 2.4 ). The highest share of farmers selling their manure to other farmers is observed in OUA-4 (p � 0.05) (Figure 2.4). Households wasting manure are found in OUA-1 and OUA-3, with a significantly higher proportion in the former than the latter cluster. In addition, a higher percentage of households in OUA-1 give livestock manure away to other crop farmers (p � 0.05) (Figure 2.4).

40 Cha ter 2

100

•••••••••••••••••••••••••• c 60 Cl) 50

Cl) 40 I

0 OUA-1, n=SO OUA-2, n=17 OUA-3, n=54 OUA-4, n=36 •use on own land% Dselling% athrowing away% �giving away% !Imultiple uses%

Figure 2.4: Manure use across four different (peri-) urban farm types in Ouagadougou (n=157).

2.4 Discussion

Four different clusters representing specific livestock production systems were identified in Ouagadougou, Burkina Faso. The main discriminant variables were related to the relative importance (in terms of herd size) of different livestock species kept, production management, and production purpose. In West Africa, previous studies have characterised livestock production systems based on production orientation, resource endowment and production strategy of different types of livestock keeping households in Kano (Nigeria), Bobo-Dioulasso (Burkina Faso) and Sikasso (Mali) (Dossa et al., 2011; Abdulkadir et al., 2012; Dossa et al., 2015a). Likewise a resource based typology of crop-livestock farmers with a focus on millet production has been carried out in cities of Gambia, Guinea, and Guinea Bissau (Somda et al., 2004).

41 Cha ter 2 2.4.1 Specialisation of the (peri-) urban livestock (dairy and pig) production systems

As already observed by Thys et al. (2005), several livestock species are kept by (peri-) urban dwellers of Ouagadougou. However some livestock keeping households tend to be more specialised towards specific livestock productions; especially farmers clustered in OUA-1 were relatively highly specialised in pig production as compared to other clusters. Specialisation was also found with some dairy farmers in OUA-3 and OUA-4, though livestock keeping was neither the main nor the only activity in most cases. In Burkina Faso, like other West Africa countries, peri-urban dairy production is supported by a set of stakeholders including governments or state institutions such as the National Pilot Dairy Development Programme (PNPDL), non-governmental organisations and farmers' and processors' organisations, collectors or middlemen and shops (Hamadou and Sanon, 2006). In the present study, households involved in dairy production were classified in OUA-3 and OUA-4 and their characteristics were similar to the ones described by Hamadou and Sanon (2006) who identified two peri-urban dairy production systems around Bobo-Dioulasso and Ouagadougou: the traditional dairy production carried out by Fulani and other people (corresponding to OUA-4), and the modern peri-urban production systems operated by a new type of farmers consisting of retired people, civil servants and monasteries (e.g., the Monastery Koubri in the periphery of Ouagadougou). Traditional poultry keeping is a common practice in West Africa in general (Ayssiwede et al., 2013) and was also found in the (peri-) urban livestock keeping households in Ouagadougou (Thys et al., 2005). Although local poultry were the most common livestock species kept by our sample households, the birds received little attention in terms of homestead feeding, health care, controlled mating and adequate housing - which all characterises the extensive traditional (rural) poultry production system that can be improved (Ouedraogo et al., 2017). As for other livestock species, a development project aiming at improving poultry production by subsidizing veterinary inputs for traditional poultry keepers (mainly in rural areas) has been vulgarised and implemented, but the investments needed for the necessary shift from the traditional scavenging system to a semi-intensive poultry production system remain relatively high (Ouedraogo et al., 2015).

2.4.2 Trends towards intensification of (peri-) urban livestock production

A study carried out in four cities located in two different agro-ecological zones of Burkina Faso (Bobo-Dioulasso and Gaoua in the Sudanian and Kaya and Dori in the

42 Cha ter 2 Sahelian region) identified three different pig production types keeping different breeds including exotic, local pig breeds and their crosses under different management intensities (intensive, semi-intensive and extensive) (Kiendrebeogo et al., 2014). In this study, of the pig keeping households, OUA-1 households provided less health care to animals and more frequently applied scavenging of pigs than households in OUA-2 suggesting that OUA-2' households constitute a more intensive segment of pig production in the study area. With respect to dairy production, health care, housing and breeding management of cattle did not differ significantly between OUA-3 and OUA-4 which is consistent with the fact that more than 90% households keeping cattle provided health care to this livestock species. However this is somehow contrasting with the findings of Hamadou and Sanon (2006) who found that traditional producers mostly rely on local zebu cattle and still practice transhumance, whereas modern producers keep a variety of breeds including local Zebu, Azawak, Gudali and crosses with high-yielding imported cattle breeds, therefore suggesting different breeding and production management strategies whereby the so called modern producers are more intensive than traditional producers. Which is in line with the observation that homestead feeding of forage to cattle was significantly higher in OUA-3 than in OUA-4 and could therefore contribute to an improved feeding management (especially during feed scarcity in the dry season). In West Africa over the past three decades the implementation of crossbreeding (through artificial insemination or natural mating) of local cattle with semen from high yielding exotic breeds has been taking place (CORAF/WECARD, 2013), and forage production coupled with stocking for dry season supplementation of livestock have equally been promoted and implemented in West Africa countries (Kiema et al., 2012; Camara et al., 2015; Dewa Kassa et al., 2018). Farmers in OUA-2, OUA-3 and OUA-4 may have been exposed to those technologies and have adopted them at different levels, resulting in differences in homestead feeding practices and production performances. Keeping animals in barns was practiced less by people with low education level, which translated into low investment in housing. This was certainly related to farmers' lack of understanding of the benefits of improved housing as well as to the extensive production practices. In addition, the low investments into dairy production in Ouagadougou might be explained by the dominance of the traditional dairy production system with its important involvement of family labour and thus reduced production costs. In the sub-region, evidence of small ruminant and beef cattle production intensification through fattening have also been described by several authors (Diogo et al., 2010; Amadou et al., 2015; Dossa et al., 2015b; Tindano et al., 2015) but

43 Cha ter 2 were less evidenced in the present study though OUA-2 depicted the highest mean TLU sold (4.3±3.01 ), frequency of health care provision (100%) and proportion of keeping animals in barns (around 80%), a couple of features characterising production intensification.

2.4.3 Urban and peri-urban mixed crop-livestock farming

Mixed crop-livestock farming is a production system in which resources (land, capital, labour) are put together to value crop and livestock simultaneously and in association. This system was frequently encountered in this study where livestock keeping was integrated with crop farming with a share of 50%, 94.1 %, 75.9%, 80.6% of households in OUA-1, OUA-2, OUA-3 and OUA-4, respectively. Yet, most pig keepers were crop farmers as well in Bobo-Dioulasso, Gaoua, Kaya and Dori, other cities of Burkina Faso (Kiendrebeogo et al., 2014), like in OUA-2. However these four cities are less urbanised than Ouagadougou where land scarcity might have prevented farmers in OUA-1 to integrate swine production with crop farming. Furthermore, OUA-1 households mostly relied on fodder markets and grinding mills for their feed supply as compared to households in the other clusters, and the use of crop residues for livestock feeding was less important, partly due to limited physiological capacities of pigs to digest fibres as compared to ruminants (Karasov and Douglas, 2013). In West Africa, feeding crop residues to livestock is gaining in importance in (peri-)urban livestock holdings as a result of the increasing urban population and the shrinking grazing areas (Tiffen, 2004). It is also established that socio-political and financial interventions aiming at securing agricultural land for farming households offers better possibilities for integrated crop and livestock production systems and enhances farmers' willingness to adopt climate-smart agricultural production systems (Amole and Ayantunde, 2016a).

Crop-livestock integration is also characterised by on-farm use of manure and crop residues and selling of the same by-products. This allows (peri-) urban households to recycle resources and improve own resource use efficiency. Though crop-livestock integration was frequent (see above), manure wastage was significantly higher in OUA- 1 as a result of the high degree of specialisation and lack of opportunities to sell the manure on manure markets or by directly linking up with crop farmers. The same observation was made in Bobo-Dioulasso where 45% of highly specialised intra-urban landless beef cattle farmers mainly dumped the manure in open spaces within the city (Dossa et al., 201 Sa). Most households wasting manure by throwing it away are found in OUA-1; they are not involved in crop farming and certainly have poor linkages with

44 Cha ter 2 (peri-) urban crop famers. With respect to organic fertilizer application by (peri-)urban livestock farmers, manure was the organic amendment mostly used (41-75%) followed by household garbage (15-41%) in Bobo-Dioulasso, Ouagadougou and Ouahigouya three cities in Burkina Faso (Nare et al., 2015). In addition according to the same study, there was a correlation between the social status (gender and education level) and organic amendment management (Nare et al., 2015).

Whereas Dossa et al. (2011) observed that none of the variables related to personal attributes of the household head and to the household's socio-demography could significantly predict (peri-) urban farming households' classification (and therefore production management or practices) in Kano (Nigeria), Bobo-Dioulasso (Burkina Faso) and Sikasso (Mali), in the present study households with household heads that only reached lower education level were more likely to integrate crop farming with livestock keeping. There is still a need to clarify whether this is a diversification strategy to improve household income and to lower vulnerability to unemployment (precarious employment) and food insecurity as compared to households where the head attained higher education level and might have better, stable and well payed occupations. According to Younoussi and Piche (2005), urban dwellers with lower education level are more likely to be employed in the informal sector, most probably receive low revenues from their employment and need to develop alternative livelihood strategies such as diversification of household income including (peri-) urban crop-livestock farming.

2.4.4 Demand-offer of livestock products and farm performance

Small scale West Africa (peri-) urban livestock production systems are ineluctably more and more market oriented as a consequence of growing city demand of livestock and livestock products by urban dwellers who do not produce food or other farm products (Gour, 2001 ). This study is in line with that statement as most households of different clusters sold animals and animal products from their farms in different quantities with OUA-2 selling more TLU per year than the other three clusters. Market orientation which is driven by making profits through selling farm products in the market on a regular basis is therefore strongly correlated to improved farm management which requires that farmers are knowledgeable about farm management (Kahan, 2013). Though all farmers provided homestead feeding to their animals, the output in terms of pig and sheep sales was not as high as expected. This can be explained by an extensive management with insufficient feed supply, not covering the animals' requirements - only one quarter of farmers in OUA-1 provided protein rich feeds to their animals. In addition, inappropriate

45 Cha ter 2 animal health care and poor housing also contributed to a low performance of this system. This is supported by the study of Kiendrebeogo et al. (2014) who also identified poor feeding and health management as main constraints for pig production in Burkina Faso. In addition, despite larger cattle herds in households of OUA-4 than in OUA-3, the output (number of TLU sold) did not significantly differ between both clusters. This suggests an improved farm management in OUA-3 in contrast to OUA-4, which is not limited to dairy production but also extends to homestead feeding of small ruminants. High milk imports by West African countries and dumping on the local market (Oxfam, 2002) probably also contributes to a low investment in dairy production. These imports are targeting almost all West African countries (Duteurtre and Corniaux, 2013) with approximately 60% of the per capita milk consumption in Burkina Faso (26 kg/year) being supplied by imported dairy products and 40% by domestic production (Groupe Agreco, 2006). The low domestic production and high production costs are not translating into higher returns for dairy farmers as a result of an unfair competition between imported and domestic dairy products (Oxfam, 2002; Dieye et al., 2007). However, Burkina Faso has witnessed less import of frozen poultry products from Europe as compared to neighbouring countries, and the demand for poultry meat and eggs is fulfilled by domestic production, mainly traditional production systems (Schneider et al., 2010). In the present study, only a small share of households sold eggs, whereas the majority preferred home consumption of their poultry products.

2.5 Conclusions

Livestock keeping in and around cities is an opportunity for the poorest and unemployed city dwellers that contribute to household's income generation and food security. As West African cities expand, it is likely that the trend towards specialised, intensive and market oriented livestock production emerges around cities and underlines the relevance of strengthening links within livestock keepers and other stakeholders such as crop farmers, feed and manure market participants, and city authorities to improve resource use through efficient recycling, reduced wastage of manure from livestock production units. However, specialisation and intensification might not be closely related phenomena or their relation depends on-farm location. Actually urban farmers might be more specialised by keeping a single species but still not be so much intensive; and integrating crop farming with livestock production may only be low due to lack of agricultural lands and poor linkages with other (peri-) urban farmers. In contrast peri urban market-oriented specialised pig and dairy production is mostly integrated with

46 Cha ter 2 crop production. In addition the diversity of livestock species in urban households and the total number of animals kept is lower than in peri-urban households. High demand for livestock products due to population growth triggers production intensification resulting in the use of improved breeds, provision of homestead feeding and healthcare and improved housing. Initiatives supporting the improvement of (peri-) urban livestock production should take into consideration the development of market linkages for livestock and animal products while enhancing both specialisation and intensification beside crop-livestock integration in order to avoid an unutilised accumulation of nutrients in the manure within and around West Africa cities and reduce the negative environmental impact of (peri-) urban livestock keeping. Also, site-specific strategies that take into consideration both innovation and farmers' socio-cultural features should be identified and implemented in order to enable modern and traditional producers to improve on production outputs such as milk, meat and eggs while decreasing the production risk related to high input costs and animal diseases. Moreover, policies that secure land for (peri-) urban livestock farms are also a key feature for the sustainable development of (peri-) urban livestock production in West Africa cities.

47 Cha ter 2

48 Cha ter 3

3. Urban and peri-urban dairy production systems at the crossroad between intensive and extensive managements in Ouagadougou, Burkina Faso

49 3.1 Introduction

In sub-Saharan Africa (SSA), rising urban demand for dairy products is currently leading to the development of relatively intensified production systems around cities (Duncan et al., 2013; Chagunda et al., 2015). In Burkina Faso, as in most West African countries, a high proportion of milk production is used for homestead consumption, and the collection and processing of milk into dairy products represents less than 3% of the local milk production (Ministere des Ressources Animales, 2012; Rouamba, 2016). In 2014, the regions with highest milk-processing were the Hauts-Bassins (1,554,500 liters), the Cascades (506,500 liters), the Centre (336,400 liters) and the Sahel (248,900 liters) (Ministere des Ressources Animales, 2015). However, domestic production fails to meet the increasing demand for milk and dairy products from the growing cities. Thus the offer of milk and dairy products remains highly dependent on imports which amount to roughly 8 billon XOF (equivalent to 12.2 million Euro) per year (Ministere des Ressources Animales, 2015). This entails a significant outflow of foreign currency for a country whose economy is already very fragile. Across SSA, (peri-) urban dairy production systems have evolved in different ways as a result of an increasing demand for milk and dairy products related to the continuous urbanisation, an ongoing creation of mini-dairies or milk processing units and a growing interest of governments and non governmental organisations in the development of the local milk sector in order to alleviate poverty and enhance food security (Bonfoh et al., 2007; Corniaux, 2015). Actually, besides its contribution to the reduction of milk and dairy product imports, (peri-) urban dairy production triggers the emergence of other economic activities such as animal health services provision, sale of animal feeds, processing and marketing of milk (Bonfoh et al., 2007). Despite these important functions, the (peri-) urban dairy sector in Burkina Faso faces certain health, genetic and especially feeding constraints that hinder its development (Chapter 1 ). Dairy cattle feeding relies mainly on crop residues, natural rangelands with spontaneous herbaceous and shrub vegetation present in the form of scattered pockets in the towns and their surroundings (Chapter 1 ). However, natural grazing areas are becoming increasingly restricted as a result of the extension of crop farming and city expansion and are characterised by qualitative and quantitative forage abundance in the rainy season and a marked feed scarcity and poor quality in the dry season (Chapter 1).

Since the last two decades, governments of SSA countries have been promoting a set of technologies to improve dairy production, either targeting private small holders or

50 Cha ter 3 investors (Bonfoh et al., 2007; Chagunda et al., 2015). In Burkina Faso as in other Sahelian countries, a set of technologies have been introduced and promoted in order to intensify production such as artificial insemination and crossbreeding, on-farm forage production, cut and carry feeding systems, fodder conservation, industrial feed processing, vaccination campaigns and improvement of extension services. Yet these technologies are not uniformly implemented and adopted by farmers of different livestock production systems (Hamadou and Sanon, 2006). The (peri-) urban dairy production systems of the capital Ouagadougou span from extensive or at best semi commercial traditional systems to intensive and modern dairy product system (Roessler et al., 2016), whereby the latter evolved from the former especially in peri-urban areas of West African cities (Bonfoh et al., 2007). The intensive dairy production system is characterised by a high level of investments in livestock infrastructure, a greater use of feed and veterinary inputs, a reasoned management of dairy cattle feeding with fodder harvest and conservation, as well as organised genetic improvement (Gnanda et al., 2016). However, the potential of local breeds to respond to improvements in the production environment is limited, and their performance in the market-oriented (peri-) urban environment remains modest (Chapter 1 ). On the other hand, there are only a few on-farm studies on the performances of exotic and crossbred cattle in West African cities. This chapter therefore analyses the impact of technology adoption in (peri-) urban dairy farms of Ouagadougou on resources use and use efficiency as well as on production performances. It is hypothesised that dairy farms that have adopted improved management technologies are more resource efficient than non-adopting farms, or, put differently, that intensification of dairy production increases resource use efficiency.

3.2 Material and methods

3.2.1 Study area

Until today, Ouagadougou experiences an important livestock development in general (Thys et al., 2005) and of dairy production in particular (Roessler et al., 2016). Two different (peri-) urban dairy production systems showing different intensification levels have been identified in Ouagadougou, namely OUA-3 (Mixed crop-(dairy) cattle production), a more intensive cluster of farms of non-Fulani owners, and OUA-4 (Mixed crop-dairy cattle production) run by Fulani owners who rear dairy cattle in a more extensive way (Chapter 2). Stall feeding and animal grazing are used as feeding

51 strategies by most (peri-) urban cattle owners (Chapter 2). Herbaceous forage is available in areas open to grazing such as post-harvest crop fields, fallow farmlands, peripheral forest or savanna areas and roadsides (Kouassi et al., 2010). Ouagadougou is located in the north-Sudanian agro-ecological zone between the 500 and 900 mm isohyets with 4 to 5 months of rainfall (Kagone, 2001 ). The rainy season (RS) extends from May to mid-October, whereby vegetation quality and quantity on fallows and rangelands around Ouagadougou are highest during the period from July 1 to October 15. The dry season is subdivided into early dry season (EDS, 16 October-15 February) and late dry season (LOS, 16 February-30 June). The vegetation period comprises 99 to 127 days. Plant communities on pastures are marked by the agricultural past, current land use pressures and bush fires. They present an agrarian vegetation whose evolution seems to be more controlled by anthropogenic than by climatic and pedological factors (Kagone, 2001) . The main units encountered are lowland and hydromorphic valley pastures, glacial pastures and plateau pastures. Lowland and hydromorphic valley pastures are linked to the alluvial stream system. Glacial pastures are the most widespread landscape and vegetation units in the north-Sudanian agro ecological zone. The vegetation type is a savanna with mostly only few trees. The herbaceous layer is dominated by Loudetia togoensis (Pilg.) C.E. Hubb, Andropogon pseudapricus Stapf, Aristida kerstingii Pilg, Dactyloctenium aegyptium (L.) Willd. and Digitaria horizontalis Willd. The woody stratum mainly comprises Combretum spp., Acacia seya/ Delite and Terminalia avicennioides Guill. & Perr. Plateau pastures develop on armored hillocks. Typical pasture types are Butyrospermum paradoxum (C.F. Gaertn.) Hepper and Schizachyrium exile (Hochst.) Pilg. woodland savanna and Butyrospermum paradoxum (C.F. Gaertn.) Hepper and Andropogon gayanus Kunth woodland savanna (Kagone, 2001 ). In the study area, fallow land is one of the main sources of forage, herbaceous and woody resources. They are open to grazing cattle year-round, but especially during the rainy season when the animals have to be kept away from crop fields and their access is generally free for all, while there are restrictions on access to forage resources on crop fields (Vall and Diallo, 2009). Crop residues, on the other hand, are increasingly harvested and stored by the individual farmer instead of being left for grazing by any herd. Fallow land and fallow vegetation is getting scarce because of the strong urbanisation and the extension of agricultural surfaces that intensify land pressure at the city outskirts.

52 Cha ter 3

3.2.2 On-farm quantification of input and output

Three and four farms were randomly selected from OUA-3 (n=14 dairy farms out of 54 cattle farms) and OUA-4 (n=36 dairy farms), respectively, for input and output quantification. The selected farms were visited on two consecutive days at 6 to 10 week intervals during 16 months (from October 2014 to February 2016), resulting in 10 visits per farm. The amount of inputs (quality and quantity of feed provided at homestead, and animal feed intake) and of output (number of offspring, body weight change and milk production) was quantified at each of these visits.

3.2.2.1 Animal identification

Based on farmers' informed consent, all cattle present at the farm were identified during the first visit, and their phenotypic traits (species, breed, sex, age, coat colour, polledness), physiological status (parity and lactation day for lactating cows that where milked) were recorded.

3.2.2.2 Animal inflow and outflow records

During each visit, animal inflow through birth, purchase, and gifting were recorded and new animals were identified as described above. Animal outflow through sale, death, stillbirth, and gifting were recorded likewise.

3.2.2.3 Weight and body condition scoring

For animals above 40 kg, live weight (LW) was assessed at each visit in the morning using a wooden weighing platform mounted on four weighing feet of an Agreto® weighing scale (AGRETO Electronic GmbH, Raabs, Austria; nominal load 1OOO kg per foot, accuracy ±1 %, used during visits 1-7) or an EziWeigh 5 weighing kit (Tru-Test Inc., Mineral Wells, USA) consisting of an EziWeigh 5 indicator and four weighing feet (total maximum load 3000 kg) during visits 8-10. For animals under 40 kg body weight (new born calves) a Profi® hanging scale (Burg Wachter TARA PS 7600, weight range 40 kg, accuracy ±1 %) hanged on weigh slings was used. In addition, the body condition score (BCS) of Sudanian zebu cattle (local zebu and Sahelian zebu) was assessed according to Vall and Bayala (2004) while the scoring of exotic cross breeds was done according to DEFRA (2001).

3.2.2.4 Daily milk yield and milk composition

At each visit, milk was quantified at every milking event by using a Profi® hanging scale (Burg Wachter TARA PS 7600, weight range 40 kg, accuracy ±1 %) for two consecutive

53 days. Thirty ml of milk was then sampled during morning milking for milk analysis. Milk samples were kept in a cooler box with ice until they were analysed few hours after collection. Milk samples were analysed for their fat, protein, lactose, and solid non-fat content using a Master Eco® milk analyser (Milkotester LTD)).

3.2.2.5 Stall feeding Animals were fed in group or individually depending on farm practice by age class, sex and physiological status of the animal and the feed type concerned. As a common picture, concentrates were distributed individually and roughages were distributed to groups. The quantification of the amount of each feedstuff offered to the monitored animals was carried out on one day at every visit by weighing. The latter was done in a weigh sling (roughages) or bucket (concentrate feeds) on a Burg Wachter hanging scale (Chapter 3.2.2.4). The fresh weight of feedstuffs and their mixtures was recorded along with the ID of the animals fed. To capture the whole day feeding patterns of monitored farms, feeding data was collected every time animals were fed during the monitoring day, that is once (morning or evening) or twice depending on feeding features prevailing on each farm. After thorough mixing of each offered feedstuff a representative sample of approximately 250 g dry matter (OM), was taken at each farm visit; the sample was weighed fresh (portable battery-driven scale, range 0-5 kg, accuracy 1 g), then air-dried under shade (dry season) or oven dried (60 °C, rainy season) to weight constancy and weighed again, to determine the air dry matter content (ADMc). Samples of a given feedstuff were pooled per farm and season (RS, EDS, and LOS). After thorough mixing pooled samples were ground to pass 1 mm sieve and stored in paper bags secured in an air-tight metal box until analysis. Refusals occurred only in some farms for cereal straws (most of the time mixed with bedding) and rarely with concentrate feedstuffs and high quality roughages such as green grass and legume hays. Upon occurrence and where possible they were quantified and their ADMc determined. The monitored animals' daily feed intake was calculated using the amount of dry matter of feedstuffs offered minus the OM of eventual refusals.

Following standard procedures of the Association of German Agricultural Analytic and Research Institutes (VDLUFA, 2012), 91 (out of 345) feed samples representing the diversity of these feedstuffs as well as the collected samples were submitted to wet chemical analysis (WCA) for dry matter (OM, method #3.1), crude ash (CA; #4.1.1), organic matter (OM = OM minus CA), neutral detergent fiber (NDF; #6.5.2) and acid detergent fiber (ADF; #6.5.3). Total phosphorus (P) concentration was determined by

54 Cha ter 3 colorimetry (Lowry and Lopez, 1946), and the nitrogen (N) concentration was determined in a VarioMax® CN analyser (Elementar Analysensysteme GmbH Hanau, Germany) and multiplied with factor 6.25 to obtain the crude protein (CP) concentration. Forty-three samples also underwent Near lnfrared Spectroscopy (NI RS), and a regression equation was established to estimate the DM concentration of N, P, NDF and ADF for further 45 samples analysed only with NIRS, thereby following the approach of Schiborra et al. (2010). Overall, results from 139 feed samples were used to calculate the animals' nutrient intake at the farm level. Due to analytical problems, nutrient concentrations of maize grain, green sorghum, millet straw and molasses were taken from Feedipedia (2017). Furthermore the metabolisable energy (ME) content of all feedstuffs was taken from Feedipedia (2017) or Close and Menke (1986).

3.2.3 Pasture use, grazing behaviour and pastoral value

Observation of animals' grazing behaviour and above ground herbaceous biomass availability on main pastures along the animals' grazing itinerary was carried out on four farms that practiced year-round grazing. All four farms belonged to cluster OUA-4 and were located in different quarters of the city (Kamboinse 1, Ponsotenga, Boassa, Kamboinse 2; see appendix). Three key periods were targeted from March 2015 to February 2016:

Rainy season (RS) with peak biomass availability of good quality (July-October 2015),

Cool or early dry season (EDS) with average biomass availability (November 2015-February 2016),

Hot or late dry season (LOS) with minimal biomass availability (March 2015-June 2015).

For the identification of the main pastures and observation of animals, GPS tracking devices were used, namely Trackstick (http://www.trackstick.com/products / supertrackstick/index. html) or Holux M-241 (Holux Technology Inc., Hsinchu, Taiwan; http://www.holux.com). These were fixed on a horn each of three dairy cows which were representative for the herd. The daily distances walked were recorded during three consecutive days (Buerkert and Schlecht, 2009; Schlecht et al., 2009), at the same time the behaviour of one of these three animals was observed by a person following from the start of moving to the pasture in the morning until the return back home in the

55 afternoon. Every five minutes the activities of feeding (grazing, browsing, walking between feeding stations), walking and resting (resting with or without ruminating, social activities and idling) were recorded as they occurred at that moment. Simultaneously the land cover classes where these activities occurred were recorded as well.

Above ground biomass on the main pastures of each of the four herds was quantified only in mid-rainy season. A two-step sampling approach was used to estimate the vegetation structure and cover (of grasses, legumes, non-legume herbs, trees and

2 shrubs). The first step was to select, per pasture, nine different plots of 10 x 10 m : three with high above-ground biomass, three with medium above-ground biomass and three with low above-ground biomass. The second step was to randomly select 4 representative vegetation spots in each 10 x 10 m2 plot (=sampling points). For each sampling point the GPS position was recorded. Then a frame of 1 m2 was placed, the plant species within the frame were identified and the stone cover and vegetation height recorded. Afterwards, the above ground herbaceous biomass within the 1 m2 frame was clipped at 1 cm height. When shrubs and/or trees were present in the sampling plot and if animals were observed to browse on shrubs and trees, about 200 g of fresh leaves and soft twigs were collected from that specimen. All vegetation samples were weighed fresh using a portable battery-driven scale (range 0-5 kg, accuracy 1 g), then put into a cotton bag and air- or oven- dried (Chapter 3.2.2.5). When completely dry, the weight of each sample was noted again.

For the species inventory of the herbaceous layer the point quadrat method was used (Daget and Poissonet, 1971 ), observing the presence of species every 20 cm along a line. In each sampling plot of 100 m2 a line of 20 m length was used, which corresponded to the sum of two sides along the diagonal of the plot. In total, 18 observation lines (i.e., 6 per biomass stratum, 2 per plot) and 900 points were used for these records.

The point quadrat method allowed the calculation of the characteristic parameters of the vegetation (Daget and Poissonet, 1971): • the specific frequency of species i (SFi) which is the number of observation points where species or taxon (i) was encountered along the observation lines; • the specific contribution of species i (SCi) is defined as the ratio between the specific frequency of species i (SFi) and the sum of specific frequencies of all the species (�SFi) indexed out of 100 sampled. It allows appreciation of the

56 Cha ter 3

equilibrium state of the land cover being studied. This is the translation of the participation of species i in aerial vegetation density: SCi (%) = SFi / �SFi x 100.

The pastoral value was obtained by calculating the products of the specific contributions (SCi) of the species by the corresponding quality index (lsi); the relative values of the species thus obtained are summed and expressed in percentage (Daget and Poissonet, 1972; Akpo and Grouzis, 2000):

PV (%) = 1/3 � (SCi x Is)

Where Is is the specific quality index of pastoral species i. This quality criterion, for the herbaceous species of the rangelands of the Sahelian zone, is established on a rating scale from Oto 3:

� Is equal to 3 corresponds to good pastoral value (Gpv);

� Is equal to 2 corresponds to average pastoral value (Apv);

� Is equal to 1 corresponds to low pastoral value (Lpv);

� Is equal to Ocorresponds to vegetation without pastoral value (Wpv).

The average biomass production (kg OM/ha) was calculated using the dry mass (OM, g) of clipped above ground biomass within each 1 m2 frame multiplied with 10.

The carrying capacity (CC) was calculated as follows (Boudet, 1975b; Sanon et al., 2014):

CC = (phytomass production (kg DM/ha)*K (%)) / (6.25 (kg DM/TLU*day) x utilisation period)

Where K (%) = Utilisation coefficient = 1/3; TLU = Tropical Livestock Unit; OM= Dry Matter.

3.2.4 Animal data calculation

A Microsoft Access database was designed to allow full and complete checking and cleansing of on-farm monitoring data. A set of filters were used to sort out insufficient data or outliers to avoid distorting statistical parameters while keeping enough consistent data for further analyses. For milk data, all inconsistent and incomplete data sets were deleted and only lactations with at least 3 complete milk records were included (i.e. milk records available for two consecutive days, with identical and

57 consistent number and time of recording, for example: four records - morning and evening on both days, only morning or evening on both days, depending on a farmer's milking practice). Thus milk records from 77 cows were used of which 42 belonged to OUA-3 (23 exotic crossbreds, 15 Sahelian crosses, 2 Fulani, 1 Bororo, 1 Gudali breed types) and 35 to OUA-4 (32 Fulani, 2 Gudali and 1 Azawak breed type). Likewise all inconsistent and incomplete LW data was deleted and missing weight data was estimated using interpolation techniques. Feed intake data was also checked against physiologically plausible intake levels to make sure that the quantified feed intake was within the biological intake range of animals (Ulbrich et al., 2004; Table A. 4 of "Appendix"). When inconsistencies of feeding records occurred the data was discarded. For age-related variables the data was either corrected considering the birth date and respective visit dates of a missing or inconsistent value. For animals without birth date records, the birth date was calculated based on the age at the first visit date. Three age categories were defined namely suckling (between 0-4 months), young (from 5-18 months) and adult (>18 months). Three physiological stages were defined for adult females, namely: neither gestating nor lactating (dry), pregnant, and lactating (irrespective if at the same time pregnant or not). Given the difference in length of the three seasons and different interval lengths between consecutive visits, weighed averages were calculated to allow meaningful comparisons of variables between seasons.

A set of calculations was applied to determine metabolisable energy (ME), crude protein (GP) and nutrient supply (N and P): the concentration (c) of N, GP (both in g/kg DM), P (mg/kg DM), and ME (MJ/kg DM) in each feedstuff consumed at the farm was related to total daily dry matter offer (TDDMO), and was calculated based on fresh matter (FM) offer as follows:

Y = kg FM x ADMc x DMc x Ne (or, GPc, Pc, MEc).

An animal's apparent daily feed (DM), crude protein (GP), nutrient (N, P) and energy (ME) intake at the homestead was calculated by dividing the total daily feed, crude protein, nutrient or energy supply (Y) by the animal's metabolic weight (MW). The latter 0 75 (kg · ) was obtained by exponentiation of live weight (LW) with 0.75:

0 75 ME requirements for maintenance (0.480 MJ ME/kg · ), for daily weight gain (DWG; 34 MJ ME/kg DWG in cows, 42.8 MJ ME/kg DWG in young females, 31.7MJ ME/kg DWG

58 Cha ter 3 in adult and young males), for milk production (5.3 MJ ME/kg 4% fat corrected milk, FCM) (Ulbrich et al., 2004; Table A. 4 of "Appendix") and for walking (kJ ME/kg LW*km) were summed up to determine the total ME requirement. Likewise, CP requirements for maintenance (3. 7 g CP/kg MW), for weight gain (380 g CP/kg DWG for adult animals and 400 g CP/kg DWG for young both females and males) and for milk production (85 g CP /kg 4% FCM) were added up to determine the total CP requirement.

To determine the adequacy of energy (or crude protein) intake of the studied dairy cattle and groups, the daily amount of ME (or crude protein) ingested (supplied) was divided by each animal's total individual ME (or crude protein) requirements. Net energy (or crude protein) supply through body weight loss was not considered. The obtained ratio was called ME (or crude protein) supply level and was stratified as: adequate energy (or crude protein) supply (0.8 - 1.2); mild energy (or crude protein) deficit (0.5 - <0.8); severe energy (or crude protein) deficit (<0.5); mild energy (or crude protein) surplus (>1.2 - 1.5) and substantial energy (or crude protein) surplus (>1.5).

To estimate the energy status of the studied lactating cows, two approaches based on milk fat-to-protein ratio (FPR) were used: 1) the calculation of milk fat-to-protein ratio deviation (FPRdev) of each cow's lactation(s) average with a high FPRdev (>0.12) indicating negative energy balance (Stoop et al., 2009) and 2) the comparison of the calculated FPR value with fixed values, whereby an FPR value > 1.5 indicates risk of ketosis (Heuer et al., 1999) and an FPR < 1.0 corresponds to a risk of acidosis (Gantner et al., 2018). For both methods comparisons were carried out for three different lactation stages in the different dairy production systems: stage 1 corresponding to postpartum or early stage lactation from 1 to 30 days of lactation, stage 2 corresponding to mid lactation from 31 to 120 days of lactation and stage 3 from 121 days of lactation onwards.

3.2.5 Statistical analyses

For each cluster (OUA-3 and OUA-4) and when applicable, descriptive analyses were used for herd structure, animal activities on pasture, the proportion of time spent on different land cover classes, (negative) energy balance, protein balance, risk for ketosis, and risk for acidosis. The Fisher's exact test (for small sample size) was used for comparing the daily activity time share (%) between the four sites studied. The rank based non-parametric Kruskal-Wallis test was used to determine if there were statistically significant differences between two or more groups (namely dry, pregnant,

59 lactating, adult male and suckling animals within or between different clusters) of the different quantitative independent variables that is performances variables (live weight, milk yield, BCS) and feeding variables (offer of proximate diet components). The same statistic test was performed to assess the influence of the season (RS, EDS, LOS) on the performances of different animal groups and quantitative pastoral variables such as the SCi of plant species groups, the PV, biomass production and carrying capacity of the different studied sites (n=4). The Pearson's Chi-squared test was used for comparison of proportion (%) of different CP and ME coverage levels. The threshold for significance was set at a P-value < 0.05. Statistical analyses were done with IBM SPSS Statistics 20.

3.3 Results

3.3.1 Herd structure and dynamic

Across seasons the average herd size in OUA-3 and OUA-4 did not differ significantly between dairy production systems. The share of the different categories (dry, lactating, pregnant cows, adult males, young and suckling animals) did not differ significantly between both production systems year round. On average one adult male was kept on farm during RS and LOS in OUA-3, whereas in OUA-4 there were 4 to 6 times more adult male cattle present on-farm during the same periods. Relative to other categories within the same dairy production system, the average percentage of dry cows in OUA-4 was more than double the average percentage of dry cows in OUA-3 across all seasons. In both dairy production systems the average number of lactating cows and suckling animals decreased from RS to EDS season and LOS while the share of pregnant cows increased (Figure 3.1 ).

Both adult and young animals died in all seasons (Table 3.1 ), though OUA-3 experienced more deaths during LOS while in OUA-4 more deaths were observed during RS. However, suckling animals died to a lesser extent than young and adult individuals in both livestock production systems across seasons. Selling occurred more frequently during LOS; young animals were preferably sold in OUA-3 as compared to OUA-4 were a considerable number of animals were only sold at adult age. In addition, animals sold at adult age in OUA-4 were predominantly males, whereas in OUA-3 the sold adult animals were mostly female. Births were recorded throughout all seasons. In OUA-3 the number of births in the LOS was twice the number of births in the two other seasons, which was not the case in OUA-4.

60 Cha ter 3

25 OUA-3 OUA-4

S 15 co � 10 co

-5 RS EDS LDS RS EDS LDS Season

• Dry cow rn Pregnant cow � Lactating cow D Adu It male cattle l!:!Young cattle ODSuckling calf

Figure 3.1: Overview of herd structure (mean±S.E. for each group) in dairy farms of clusters OUA-3 and OUA-4 during the rainy season (RS), the early dry season (EDS) and the late dry season (LOS) (October 2014 - February 2016).

Table 3.1: Animal inflow and outflow in dairy farms of clusters OUA-3 and OUA-4 during the rainy season (RS), the early dry season (EDS), and the late dry season (LOS) (October 2014 2014 - February 2016). Inflow/Outflow RS EDS LOS type Group OUA-3 OUA-4 OUA-3 OUA-4 OUA-3 OUA-4 Dry cow 0 3 0 1 2 1 Died (n) Young cattle (m, f) 0 1 1 1 3 0 Suckling calf (m, f) 0 0 1 1 1 0 Dry cow 4 2 5 5 5 4 Sold (n) Adult male 2 6 0 4 2 11 Young cattle (m, f) 1 1 8 8 19 5 Born (n) Suckling calf (m, f) 11 11 10 13 20 9 n: number of animals, m: male, f: female.

3.3.2 Feeding strategies on (peri-) urban dairy farms

Dairy farmers in and around Ouagadougou have adopted different feeding strategies that are season dependent and related to their livestock production system (Table 3.2). As a general observation during RS OUA-3 farmers took advantage of the abundant

61 forage availability to provide their animals with green forage either on pasture or on farm through cut and carry. Depending on the availability of resources they also engaged in forage production. At the onset of EDS they started hay harvesting or purchasing hay for storage while still partially feeding their animals with fodder from open access spaces. Later on they applied a zero grazing system and, for roughage provision, relied on fodder stored until mid to end EDS. They additionally provided their animals with agro-industrial by-products and commercial dairy feed year round. OUA-4 farmers relied on pasture all year round for feeding and also provided their dairy cattle with agro-industrial by-products, whereas hay harvesting and/or purchase and storage was applied to a lesser extent than in OUA-3 (data not shown).

3.3.2.1 Homestead feeding

Irrespective of season and farm cluster, 21 different feed types (Table 3.3) were provided to (peri-) urban dairy cattle in Ouagadougou. More than a half (12 out of 21) of the feed types observed were roughages (protein roughages, straws and hays). The remaining feeds were proteins feeds (4), energy feeds (2) and protein energy mixes (2). All the 21 identified feed types were provided to dairy cattle by OUA-3 farmers but only 5 feed types were used by OUA-4 dairy producers. During the rainy season dairy farmers used less feed types (5 types in OUA-3 and 3 types in OUA-4) as compared to the two other seasons. In OUA-3 the most used feed types were commercial dairy feed with high protein and energy content and protein rich brewers' grain in all three seasons, and green grass or hay in RS and EDS/LOS, respectively. In OUA-4 protein rich feed types like cotton seed cake, maize grain residues and sesame chaff were frequently used concentrates during all seasons, and hay was only provided during LOS. Table A.3 ("Appendix") presents the proximate composition of feedstuffs offered to dairy cattle in the (peri-) urban area of Ouagadougou.

Across production systems, stall feeding intensity depended on physiological status, age and sex. Thus for most animal categories homestead offer of feed dry matter, OM, CP, P, NDF, ADF and ME per kg of metabolic body weight (MW) differed significantly between the two production systems and the three different seasons (p � 0.05). For pregnant cows the daily offer of proximate diet components did not differ significantly between seasons in OUA-4 unlike in OUA-3. Yet the difference between the two dairy production systems was highly significant (p � 0.001, Table 3.4): animals were more intensively fed during LOS than in the other two seasons, and much more intensively fed in OUA-3 than in OUA-4. The same situation was observed for lactating cows. Thus,

62 Cha ter 3 apart from the offer of P, the feeding intensity of proximate diet components was significantly different between the three seasons and two production systems, with highest offer observed during LOS and lowest during RS. Obviously, being visibly productive animals, lactating cows were fed more intensively than pregnant cows in each dairy production system. Homestead feeding intensity did not differ significantly between seasons for dry cow in OUA-3 unlike in OUA-4. Yet the feeding intensity for dry cows was significantly higher in the former than the latter system (p � 0.001 ). In adult males of OUA-3, the offer of OM, CP, P, and NDF differed significantly between the three seasons unlike in OUA-4 where the offer of all proximate diet components differed significantly between all seasons.

63 Table 3.2: General overview of feeding strategies and feeding related activities used by (peri-) urban dairy farmers in Ouagadougou Burkina Faso

Season RS EDS LOS

Month Jun Jui Au Oct Nov Dec Jan Feb Mar A

Cereal cultivation for silage

Fresh fodder harvesting and feeding

Hay harvesting and purchase

Silage feeding OUA-3 Hay storage and feeding

Agro-industrial by-products � Industrial dairy feed

Pasture or exercise (in the farm surroundings)

Crop cultivation

Hay/straw harvesting OUA-4 Hay/straw storage and feeding

Agro-industrial by-products

Pasture RS: Rainy Season, EDS: Early Dry Season, LDS: Late Dry Season Chapter 3 Dry cows and adult males were often fed more intensively during LOS and less intensively during RS in OUA-4 (p � 0.01, except for ME offer to dry cows (p = 0.078, n.s.) and p � 0.001 for adult males, Table 3.5). For feeding intensity of suckling calves in OUA-3 there was no difference between the three seasons unlike in OUA-4 where suckling animals were more intensively fed during EDS than during RS and LOS (p � 0.001). Yet the offer of all proximate diet components followed the same trend as for the other categories, with suckling cattle being more intensively fed in OUA-3 than in OUA- 4. Feeding of young cattle followed the same trends as in suckling cattle except for CP and P which differed significantly between the three seasons in OUA-3. Again, the lowest feed offer to young cattle was observed during RS in OUA-4 (Table 3.6).

3.3.2.2 Dairy cattle activities on pasture The total length of time spent on pasture, that is the time interval between the departure of the herd in the morning and its return to the farm in the evening, was relatively constant throughout the year and not influenced by season but rather by site (p � 0.001) (Table 3. 7). The daily distance covered by dairy cattle was significantly affected by site (p � 0.05) and season (p � 0.001 ), with average distance being significantly higher during LOS than in the two other seasons (p � 0.001 ). The average time spent walking did not differ between RS (14.4%) and EDS (16.6%) but significantly (p � 0.001) increased in LOS (24.4%). The time spent on grazing differed significantly between the three seasons and was higher in RS and lower in LOS (p � 0.001). Thus at all sites, cows spent on average three quarters of the total observation time on grazing in RS and EDS but less than half of the observation time in LOS (p � 0.001 ). In RS, the time devoted to watering was very low (0.7%), time devoted to simple resting and resting with rumination was 0.9% and 4.4% respectively in RS. Resting was observed around water points and lasted 5 to 15 minutes. However, there was no significant difference in resting time whatever the period and the site. The time allocated to social interactions was also low (1.0%). No significant effect of the season or site on social behavior was observed. Social behavior mainly comprised cow-to-calf interactions (suckling, licking) and was observed in the morning before departure and in the evening after grazing. Browsing significantly increased from 0.6% in RS to more than a quarter of the observation time in LOS (p � 0.001). Browsing was the most important daily activity at sites Kamboinse 1 and 2 during LOS and was significantly higher at those two than at the two other sites.

65 Table 3.3: Feed categories and percentages of feedstuffs ( on dry matter basis) offered to (peri-) urban dairy cattle in Ouagadougou during rainy season (RS), early dry season (EDS), and late dry season (LOS) (October 2014 - February 2016). OUA-3 OUA-4 Feed category 1 Feed category 2 Feed type RS EDS LOS RS EDS LOS Energy feed Cereal grains Maize grain 0.46 Energy feed Other Molasses 0.66 Energy/protein mix Commercial feed Commercial dairy feed 34.79 30.15 23.77 Energy/protein mix Other Tree fruit (wild) 0.64 0.46 Protein feed Protein feed Brewers grain (moist) 21.54 14.27 9.69 Protein feed Protein feed Cotton seed cake 4.55 1.20 1.26 36.75 25.43 28.73 Protein feed Protein feed Maize grain residues 14.17 7.03 4.71 30.32 72.96 35.88 Protein feed Protein feed Sesame chaff 0.96 32.93 13.16 Roughage Protein roughage Bean hulls 0.03 0.23 Roughage Protein roughage Green grass 24.95 2.97 Roughage Protein roughage Green maize 3.54 4.24 Roughage Protein roughage Green sorghum 0.23

(j) Roughage Protein roughage Maize silage 2.05 (j) Roughage Straws and hays Cereal straw 1.19 Roughage Straws and hays Grass hay 23.74 32.33 1.61 21.27 Roughage Straws and hays Maize straw 1.02 0.96 Roughage Straws and hays Millet straw 0.49 Roughage Straws and hays Rice husks 4.13 Roughage Straws and hays Rice straw 0.63 2.27 Roughage Straws and hays Sorghum silage 7.02 16.35 Roughage Straws and hays Sorghum straw 2.27 Cha ter 3 Table 3.4: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S. D.) offered to pregnant (a) and lactating (b) dairy cows in the (peri-) urban area of Ouagadougou (October 2014- February 2016). a) Farm cluster Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (25) 25.3±32.3 22.5±28.8 3.8±5.3 0.19±0.28 15.4±19.3 9.1±12.1 637± 791 OUA-3 EDS (76) 57.3±47.4 52.7±43.2 5.6±6.5 0.24±0.35 37.2±29.9 22.1±1 8.2 497± 674 LOS (73) 86.3±46.3 78.7±42.9 8.3±6.1 0.43±0.38 56.9±29.4 33.4±17.8 1035±1170 p:5: 0.001 0.001 0.001 0.001 0.001 0.001 0.05 RS (42) 3.1±14.8 2.7±12.8 0.6±2.8 0.01±0.07 1.0± 4.7 1.0± 5.1 0± 0 OUA-4 EDS (112) 3.4± 8.7 3.1± 8.1 0.5±1.1 0.04±0.10 1.1± 2.6 0.5± 1.2 51± 156 LOS (86) 4.2±12.5 3.9±11.6 0.6±2.2 0.03±0.11 1.7± 5.1 1.1± 3.4 48± 180 p:5: n.s. n.s. n.s. n.s. n.s. n.s. 0.05 p:5: 0.001 0.001 0.001 0.001 0.001 0.001 0.001

-..J Farm cluster Season(n) OM OM CP p NDF ADF ME kJ/kg MW RS (47) 111.1±48.9 99.8±44.1 22.6±10.1 1.51±0.95 54.9±24.2 30.9±13.6 909±1128 OUA-3 EDS (118) 139.4±90.3 126.7±76.8 22.4±11.5 1.24±0.74 76.1±55.7 42.7±36.6 756± 945 LOS (77) 157.1±79.1 142.6±71.5 22.8±12.9 1.56±0.95 90.9±44.7 53.7±27.0 1167±1221 p :5: 0.01 0.01 n.s. 0.05 0.001 0.001 0.05 RS (53) 17.5±34.4 16.3±31.9 3.2± 6.4 0.09±0.18 7.4±14.2 5.1±10.5 23± 143 OUA-4 EDS (114) 22.7±22.6 21.3±21.2 3.4± 3.6 0.22±0.23 8.3± 9.1 4.1± 4.8 237± 414 LOS (73) 27.4±24.3 25.8±22.9 4.7± 4.3 0.25±0.23 11.5±10.3 7.0± 6.5 417± 765 p :5: 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p:5: 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners). Table 3.5: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry (a) and adult male (b) dairy cattle in the peri- urban area of Ouagadougou (October 2014 - February 2016). a) Farm cluster Season(n) OM OM CP p NDF ADF ME kJ/kq MW RS(3) 45.2±39.5 39.6±34.6 3.7±3.2 0.20±0.18 29.4±25.7 20.5±17.8 0±0 OUA-3 EDS(18) 24.9±35.5 22.8±32.4 1.4±2.3 0.04±0.05 17.9±25.8 11.4±16.7 342±527 LDS(11) 62.3±64.8 57.5±59.8 6.7±7.8 0.41 ±0.48 40.0±42.2 23.3±25.0 620±940 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS(29) 2.8±15.3 2.4±12.8 0.6±3.0 0.1± 0.7 0.8± 4.3 1.0± 5.5 32± 17 OUA-4 EDS(69) 0.3± 2.1 0.2± 1.2 0.0±0.3 0.00±0.03 0.1± 0.6 0.0± 0.2 4± 18 LDS(48) 5.6±13.1 4.8±11.2 1.0±2.6 0.03±0.06 1.8± 3.9 1.8± 4.6 33±100 p� 0.01 0.01 0.01 0.01 0.01 0.01 n.s. p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001

b) en co Farm cluster Season (n) OM OM CP p NDF ADF ME kJ/k_g MW RS (4) 34.1±37.3 30.2±33.2 6.5±9.9 0.43±0.67 21.9±23.8 12.9±14.0 291±323 OUA-3 EDS (14) 27.5±39.4 25.8±37.0 2.0±4.6 0.09±0.23 19.2±27.5 11.8±16.7 228±327 LOS (10) 77.8±62.1 71.8±58.0 4.9±3.4 0.27±0.21 55.7±46.7 33.0±29.0 383±473 p� n.s. 0.05 0.05 0.05 0.05 n.s. n.s. RS (33) 0.0± 0.0 0.0± 0.0 0.0±0.0 0.00±0.00 0.0± 0.0 0.0± 0.0 0± 0 OUA-4 EDS (82) 0.5± 1.8 0.5± 1.6 0.1±0.2 0.01±0.02 0.1± 0.5 0.1± 0.2 27± 82 LOS (75) 20.3±43.8 18.8±40.9 1.9±3.8 0.07±0.15 13.0±30.2 8.7±19.6 160±489 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners). Chapter 3 Table 3.6: Seasonal variation of the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to suckling (a) and young (b) dairy cattle in the peri- urban area of Ouagadougou (October 2014 - February 2016). a) Farm cluster Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (20) 81.8± 49.1 73.1± 43.3 9.0± 8.2 0.75±0.53 48.9± 37.1 26.4±21.6 72±190 OUA-3 EDS (50) 115.0±125.4 106.6±116.6 11.0±18.1 0.68±1.13 79.4± 84.5 48.2±51.2 320±243 LOS (35) 114.3± 77.7 105.7± 72.0 9.5± 7.2 0.59±0.47 77.4±56.25 46.7±34.3 173±361 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS (16) 0.0± 0.0 0.0± 0.0 0.0± 0.0 0.00±0.00 0.0± 0.0 0.0± 0.0 0± 0 OUA-4 EDS (37) 13.1± 18.8 12.2± 17.6 1.6± 2.3 0.16±0.23 3.6± 5.2 1.4± 2.0 132±299 LOS (21) 2.8± 4.4 2.6± 4.1 0.3± 0.5 0.03±0.05 0.8± 1.2 0.3± 0.5 28± 62 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.05 p� 0.001 0.001 0.001 0.001 0.001 0.001 n.s.

b) CD Farm cluster Season(n) OM OM CP p NDF ADF ME kJ/k MW RS(25) 63.9±44.7 57.0±39.8 7.0±7.1 0.46±0.47 39.0±29.8 24.4±19.7 133±323 OUA-3 EDS(132) 67.8±69.8 62.5±63.6 3.9±6.6 0.28±0.72 47.4±49.1 29.4±31.3 208±438 LDS(79) 77.8±68.4 71.5±63.5 7.3±7.2 0.42±0.46 51.5±45.1 30.4±26.7 271±633 p� n.s. n.s. 0.001 0.001 n.s. n.s. n.s. RS(56) 4.1±21.5 3.4±17.9 0.8±4.2 0.02±0.09 1.1± 6.0 1.5± 7.8 13± 99 OUA-4 EDS(106) 11.5±16.6 10.8±15.5 1.4±2.0 0.14±0.20 3.2± 4.6 1.2± 1.8 100±256 LDS(65) 10.0±11.8 9.2±10.6 1.5±2.0 0.10±0.11 3.1± 3.7 1.8± 3.3 156±413 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.05 p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners). Table 3.7: Characteristic activities of dairy cattle from cluster OUA-4 on pastures in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) (RS: July 2015- October 2015, EDS: November 2015-February 2016, LOS: March 2015-June 2015). Italic values indicate averages across sites.

Season Site Day Itinerary Daily activity time share(%) (n) Distance {km} Duration {h} A B G R RR s w RS K2 3 10.6 7.7 1.1 a 1.1 a 75.7 a 0.4 a 0.0 a 1.8 a 19.9 a p 3 6.7 6.5 1.2 a o.o a 66.9 ac 16.2 a 3.1 a 2.3 a 10.4 b B 3 9.1 0.7 a 0.7 a 82.2 a 0.0 a 0.0 a 0.3 a 16.2 ab K1 3 7.6 6.9 o.o a 0.4 a 84.9 ab 2.5 a 0.0 a 1.3 a 10.9 b 12 8.3 7.5 0.7 0.6 77.5 4.4 0.9 1.4 14.4 EDS K2 3 8.2 6.3 2.7 a 0.9 a 67.9 a 2.2 a 0.9 a 1.8 a 23.7 a p 3 8.9 8.7 1.9 b 1.0 a a 2.9 a 1.9 a 3.2 a 11.8 a 77.4 B 3 8.5 1.9 ab 9.4 a 67.7 ab 3.5 a 0.0 a 0.3 a 17.2 a

K1 3 7.4 6.8 C 1.3 o.o a 81.0 ac 1.7 a 0.8 a 1.7 a 13.5 a 12 8.2 7.6 1.9 2.8 73.5 2.6 0.9 1.7 16.6 LOS K2 3 10.6 7.2 1.1 a 53.2 a 15.4 a 0.0 a 0.0 a 2.3 a 28.1 a p 3 8.3 6.0 1.4 a 8.4 b 4.5 b a 0.0 a 0.5 a 21.0 a 6 4.2 B 3 8.9 1.6 a 9.0b 69.0 b 0.0 a 0.0 a 0.9 a 19.5 a K1 3 12.4 7.1 a ab ab a a a a 1.7 33.5 35.6 0.0 0.0 0.4 28.8 12 10.4 7.3 1.5 26.2 45.9 1. 1 0.0 1.0 24.4 Independent variable df* p� Site 3 0.05 0.001 n.s. n.s. n.s. 0.05 n.s. 0.05 0.05 Season 2 0.01 n.s. 0.01 0.001 0.001 n.s. n.s. n.s. 0.001 p: p-value, A: Drinking water, B: Browsing, G: Grazing, R: Resting without rumination, RR: Resting with rumination, S: Social interactions, W: Walking. a.b.c: significant differences between means in the same column within one season according to post-hoe tests, df*: degree of freedom (equal 2 for distance). Cha ter 3 The analysis of the proportion of time spent on different land cover classes by (peri-) urban dairy cattle herds showed that concerning feeding activities (browsing and grazing) fallows accounted for 56.5% of the feeding time of cows during the rainy season and decreased to 12.3% in the early dry season as the proportion of feeding time spent on crop fields increased (Figure 3.2). During rainy season the proportion of feeding time in settlements represented 22.3% and was of highest importance in Ponsontenga (83.8%). The proportion of feeding time on backyard farms (the fields near the housing) was almost negligible. During the early dry season the cows devoted half of their feeding time to crop fields. During the same season the proportion of feeding time on fallows, savanna and valleys remained the same. Fields were no more used for feeding during the late dry season and more feeding time was spent on fallows, savanna and valleys. Feeding in settlement areas was only observed in Ponsontenga in LOS. During RS walking activities mainly took place in the savanna except for Ponsontenga where the proportion of time devoted on walking in settlement areas represented 88.9%. Almost the same results were obtained during EDS with the exception that settlement areas were replaced by fallows and fields in Ponsontenga. During LOS walking was observed in 5 different land cover classes: savanna, fields and valleys. During RS resting took place solely on fallows and crop fields in Kamboinse 2 and Kamboinse 1, respectively. In Ponsontenga resting took place in the farm yard and in settlement areas whereas no resting activity was recorded for Boassa during the rainy season. Fields and savanna were used for resting during EDS in Kamboinse 2 and Kamboinse 1, respectively, while farm yard and savanna were the main land cover classes used for resting in Ponsontenga and Boassa. During the late dry season resting was only recorded in Ponsontenga and took place in settlement areas.

Three plant species contributed to more than half of the total time spent for browsing, namely Guiera senegalensis G.F.Gmel., Securinega virosa (Roxb. Ex Willd.) Baill., and Combretum acu/eatum Vent. which were mainly consumed during LOS (Table 3.8).

71 Cha ter 3

RS (1) 100 EDS LDS E ;,::; O> 80 C

(1) 60 � .....0 C 40 0 t 0 20 eQ_ Cl. 0 K2 p B Kl K2 P B Kl K2 P B Kl D Fallow II Farm yard D Field (;I) Road side D Savanna � Settlement area • Valley

RS (1) 100 EDS LDS E

O> 80 C :S2 ro 60 � .....0 C 40 0 t 0 20 eQ_ Cl. 0 K2 p B Kl K2 P B Kl K2 P B Kl D Fallow II Farm yard [J Field EH Road side D Savanna � Settlement area • Valley b)

RS 100 EDS LDS (1) E ;,::; 80 O> C ;,::;

(1) 60

0 C 40 0 0 20 eQ_ Cl. 0 K2 p B Kl K2 P B Kl K2 p B Kl D Fallow II Farm yard [] Field D Savanna � Settlement area c)

Figure 3.2: Proportion of daily a) feeding, b) walking and c) resting (resting without and with rumination) time (%) spent by dairy cattle of cluster OUA-4 in Kamboinse 2 (K2), Ponsotenga (P), Boassa (B) and Kamboinse 1 (K1) on different land cover classes. Data are shown for rainy season (RS: Jui. - Oct. 2015), early dry season (EDS: Nov. 2015 - Feb. 2016 and late dry season (LOS: Mar. - Jun. 2015).

72 Table 3.8: Relative contribution (%) of functional woody plant to daily browsing time of dairy cattle in cluster OUA-4. Data are shown for the rainy season (RS: July 2015 - October 2015), the early dry season (EDS: November 2015 - February 2016), and the late dry season (LOS: March 2015 - June 2015).

Season Site Acacia Combretum Dichrostachys Gardenia Guiera Pi/iostigma Securinega Ximenia macrostachya aculeatum cinerea erubescens senegalensis reticulatum virosa americana DC. Vent. (L.) Wight & Stapf & G.F.G mel. (DC.) (Roxb. Ex L. Arn. Hut ch. Hochs t. Willd.) Baill RS K2 p B 50.0 K1 100.0 66.7 EDS K2 100.0 w'""" p 66.7 B 34.5 6.9 24.1 13.8 K1 29.4 5.9 26.5 5.9 11.8 LOS K2 1.4 27.0 1.4 16.3 3.6 31.2 18.4 p 50.0 22.2 B 11.5 3.9 11.5 34.6 15.4 11.5 K1 3.9 3.9 11.5 44.9 20.5 10.3 3.0 16.0 4.2 9.9 22.1 6.1 22.8 11.0 Tot al 6.0 14.0 4.3 8.7 22.3 6.0 22.0 9.7 K2 : Kamboinse 2, P: Ponsotenga, B: Boas sa, K1 : Kamboinse 1. Cha ter 3

3.3.3 Pasture quality, biomass production and carrying capacity

At all sites the highest number of species was found for annual grasses and forbs. Perennial legumes were limited to a single species per grazing site. The specific contribution of annual grasses was above 50% in most cases and there was a significant difference between sites with regards to specific contribution of annual grasses (p � 0.001). The specific contribution of annual legumes was significantly lower in Boassa than the other three sites (p � 0.001 ). In the latter the average specific contribution ranged from 26 to 32%. The specific contributions of perennial grasses, perennial legumes and forbs were less than 10% on average with huge variation within sites and a significant difference between sites (Table 3.9). All sites were dominated by plant species of low to average pasture quality (Table 3.10). Nevertheless, the pastoral value of the different studied sites was above 50%, with significant differences between sites. The highest pastoral value (72.1±7.2) was found in Kamboinse 1 and the lowest (50.9±5.0) was found in Boassa (Table 3.10).

The average biomass production did not differ significantly between the different sites. In addition apart from forbs, the relative contribution of the four other species groups to biomass production did not differ significantly between the four sites. Yet, the contribution to biomass production was dominated by annual grasses (form 38 to 50% on average). On average the annual carrying capacity of each site approximated 0.5 TLU/ha*year. Thus, the annual and seasonal carrying capacity did not differ between the different grazing sites (Table 3.11 ).

3.3.4 Body weight development

Across seasons and irrespective of sex, age group and physiological status, the LW of cattle in OUA-3 was higher than in OUA-4 (p � 0.05). Apart from suckling animals the same trend was observed with BCS that was higher in OUA-3 than in OUA-4 (p � 0.05). In adult animals, ADG was significantly higher in lactating cows in OUA-3 than in OUA- 4, mainly during EDS and LOS. Likewise, young cattle had higher ADG in OUA-3 than in OUA-4 during the same seasons and the same holds true for BCS in EDS and LOS. Except for adult males in EDS, there was no significant difference in ADG within suckling animals and adult male animals, respectively (Table 3.12; Table 3.13; Table 3.14).

Site Average biomass Relative contribution of species groups Annual carrying Seasonal carrying (kg OM/ha) to biomass production (%) capacity capacity (TLU/ha* (n=3) (TLU/ha*year) 120 days)

AG PG AL PL F

K2 3984±2278 49.9 a ±44.4 - 11.2 a ±17.6 7.8 a ±13.5 31.2 aE ± 32.4 0.6 a ±0.3 1.8 a ±1.0

a a a a a a p 2767±1364 46.9 ±23.2 11.4 ± 9.9 23.0 ±13.0 12.4 ±17.9 6.3 ab ± 5.8 0.4 ±0.2 1.2 ±0.6

a a a a a a B 3560±1312 47.6 ±28.7 19.9 ± 6.1 1.3 ± 2.3 25.7 ±16.9 5.6 b ± 5.1 0.5 ±0.2 1.6 ±0.6

K1 2919± 547 37.5 a ± 6.5 7.5 a ±13.0 7.2 a ±12.5 2.4 a ± 4.1 45.4 a ± 9.0 0.4 a ±0.2 1.3 a ±0.2 p� n.s. n.s. n.s. n.s. n.s. 0.05 n.s. n.s.

p: p-value, AG: Annual grass, PG: Perennial grass, AL: Annual legume, PL: Perennial legume, F: Forbs, TLU: Tropical Livestock Unit, OM: dry matter, n.s.: not significant. Cha ter 3 In OUA-3 the season had a significant effect on LW development of pregnant and young animals. Thus LW of both groups was highest during LOS and lowest during EDS. Apart from dry cows and adult males, ADG was significantly different between seasons in the four other groups (pregnant, lactating, young and suckling animals). In those groups ADG was highest during EDS and lowest during LOS. BCS was not significantly affected by the season in OUA-3 (Table 3.12; Table 3.13; Table 3.14).

In OUA-4, the season had a significant effect on LW of pregnant, lactating, young and suckling animals. In those four groups LW was significantly lower during RS whereas there were no differences in LW between EDS and LOS (except for young animals). In all groups ADG was significantly affected by season. Similarly to OUA-3, ADG was highest during EDS and lowest during LOS in all groups, except for suckling animals. Unlike in OUA-3, BCS was significantly affected by season in pregnant and lactating cows, suckling animals and adult males: values were highest in EDS for pregnant and lactating cows and adult males, and in LOS for suckling animals (Table 3.15).

3.3.5 Milk production and composition

Irrespective of season the average daily milk yield was significantly higher in OUA-3 than in OUA-4 (p � 0.001 ), and the average lactation day of lactating cows was lower in OUA-3 than in OUA-4 in all three seasons (p � 0.01 ). There was no significant difference for milk fat and milk protein content between dairy production systems. Only in RS was milk lactose concentration significantly higher in OUA-4 than in OUA-3 (p � 0.01 ), while the concentration of solid non-fats was higher in OUA-4 than in OUA-3 during RS and LOS (p � 0.01 and p � 0.05, respectively) (Table 3.16).

Irrespective of season there was no significant difference for average lactation day in both dairy production systems. Likewise the average daily milk yield did not differ significantly in OUA-3 and OUA-4. There was a significant difference for milk fat content between the three seasons in OUA-3 and only between EDS and LOS in OUA-4. In both production systems milk fat content was lowest in EDS, and no significant difference was observed between LOS and RS. Milk protein content did not differ significantly between seasons in both production systems, but season significantly affected the lactose content of milk in both dairy production systems. The difference between seasons for solid non-fat was significant for EDS and RS (Table 3.17).

77 Table 3.12: Seasonal variation in live weight (LW) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016).

Season Group OUA-3 OUA-4 p� Age (month) LW (kg) Age (month) LW (kg) Mean S.D. Mean S.D. n Mean S.D. Mean S.D. n Pregnant cow 81 36 412 111 24 70 31 204 54 42 0.001 Lactating cow 83 38 410 93 51 84 21 226 38 63 0.001 RS Dry cow 77 22 198 27 10 Adut male cattle 37 2 412 145 4 48 16 201 57 21 0.01 Young cattle (m,f) 11 4 124 69 34 11 4 72 19 56 0.001 Suckling calf (m ,f) 3 1 42 18 17 3 1 23 8 11 0.01 Pregnant cow 66 40 320 83 57 67 31 233 65 110 0.01 Lactating cow 84 33 400 8 120 86 21 262 83 119 0.001 EDS Dry cow 29 21 217 54 6 67 29 218 48 25 n.s. Adut male cattle 34 7 345 146 11 46 16 228 68 68 0.01 -....I Young cattle (m,f) 10 4 110 48 125 11 4 87 28 120 0.001 Suckling calf (m ,f) 3 1 47 18 64 2 1 36 10 46 0.05 Pregnant cow 75 43 366 115 74 65 28 227 60 87 0.001 Lactating cow 78 31 416 111 76 87 20 252 42 100 0.001 LOS Dry cow 44 29 329 238 2 66 29 204 51 29 n.s. Adut male cattle 35 6 386 152 11 42 15 211 62 66 0.001 Young cattle (m,f) 12 4 137 54 68 10 3 79 19 102 0.001 Suckling calf (m ,f) 2 1 47 18 36 2 1 39 14 30 n.s. p: p-value (for LW), RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male, f: female, n.s.: not significant. Cha ter 3 Table 3.13: Seasonal variation in average daily weight gain (kg) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016).

Season Group OUA-3 OUA-4 p� Mean S.D. n Mean S.D. n Pregnant cow -0.08 0.23 21 0.04 0.31 42 n.s. Lactating cow -0.07 0.33 40 0.04 0.33 63 n.s. RS Dry cow -0.02 0.29 10 Adut male cattle 0.28 0.31 3 -0.04 0.31 21 n.s. Young cattle (m, f) 0.18 0.27 15 0.10 0.27 55 n.s. Suckling calf (m, f) 0.13 0.23 15 -0.01 0.12 9 n.s. Pregnant cow 0.25 0.43 53 0.23 0.30 86 n.s. Lactating cow 0.13 0.52 112 0.07 0.34 83 n.s. EDS Dry cow -0.02 0.29 5 0.09 0.22 11 n.s. Adut male cattle 0.43 0.42 9 0.13 0.27 42 0.05 Young cattle (m, f) 0.38 0.29 112 0.20 0.21 85 0.001 Suckling calf (m, f) 0.33 0.19 33 0.29 0.13 16 n.s. Pregnant cow 0.07 0.38 64 -0.20 0.29 85 0.001 Lactating cow -0.22 0.66 64 -0.34 0.45 99 n.s. LOS Dry cow 0.00 1 -0.19 0.36 29 n.s. Adut male cattle 0.09 0.59 10 -0.00 0.40 63 n.s. Young cattle (m, f) 0.26 0.32 61 0.09 0.17 101 0.001 Suckling calf (m, f) 0.30 0.16 15 0.26 0.12 21 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male, f: female, n.s.: not significant.

79 Table 3.14: Seasonal variation in body condition score (BCS) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016).

Season Group OUA-3 OUA-4 p� Mean S.D. n Mean S.D. n Pregnant cow 3.2 0.3 24 2.5 0.3 42 0.001 Lactating cow 3.3 0.4 51 2.5 0.4 62 0.001 RS Dry cow 2.3 0.4 9 Adut male cattle 3.6 0.3 4 2.7 0.6 17 0.01 Young cattle (m, f) 3.2 0.3 34 2.7 0.3 48 0.001 Suckling calf (m, f) 2.9 0.4 15 2.7 0.3 11 n.s. Pregnant cow 3.2 0.5 57 2.8 0.3 109 0.001 Lactating cow 3.4 0.5 120 2.7 0.4 118 0.001 EDS Dry cow 3.1 0.2 6 2.6 0.3 25 0.01 Adut male cattle 3.5 0.3 11 3.0 0.4 68 0.001 Young cattle (m, f) 3.0 0.4 125 2.7 0.3 120 0.001 Suckling calf (m, f) 3.0 0.3 63 2.8 0.2 46 0.01 Pregnant cow 3.2 0.6 74 2.6 0.3 86 0.001 Lactating cow 3.2 0.5 76 2.5 0.3 100 0.001 LOS Dry cow 2.8 0.4 2 2.4 0.3 29 n.s. Adut male cattle 3.3 0.3 11 2.9 0.5 66 0.01 Young cattle (m, f) 3.1 0.4 68 2.8 0.3 100 0.001 Suckling calf (m, f) 3.1 0.3 36 2.9 0.2 30 0.01 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male, f: female, n.s.: not significant.

80 Cha ter 3 Table 3.15: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) of cattle of different physiological status kept in two (peri-) urban dairy production systems in Ouagadougou (October 2014 - February 2016). Variable Season p Group Dry Pregnant Lactating Young Suckling Adult male cow cow cow cattle calf cattle (m, f) (m, f) OUA-3 EDS-LDS-RS <0.01 0.564 <0.001 0.504 0.609 LW EDS-LOS <0.001 0.298 <0.001 0.741 0.491 EDS-RS 0.505 <0.05 0.574 <0.001 0.299 0.361 LDS-RS 0.166 0.720 0.71 0.282 0.695 EDS-LDS-RS <0.01 <0.01 <0.01 <0.05 0.453 ADG EDS-LOS <0.001 <0.01 <0.05 0.586 0.205 EDS-RS 0.351 <0.05 0.053 <0.01 <0.01 0.644 LDS-RS 0.076 0.741 0.318 <0.05 0.735 EDS-LDS-RS 0.062 0.131 0.178 0.133 0.106 BCS EDS-LOS <0.05 0.050 0.198 0.107 0.104 EDS-RS 0.127 <0.05 0.568 0.178 0.452 0.481 LDS-RS 0.941 0.217 0.605 0.072 0.074 OUA-4 EDS-LDS-RS 0.151 0.05 0.001 <0.01 <0.01 0.117 LW EDS-LOS 0.231 0.263 0.055 <0.05 0.570 0.124 EDS-RS 0.053 0.01 0.001 0.001 <0.01 0.084 LDS-RS 0.368 0.05 0.001 <0.05 <0.01 0.329 EDS-LDS-RS <0.05 0.001 0.001 <0.01 <0.001 <0.05 ADG EDS-LOS <0.01 0.001 0.001 <0.01 0.407 <0.05 EDS-RS 0.245 0.01 0.439 <0.05 <0.001 <0.05 LDS-RS 0.101 0.001 0.001 0.805 <0.001 0.687 EDS-LDS-RS 0.111 0.001 0.001 0.186 <0.01 <0.05 BCS EDS-LOS <0.05 0.001 0.001 0.209 <0.05 <0.05 EDS-RS 0.165 0.001 0.001 0.414 0.231 <0.05 LDS-RS 0.908 0.283 0.658 0.083 <0.01 0.322 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners), m: male, f: female.

81 The milk fat-to-protein ratio (FPR) averaged 1.56±0.07 (204 milk samples) and 1.53±0.07 (164 milk samples) in OUA-3 and OUA-4, respectively (Figure 3.3). Between dairy production systems FPR was not significantly different for the same lactation stage. However, within dairy production systems, there was a significant difference in FPR between the 2nd and 3rd stage of lactation (p � 0.05).

The percentage of animals with a negative energy balance (NEB) and at risk of ketosis (RK) followed the same trend and increased from early lactation (stage 1) through mid lactation (stage 2) to late lactation (stage 3) in both dairy production systems (Figure 3.4). Yet the increase was more important in OUA-4 where the percentage of NEB and RK in stage 3 was 7.5 and 2.3 times the percentage in stage 1, respectively, while it was only 1.1 and 1.2 times that level in OUA-3. In OUA-3 the percentage of NEB and RK was always above 50% in each stage of lactation while in OUA-4 it was above 50% only for stage 3. The percentage of animals at risk of acidosis (RA) was low across all stages (�15% except for stage 2 in OUA-4 where it reached 30%).

3.3.6 Feed use efficiency

Throughout the study period the average daily intake of feed dry matter, nutrients and metabolisable energy (ME) at homestead by cows, adult males and young animals, all expressed per tropical livestock unit (TLU), was significantly different for different groups within the same dairy production system, as well as within the same group between the two dairy production systems (Table 3.18). In OUA-3 in most cases the daily intake at the homestead was highest in lactating cows followed by young animals and the same holds true for OUA-4. The average OM intake per TLU of lactating and young animals was, respectively, 6 and 8 times higher in OUA-3 than in OUA-4, and the magnitude of difference was similar for nutrients and ME. In both dairy production systems the average daily intake was lowest in dry animals, followed by adult males. Animal group had a significant effect on the average crude protein to energy ratio in both production systems (Table 3.18). Apart from lactating cows and dry cows, the dairy production system had no effect on the average crude protein to energy ratio in the homestead feed of pregnant, male and young animals (Table 3.18). The average BCS of all groups was significantly higher in OUA-3 than in OUA-4. In contrast to OUA-3 the BCS was significantly different between the different groups in OUA-4 (p � 0.001 ). In the latter cluster, young animals had the highest body condition score (Table 3.18).

82 Table 3.16: Milk yield and milk composition of cows kept in two different production systems in the (peri-) urban area of Ouagadougou (October 2014 - February 2016). RS EDS LOS OUA-3 OUA-4 p� OUA-3 OUA-4 p� OUA-3 OUA-4 p� Criteria Mean±S.D. (n) Mean±S.D. (n) Mean±S.D. (n) Mean±S.D. (n) Mean±S.D. (n) Mean±S.D. (n)

LacD(day) 144±112(40) 221±126 (38) 0.01 149±91 (111) 198±113 (89) 0.01 171 ±111(78) 221±93 (56) 0.01

LacN 3.4±2.54 (40) 3.0±1.47 n.s. 3.3±2.06 (111) 2.7±1.49(89) n.s. 3.2±2.00 (78) 2.8±1.24(56) n.s.

DMY (g) 10121±4810 (40) 1565±833 (38) 0.001 8384±4555 (111) 1621 ±807 (89) 0.001 8369±4531 (78) 1389±665 (56) 0.001

Fat(%) 5.3±1 .13 (29) 5.0±1.62 (37) n.s. 4.7±1.27 (101) 4.6±1.27(71) n.s. 5.1 ±1.61(75) 5.1 ±1.6(56) n.s.

Protein(%) 3.1±0.20 (29) 3.2±0.17 (37) 0.01 3.2±0.19 (101) 3.2±0.12(71) n.s. 3.1 ±0.23(75) 3.2±0.22(56) n.s.

Lactose(%) 4.4±0.46 (29) 4.8±0.29 (37) 0.001 4.7±0.29 (101) 4.7±0.21(71) n.s. 4.7±0.32 (75) 4.8±0.34(56) n.s. co (.,,) SNF (%) 8.6±0.57 (29) 9.0±0.44 (37) 0.01 8.8±0.50 (101) 8.8±0.32 (71) n.s. 8.7±0.55 (75) 9.0±0.60 (56) n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard deviation, n: number of measurements, LacD: Lactation day, LacN: Lactation number, DMY: Daily milk yield, SNF: Solid non-fats, n.s.: not significant, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production(Fulani owners). Chapter 3 Table 3.17: Effects of season on milk yield and milk composition in two different (peri-) urban dairy production systems of Ouagadougou (October 2014 - February 2016). Criteria Season p OUA-3 OUA-4 Lactation day EDS-LDS-RS 0.318 0.268 EDS-LOS 0.229 0.148 EDS-RS 0.558 0.248 LDS-RS 0.194 0.746 Average daily milk yield EDS-LDS-RS 0.118 0.234 EDS-LOS 0.904 0.085 EDS-RS 0.055 0.580 LDS-RS 0.061 0.416 Percent fat EDS-LDS-RS <0.05 0.088 EDS-LOS 0.110 <0.05 EDS-RS <0.01 0.209 LDS-RS 0.177 0.632 Percent protein EDS-LDS-RS 0.189 0.058 EDS-LOS 0.405 0.170 EDS-RS 0.064 <0.05 LDS-RS 0.294 0.365 Percent lactose EDS-LDS-RS <0.01 <0.05 EDS-LOS 0.946 0.132 EDS-RS <0.001 <0.01 LDS-RS <0.01 0.257 Solid non-fats EDS-LDS-RS 0.066 <0.05 EDS-LOS 0.996 0.055 EDS-RS <0.05 <0.05 LDS-RS 0.053 0.580 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners).

84 Cha ter 3

0 0

0 0 0 0

0 0 0 €l � ·a5 2

0 o,�---�---�---�---�----�---�--� 1-0UA-3 1-0UA-4 2-0UA-3 2-0UA-4 3-0UA-3 3-0UA-4 Lactation stage-farm cluster

85 Cha ter 3

•NEB CJRK CJRA 100 90

80 70

� 60 Q) 0) $ 50 Q) � Q) 40

30 20 10 0 1-0UA-3 2-0UA-3 3-0UA-3 1-0UA-4 2-0UA-4 3-0UA-4 Lactation stage-farm cluster

Figure 3.4: Percentage of cows depicting a negative energy balance (NEB), a risk for ketosis (RK), and a risk for acidosis (RA) at least once in different lactation stages (1: 1- 30 days, 2: 31-120 days, 3: >120 days) in two different (peri-) urban dairy production systems in Ouagadougou (OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners)) (October 2014 - February 2016).

Nutrient and ME requirements were affected by the dairy production system in pregnant, lactating and young animals contrarily to dry cows and adult males where the production system had no effect on nutrient and ME requirements (Table 3.19). Obviously the group effect on nutrient and ME requirements was significant.

Apart from one group (young animals for ME and dry cows for CP), the dairy production system effect on the coverage of ME and CP requirements though homestead feeding was significant with higher values being observed in OUA-3 (p � 0.05; Figure 3.5; Figure 3.6). The group effect over the coverage of ME and CP requirements varied considerably. Thus, in OUA-3, the coverage of ME requirements was significantly different (p � 0.05) between dry and lactating, dry and male, lactating and pregnant, male and pregnant, pregnant and young animals (half of the pairwise comparisons). Still

86 Cha ter 3 in the same dairy production system, the coverage of GP requirements was significantly different (p � 0.05) only between young animals and productive cows (pregnant and lactating) (one fifth of the pairwise comparisons). In OUA-4, the coverage of ME requirements was significantly different (p � 0.05) between dry and young animals, lactating cows and the three other groups except for dry cows, young and pregnant, young and male animals (two thirds of the pairwise comparisons). The coverage of GP requirements was significantly different (p � 0.05) between lactating and pregnant, lactating and male, young and male animals.

Within and across dairy production systems, the coverage of ME and GP requirements through intake at the homestead displayed a wide range of variation from severe ME (or GP) deficit to substantial ME (or GP) supply (Table 3.20; Table 3.21 ). In OUA-3 the proportion of substantial ME supply situations was highest and pregnant cows were concerned in more than half of the cases. Half of the cases were shared between adequate ME supply, moderate ME deficit, and moderate ME supply with no significant differences in the share of ME supply levels. Within the same cluster, a severe ME deficit was observed in almost half (44%) of the cases in lactating cows. In OUA-4, for more than 75% of the cases a severe ME deficit was observed, and no case of substantial ME supply (Table 3.20). In parallel to the adequacy of ME coverage, adequacy of coverage of crude protein requirements was observed in OUA-3 (Table 3.21 ). In OUA-4, however, there were less than 10% cases of substantial crude protein supply observed, of which more than two thirds occurred in lactating cows.

87 Cha ter 3

* 0 *

0 0

Q) E � -s * 0 c:r Q) 0 W1 � * 0 * j

0) (0 8 0 3-Preg 3-Lact 3-Dry 3-Male 3-You 4-Preg 4-Lact 4-Dry 4-Male 4-You Group

Figure 3.5: Coverage of metabolisable energy (ME) requirements (intake (MJ ME d- 1 ):requirements (MJ ME d-1 )) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou, Burkina Faso (October 2014 - February 2016), 3: OUA-3 cluster, mixed crop-(dairy) cattle production, 4: OUA-4 cluster, mixed crop-dairy production (Fulani owners), Preg: pregnant cow, Lact: lactating cow, Dry: dry cow, Male: adult male cattle, you: young cattle.

88 Table 3.18: Intake at the homestead (mean±S.D. for the 16 month study period) of feed dry matter, crude protein, phosphorus and metabolisable energy by cows, males and young tropical livestock units (TLU) in two dairy production systems in Ouagadougou (October 2014 - February 2016). Farm Group n Body weight BCS Dry matter Crude protein Phosphorus Metabolisabl Crude protein cluster (kg) (kg) (g) (g) e energy to energy ratio (MJ) (%) Pregnant cow 62 358 8e ± 87 3.2 8 ±0.5 3.9 e ±2.3 405.5 e ±322.6 20.3 e±19.1 40.0 8e ±33.8 9.6 8 ± 7.1 Lactating cow 51 389 8 ± 90 3.3 8 ±0.4 8.6 8 ±4.3 1338.7 8 ±443.4 84.9 8 ±31.5 51.0 8 ±40.5 75.8 b ±259.5 OUA-3 Dry cow 22 231 b ± 71 3.1 8 ±0.4 1.7 b ±2.4 148.8 b ±232.8 7.6 b ±12.9 23.8 b ±35.7 6.2 8 ± 3.7 Adult male cattle 6 271 8be ±150 3.4 8 ±0.2 2.8 be ±1.8 225.4 be ±174.7 12.0 bc±10.3 17.6 be±16.2 28.0 8b ± 25.2 d 8 bd b 8 Young cattle(m, f) 79 113 ± 43 3.2 ±0.4 6.1 ±5.4 425.9 bd ±469.8 32 .5 cd±60 .8 22.2 ±37.3 31.0 ± 53.4 p� 0.001 n.s. 0.001 0.001 0.001 0.001 0.001 Pregnant cow 75 240 8e ± 64 2.7 8e ±0.3 0.2 e ±0.6 35.5 e ± 84.7 2.5 e ± 6.8 3.3 be ± 1.1 15.5 d ± 13.3 Lactating cow 54 260 8 ± 49 2.6 8 ±0.3 1.5 8 ±1.1 242.7 8 ±167.7 13.1 8 ± 9.9 15.4 8 ±17.4 23.7 8 ± 23.5 OUA-4 Dry cow 47 202 b ± 53 2.6 8 ±0.4 0.1 b ±0.4 19.0 b ± 63.0 0.7 b ± 2.6 1.0 b ± 2.6 18.2 8bd ±10.6 de b b c.oCXl Adult male cattle 35 223 ± 66 2.9 ±0.4 0.6 e ±1.1 52.8 e ± 86.4 2.1 e ± 3.5 5.5 c± 9.5 20.3 e± 43.1 Young cattle(m, f) 63 90 e ± 25 2.7 e ±0.2 0.8 e ±1.1 109.2 e±149.4 9.2 d ±13.0 9.2 de ±18.1 8.6 e± 6.3 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 Pregnant cow 0.001 0.001 0.001 0.001 0.001 0.001 n.s. Lactating cow 0.001 0.001 0.001 0.001 0.001 0.001 0.01 p� Dry cow n.s. 0.001 0.01 0.01 0.01 0.001 0.05 Adult male cattle n.s. 0.01 0.05 0.05 0.05 0.05 n.s. Young cattle(m, f) 0.001 0.001 0.001 0.001 0.001 n.s. n.s. p: p-value, BCS: Body condition score, n: number of animals per group, n.s.: not significant, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners). Different superscript letters indicate a significant difference (p�0.05) between groups of the same dairy production system (in the same column) as determined by Kruskal-Wallis test, m: male, f: female. Table 3.19: Energy and protein requirements for maintenance, milk synthesis and growth of different groups of cattle in two dairy production systems in Ouagadougou (October 2014 - February 2016). Per formance Energy requirement Protein requirement (MJ ME Animar1 d-1 ) (g CP Animar1 d-1 ) Farm Group Growth Milk Maint. Growth Activity Milk Maint. Growth Milk cluster (g d-1 ) offtake (walking) synth. synth. (kg d-1) Pregnant cow -2.6 ab ±352 39.2 bd ± 7.2 0.6 ab ±8.8 302 bd ± 55 6 ab ±99 Lactating cow 0.1 a ±201 8±4 41.9 8c ± 7.3 -0.1 b ±6.1 87±20 323 ac ± 57 -1 b ±45 1389±314 OUA-3 Dry cow 15.8 8 ±117 28.2 8 ±61.9 a.s ad ±4.0 217 8 ± 48 6 8 ±45 bc cd bc bc Adult male cattle 210.8 ±211 31.2 ±13.0 5.0 ±S.O 240 cd ±101 59 ±60 Young cattle(m ,f) 262.8 C ±242 16.0 e ±48.0 9.2 C ±8.4 127 e ± 40 98 C ±01 p� 0.001 0.001 0.001 0.001 0.001 Pregnant cow -69.8 ac ±270 29.0 bd ± 6.1 -2.2 ad ±8.6 5.4 be ±2.0 224 bd ± 47 -24 ad ±96 Lactating cow -94.9 ab ±208 2±1 30.9 b ± 4.4 -2.8 8 ±6.0 6.0 b ±1.9 43± 3 238 b ± 34 -31 8 ±67 1069±423 OUA-4 Dry cow -7.7 8c ±105 25.6 8 ± 5.2 -0.3 ab ±3.0 4.2 8 ±1.7 197 a ± 40 -4 ab ±33 ad bd ac ad bd Adult male cattle 35.3 C ±190 27.4 ± 6.3 1.4 ±5.5 4.5 ±2.0 211 ± 48 17 ±66 b c d c Young cattle(m, f) 99.7 ± 96 13.9 C ± 3.0 3.6 ±3.7 2.1 ±0.7 107 C ± 23 38 ±38 p� 0.001 0.001 0.001 0.001 0.001 0.001 Pregnant cow n.s. 0.001 n.s. 0.001 n.s. p� Lactating cow 0.05 0.001 0.001 0.05 0.001 0.001 0.05 0.001 Dry cow n.s. n.s. n.s. n.s. n.s. Adult male cattle 0.05 n.s. n.s. n.s. n.s. Young cattle(m, f) 0.001 0.001 0.001 0.001 0.001 p.- p-value, n.s.: not significant, OUA-3: Mixed crop-(dairy) cattle production, OUA-4: Mixed crop-dairy production (Fulani owners), Different superscript letters indicate significant differences (p�0.05) between groups of the same dairy production system (in the same column) as assessed by Kruskal-Wallis test, m: male, f: female, Maint.: Maintenance, Synth.: Synthesis. 91 Cha ter 3 Table 3.20: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1 ): requirements (MJ ME d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou (October 2014 - February 2016).

Farm Group n Ade. Mod. Mod. Sev. Sub. Total cluster % supply deficit supply deficit supply OUA-3 Pregnant n 7 8 11 5 21 52 ab b d ac cow % 13.5 15.4 21.2 C 9.6 40.4 100.0 Lactating n 10 11 3 16 7 47 cow % 21.3 a 23.4 a 6.4 a 34.0 a 14.9 a 100.0 Dry n 2 1 0 0 5 8 cow % 25.0 ab 12.5 ab o.o ab O.O b 62.5 a 100.0 Adult male n 1 1 0 2 0 4 cattle % 25.0 a 25.0 a o.o a 50.0 a o.o a 100.0 Young n 8 5 1 13 8 35 cattle % 22.9 a 14.3 a 2.9 a 37.1 a 22.9 a 100.0 (m, f) Total n 28 26 15 36 41 146 OUA-3 % 19.2 ab 17.8 b 10.3 ab 24.7c 28.1 a 100.0 OUA-4 Pregnant n 0 2 2 21 0 25 cow % o.o a 8.0 a 8.0 a 84.0 a o.o a 100.0 Lactating n 2 6 1 39 0 48 cow % 4.2 a 12.5 ab 2.1 ab 81.3 b o.o a 100.0 Dry n 0 0 0 6 0 6 cow % o.o a o.o a o.o a 100.0 a o.o a 100.0 Adult male n 1 3a oa 21 a oa 25 cattle % 4.0 a 12.0 0.0 84.0 0.0 100.0 Young n 3 6 0 9 0 18 cattle (m, f) % 16.7 a 33.3 a o.o a 50.0 a o.o a 100.0 Total n 6 17 3 96 0 122 OUA-4 % 4_9 ab 13.9 b 2.5 ab 78.7 c o.o a 100.0 Total n 34 43 18 132 41 268 % 12.7 16.0 6.7 49.3 15.3 100.0 Different superscript letters indicate significant differences (p$'.0.05) between levels of coverage of metabolisable energy requirements within the same group (in the same row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

92 Cha ter 3

* * *

.. � .9 4 * .::.:.Q) cco 0

� 3 * E * ·s * Q) * 0.. 2 0 * 0 Q) Ol co � 1

80�� 3-Preg 3-Lact 3-Dry 3-Male 3-You 4-Preg 4-Lact 4-Dry 4-Male 4-You Group Q Figure 3.6: Coverage of crude protein (CP) requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou (October 2014 - February 2016), 3: OUA-3 cluster, mixed crop-(dairy) cattle production, 4: OUA-4 cluster, mixed crop-dairy production (Fulani owners), Preg: pregnant cow, Lact: lactating cow, Dry: dry cow, Male: adult male cattle, You: young cattle.

93 Cha ter 3 Table 3.21: Adequacy of coverage of CP requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in dairy cattle of two different dairy production systems in Ouagadougou (October 2014 - February 2016).

Farm Group n Ade. Mod. Mod. Sev. Sub. Total cluster % supply deficit supply deficit supply OUA-3 Pregnant n 7 9 4 7 25 52 cow % 13.5 abc 17.3 c 7.7 abc 13.5 b 48.1 ac 100.0 Lactating n 13 3 14 1 16 47 cow % 27.7 ab 6.4 cd 29.8 b 2.1 d 34.o ac 100.0 Dry n 0 1 2 2 3 8 cow % o.o a 12.5 a 25.0 a 25.0 a 37.5 a 100.0 Adult male n 0 2 0 1 1 4 cattle % o.o a 50.0 a o.o a 25.0 a 25.0 a 100.0 Young cattle n 4 7 1 15 8 35 (m, f) % 11.4 a 20.o a 2.9 a 42.9 a 22.9 a 100.0 Total n 24 22 21 26 53 146 ab b a ab OUA-3 % 16.4 15.1 14.4 17.8 C 36.3 100.0 OUA-4 Pregnant n 2 2 0 19 2 25 cow % 8.0 a 8.0 a o.o a 76.0 a 8.0 a 100.0 Lactating n 2 6 0 33 7 48 cow % 4.2 ab 12.5 ab O.O b 68.8 a 14.6 ab 100.0 Dry n 1 0 0 4 1 6 cow % 16.7 a o.o a o.o a 66.7 a 16.7 a 100.0 Adult male n 2 4 0 19 0 25 cattle % 8.0 ab 16.0 ab o.o ab 76.0 b o.o a 100.0 Young cattle n 3 2 1 12 0 18 (m, f) % 16.7 a 11.1 a 5.6 a 66.7 a o.o a 100.0 Total n 10 14 1 87 10 122 ab b a ab OUA-4 % 8.2 11.5 0.8 71.3 C 8.2 100.0 Total n 34 36 22 113 63 268 % 12.7 13.4 8.2 42.2 23.5 100.0 Different superscript letters indicate significant differences (p:5:0.05) between levels of coverage of crude protein requirements within the same group (in same the row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

94 Cha ter 3

3.4 Discussion

In this long term study (16 month monitoring), spatio-temporal changes in herd management, homestead feeding strategies, animal feeding behavior on pasture, pasture quality and quantity, use and adequacy of resource use and production performance of (peri-) urban dairy cattle were assessed in two production systems. The average daily distance covered by grazing cattle herds and the time spent on different activities on pasture was determined for each study period and site. Energy and protein requirements for maintenance, activities and production were determined for each animal and category and their fulfilment through homestead feeding was evaluated, use efficiency analysed and the relative impact of feeding on energy balance assessed, along with consequences for health status.

3.4.1 Herd management

Unlike the findings by Roessler et al. (2016) (and Chapter 2) where herd size on farms in cluster OUA-4 was higher than in cluster OUA-3, the average herd size in OUA-3 and OUA-4 dairy farms was comparable but showed important intra-system variations across seasons. The remarkable difference in the number of animals belonging to the groups of dry cows and adult males (higher in OUA-4) is an interesting discriminating feature of intensive and extensive dairy farms. OUA-4 farms can be considered as pastoral derived, mixed dairy-meat production system. Traditionally this system is characterised by an important stocking of adult animals that have different functions (Baroin and Boutrais, 2008). Here milk plays an important role in feeding the household and is mostly handled by women who decide on the amount of milk to keep for the suckling calves and therefore have a certain control over herd viability (Bruggeman, 1994). However, in the present study dairy farmers sold most of their milk production though in OUA-4 a small part of the production was kept for homestead consumption (data not shown) with less implication of women in dairy production observed at farm level. The same holds true for the age at which animals are sold. Selling animals at maturity depends on the farmer's willingness to sell, and on marketing factors whereby farmers search for a good price for their animals (Bechir, 2010). Slower animal growth due to genetic and environmental factors obliges farmers to sell their animals when they reach a certain weight later on. In this regards the results of this study are consistent with findings showing that local zebu, the main breed type kept by farmers in OUA-4, had a lower growth as compared to other breeds (exotic and Sahelian crossbreds) mainly kept by OUA-3 farmers. As a consequence, animals of the same age class were

95 Cha ter 3 sold for twice the price in OUA-3 (improved breeds) compared to OUA-4 (mainly local breeds) (Gnanda et al., 2016).

3.4.2 Feeding of dairycattle

The more diverse feeding-related activities in cluster OUA-3 are a result of the adoption and implementation of technologies promoted by the government, non-government, international development organisations that were adopted by farmers who, less than two decades ago, reared mostly local breeds in a traditional pastoral or extensive way. In SSA, the chronological development, promotion, and implementation of these technologies have been described by various authors and follow the same scheme in most West African countries, including i) agro-ecological intensification through animal housing, cut-and-carry feeding systems, fodder cultivation, ii) genetic improvement strategies by the importation of exotic dairy breeds and cross-breeding of the local breeds with imported exotic breeds, iii) training of tamers and various value chain stakeholders and capacity-building activities to insure sustainable livelihoods for farmers and enhance food security (Bonfoh et al., 2007; Duncan et al., 2013; Chagunda et al., 2015). The reason for the adoption or non-adoption of these technologies and strategies and their adaptation by (peri-) urban dairy farmers are related to the farmers' socio cultural status and mindset, availability and costs of promoted feeds, crop-livestock integration and production practices (Millogo et al., 2008; Kouame-Sina et al., 2012; Toure et al., 2015). Dairy cattle in (peri-) urban dairy farms of Ouagadougou rely on homestead feeding and pasturing to different degrees. According to Gnanda et al. (2016) in Ouagadougou (Burkina Faso) 57% of dairy cattle are fed on pasture year round (like in OUA-4) and 41 % of the animals are fed on pasture only part of the year (as dairy cattle in OUA-3).

3.4.2.1 Homestead feeding

The variety of feedstuffs used by (peri-) urban dairy producers in this study are similar to those described by Roessler et al. (2016), who showed that homestead feeding was used to a higher percentage by households of OUA-3 cluster. This study provides more insights into the types of feedstuffs used, their usage by dairy enterprises of the different urban and dairy production systems (more diverse in OUA-3 than in OUA-4) and the seasonal variation of the relative share of each feed type in the diets provided to dairy cattle in Ouagadougou. The changes observed throughout the seasons correspond to opportunistic and adaptive feeding strategies used by farmers of the different clusters

96 Cha ter 3 due to changes in feed availability. In this context it is worth to mention the greater use of hays and straws during the late dry season as compared to the two other seasons. The seasonal variation in feed availability and homestead feed provision to animals also translates into changes of the daily offer of proximate diet components in both dairy production systems for almost all animal categories or groups and is characterised by an increase of homestead supply of nutrients and energy from the rainy season to the late dry season. In (peri-) urban areas of West Africa, similar studies have pointed out the seasonal variation of feed supply to cattle and therefore the offer of proximate diet components in low/high input systems and in grazing only/grazing and supplementing cattle enterprises (Diogo et al., 201O; Amadou et al., 2015). The crude protein to energy ratio of the dairy cattle diet varies between dairy production and groups, with higher values observed in OUA-3, mainly due to the use of compound dairy feed (an energy protein mix accounting for almost one third of the diet across the year in OUA-3) from feed processing companies which also give recommendations to farmers on how to use their products. In the case of lactating cows, feed processors' recommendations to farmers are based on milk yield without considering animal body weight changes or other production requirements such as pregnancy and/or activities (walking if any). In addition this study shows that beside compound dairy feeds tamers in OUA-3 also offer other proteins feed such as brewers' grains, cotton seed cake, and corn residues, which together account for about more than a quarter of the diet. This common practice might have negative implications on resource use efficiency given the intra- and inter-system variations in cattle phenotypes and the managerial differences. It has been shown that reducing protein feeds in the diet of lactating cows improves feed efficiency and reduces feeding related costs (Chase et al., 2012).

3.4.2.2 Activities on pasture of year round grazing (peri-) urban dairy cattle in Ouagadougou (OUA-4 cluster)

The monitoring of animals on pasture allows the assessment of the composition of diets and the proportion of time spent for each activity on pasture. Similar methods of determining the feeding behavior of grazing animals have already been used by other authors (Bechir, 201O; Zampaligre, 2012). Our results for the proportion of time devoted to walking are similar to those obtained in Chad (Bechir, 2010) and in the sub-humid zone (Ouedraogo-Kone et al., 2006) as well as in the west of Burkina Faso (Botoni, 2003) where walking was more important during the dry season when fodder is scarce and of poor quality (Diarra et al., 1995). Increasing travelling distance is translated into

97 Cha ter 3 an increasing energy requirement for locomotion which may be adequately supplied during the rainy season but not during the late dry season when pastures are qualitatively and quantitatively poor (Bechir, 201O; Ayantunde and Amole, 2016). The proportion of time spent on grazing activity also takes into account the grazing of crop residues after harvesting of crop fields. In West Africa in general and in Burkina Faso specifically, crop residues play an important role in ruminant feeding (Savadogo et al., 1999). The proportion of time spent on grazing decreases with the decrease of herbaceous forage offer from the rainy to early and late dry season. In the present study, the proportion of time spent grazing was higher in the rainy season matching the results of Zampaligre (2012) who recorded 65-69% of grazing time per day in the northern Sudanian zone. In the rainy season, the herbaceous layer is abundant and cows spend more time grazing fresh herbaceous plants. However the proportion of time spent on grazing can be low when restrictions are imposed, such as housing or corralling animals during the cropping season. The availability of crop residues in crop fields during the early dry season allows for a high proportion of time spent on grazing. Later on, it decreases with the disappearance of herbaceous biomass, crop residues and preparation of crop fields in the late dry season (Bechir, 201 O; Ayantunde and Amole, 2016). From the rainy season to the late dry season the proportion of time spent on browsing increases, while the proportion of time spent on grazing decreases. As herbaceous plants become scare and poor in quality in the late dry season, woody plants are preferably ingested by cows as their nitrogen content is higher than that of the grasses (lckowicz and Mbaye, 2001). Similar observations have been made in Chad (Bechir and Kabore-Zoungrana, 2012) and in Cameroon (Onana and Devineau, 2002).

The proportion of time spent on drinking water follows the same trend as that of grazing due to the high water content of forage during the rainy season and the abundance of watering points across grazing lands. In the early dry season and late dry season, respectively, cattle are watered on average 1 to 2 times per day at temporary water reservoirs (rivers, ponds) which in the late dry season fall dry due to human activities (mainly irrigation) and hot temperatures (irrigation). The proportion of time spent on resting is important during the period of high forage availability. Resting is often influenced by the herder who, in the afternoon, forces the animals to rest for personal reasons (eating or praying). On the other hand, the poor quality of rangelands and the scarcity of feed resources in the dry season oblige the cows to take long walks to explore a greater diversity of grazing lands and therefore reduces the proportion of time devoted to resting. Social behavior was dominated by mother and calf interactions in the

98 Cha ter 3 morning and late in the evening, which is explained by the fact that the calves remain on the farm under the surveillance of another herder, so the interaction with the cow is limited to the milking times. Licking between adults mainly takes place during daytime rest and depends on the hierarchical relationships within the herd (Dumont et al., 2001; Bouissou and Boissy, 2005) which also determines the onset of competitions between individuals for the access to feed resources when these are significantly limited (not observed during the present study).

The duration and distance covered while on pasture are two important features of grazing management. The duration depends on the proportion of time spent by the cows walking and grazing. Long journeys induce high energy expenditure, less time spent on grazing and a reduced efficiency of energy use on pasture. The duration of time spent on pasture in the present study is comparable to observations made by Zampaligre (2012) in the northern Sudanian zone of Burkina Faso. The seasonal variation in the daily travelling distance is similar to observations by Bechir (2010) and Zampaligre (2012) who also recorded short distances in the rainy season and long distances in the early and late dry season. The scarcity of fodder resources explains the lengthening of the distances traveled during the late dry season. With the installation of rains and grass growth on natural pastures, cows had to move very little and spent more time grazing. Land cover preferences of the herds greatly depended on season, location of the farm and its surroundings, with urbanisation and construction sites reducing access to grass lands. During the early dry season when almost all fields are harvested, the animals benefit from easy crop residue grazing on fields (Savadogo et al., 1999), and crop field visits become more important when forage mass and quality decrease on rangelands (Ouedraogo-Kone et al., 2006). The contribution of the different land cover classes to feed intake is however influenced by the herder as has been observed in other sub-humid regions of West Africa (Petit and Mallet, 2001; Bechir, 2010). Although their decision-making might depend on several factors, the main motivation is always the search of good quality and sufficient quantity forage for their cattle.

3.4.2.3 Pastoral value

The pastoral value is an index characteristic of the value of pasture land that takes into account the abundance of species measured by their specific contribution and of their quality measured by a specific index (which is most often determined though surveys of livestock owners or herders) (Le Floc'h, 2007). It depends on non woody plant species

99 Cha ter 3 composing the herbaceous carpet. These species are distinguished into four groups; species of good, average and weak as well as without pastoral value (Akpo and Grouzis, 2000). In this study apart from one study site the pastoral values of the different units did not reach those indicated for pastures in the North-Sudanian areas. Indeed, for this region a given pasture is of excellent quality only if its pastoral value reaches and exceeds 65% (Cesar, 2005). However, pastoral values obtained during our study are clearly superior to those found by Yameogo et al. (2013) in Vipalgo, a village localised in the peri-urban area of Ouagadougou. A good pastoral value is explained by the dominant annual grasses of good to medium pastoral values (Andropogon pseudapricus Stapf and Elionurus elegans Kunth) and annual legumes such as Zamia glochidiata DC. The low pastoral value could be explained by the absence or low specific contribution of legumes. The average phytomass production obtained in this study is similar to those obtained by Sawadogo et al. (2005) and Yameogo et al. (2013). These authors found 2588±330 kg DM ha-1, 4010±151 kg DM ha-1 and 2205 kg DM 1 ha- , respectively, in the peri-urban area of Ouagadougou (within a radius of fifty kilometers from Ouagadougou city center), in the Tioga classified forest belonging to the northern Sudanese zone, and in Vipalgo. Carrying capacities of less than 1 TLU ha-1 yea(1 are frequently reported in (peri-) urban livestock settings such as of 0.73 TLU ha-1 yea(1 in Vipalgo (Yameogo et al., 2013) and 0.96 TLU ha-1 for a period of use of 118 days in the peri- urban areas of Port Bouet and Grand Bassam (Cote d'Ivoire) (Kouassi et al., 2014). The calculated carrying capacities are very low as compared to the number of animals actually grazing on the main pastures. This might result in overgrazing, a phenomenon that could further increase the degradation of grazing lands in the years to come due to important biomass removal and trampling pressure (static load) exerted by the animals of which the deleterious impact increases as stocking rate increases (Savadogo et al., 2007). Thus given human and animal pressure over pastures in the (peri-) urban grazing lands of Ouagadougou and West Africa cities, one would expect an important degradation of the pasture lands year after year. However it must be noticed that most herders or farmers systematically tend to shift grazing lands whenever possible as soon as they feel insecure or realise that grazing animals lose weight cannot meet their needs (not necessarily nutrient requirements of the animals which are rarely adequately met). Yet there is still a lot to learn about the decision making process of this pasture land shifting in the context of (peri-) urban dairy production.

100 Cha ter 3 3.4.3 Weight development and body condition

In general cattle of all groups in cluster OUA-3 had higher LW, ADG and BCS than animals in cluster OUA-4 as a result of differences in the breeds reared by farmers in OUA-3 (mainly European/ international x local zebu crosses, Sahelian zebu/ transboundary x local zebu crosses) as compared to the breeds on OUA-4 farms (local zebu) and environmental factors, among which feeding management plays a core role. These differences influence decision making of farmers as to when to replace (old) cows and to cull or to sell animals. Roessler et al. (2019) have shown for (peri-) urban dairy production systems that crossbred female cattle had higher bodyweight at maturity and a higher maturity rate. If those two features are positively correlated to lifetime productivity, lower body weight and poor body condition (as observed in OUA-4) result in lengthening the unproductive rearing phase of replacement heifers, increase the risk of metabolic diseases (Heuer et al., 1999) and delay the age at which the animal reaches it breeding or selling weight (Gnanda et al., 2016). In addition it must be noted that poor body condition is also strongly correlated with production and reproduction diseases (Lanyasunya et al., 2005), which depresses animal and farm performance and therefore resource use efficiency. The season effect on the body condition of animals was more pronounced in OUA-4 than in OUA-3, with OUA-4 animals being in poor condition during late dry season as compared to rainy season. Yet, dairy farmers in OUA-4 exploit an adaptive trait of West Africa local breeds to seasonal feed shortage, namely the mechanism known as compensatory growth or catch growth. This means that animals have an accelerated growth if a period of feed abundance (such as RS) follows a period of feed scarcity (such as LOS) (Hector and Nakagawa, 2012). In congruency with the present findings, Diogo et al. (2010) recorded a high ADG of cattle in the early dry season. In both dairy production systems of Ouagadougou, the higher ADG obtained during the early dry season can be explained by a combination of good feeding management/ feed availability and health status. In fact, animals in OUA-4 have access to freely available crop residues during early dry season (Chapter 3.4.4) present on harvested farms (Savadogo et al., 1999), while in OUA-3 good quality fodder (conserved for a few weeks or just harvested) is provided. In addition, the challenge of animal diseases such as parasite infestations decrease from the rainy to the late dry season (Wymann, 2005).

101 Cha ter 3 3.4.4 Milk offtake

In this study the average daily milk offtake in OUA-3 was 5 to 7 times higher than that in OUA-4. Roessler et al. (2019) explained this important difference by the genetic potential of breeds used in different dairy production systems and diverging environmental factors. The latter, in our study, are shaped by housing, health care, labour input, feeding management, and (re-)production management on the two farm types. Whereas OUA-3 dairy farmers use high-yielding crossbreds, OUA-4 farmers mainly rely on local breeds. Several authors have underlined the higher milk yield of improved crossbreds both on station and on-farm (Chapter 1 ). In addition, environment factors clearly discriminate both production systems. In comparison to farmers in cluster OUA-4, OUA-3 farmers invest more in good housing, continuous health care, cattle oriented labour and diverse feeding management strategies, a set of features characteristic for intensive herd management (Gnanda et al., 2016; Marshall et al., 2016). Across (peri-) urban farm households in Ouagadougou it has been shown that factors predicting intensification of livestock keeping (including dairy cattle farming) are site-specific and influenced by the education level of the farmer and the security of land ownership; however, higher inputs in livestock systems do not always translate into higher outputs (Roessler et al., 2016). The breeding strategies used by farmers in OUA- 3 and OUA-4 did not in qualitative terms significantly differ (Chapter 1; Roessler et al., 2016) but differed in terms of intensity or frequency. Thus farmers in OUA-3 more frequently use cross breeding for genetic improvement of their stock whereas farmers in OUA-4 only (rarely) apply cross breeding under genetic improvement programs promoted by diverse institutions (government or NGOs) and later on give up by getting rid of crossbred animals less adapted to their grazing-based herd management (own discussion with OUA-4 farmers) at a very early stage of the process. This results in different levels or stages of genetic improvement through artificial insemination or cross breeding at farm level (Gnanda et al., 2016). Thus farmers in OUA-3 cluster have fully adopted genetic improvement technologies since decades whereas the ones belonging to OUA-4 cluster have poorly adopted these technologies. It is thus important to consider farmers' production purposes and expectations, production management and capacities when planning for genetic improvement, and also consider local breeds' traits that are important to farmers but poorly expressed in imported cattle and their crossbreds. The absence of a significant effect of season on the average daily milk yield (calculated from lactating cows being in different lactation stages) in the two dairy production systems is a result of the management strategies put in place by the

102 Cha ter 3 farmers, which aims at maintaining a certain level of milk production irrespective of the season, in order to supply the market across the year. Yet, this does not mean that milk production at animal level is similar across seasons as the lactation curve follows a given physiological trend (Roessler et al., 2019) and the number of lactating cows at a specific lactation stage varies as well at the farm level. In this regard it has been shown by several authors that the milk yield and therefore market supply with milk is higher during the rainy season as a result of seasonal calving of cows, showing a peak of births in the last months of the dry season and the first month of the rainy season. Reasons for this seasonality are the availability and quality of feed resources, but also herders' decisions, in particular about sparing milk for the calves (Diop et al., 2009; Kassa et al., 2016; Zezza et al., 2016). The extreme heat stress conditions as well as feed and water scarcity prevailing in the late dry season in sub Saharan Africa (Ayantunde and Amole, 2016) require that farmers implement measures to keep their cows productive and reduce thermal stress by providing shelter, sprinkling cool water over animals, providing clean and fresh drinking water, feeding in the cooler early morning and late evening hours. In addition, adequate feeding must be assured especially for heat sensitive and highly demanding improved crossbreds (Hansen and Arechiga, 1999; West, 1999; Staples, 2007).

3.4.5 Milk composition and energy balance

In this study the milk composition measured in the two dairy production systems of Ouagadougou was similar to the composition reported by Millogo et al. (2008) in Burkina Faso. Several genetic and non-genetic factors such as breed, lactation number, lactation stage, body condition score, pregnancy, as well as seasonal and managerial factors affect milk composition (Baro et al., 2016). In this study the production system had an effect on milk protein, lactose and SNF content only during the rainy season, while there was a seasonal effect on milk fat, lactose and SNF content. The seasonal effect on milk fat content can be explained by a higher fiber (NDF) intake of dairy cattle during the late dry season as compared to the rainy and early and dry season (Chapter 3.4.3). A higher NDF content of the diet leads to a linear increase in milk fat content, due to a higher acetic acid production in the rumen (Arelovich et al., 2008). Hence the production of abnormally high levels of milk fat usually indicates that total milk offtake is low and therefore farmers' return to investment is also low in comparison to situations of normal milk fat content (Bailey et al., 2005). However despite the significant increase in milk fat in the late dry season, the milk yield obtained in this study did not decrease

103 Cha ter 3 significantly that period of the year. Substantial or excessive energy intake, such as occurring when overfeeding concentrates, may reduce milk fat content and increase milk protein content (Arelovich et al., 2008). Normal milk protein content can be achieved when energy requirements are being met as milk protein content is more under control of the mammary gland than protein content of the diet (Bequette et al., 1998; Chase et al., 2012). It is important to optimise feed intake in order to minimize situations of negative energy balance during early lactation and other lactation stages (Bruinenberg et al., 2002; Bjerre-Harp0th et al., 2012). In the present study the proportion of animals in negative energy balance (NEB) and thus at risk of ketosis (RK) was above 50% in all lactation stages in OUA-3 and reached a peak around 80% in OUA-4 for the last stage of lactation. Thus, during the last stage of lactation, the energy deficit was 1.6 fold higher in cows of OUA-4 than in cows of OUA-3. This was most likely the result of a stronger dependence of OUA-4 on grazing quantitatively and qualitatively poor pastures during the late dry season. The relatively low increase of RK from early stage of lactation to late stage of lactation in OUA-3 is a consequence of a better feeding management through fodder conservation (yet the forage kept might have low quality during the late dry season as compared to early dry season and rainy season). Thus the milk fat content significantly increased while milk protein content did not significantly change from the rainy season to the late dry season. Keeping in mind the high percentage of animals calving during the last months of the late dry season and the first month of the rainy season, those animals are in late lactation during the late dry season. In addition, in the same context, it has been shown that cows in first lactation have a higher milk protein content than multiparous cows (Roessler et al., 2019). Risk factors of ketosis include high milk yield, increasing parity, over-feeding of pre-partum animals, season of calving, and dry period length; high yielding breeds such as exotic cattle and their crosses (OUA-3) are therefore bearing a higher risk of ketosis. A subclinical ketosis prevalence of 24.1 % (ranging from 8.3 to 40.1 % ) has been reported for early stage lactating cows from different regions worldwide, with always higher prevalence in multiparous than primiparous cows (Brunner et al., 2018). One case of a cow that depicted symptoms of clinical ketosis during early lactation was observed during the monitoring phase of the present study. Still the risk of other production disorders due to negative energy balance such as milk fever, retained placenta, mastitis, metritis, displaced abomasum, claw disease and clinical ketosis, dystocia, and decreased fertility rate should be considered as well. Overfeeding grains and underfeeding fibers is the main risk factor of acidosis in dairy cows (Oba and Wertz-

104 Cha ter 3 Lutz, 2011). However, grains are less frequently provided to dairy cattle in the (peri-) urban dairy farms in Ouagadougou (Chapter 3.3.2.1; Table 3.3), supporting the notion of a comparatively low percentage of dairy cows at risk of acidosis.

3.4.6 Feeds and feeding efficiency

The high proportion of animals oversupplied with crude protein and metabolisable energy in OUA-3 (about 50%) indicates a very low conversion of homestead feed dry matter, nutrient and energy intake into weight gain and/or milk yield. The observed inefficiency is worst during the rainy season and the early months of dry season as animal might further have access to nearby available grazing grasslands (Schlecht et al., 2019). The situation in OUA-3 is contrasting with that one in OUA-4: in the latter system almost all animals are undersupplied with crude protein (82.8% of animals) and metabolisable energy (92.6% of animals), which translates into a very low expression of production traits. OUA-4 dairy farms are still highly dependent on the quality and quantity of fodder present on pasture, and the results obtained in the present study support the notion that (peri-) urban pasture land in Ouagadougou is not adequately covering the nutrient and energy requirements of dairy cattle. This is supported by the strong evidence of poor pasture quality (Chapter 3.3.3) and the high proportion of lactating cows at negative energy balance (Chapter 3.3.5). Efficient use of ingested nutrients and metabolisable energy is only achieved when animals are fed at an adequate level. In contrast, other supply levels such as mild and severe deficit and mild and substantial oversupply, respectively, are cases of inefficient utilization (Chapter 3.3.6). Efficient use of the metabolisable energy provided at the homestead was only noted in 19.2% and 4.9% of the observations in OUA-3 and OUA-4, respectively. For crude protein, an efficient utilization was observed in 16.4% and 8.2% of the cases in OUA-3 and OUA-4, respectively. By efficiency we mean optimal provision of feed to animals that then adequately convert homestead feed crude protein and energy intake into growth and milk output while inefficiency includes all imbalance or discrepancy of intake versus requirements (Bruinenberg et al., 2002; Bjerre-Harp0th et al., 2012), leading to physiological imbalance. Inefficient use of energy and nutrients due to their oversupply is not economically profitable for tamers and leads to a substantial nutrient accumulation in the city environment (Schlecht et al., 2019), especially in the context of poor manure handling and storage. Inefficient nutrient and energy supply have also been reported from (peri-) urban livestock farms in Bobo-Dioulasso (Dossa et al., 2015a), Niamey (Diogo et al., 2010) and Sikasso (Amadou et al., 2015).

105 Cha ter 3 3.5 Conclusions

The present study indicates that (peri-) urban dairy enterprises in Ouagadougou are characterised by both an oversupply of nutrients to dairy cattle in commercial and intensive dairy production systems, and an undersupply in the semi-commercial and extensive dairy production systems. Oversupply is due to poor feeding management caused by farm internal and external factors; it results in inefficient nutrient. Underfeeding dairy cattle results in low outputs (low growth rate, low milk yield, poor body condition, fertility problems). Undersupply arises when homestead feed supply is insufficient and grazing lands supply to little feed of poor quality as observed in the present study.

Consequently most lactating cows in the extensive dairy production system experience a negative energy status which can result in metabolic diseases and production diseases, affecting the farms' profitability. Therefore, in the context of a challenging (peri-) urban environment, trends towards dairy production intensification bring about new challenges for farmers and animal health care professionals as the risk for metabolic and production diseases increases and farm profitability decreases. Since (peri-) urban dairy production intensification is likely to further increase, all bottlenecks preventing more efficient dairy production should be identified and mitigation strategies implemented accordingly. Policies and strategies that aim at supporting the dairy sector in West Africa should therefore address the question of efficient resource use in a holistic way. Only then will interventions be relevant and cost-effective and contribute to food security enhancement with regards to milk and dairy products.

106 Cha ter 3

107 Cha ter 4

4. Performancesand efficiency of (peri-) urban pig and beef cattle breeds under different production managements in Ouagadougou, Burkina Faso

108 Cha ter 4 4.1 Introduction

Globally, the growth in demand for pig meat is supposed to reach 34.6 million metric tons, representing a 66% increase from the year 2000 as baseline until 2030 (FAO, 2011). In sub-Saharan Africa, it will reach 1.1 million metric tons, corresponding to a 155% increase for the same period. This translates to a growth in per capita demand for pork of 44% worldwide (an absolute increase of 1.9 kg/person) and of 47% in sub Saharan Africa (an absolute increase of 0.6 kg/person) (FAO, 2011 ). Within the same period the importance of urbanisation for growth patterns in the demand of pig meat in sub-Saharan Africa is such that urban demand will increase for about 235%. This is more than twice the growth of demand in the rural areas (103%), while domestic production is expected to increase by about 162% (equivalent to an absolute increase of 1079 million metric tons) (FAO, 2011). Since pig consumption is limited in some areas by cultural and religious considerations, the increase in domestic pig production will depend both on (peri-) urban and rural production in areas where religious believes are not an obstacle to that growth (Kiendrebeogo et al., 2008). Several pig production systems operate in the vicinity of West African cities and rely on different breeds or types and production strategies, resulting in different production performances ( Oke et al., 2006; Fualefac et al., 2014; Kiendrebeogo et al., 2008, 2012b, 2014; Kouamo et al., 2015). Crossbreeding of local and exotic breeds has been more or less important in different countries (Oke et al., 2006; Kiendrebeogo et al., 2008, 2012b; Fualefac et al., 2014) such that the average carcass weight of animals slaughtered is highly variable, ranging from less than 30 kg in a first group of countries (Guinea, Ghana, Togo, Benin, Cameroon, Gabon), from 30 kg to 40 kg in a second group of countries (Guinea Bissau, Liberia), and up to more than 50 kg in a third group of countries (Cote d'Ivoire, Equatorial Guinea, Nigeria, Sierra Leone) (D'Orgeval, 1997).

Within the same time frame (from 2000 till 2030) the demand of cattle meat will also remarkably increase in sub-Saharan Africa, presumably reaching about 113% with still the highest increase observed in cities and their vicinities (200%) as compared to rural areas (64%) (FAO, 2011 ). In West African capital cities such as Ouagadougou in Burkina Faso, mutton and beef meat are an alternative to pork and are highly appreciated as a source of animal protein by both people that do (and do not) consume pork for cultural or religious reasons (FAO, 2007). In addition, Sahelian countries such as Burkina Faso are known to be beef cattle providers to markets of West African coastal countries (SWAC-OECD/ECOWAS, 2008). For the (peri-) urban areas of West

109 Cha ter 4 African cities, several beef cattle production systems with variable ranges of resource use and use efficiency have been described (Diogo et al., 2010; Amadou et al., 2012) and most often different cattle breeds or types of different production potential are used in those settings. Trends towards specialisation and intensification of livestock production in West Africa are country, city and site specific (Roessler et al., 2016) and production performances involve both environmental or non-genetic factors (such as feeding intensity and resource use over successive seasons, production management, animal health and disease) and genetic factors (breed for instance). Whereas the efficiency of cattle production in (peri-) urban settings has already been studied in a few West African cities such as Niamey (Niger) (Diogo et al., 2010), Sikasso (Mali) (Amadou et al., 2012) and Bobo-Dioualasso (Burkina Faso) (Dossa et al., 2015a) less is known about the efficiency of pig farming systems. Being important sources of protein of animal origin, the question therefore arises about the situation of pig and beef cattle breed types performances and their resource use efficiency in and around Ouagadougou. We hypothesised that in this West African city (which is taken as an example for other cities in the region), for both species (beef cattle and pig), production performances and resource use efficiency are positively correlated to the use of breeds of higher production potential. The latter is thereby defined as an attempt towards intensification of livestock production. Consequently, the objective of the current study is to determine the effects of different pig and beef cattle breeds/genotypes under different husbandry and feeding management systems on production performance and resource use.

4.2 Material and methods

4.2.1 Study site

This study was carried out in Ouagadougou, the capital city of Burkina Faso. It is located in the hot semi-arid Sudano-Sahelian West Africa. The annual rainfall was about 731 and 821 mm in 2014 and 2015, respectively. The rainy season stretches from May to October when vegetation quality and quantity on fallows and rangelands in and around Ouagadougou are highest. The cool dry season (or early dry season) with lowest temperatures (minima of 14° C in February during the study period) runs from December to February and the hot dry season (or late dry season) with highest temperatures (average maxima of 40°C in April during the same period) runs from March to May (UrbanFoodPJus Project own records).

110 Cha ter 4 4.2.2 Households and animals

The study focused on 7 households involved in beef cattle farming and 6 households involved in pig production. Farms were purposively selected after prior clustering of 157 surveyed livestock keeping households into 4 major livestock production systems (Chapter 2; Roessler et al., 2016). The pig farms belonged to 2 clusters namely OUA-1 (pig production; n=50) and OUA-2 (mixed crop-pig and non-dairy ruminant production; n=17) whereas beef cattle farms were selected from OUA-2 and OUA-3. OUA-3 was composed of crop-cattle farms (n=54 of which 14 and 40 farms focused on dairy and beef production, respectively) and OUA-4 exclusively grouped dairy farms (n=36) owned by Fulani and not considered in the present study. For input (homestead feed offer and animal intake) and output quantification (animal weight changes, offspring) each farm was visited ten times every 6 - 10 weeks from October 2014 until February 2016 that is during a period of 16 months. Three of the monitored pig farms kept exclusively local pigs and the three others kept crossbred pigs (of which one farmer kept both, but with a strong focus on crossbred). Beef cattle where classified into two major breeds namely local zebu or local breed ("zebu Peul") and Sahelian zebu or Sahelian breed ("zebu Gudali"). Two out of the seven beef cattle farms were not supplementing their animals (mainly relying on pasture year round to feed their animals) and were designated as non-supplementing households, while the five other providing homestead feeding were grouped into supplementing households. Only the latter farmers kept Sahelian zebu and both household types (supplementing and non-supplementing) kept local zebu.

4.2.3 Data collection

On-farm monitoring of beef cattle started with animal identification (breed, sex, age group, physiological status) and regular quantification of the animal intake and offtake data, body weight development, body condition score and quantification of stall feeding. The data was collected in the same way as described for dairy cattle (Chapter 3.2). For pig farms a few differences (as compared to cattle farms) in data collection concerned weighing of animals: individuals heavier than 40 kg were weighed in a wooden box rather than on a platform (used for cattle) and body condition scoring was done according to DEFRA (1998). In addition, with respect to the definition of age groups of pigs, distinction was made between suckling piglets (pigs <2 months), young animals (pigs: >2-11 months for crossbred pigs, >2-12 months for local pigs) and adult animals (>11 months for crossbred pigs, >12 months for local pigs). The total number of animals

111 Cha ter 4 per species, breed and group at the start of the study are given in Table 4.1 and Table 4.2.

Table 4.1: Number of pigs per group and breed at the start of the study. Group Local breed (n) Crossbred (n) Dry sow 5 5 Lactating sow 3 1 Pregnant sow 1 3 Adut male pig 0 0 Young pig (m, f) 65 28 Suckling piglet (m, f) 0 14 n: number of animals, m: male; f: female.

Table 4.2: Number of cattle per group, feeding intensity (non-/supplemented) and breed at the start of the study. Non-suppmemented Supplemented Group Local breed (n) Local breed (n) Sahelian breed (n) Dry cow 11 1 0 Lactating cow 3 0 0 Pregnant cow 2 2 0 Adut male cattle 4 25 2 Young cattle (m, f) 3 10 0 Suckling calf (m, f) 3 0 0 n: number of animals, m: male, f: female.

4.2.4 Data calculation

The same methodology used for data calculation in dairy farms (Chapter 3.2.2.5) was applied for beef cattle. For pigs a few differences were considered as far as the calculation of ME and digestible protein (DP) requirements were concerned. Thus, to determine excess feeding and feed scarcity from the share of stall-feeding, values for DM and OM digestibility (DMD, OMO) of all feedstuffs were taken from Feedipedia (2017) or Close and Menke (1986) (Table A.3 of "Appendix"). In addition the pre-caecal CP digestibility in pigs was also derived from these two sources to calculate a feedstuffs concentration of digestible protein (DP). ME and DP requirements for maintenance and growth were calculated using pig-specific values (Ulbrich et al., 2004; Table A.4 of "Appendix"). The determination of the adequacy of metabolisable energy (or crude / digestible protein) intake of both pigs and beef cattle was carried out following the same procedure described for dairy cattle (Chapter 3.2.2.5).

112 Cha ter 4 4.2.5 Statistical analyses

As for dairy cattle (Chapter 3.2.5) a similar methodology was used for statistical analyses. For each species comparative analyses where carried out between breeds and/or feeding intensities. Thus descriptive analyses were used for herd structure, animal intakes and offtakes. The rank-based non-parametric Kruskal-Wallis test was used to compared two or more groups (namely dry, pregnant, lactating, adult male and suckling animals within or between breeds and feeding intensities whenever applicable) of the different quantitative independent variables namely performance variables (reproduction performance, live weight, average daily weight gain, body condition score) and feeding variables (offer of proximate diet components). The season's (RS, EDS, LOS) impact over group performances was tested using the same statistic test. For proportion comparisons of the different CP and ME coverage levels the Pearson's Chi squared test was applied. Logarithmic, polynomial and power functions (not following the biological principles of mammal's growth) were used to depict growth curves for different groups of both species, and were selected according to the best fit. A threshold for significance was set at a p-value � 0.05. All the statistical analyses were done with IBM SPSS Statistics 20.

4.3 Results

4.3.1 Pig production

4.3.1.1 Herd structure and dynamic

The total number of animals for each breed and group is depicted in Table 4.2. Households did not keep adult male pigs on-farm. Young pigs represented 87.8% of animals while the share of adult females was 12.2% of local pigs. Crossbred pig herds were composed of 17.6% adult females, 54.9% young pigs and 27.5% piglets. In Table 4.4 the main reasons for herd inflows and outflows, namely births, purchases, sales and deaths, are depicted for the three seasons. The mortality rate was higher for local pigs than for crossbred pigs (p � 0.001). Throughout the study period, 61 % of the registered local animals died, with a range of losses between 54 and 70% per farm, while only 14% of the crossbred pigs died. Thus there occurred more deaths than births for local pigs. Consequently, the number of animals kept slightly decreased from the start of data collection until the end. Died animals in local pigs were mainly composed of young animals (61 %), followed by suckling piglets (21 %) and sows (15%). The same trend was observed in crossbred pigs where dead animals were mainly young animals (71 %),

113 Cha ter 4 sows (13%), adult males (10%) and suckling piglets (6%). For local pigs the mortality was significantly influenced by the season and was highest during LOS (50%, p � 0.001), whereas the mortality rate was not affected by season in crossbred pigs. As far as births are concerned, most pregnancies started during EDS, so that more than half of the piglets were born during LOS. For the local breed a seasonal effect on the production cycle was observed (p � 0.001) such that two thirds of births occurred during LOS and only 10% of births were observed during EDS. Though also affected by the season, the situation was slightly different and less pronounced for crossbred pigs where two fifths of births happened during RS and a quarter of births happened during EDS. With respect to marketing, 62% of sales where reported during EDS and LOS in local pigs and crossbred pigs, respectively, showing an effect of the season on selling with animals being mainly sold during the long dry season (EDS plus LOS).

4.3.1.2 Reproduction performances

Reproduction performances in both breed types were similar (Table 4.3). During the whole investigation period, 92 and 96 births (including births recorded at the start of the study) were registered from 17 crossbred and 19 local reproductive sows. This means that respectively 5.4±2.3 versus 5.1 ±1.9 piglets were registered alive during the sixteen months study period. For both groups the overall alive piglets to sow ratio averaged 3. The inter-farrowing interval averaged 206±42 days and there was no significant difference between both groups for farrowing interval.

Table 4.3: Reproduction performance of pigs in Ouagadougou. Breed type Births Adult Farrowing Farrowing interval Average litter Overall alive (n) sows (n) (days) size (n) piglets to sow (n) Mean±SE (n) ratio Local breed 96 29 19 194±46 (5) 5.1±1.9 3.3 Crossbred 92 32 17 227±29 (3) 5.4±2.3 2.8 Total 186 61 36 206±42 (8) 5.2±2.1 3.0 p� n.s. n.s. p: p-value, n: number of observations, n.s.: not significant.

114 Cha ter 4 Table 4.4: Animal intakes and offtakes in two different pig breed types in Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LOS) (October 2014 - February 2016). Breed ty�e Local breed {n} Crossbred {n} Season RS EDS LOS Total p� RS EDS LOS Total p� Inflow/Out Group flow e Dry sow 5 1 8 14 0 0 2 2 Lactating sow 0 0 0 0 0 0 0 0 Died Pregnant sow 0 0 0 0 0.001 0 0 0 0 n.s. Adut male pig 2 0 1 3 0 1 1 2 Young pig (m, f) 12 9 34 55 3 3 5 11 Suckling piglet (m, f) 4 6 9 19 0 1 0 1 Dry sow 0 2 2 4 2 1 15 18 Lactating sow 0 0 0 0 0 0 0 0 Sold Pregnant sow 0 0 0 0 n.s. 0 0 0 0 0.001 Adut male pig 0 3 2 3 1 2 1 4 Young pig (m, f) 2 10 1 13 11 7 21 39 Suckling �iglet {m, :Q 0 0 0 0 0 0 0 0 Dry sow 0 0 0 0 0 0 0 0 Lactating sow 0 1 0 1 0 0 1 1 Newly Pregnant sow 0 0 0 0 0.001 1 1 0 2 0.001 registered Adut male pig 0 0 0 0 0 1 0 1 Young pig (m, f) 5 0 4 9 6 0 6 12 Suckling piglet (m, f) 18 10 53 81 29 13 24 66 Dry sow 0 1 2 3 0 1 0 1 Lactating sow 0 0 1 1 0 0 0 0 Other Pregnant sow 0 0 0 0 n.s. 0 0 0 0 n.s. Outflows Adut male pig 0 0 0 0 1 1 1 3 Young pig (m, f) 1 9 2 12 8 9 5 22 Suckling piglet (m, f) 0 0 3 3 4 0 0 4 p: p-value, n: number of animals, m: male, f: female, n.s.: not significant.

4.3.1.3 Feed quantity and quality

There were differences in feeding patterns between breeds and between seasons (Table 4.5). In contrast to LOS and EDS, less feedstuffs were used in the diet during RS for both breed types. The evaluation of the relative percentage of feedstuffs (on dry mater basis) offered to pigs showed that brewers spent grain was used in the diet of both breeds but had a higher share in the diet of local pigs during all three seasons. Feedstuffs such as commercial pig concentrates, fishmeal and rice (cereal) husk were used only in the diet of crossbred pigs. Green grass was only marginally used during RS in the diet of local pigs. Other feedstuffs such as fruits (mainly mango) were used during LOS for the local pigs (3% of the diet). The share of meal leftovers in the diet was similar in both breeds during RS unlike the two other seasons when meal leftovers were not used in the diet of crossbred pigs. Starting with a marginal share (4%) in EDS, one

115 Cha ter 4 third of the diet of crossbred pigs was made up of straws and hays during LOS. In both breeds the use of cereal grain residues in the diet increased from RS to LOS, starting from 5% in the RS and reaching 16% and 25% in the LOS diet of local and crossbred pigs, respectively. Table A.3 ("Appendix") presents the proximate composition of feedstuffs offered to pigs in the (peri-) urban area of Ouagadougou. For the same breed, apart from daily P offer in local pregnant sows, the season did not influence the daily offer of proximate diet components for pregnant and lactating sows. However the daily offer of proximate diet components was significantly higher in crossbred sows' diet than in local sows' for both lactating and pregnant groups (p � 0.001; Table 4.6). Though similar for both pig breeds, the daily offer of proximate diet components decreased from RS to EDS in crossbred dry sows unlike in their local breed counterparts where it was constant (Table 4.7). For adult male pigs the season and the breed had no effect on the daily offer of proximate diet components. Except for daily DP and ME offer in local young pigs, and daily OM, OM, NDF and ADF offer in crossbred young pigs, the season and the breed significantly influenced the daily offer of proximate diet components in young and suckling pigs (Table 4.8). The highest daily offers were observed in crossbred young and sucking pigs. For crossbred young pigs the highest daily offers were observed during LOS, whereas local young pigs received the highest offer during RS. Compared to young pigs a difference was observed in crossbred suckling piglets with the highest daily supplies provided during EDS.

116 Table 4.5: Feed categories and percentages of feedstuffs (on dry mater basis) offered to pigs in Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LOS) (October 2014 - February 2016). Local breed Crossbred Feed category 1 Feed category 2 Feed type RS EDS LOS RS EDS LOS Energy feed Cereal grain Brewers grain, Meal leftover 33.74 11.33 14.35 26.45 Energy feed Cereal grain Meal leftover 3.45 Energy feed Other Fruit 2.98 Energy/Protein mix Commercial feed Pig concentrate (commercial) 26.38 27.65 5.03 Protein feed Protein feed Brewers grain 59.24 44.12 38.49 11.76 19.32 6.50 Protein feed Protein feed Brewers grain, Cereal chaff 26.54 24.66 30.72 20.34 11.16 Protein feed Protein feed Cereal grain residues 6.99 18.01 16.08 4.70 10.48 25.16 Protein feed Protein feed Fishmeal 1.29 Protein feed Protein feed Brewers grain, Fishmeal, Cereal grain residues 18.45 Protein feed Protein feed Brewers grain, Fishmeal, Cereal grain residues, Cereal husks 17.47 Roughage Straws and hays Brewers grain, Cereal husks 3.75 12.09 � � Roughage Straws and hays Cereal husks 21.31 Roughage Protein roughage Green grass 0.02 Table 4.6: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to pregnant (a) and lactating (b) sows of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). a) Breed type Season (n) OM OM DP p NDF ADF ME kJ/kq MW RS (5) 59.1 ±48.6 54.3±43.2 9.8± 5.8 0.61±0.39 23.2± 14.6 10.4± 6.4 632.0±563.1 Local breed EDS (21) 42.9±25.4 33.3±22.7 6.0± 3.8 0.42±0.24 18.0± 11.2 8.7± 7.1 448.5±261.6 LOS (13) 24.8±19.4 21.9±17.1 4.± 3.1 0.18±0.18 10.7± 7.8 6.0± 4.5 266.8±211.1 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS (3) 121.3±59.9 102.2±46.8 21.5±12.5 1.11±0.69 62.1± 34.4 38.5± 24.0 1350.6±691.6 Crossbred EDS (22) 169.2±98.1 155.1±88.6 23.5±13.5 1.9±1.40 63.2± 47.5 30.0± 26.6 1777.2±940.6 LOS (10) 231.7±301.1 254.1 20.0±19.1 1.8±1.80 103.3±163.5 60.7±107.2 1766.9±1302.6 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. E.� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 b)

_,_ _,_ Breed type Season (n) OM OM DP p NDF ADF ME co kJ/kq MW RS (?) 82.2±24.4 70.6±18.0 13.0±4.3 0.75±0.20 38.2±16.2 21.0±11.9 894.1 ±284.2 Local breed EDS(?) 80.7±41.4 72.7±37.9 11.9±5.8 0.47±0.24 30.8±18.1 15.9± 9.1 879.0±443.8 LDS(17) 58.7±55.1 54.2±51.5 8.3±7.8 0.54±0.69 22.8±21.1 10.5± 9.6 631.7±595.4 p � n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS (6) 155.9±53.5 141.3±51.0 23.3±6.6 1.73±0.71 47.4± 7.9 22.4±24.4 1798.9±649.8 Crossbred EDS(?) 127.4±81.6 116.8±74.7 17.8±11.2 1.09±1.04 48.4±35.9 22.1±19.3 1400.2±878.3 LOS (2) 138.2±74.0 126.3±65.3 18.2±13.6 0.90±0.18 58.2±44.1 29.7±26.5 1482.8±759.7 p � n.s. n.s. n.s. n.s. n.s. n.s. n.s. p� 0.01 0.01 0.01 0.01 0.05 0.05 0.01 p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, DP: digestible protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW= metabolic body weight. Table 4.7: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry sows (a) and adult males (b) of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). a) Breed type Season (n) OM OM DP p NDF ADF ME kJ/kq MW RS (13) 37.1± 36.3 33.2± 31.3 6.1± 5.6 0.36±0.34 16.7±17.7 8.0± 9.4 385.1± 382.3 Local breed EDS (23) 50.5± 51.3 46.4± 47.4 7.5± 7.3 0.47±0.58 19.1±18.6 8.9± 8.5 541.3± 553.7 LOS (18) 51.6±126.9 46.8±114.9 7.6±18.5 0.23±0.52 23.4±59.5 12.6±32.3 549.7±1345.3 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS (6) 106.6±605.3 98.0± 56.2 15.4± 8.8 0.94±0.64 40.0±26.2 18.8±12.8 1171.0± 649.7 Crossbred EDS (25) 86.9± 94.5 79.8± 86.8 14.1±1 7.4 0.87±1.03 30.6±34.0 13.0±15.6 1021.0±1143.6 LOS (32) 26.4± 78.2 23.9± 70.6 2.2± 6.1 0.12±0.31 13.4± 4.3 7.4±24.8 175.4± 474.3 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 E.� n.s. n.s. n.s. n.s. n.s. n.s. n.s. b) � � Breed type Season (n) OM p NDF ADF ME CD DM DP kJ/ka MW RS (4) 26.1± 33.3 24.4± 31.1 5.2± 6.6 0.31 ±0.39 11.9±15.2 5.4± 6.9 263.7± 336.3 Local breed EDS (7) 50.2± 20.7 56.2± 26.7 7.4± 4.6 0.42±0.22 19.7±11.5 9.4± 6.0 538.2± 325.2 LOS (6) 25.0± 34.0 22.7± 31.1 4.4± 6.2 0.16±0.21 12.5±18.8 6.4± 9.7 259.0± 345.1 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS (5) 45.5± 65.4 42.1± 60.6 6.2± 8.9 0.28±0.40 19.6±28.2 9.4±13.6 481.9± 693.5 Crossbred EDS (18) 85.5±121.3 79.3±112.8 13.3±20.7 0.60±0.75 36.1±55.4 17.1±26.6 986.1±1470.4 LOS (10) 84.5±113.3 77.2±102.5 9.2±14.4 0.50±0.56 37.8±57.5 20.0±32.5 771.2±1059.4 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, DP: digestible protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW= metabolic body weight. Table 4.8: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to young pigs (a) and suckling piglets (b) of two different pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). a) Breed type Season (n) OM OM DP p NDF ADF ME kJ/kq MW RS (79) 53.1± 64.7 47.2± 58.6 6.6± 8.7 0.47±0.49 28.4± 50.1 17.1± 46.5 410.1± 546.1 Local breed EDS (204) 35.8± 31.7 32.8± 28.6 4.9± 4.5 0.28±0.27 15.0± 13.4 7.6± 7.6 375.6± 331.7 LOS (117) 29.1± 39.6 26.1± 36.1 4.6± 5.9 0.20±0.30 12.4± 17.6 6.6± 9.3 311.1± 422.1 p� 0.001 0.001 n.s. 0.001 0.001 0.01 n.s. RS (76) 123.5±117 .4 110.8±106.7 18.3±17.5 1.11±0.91 49.5± 52.9 25.6± 28.9 1368.2±1267.5 Crossbred EDS (170) 109.7±117.6 100.9±108.4 16.8±19.2 1.08±1.16 39.7± 48.8 17.9± 23.8 1229.4±1346.8 LOS (114) 179.9±312.2 160.6±273.7 17.1±32.3 1.13±1.82 88.2±170.0 52.3±107.3 1350.6±2170.4 p� n.s. n.s. 0.01 0.05 n.s. n.s. 0.05 E.� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 b) � Breed type Season (n) OM OM DP p NDF ADF ME � 0 kJ/kq MW RS (23) 23.0± 35.6 21.5± 33.2 4.58± 7.09 0.27±42.0 10.5±16.3 4.8± 7.4 232.3± 359.4 Local breed EDS (32) 4.2± 14.6 3.9± 13.9 0.56± 1.89 0.04±0.14 1.7± 5.6 0.7± 2.6 44.4± 154.7 LDS(103) 7.8± 19.9 7.2± 18.4 1.23± 3.20 0.05±0.13 2.4± 6.0 1.1± 2.8 90.7± 232.5 p � 0.05 0.05 0.05 0.05 0.05 0.05 0.05 RS(45) 50.9± 91.7 43.3± 76.9 8.52±16.50 0.52±0.91 24.6±47.2 15.0±30.0 568.7±1028.0 Crossbred EDS (84) 103.9±134.3 94.9±122.4 13.23±16.37 1.02±1.51 40.0±59.0 19.8±30.9 1035.2±1280.1 LOS(28) 25.5± 63.7 23.5± 58.6 3.11± 7.77 0.27±0.68 6.4±16.0 3.0± 7.6 296.3± 740.0 p � 0.01 0.01 0.01 0.05 0.01 0.05 0.01 p� 0.001 0.001 0.001 0.001 0.001 0.001 0.001 p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, DP: digestible protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW= metabolic body weight. Cha ter 4 4.3.1.4 Weight development

Crossbred pigs were in general about 0.5 to two times heavier than local pigs, irrespective of age and physiological group (Table 4.9). The difference was more pronounced for adult male animals with crossbred adult male pigs being about three times as heavy as adult male pigs of the local breed. However the variation (standard deviation or variance) of live weight in crossbred pigs was higher than the variation of live weight in local pigs, which also varied substantially within the same group. Therefore, even though average live weight for specific groups was significantly different between both breeds (p � 0.05) there were overlaps such that some crossbred pigs were three times as heavy as local pigs of the same group, and others fell within the range of the average live weight of local pigs. A seasonal effect on live weight was observed only in piglets and young pigs of both breeds (p � 0.05). Young pigs had the highest live weight during EDS (25.67±21.13 kg) and LOS (16.33±10.53 kg) in crossbred and local pigs, respectively. For suckling pigs the highest live weights were measured during EDS (5.74±3.91 kg) and RS (3.72±2.82 kg) in crossbred and local pigs, respectively. In Table 4.10 the average daily weight gain (ADG) is shown for different age and physiological groups of both breeds for the three different seasons. Overall ADG in both breeds showed a wide variation with standard deviations exceeding the means in most cases. Apart from pregnant sows in LOS, where the average daily gain was significantly higher in crossbred than in local pigs, there was no significant difference in other groups as far as ADG is concerned. Apart from lactating sows, the highest values of ADG in other groups were observed for crossbred pigs (with no significant difference) with in general local pigs showing about 65% of the ADG of crossbred pigs. Apart from the RS, pregnant sows showed the highest ADG followed by adult male pigs during EDS and LOS. Season had a significant effect on ADG in pregnant sows of both breeds and on young pigs of local breed only. Pregnant sows had the highest ADG during LOS (0.38±0.23 kg) and EDS (0.18±0.13 kg) in crossbred and local pigs, respectively. For young local pigs the highest values of ADG were recorded during EDS (0.08±0.09 kg). In Table 4.11 the average body condition score (BCS) is shown for different age and physiological groups of both breeds for the three different seasons. Average BCSs were significantly higher in lactating sows, young pigs/piglets and pregnant sows of crossbred pigs than in the comparative groups of local breed pigs during RS, EDS and LOS, respectively. Whatever the season, in other comparative groups there was no significant difference in the average BCS between both breeds despite slightly higher values for

121 Cha ter 4 crossbred pigs in all groups (except adult male pigs). In both breeds and throughout the study period pregnant sows, followed by lactating sows, had the highest average BCS of around 3.5 and 3, respectively. Dry sows and adult male pigs had the lowest average BCS in both breeds (less than 2 except for EDS). Apart from dry and lactating sows of the local breed, the season had no effect on BCS (Table 4.12). The average BCS was only affected by the breed for pregnant sows, whereby pregnant sows of the crossbred type were in better condition than local pregnant sows. Within the same pig breed the average BCS was the highest for pregnant sows and the lowest for dry sows and adult males. In fact the latter were remarkably in poor conditions with an average BCS inferior to 2.

Table 4.9: Seasonal variation in live weight (kg) of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

Season Group Local breed Crossbred p� Mean S.D. n Mean S.D. n Pregnant sow 72 18 5 79 43 3 n.s. Lactating sow 53 15 7 109 60 6 n.s. RS Dry sow 46 17 13 82 34 6 0.05 Adut male pig 21 12 4 67 19 5 0.05 Young pig (m, f) 9 6 96 19 19 76 0.01 Suckling piglet (m, f) 4 3 23 4 3 45 n.s. Pregnant sow 62 14 21 95 29 22 0.001 Lactating sow 47 19 7 80 40 7 n.s. EDS Dry sow 47 19 23 75 30 25 0.01 Adut male pig 36 10 7 86 27 18 0.001 Young pig (m, f) 14 8 204 26 21 170 0.001 Suckling piglet (m, f) 3 7 32 6 4 84 0.001 Pregnant sow 50 21 13 124 36 10 0.001 Lactating sow 48 14 17 107 48 2 0.05 LOS Dry sow 44 17 18 67 23 32 0.01 Adut male pig 35 4 6 95 36 10 0.01 Young pig (m, f) 16 11 118 22 12 114 0.001 Suckling piglet (m, f) 3 2 103 4 2 28 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male, f: female, n.s.: not significant.

122 Cha ter 4 Table 4.10: Seasonal variation in average daily weight gain (kg/day) of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

Season Group Local breed Crossbred p� Mean S.D. n Mean S.D. n Pregnant sow -0.03 0.11 2 -0.17 0.05 2 n.s. Lactating sow -0.14 0.27 2 -0.20 0.20 5 n.s. RS Dry sow 0.02 0.09 5 -0.01 0.22 2 n.s. Adut male pig 0.06 0.04 2 0.22 0.29 2 n.s. Young pig (m, f) 0.04 0.04 36 0.10 0.12 46 n.s. Suckling piglet (m, f) 0.03 1 0.08 0.09 13 n.s. Pregnant sow 0.19 0.13 18 0.22 0.19 13 n.s. Lactating sow -0.05 0.37 4 -0.14 0.26 6 n.s. EDS Dry sow 0.07 0.14 16 0.00 0.20 12 n.s. Adut male pig 0.10 0.10 6 0.21 0.18 7 n.s. Young pig (m, f) 0.07 0.07 130 0.11 0.12 95 n.s. Suckling piglet (m, f) 0.11 0.11 2 0.10 0.06 43 n.s. Pregnant sow 0.14 0.16 13 0.38 0.23 8 0.05 Lactating sow -0.09 0.17 16 LOS Dry sow -0.01 0.11 7 0.8 0.8 14 n.s. Adut male pig 0.12 0.07 3 0.17 0.11 8 n.s. Young pig (m, f) 0.08 0.09 79 0.11 0.11 72 n.s. Suckling piglet (m, f) 0.06 0.04 39 0.06 0.03 4 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male; f: female, n.s.: not significant.

123 Cha ter 4 Table 4.11: Seasonal variation in body condition score of pigs of different physiological status in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

Season Group Local breed Crossbred p� Mean S.D. n Mean S.D. n Pregnant sow 3.3 0.5 5 3.2 0.3 3 n.s. Lactating sow 2.2 0.4 7 2.8 0.6 6 0.05 RS Dry sow 1.9 1.6 13 1.8 1.5 6 n.s. Adut male pig 1.5 1.7 4 1.0 1.4 5 n.s. Young pig (m, f) 2.3 1.1 96 1.8 1.5 76 n.s. Suckling piglet (m, f) 2.5 1.2 23 2.7 1.0 45 n.s. Pregnant sow 3.2 0.4 21 3.4 0.5 22 n.s. Lactating sow 2.8 0.7 7 2.8 0.9 7 n.s. EDS Dry sow 2.2 1.3 23 2.2 1.2 25 n.s. Adut male pig 2.6 1.1 7 1.6 1.8 18 n.s. Young pig (m, f) 1.9 1.3 204 2.2 1.4 170 0.05 Suckling piglet (m, f) 2.0 1.5 32 2.8 0.7 84 0.05 Pregnant sow 3.0 0.3 13 3.7 0.4 10 0.01 Lactating sow 2.7 0.8 17 3.3 0.4 2 n.s. LOS Dry sow 0.9 1.4 18 1.4 1.5 32 n.s. Adut male pig 1.6 1.7 6 2.0 1.7 10 n.s. Young pig (m, f) 1.8 1.3 118 2.0 1.4 114 n.s. Suckling piglet (m, f) 2.7 1.0 103 3.0 0.5 28 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, m: male, f: female, n.s.: not significant.

124 Cha ter 4 Table 4.12: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) in pig groups of different breeds in Ouagadougou (October 2014 - February 2016). Variable Season p Group Dry Pregnant Lactating Young Suckling Adult sow sow sow pig piglet male (m, f) (m, f) pig Local breed EDS-LDS-RS 0.883 0.091 0.681 <0.001 <0.001 0.083 LW EDS-LOS 0.655 0.111 0.727 0.266 <0.001 0.566 EDS-RS 0.843 0.241 0.482 <0.001 <0.05 0.088 LDS-RS 0.718 0.061 0.446 <0.001 0.764 <0.05 EDS-LDS-RS 0.594 <0.05 0.907 <0.01 0.116 0.545 ADG EDS-LOS 0.316 0.207 0.636 0.540 0.053 0.606 EDS-RS 0.772 <0.05 0.639 <0.01 0.221 0.505 LDS-RS 0.567 0.069 0.888 <0.01 0.461 0.248 EDS-LDS-RS <0.05 0.315 <0.05 0.071 0.063 0.537 BCS EDS-LOS <0.01 0.161 0.660 0.421 <0.05 0.455 EDS-RS 0.675 0.832 0.075 0.085 0.186 0.223 LDS-RS 0.084 0.259 <0.05 <0.05 0.539 0.814 Crossbred EDS-LDS-RS 0.339 0.055 0.444 <0.01 <0.01 0.248 LW EDS-LOS 0.392 <0.05 0.380 0.745 <0.05 0.828 EDS-RS 0.497 0.427 0.253 0.01 <0.01 0.071 LDS-RS 0.147 0.091 0.867 <0.001 0.379 0.293 EDS-LDS-RS 0.488 <0.05 0.463 0.610 0.133 0.887 ADG EDS-LOS 0.236 0.119 0.840 0.147 0.640 EDS-RS 0.784 <0.05 0.463 0.521 0.105 0.883 LDS-RS 0.634 <0.05 0.257 0.821 0.792 EDS-LDS-RS 0.229 0.117 0.623 0.205 0.768 0.588 BCS EDS-LOS 0.086 0.143 0.550 0.327 0.470 0.699 EDS-RS 0.716 0.214 0.821 0.084 0.922 0.508 LDS-RS 0.547 0.069 0.211 0.409 0.567 0.243 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, m: male, f: female.

125 Cha ter 4

120 .t...... · ... ·· .t. .t. t . · ...... · · 100 .·· · ... .t. • · ... .t. .· ... .,.... ·· ... .t. .. " .t. .t. ..•· 80 '\• ... : . · ...... • ...... ,..; · • • 0) • :.-1 • . • ...... • • • ,.·· .t. ..., • ..., .t.. • • 60 ! ...• • • 0) ...... • ·a, ...... � . . : . • .t. , . • . . . I • • • ...... • : • .I : • • • • •: .·1 • • • 40 • ... I ....· • • ...... • I • • • ... 2 I. y = -0.0254x + 5.0876x - 3.6163 • ...... · ...I . t 2 • I • • R = 0.80 .t.• .t. -�• ·• I • • • .t. I ...• . J· t • • 20 • y = -0.0228x2 + 2.6908x - 1 .1646 R2 = 0.82

0 0 5 10 15 20 25 30 35 40 Age (month)

.t. CB • LB •• •• •• • Poly. (CB) --Poly. (LB)

Figure 4.1: Weight-age diagram for crossbred (CB) and local breed (LB) pigs in Ouagadougou, Burkina Faso.

In Figure 4.1 trend lines (best fit observed with polynomial functions) show that crossbred pigs were heavier than local pigs whatever the age. The coefficient of determination (R2) was about 0.80, meaning that a high proportion of the variance in live weight is predictable by age in both breeds according to the equations. The trend lines of growth depicted an almost linear curve with increasing age, particularly for crossbred pigs.

4.3.1.5 Feed use efficiency

Apart from the offer of metabolisable energy (ME), the average daily offer of feed dry matter and nutrients (all expressed per tropical livestock unit, TLU) to local pigs during the sixteen months study period was significantly different between groups within the same pig breed, as well as between comparative groups of both pig breeds (Table 4.13). In local pigs the highest average daily offers were observed in lactating sows,

126 Cha ter 4 whereas in crossbred pigs pregnant and lactating sows were both supplied significantly higher than all other groups. In local pigs, lactating sows received the second highest amounts, whereas adult male and dry sows received the lowest daily offers among adult animals. The average daily offer of DM and nutrients in adult males and dry sows was similar in both pig breeds. Average daily ME offer was higher in crossbred pigs than local pigs for all groups. Apart from suckling piglets the breed had no effect on the average ratio of digestible protein to energy (Table 4.13). The latter was the higher for adult animals as compared to young and suckling pigs in both breeds. The live weight was significantly affected by pig breed and group, which naturally affected the nutrient and ME requirements for maintenance as depicted in Table 4.14. ME requirements for growth were affected by breed only for suckling piglets, while DP requirement for growth was affected by breed for pregnant, young and suckling pigs, with highest values observed in crossbred pigs (Table 4.14 ).

Overall the breed had a significant effect on the coverage of ME and DP requirements (p � 0.001). Thus between both breeds, and apart from dry sows and adult male pigs, the coverage of ME and DP requirements was significantly different between groups (p � 0.01) (Figure 4.2; Figure 4.3). However, the group effect on the coverage of ME and DP requirements varied considerably. Thus in crossbred pigs, the coverage of ME requirements was significantly different (p � 0.05) between dry sows and the other groups except for male pigs; between lactating sows, male and young pigs, between adult male pigs, pregnant sows and between pregnant sows and young pigs. Moreover, the coverage of DP requirements almost followed the same trends as the coverage of ME requirements although there was no significant difference between lactating sows and young pigs and between adult male and young pigs. In local pigs, the coverage of ME and DP requirements was significantly different (p � 0.05) between lactating sows and all other groups. In addition, the coverage of DP requirements was also significantly different between pregnant sows and young pigs.

In local pigs, the proportion of animals experiencing a severe ME deficit was slightly above half of the cases whereas adequate and moderate deficit and supply, respectively, concerned one third of the cases. In crossbred pigs the proportion of substantial ME supply and severe ME deficit situations were highest with 39% and 48% of the cases for both supply levels respectively (Table 4.15).The adequacy of coverage of DP requirements was characterised by a high proportion of severe (and substantial) DP deficit (supply) in local pigs with a share of about two fifths of the cases for each of

127 Cha ter 4 the two supply levels (Table 4.16). Yet, the share of substantial DP supply concerned about half of the cases in crossbred pigs (81.7% of adult male pigs and 91.4% lactating sows).

0 'o w * 2 0 -, 8 � cV, * 0 Q) 7 E 0 � ·s O" � .. 6 * w 2 -, 5 0 � * Q) :Y. co 4 0 -=-c * cV, Q) E 3 � ·s O" � w 2 2 0 Q) 0) 1 co 0 Q) u 0 LB-Dry LB-Lact LB-Preg LB-Male LB- CB-Preg CB-Lact CB-Dry CB-Male CB- Young Young Group

Figure 4.2: Coverage of metabolisable energy (ME) requirements (intake (MJ ME d-1): requirements (MJ ME d-1 )) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LB: local breed pig, CB: crossbred pig, Preg: pregnant sow, Lact: lactating sow, Dry: dry sow, Male: adult male pig, Young: young pig.

128 Table 4.13: Intake at the homestead (mean±S.D. for the 16 month study period) of feed dry matter, digestible protein, phosphorus and metabolisable energy per tropical livestock unit (TLU) of different groups of two pig breed types in Ouagadougou (October 2014 - February 2016). Breed type Group n Live weight Dry matter Digestible protein Phosphorus Metabolisable DP (g) to ME (kg) (kg) (g) (g) energy (MJ) (MJ) ratio(%) 6 a a a a Pregnant sow 39 59 ±18 3.5 ± 2.5 366.1 ± 262.5 22.8 b ±17.5 47.8 ± 35.6 18.6 C ± 9.5 Lactating sow 31 49 a ±15 6.6 b ± 4.5 638.4 b ± 432.2 35.8 bd ±33.4 36.2 ab ± 31.8 29.0 a ±21.8 Local breed Dry sow 54 46 a ±18 4.7 a ± 7.9 453.7 a ± 741.4 22.8 ad ±32.3 47.8 ab ± 35.6 18.2 ac ± 8.6 a a abd ab abc Adut male pig 17 32 C ±10 3.7 ± 3.5 366.4 ± 346.0 19.1 ±17.2 36.3 ± 20.8 17.9 ± 6.7 Young pig (m, f) 418 14 d ± 9 4.6 a ± 5.0 326.2 a ± 385.3 17.7 d ±21.7 39.7 b ± 33.8 16.9 b ±15.0 e ab ac Suckling piglet (m, f) 158 3 ± 4 1.4 C ± 3.5 99.7 C ± 251.6 5.1 C ±13.4 36.0 ± 16.9 18.6 ± 5.2 p� 0.001 0.001 0.001 0.001 n.s. 0.001 Pregnant sow 35 102 b ±34 14.9 b ±14.5 1403.6 b ± 934.7 111.2 b ±92.3 74.3 a ± 47.2 20.1 a ± 3.6 Lactating sow 15 95 ab ±49 11.6 b ± 5.5 1260.3 b ± 597.0 83.0 be ±55.2 57.1 a ± 30.9 27.0 a ±22.6 Crossbred Dry sow 63 72 a ±27 4.9 a ± 7.6 516.0 a ± 845.2 31.0 a ±50.4 71.7 a ± 35.2 17.3 a ± 6.4 a ad a a a N Adut male pig 33 86 C ±30 6.6 ± 9.5 689.9 ±1095.4 32.8 ±40.6 70.6 ± 55.0 21.2 ± 6.7 d b cd Young pig (m, f) 360 23 ±18 16.6 ±24.1 1079.9 ±1497.0 69.7 C ±85.6 133.5 C ±114.6 13.6 C ± 5.2 Suckling piglet (m, f) 157 5 e ± 4 11.8 a ±19.1 633.9 a ± 983.4 46.8 a ±80.0 193.5 b ±122.2 10.6 b ± 2.0 p� 0.001 0.001 0.001 0.001 0.001 0.001 Pregnant sow - 0.001 0.001 0.001 0.001 0.001 n.s. Lactating sow - 0.01 0.01 0.01 0.01 0.05 n.s. p� Dry sow - 0.001 n.s. n.s. n.s. 0.01 n.s. Adut male pig - 0.001 n.s. n.s. n.s. 0.05 n.s. Young pig (m,f) - 0.001 0.001 0.001 0.001 0.001 n.s. Suckling piglet (m, f) - 0.001 0.001 0.001 0.001 0.001 0.001 p: p-value, n: number of measurements, n.s.: not significant, Different superscript letters indicate a significant difference (p�0.05) between groups of the same pig breed type(in the same column) as determined by Kruskal-Wallis test, m: male, f: female. Table 4.14: Energy and digestible protein requirements (mean±S.D.) for maintenance and growth in groups of two pig breed types in Ouagadougou (October 2014 - February 2016). Performance Energy requirement (MJ ME Animar1 Protein requirement (g DP Animar1 d-1) d-1) Breed type Group Growth Maintenance Growth Maintenance Growth (g d-1) Pregnant sow 15 a ±151 9.3 e ±2.2 3.5 e ±3.7 55.2 e ±13.2 52.86 ±40.8 Lactating sow -80 b ±207 8.1 8 ±1.9 0.8 8 ±2.4 48.1 8 ±11.2 36.2 ab ±47.6 Local breed Dry sow 40 8 ±126 7.6 8 ±2.2 0.9 8 ±2.1 45.4 8 ±13.3 38.1 ab ±31.7 Adult male pig 100 acd ± 80 5.8 b ±1.5 1.7 be ±2.1 34.6 b ± 8.9 34.0 ab ±22.3 Young pig (m,f) 70 ee ± 69 3.0 e ±1.4 1.1 b ±1.4 18.0 e ± 8.6 23.8 8 ±17.1 Suckling piglet (m ,f) 60 ae ± 36 1.1 d ±0.7 0.4 d ±0.8 6.3 d ± 4.1 19.2 8 ± 9.7 p� 0.001 0.001 0.001 0.001 0.01 Pregnant sow 240 e±244 14.0 b ±3.6 4.5 b ±5.5 82.9 b ±21.2 101.6 b ±52.8 Lactating sow -170 b ±222 13.0 abd ±5.1 0.3 8 ±0.7 77.4 abd ±30 .4 16.6 ad ±13.8 8 8 8 8 ed (.,,) Crossbred Dry sow 40 ±153 10.7 ±3.0 0.9 ±1.8 63.6 ±17.9 36.1 a ±23.3 0 Adult male pig 190 e ±140 12.3 d ±3.2 2.6 be ±3.6 72.9 d ±18.9 62.0 8 ±39.5 Young pig (m,f) 110 8 ±116 4.4 e ±2.6 1.7 e ±2.7 26.1 e ±15.4 38.5 c ±33.4 8 e 8 ed Suckling piglet (m ,f) 90 ± 62 1.4 ±0.8 0.9 ±1.5 8.3 C ± 4.5 28.7 ±18.3 p� 0.001 0.001 0.001 0.001 0.001 Pregnant sow n.s. 0.001 n.s. 0.001 0.01 p� Lactating sow n.s. 0.01 n.s. 0.01 n.s. Dry sow n.s. 0.001 n.s. 0.001 n.s. Adult male pig n.s. 0.001 n.s. 0.001 n.s. Young pig (m,f) 0.05 0.001 n.s. 0.001 0.001 Suckling piglet (m ,f) 0.01 0.001 0.01 0.001 0.05 p: p-value, n.s.: not significant, Different superscript letters indicate a significant difference (p�0.05) between groups of the same pig breed type (in the same column) as determined by Kruskal-Wallis test, m: male, f: female. Cha ter 4 Table 4.15: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Breed type Group n Ade. Mod. Mod. Sev. Sub. Total % supply deficit supply deficit supply Local breed Pregnant n 8 8 3 16 4 39 sow % 20.5 a 20.5 a 7.7 a 41.0 a 10.3 a 100.0 Lactating n 4 2 3 7 15 31 sow % 12.9 a 6.5 a 9.7 a 22.6 a 48.4 b 100.0 Dry n 4 3 6 30 11 54 sow % 7.4 b 5.6 b 11 .1 ab 55.6 ab 20.4 a 100.0 Adult male n 2 3 3 8 1 17 pig % 11.8 a 17.6 a 17.6 a 47.1 a 5.9 a 100.0 Young n 60 52 59 215 32 418 pig (m,f) % 14.4 a 12.4 a 14.1 a 51.4 a 7.7 b 100.0 Total local n 78 68 74 276 63 559 breed % 14.0 C 12.2 b 13.2 C 49.4 a 11.3 d 100.0 Crossbred Pregnant n 0 0 1 3 31 35 sow % o.o ab o.o ab 2.9 ab 8.6 a 88.6 b 100.0 Lactating n 1 0 0 1 13 15 sow % 6.7 ab o.o ab o.o ab 6.7 a 86.7 b 100.0 Dry n 4 2 0 38 19 63 ab a c a be sow % 6.3 3.2 o.o 60.3 30.2 100.0 Adult male n 0 0 1 18 14 33 pig % o.o a o.o a 3.0 a 54.5 a 42.4 a 100.0 Young n 19 2 32 139 168 360 pig (m,f) % 5.3 ab 0.6 a 8.9 b 38.6 a 46.7 a 100.0 Total n 24 4 34 199 245 506 crossbred % 4.7 C 0.6 b 6.7 C 39.3 a 48.4 d 100.0 Total n 102 72 108 475 308 1065 % 9.6 6.8 10.1 44.6 28.9 100.0 Different superscript letters indicate significant differences (pS:0.05) between levels of coverage of metabolisable energy requirements within the same group (in the same row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

131 Cha ter 4

LU 0

18 .9- * Q) 16 * � ·s 14 ..� � .9 12 ro 0 10

� 8 0 ·s Q) 0

6 0 0 Q) ro 4 Q)

u 2 0

0 LB-Preg LB-Lact LB-Dry LB-Male LB- CB-Preg CB-Lact CB-Dry CB-Male CB- Young Young Group

1 Figure 4.3: Coverage of digestible protein (DP) requirements (intake (g d- ): 1 requirements (g d- )) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LB: local breed pig, CB: crossbred pig, Preg: pregnant sow, Lact: lactating sow, Dry: dry sow, Male: adult male pig, Young: young pig.

132 Cha ter 4 Table 4.16: Adequacy of coverage of DP requirements (intake (g d-1 ): requirements (g d-1 )) through homestead feeding in two pig breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Breed type Group n Ade. Mod. Mod. Sev. Sub. Total % supply deficit supply deficit supply Local breed Pregnant n 1 4 1 25 23 54 sow % 1.9 b 7.4 ab 1.9 ab 46.3 a 42.6 a 100.0 Lactating n 1 0 1 6 23 31 sow % 3.2 8 0.0 8 3.2 8b 19.4 8 74.2 b 100.0 Dry n 3 1 0 6 7 17 sow % 17.6 8 5.9 a 0.0 8 35.3 a 41.2 a 100.0 Adult male n 5 7 2 8 17 39 pig % 128 ab 17.9 b 5.1 ab 20.5 8 43.6 8b 100.0 Young n 52 32 21 171 142 418 pig (m,f) % 12.4 8 7.7 8b 5.0 8b 40.9 8 34.0 b 100,0 Total local n 62 44 25 216 212 559 breed % 11.1 7.9 4.5 38.6 37.9 100.0 Crossbred Pregnant n 2 1 38 22 63 sow % 3.2 8b 1.6 8b 60.3 8 34.9 b 100.0 Lactating n 0 1 1 13 15 sow % 0.0 8b 6.7 b 6.7 8 86.7 b 100.0 Dry n 1 0 18 14 33 sow % 3.0 8 0.0 8 54.5 8 42.4 a 100.0 Adult male n 0 0 3 32 35 pig % 0.0 8b 0.0 8b 8.6 a 91.4 b 100.0 Young n 10 10 139 201 360 pig (m,f) % 28 8 2.8 8 38.6 8 55.8 8 100.0 Total n 13 12 199 282 506 crossbred % 2.6 2.4 39.3 55.7 100.0 Total n 75 44 37 415 494 1065 % 7.0 4.1 3.5 39.0 46.4 100.0 Different superscript letters indicate significant differences (pS:0.05) between levels of coverage of digestible protein requirements within the same group (in the same row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

133 Cha ter 4 4.3.2 Beef cattle production

4.3.2.1 Herd structure and dynamic

All seven monitored beef cattle farms kept local zebus, two of them additionally kept a small number of Sahelian zebu. The two farms keeping Sahelian zebu and three other farms keeping only local zebu provided supplementation to their animals at the homestead. Households supplementing their beef cattle at the homestead had a higher share of male animals than households not supplementing their animals. Throughout the study period the different households depicted the same patterns as far as the herd structure is concerned (Table 4 .17). Households feeding their beef cattle at homestead had a much stronger seasonal production cycle in comparison to those not supplementing their animals. Also regarding animal sales, households supplementing their animals sold more than the ones not supplementing. The latter were therefore less market oriented than the former. For supplementing households 14%, 24% and 62% of the sales happened respectively during RS, EDS, and LOS and the share of adult males in the total sales was 84%. In addition they also acquired more replacement animals (mainly adult male cattle and young) trough purchases, of which 60% and 40% happened during EDS and LOS, respectively; 56% of the purchased animals were adult male cattle. In non supplementing households inflows of animals occurred mainly through births (25% during RS and 75% during LOS) and deaths represented the main cause of outflows. A mortality rate of 4.1 % and 6.7% was registered during the sixteen months study period in non-supplementing and supplementing households, respectively. Animals died on average at about 48±67 and 47±12 months of age in non-supplementing and supplementing households.

134 Cha ter4 Table 4.17: Animal intake and offtake in supplementing and non-supplementing beef cattle farms during the rainy season (RS), the early dry season (EDS), and the late dry season (LOS) (October 2014 - February 2016). Feeding intensity Non supplemented (n) Supplemented (n Season RS EDS LOS Total RS EDS LOS Total Breed type Inflow/ Age category Outflow Dry cow 1 0 0 1 0 1 0 1 Died Adut male cattle 0 0 0 0 0 1 0 1 Suckling calf {m, f) 0 0 1 1 0 0 0 0 Dry cow 0 0 0 0 1 0 3 4 Sold Adut male cattle 0 1 0 1 3 6 14 23 Young cattle {m, :Q 0 0 0 0 0 1 0 1 Dry cow 0 0 0 0 0 1 1 2 Lactating cow 0 0 0 0 0 0 0 0 Local breed Newly Pregnant cow 0 0 0 0 0 1 0 1 registered Adut male cattle 0 0 0 0 0 10 3 13 _. Young cattle (m, f) 0 0 0 0 0 2 4 6 Suckling calf {m, f} 1 0 3 4 0 1 1 2 Dry Cow 0 0 0 0 0 0 1 1 Other Adut male cattle 0 0 1 1 1 1 5 7 outflows Young cattle (m, f) 0 0 1 1 0 0 2 2 Suckling calf (m, f) 0 0 0 0 0 0 1 1 Sold Adut male 0 0 0 0 0 1 1 2 Newly Lactating cow 0 0 0 0 0 0 1 1 Sahelian registered Adut male cattle 0 0 0 0 0 0 6 7 breed Suckling calf (m, f) 0 0 0 0 0 0 1 1 Other Adut male cattle 0 0 0 0 1 0 0 1 outflows n: number of animals, m: male, f: female. Cha ter 4

4.3.2.2 Homestead feeding

In total supplemented beef cattle received 8 different unique feed types and 2 mixtures in the (peri-) urban beef cattle farms of Ouagadougou (Table 4.18). Half of the feed types observed were roughages (protein roughage, straws and hays). The remaining diversity was represented by proteins feeds ( 4) and energy feed ( 1 ). No energy mix or commercial feed was used for beef cattle. Local zebu beef cattle received all observed feed types whereas Sahelian beef zebu received only half of them. In addition all farms provided mineral feed additives such as salt and licking blocks to their animals. An increase of the variability of supplied feed types was observed from RS to LOS for both beef breed types, starting from 4 to 9 and 1 to 5 feed types provided to local zebu beef and Sahelian zebu beef cattle, respectively. Cereal grain residue (protein rich feed types) was the main feed type provided to both beef cattle breeds together with cereal straws (with high fiber content). During RS green grass (43% of the diet) and brewer spent grain (mixed with cereal grain residues or not) were the main feed types provided to local zebu beef cattle whereas Sahelian zebu beef cattle were supplied only with cereal grain residues during the same season. With respect to feeding intensity, homestead offer of feed dry matter, OM, CP, P, NDF, ADF and ME per kg of metabolic body weight (MW) differed significantly across the three different seasons, but only for adult males and young animals (Table A.6, Table A.7, "Appendix"). Adult males were more intensively fed during LOS than during the other two seasons in both breed types, whereas young beef cattle (of local breed) where more intensively feed during RS. During LOS adult male beef cattle received three times as much nutrients and ME as during RS. Apart from P offers there was no significant difference in feeding intensity between both beef cattle breed types. Table A.3 ("Appendix") presents the proximate composition of feedstuffs offered to beef cattle in the (peri-) urban area of Ouagadougou.

136 Chaoter4 Table 4.18: Feed categories and percentages of feedstuffs offered to beef cattle (on dry mater basis) in the (peri-) urban area of Ouagadougou during the rainy season (RS), the early dry season (EDS), and the late dry season (LOS) (October 2014 - February 2016). Local breed Sahelian breed Feed category 1 Feed category 2 Feed type RS EDS LOS RS EDS LOS Energy feed Cereal grain Meal leftover 0.57 8.33 Protein feed Protein feed Brewers grain 17.79 16.92 7.35 Protein feed Protein feed Brewers grain, Cereal grain residues 2.45 Protein feed Cereal byproducts Cereal grain residues 19.55 33.62 11.58 100.00 53.22 43.43 Protein feed Protein feed Cotton seed cake 5.47 22.68 6.74 8.07 Roughage Straws and hays Cereal straw 22.98 23.73 33.35 36.06 Roughage Straws and hays Grass hay 12.11 24.28 6.69 4.10 Roughage Protein roughage Green grass (legume) 42.86 8.91 7.01 Roughage Protein roughage Legume seed hulls 0.34 Roughage Protein roughage Legume seed hulls, Brewers grain 19.80 Cha ter 4 4.3.2.3 Weight development

Given the focus of beef farms on adult males and young animals the comparison between breeds is mainly relevant for those groups. Apart from the RS, adult male beef cattle of supplementing households were heavier than those of non-supplementing households during the other seasons. Moreover in the supplementing households adult male beef cattle of the Sahelian breed were heavier than those of the local zebu during EDS (Table 4.19). In addition the live weight of young beef cattle also depicted the same features observed for adult males. During RS the average daily weight gain (ADG) was higher in adult male beef cattle of Sahelian zebu than in local zebu (supplemented and non-supplemented) (Table 4.20). Supplemented and non supplemented local zebu had similar ADG during RS. The ADG was higher for adult male cattle of Sahelian zebu than of non-supplemented local zebu of the same group during EDS and supplemented adult males of both breed types had a higher ADG than the non-supplemented animals of the comparative group during LOS. The ADG in supplemented young Sahelian beef cattle was higher than of non-supplemented young local beef cattle. In non-supplemented local zebu beef cattle the highest BCSs were observed in pregnant and suckling animals. Apart from lactating and dry cows that had the lowest average BCS, average BCS was slightly above 3 in the other groups of local zebu and Sahelian zebu beef cattle. The BCS in adult male Sahelian beef cattle (3.59±0. 71) was significantly higher than the one of local zebu cattle (3.13±0.63 and 2.92±0.43 for supplemented and non-supplemented animals, respectively) during EDS. Likewise, the BCS of supplemented young local zebu beef cattle was higher than the one of non-supplemented young beef cattle of the same breed during both dry seasons (Table 4.21). For supplemented animals the seasonal effect on live weight and BCS was marked for adult Sahelian type males (Table 4.22). Season highly affected live weight and body condition of non-supplemented local beef zebu, but not of supplemented local cattle. Thus in non-supplemented local zebu beef cattle a seasonal effect on live weight, ADG and BCS was observed in dry and pregnant cows, lactating cows, adult male and young cattle (Table 4.22). In Figure 4.4 the weight at a given age is depicted for cattle breeds utilized as beef cattle in Ouagadougou and for the different feeding intensities (homestead supplemented and non-supplemented). Trend lines (best fit observed with a power function for Sahelian breed cattle and polynomial function for local zebu cattle) showed that supplemented beef cattle belonging to the Sahelian breed are heavier than supplemented and non-supplemented local zebu beef cattle whatever their age.

138 Cha ter 4 However, local zebu beef cattle from the supplemented group were slighter than local breed cattle of the non-supplemented group before 70 months of age. The coefficients of determination (R2) were 0.61, 0.53 and 0.66 for Sahelian zebu, non-supplemented local zebu and supplemented local zebu, respectively, indicating a relatively high variability of live weight at a specific age independent of breed and feeding intensity. Furthermore, apart from the CP requirement for growth, the CP and ME requirements were influenced by the group as well (Table 4.26).

4.3.2.4 Feed use efficiency

Overall the average daily homestead offer of feed dry matter (DM), nutrients and ME in supplemented beef cattle cows, adult males, young animals and suckling animals, all expressed per tropical livestock unit (TLU), was significantly different between groups of the local zebu breed. In this breed the average daily offer of DM, nutrients and ME was the highest for pregnant and lactating cows followed by adult males and young cattle. However, there was no significant difference between daily offer of nutrients and ME between adult males and young cattle of the Sahelian breed. For both age groups, breed had no effect on the supply of DM, nutrients and ME (Table 4.24 ). However, group had a significant effect on the average crude protein to energy ratio in local zebu but not in Sahelian zebu for adult males versus young beef cattle. The average crude protein to energy ratio was higher in adult males of local zebu than in the comparative group of Sahelian zebu. In supplemented lactating and young beef cattle, CP and ME requirements for maintenance were significantly influenced by group and breed such that requirements for maintenance were about one third higher in lactating cows and one third lower in young cattle of Sahelian zebu than in their local counterparts (Table 4.25). ME requirements for growth were only affected by the breed for adult male cattle, while CP requirements for growth were affected by breed for both adult male and young cattle. Here, the highest values were observed in adult male Sahelian zebu and in young local zebu (Table 4.25).

139 Table 4.19: Seasonal variation in live weight (kg) of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Season Group Non supplemented Supplemented Local breed Local breed Sahelian breed p*$'. p**$; p***$; Mean S.D. n Mean S.D. n Mean S.D. n Pregnant cow 186 17 2 170 8 2 n.s. Lactating cow 231 28 4 169 21 3 269 3 2 0.05 n.s. n.s. RS Dry cow 177 23 11 95 134 2 n.s. Adut male cattle 214 84 6 218 97 33 265 61 14 n.s. n.s. n.s. Young cattle (m, f) 88 9 3 104 41 11 n.s. Suckling calf (m, f) 44 9 3 57 1 48 4 2 n.s. Pregnant cow 226 44 33 245 64 7 n.s. Lactating cow 227 25 8 189 1 251 6 2 n.s. EDS Dry cow 197 33 40 204 19 7 n.s. Adut male cattle 219 77 28 266 64 104 319 73 20 0.01 0.001 0.01 Young cattle (m, f) 94 30 20 139 39 41 71 9 2 0.001 n.s. 0.05 Suckling calf (m, f) 36 12 7 18 1 Pregnant cow 250 47 13 190 40 3 0.05 Lactating cow 207 26 6 191 23 2 277 1 n.s. LOS Dry cow 166 24 49 210 16 8 0.001 Adut male cattle 196 82 20 267 72 74 247 79 17 0.01 0.05 n.s. Young cattle (m, f) 81 29 16 122 35 15 0.01 Suckling calf (m, f) 19 6 5 34 13 4 39 1 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, n.s.: not significant, p*: statistics for supplemented cattle of local breed versus not supplemented cattle of local breed, p**: statistics for not supplemented cattle of local breed versus supplemented cattle of Sahelian breed, p***: statistics for cattle of local breed versus Sahelian breed under homestead supplementation, m: male, f: female. Table 4.20: Seasonal variation in average daily weight gain (kg) of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Season Group Non supplemented Supplemented Local breed Local breed Sahelian breed p*:::,;; p**:::,;; p***:::,;; Mean S.D. n Mean S.D. n Mean S.D. n Pregnant cow 0.48 0.11 2 0.27 0.00 2 n.s. Lactating cow 0.31 0.36 4 0.10 0.00 2 -0.11 0.00 2 n.s. n.s. n.s. RS Dry cow 0.46 0.08 10 Adut male cattle 0.21 0.09 4 0.33 0.37 16 0.83 0.30 10 n.s. 0.01 0.01 Young cattle (m, f) 0.38 0.05 3 0.52 0.10 9 0.05 Suckling calf (m, f) 0.28 0.02 2 0.38 1 0.13 0.00 2 n.s. Pregnant cow 0.16 0.29 30 0.23 0.10 3 n.s. Lactating cow -0.09 0.18 5 0.10 1 -0.14 0.05 2 n.s. EDS Dry cow -0.02 0.32 28 Adut male cattle 0.15 0.15 16 0.29 0.33 20 0.48 0.31 12 n.s. 0.01 n.s. ->. Young cattle (m, f) 0.15 0.16 17 0.48 0.19 15 0.21 0.11 2 0.001 n.s. n.s. ->. Suckling calf (m, f) 0.28 0.06 3 Pregnant cow -0.06 0.39 12 0.23 0.11 3 n.s. Lactating cow -0.84 0.42 6 -0.68 0.17 2 n.s. LOS Dry cow -0.13 0.29 48 0.00 1 Adut male cattle -0.07 0.34 15 0.33 0.51 44 0.66 0.35 7 0.01 0.01 n.s. Young cattle (m, f) 0.09 0.13 16 0.61 0.13 7 0.001 Suckling calf (m, f) 0.21 1 0.15 1 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, n.s.: not significant, p*: statistics for supplemented cattle of local breed versus not supplemented cattle of local breed, p**: statistics for not supplemented cattle of local breed versus supplemented cattle of Sahelian breed, p***: statistics for cattle of local breed versus Sahelian breed under homestead supplementation, m: male, f: female. Table 4.20: Seasonal variation in average daily weight gain (kg) of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Season Group Non supplemented Supplemented Local breed Local breed Sahelian breed p*:::,;; p**:::,;; p***:::,;; Mean S.D. n Mean S.D. n Mean S.D. n Pregnant cow 0.48 0.11 2 0.27 0.00 2 n.s. Lactating cow 0.31 0.36 4 0.10 0.00 2 -0.11 0.00 2 n.s. n.s. n.s. RS Dry cow 0.46 0.08 10 Adut male cattle 0.21 0.09 4 0.33 0.37 16 0.83 0.30 10 n.s. 0.01 0.01 Young cattle (m, f) 0.38 0.05 3 0.52 0.10 9 0.05 Suckling calf (m, f) 0.28 0.02 2 0.38 1 0.13 0.00 2 n.s. Pregnant cow 0.16 0.29 30 0.23 0.10 3 n.s. Lactating cow -0.09 0.18 5 0.10 1 -0.14 0.05 2 n.s. EDS Dry cow -0.02 0.32 28 Adut male cattle 0.15 0.15 16 0.29 0.33 20 0.48 0.31 12 n.s. 0.01 n.s. ....>. Young cattle (m, f) 0.15 0.16 17 0.48 0.19 15 0.21 0.11 2 0.001 n.s. n.s. Suckling calf (m, f) 0.28 0.06 3 Pregnant cow -0.06 0.39 12 0.23 0.11 3 n.s. Lactating cow -0.84 0.42 6 -0.68 0.17 2 n.s. LOS Dry cow -0.13 0.29 48 0.00 1 Adut male cattle -0.07 0.34 15 0.33 0.51 44 0.66 0.35 7 0.01 0.01 n.s. Young cattle (m, f) 0.09 0.13 16 0.61 0.13 7 0.001 Suckling calf (m, f) 0.21 1 0.15 1 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, n.s.: not significant, p*: statistics for supplemented cattle of local breed versus not supplemented cattle of local breed, p**: statistics for not supplemented cattle of local breed versus supplemented cattle of Sahelian breed, p***: statistics for cattle of local breed versus Sahelian breed under homestead supplementation, m: male, f: female. Table 4.21: Seasonal variation in body condition score of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - Februar y 2016). Season Group Non Supplemented Supplemented Local breed Local breed Sahelian breed p*:5: p**:5: p***:5: Mean S.D. n Mean S.D. n Mean S.D. n Pregnant cow 2.8 0.4 2 3.8 0.4 2 n.s. Lactating cow 2.8 0.3 4 2.3 0.3 3 2.5 0.0 2 n.s. n.s. n.s. RS Dry cow 2.7 0.5 10 Adut male cattle 3.0 0.4 4 3.0 0.6 23 3.3 0.4 12 n.s. n.s. n.s. Young cattle (m, f) 2.8 0.3 3 2.9 0.2 5 n.s. Suckling calf (m, f) 3.0 0.0 3 3.5 1 3.3 0.4 2 n.s. Pregnant cow 3.1 0.3 33 3.4 0.6 7 n.s. Lactating cow 2.6 0.2 8 2.5 1 2.5 0.0 2 n.s. EDS Dry cow 2.7 0.3 40 2.2 0.6 3 n.s. Adut male cattle 2.9 0.4 33 3.1 0.6 77 3.6 0.7 16 n.s. 0.01 0.01 +so Young cattle (m, f) n.s. n.s. (,.) 2.9 0.4 20 3.2 0.5 37 2.5 0.0 2 0.05 Suckling calf (m, f) 2.9 0.2 7 3.0 1 Pregnant cow 2.9 0.3 13 3.3 0.8 3 n.s. Lactating cow 2.3 0.3 6 2.5 0.0 2 3.0 1 n.s. LOS Dry cow 2.4 0.4 49 1.8 0.4 2 0.05 Adut male cattle 2.7 0.6 19 3.3 0.6 48 3.0 0.7 13 0.01 n.s. n.s. Young cattle (m, f) 2.6 0.3 15 3.2 0.5 11 0.01 Suckling calf (m, f) 2.9 0.3 4 3.3 0.4 2 3.5 1 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, n.s.: not significant, p*: statistics for supplemented cattle of local breed versus not supplemented cattle of local breed, p**: statistics for not supplemented cattle of local breed versus supplemented cattle of Sahelian breed, p***: statistics for cattle of local breed versus Sahelian breed under homestead supplementation, m: male, f: female. Table 4.21: Seasonal variation in body condition score of beef cattle of different physiological status in two cattle breed types in Ouagadougou, Burkina Faso (October 2014 - Februar y 2016). Season Group Non Supplemented Supplemented Local breed Local breed Sahelian breed p*:5: p**:5: p***:5: Mean S.D. n Mean S.D. n Mean S.D. n Pregnant cow 2.8 0.4 2 3.8 0.4 2 n.s. Lactating cow 2.8 0.3 4 2.3 0.3 3 2.5 0.0 2 n.s. n.s. n.s. RS Dry cow 2.7 0.5 10 Adut male cattle 3.0 0.4 4 3.0 0.6 23 3.3 0.4 12 n.s. n.s. n.s. Young cattle (m, f) 2.8 0.3 3 2.9 0.2 5 n.s. Suckling calf (m, f) 3.0 0.0 3 3.5 1 3.3 0.4 2 n.s. Pregnant cow 3.1 0.3 33 3.4 0.6 7 n.s. Lactating cow 2.6 0.2 8 2.5 1 2.5 0.0 2 n.s. EDS Dry cow 2.7 0.3 40 2.2 0.6 3 n.s. Adut male cattle 2.9 0.4 33 3.1 0.6 77 3.6 0.7 16 n.s. 0.01 0.01 Young cattle (m, f) n.s. n.s. 2.9 0.4 20 3.2 0.5 37 2.5 0.0 2 0.05 Suckling calf (m, f) 2.9 0.2 7 3.0 1 Pregnant cow 2.9 0.3 13 3.3 0.8 3 n.s. Lactating cow 2.3 0.3 6 2.5 0.0 2 3.0 1 n.s. LOS Dry cow 2.4 0.4 49 1.8 0.4 2 0.05 Adut male cattle 2.7 0.6 19 3.3 0.6 48 3.0 0.7 13 0.01 n.s. n.s. Young cattle (m, f) 2.6 0.3 15 3.2 0.5 11 0.01 Suckling calf (m, f) 2.9 0.3 4 3.3 0.4 2 3.5 1 n.s. p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, S.D.: Standard Deviation, n: number of observations, n.s.: not significant, p*: statistics for supplemented cattle of local breed versus not supplemented cattle of local breed, p**: statistics for not supplemented cattle of local breed versus supplemented cattle of Sahelian breed, p***: statistics for cattle of local breed versus Sahelian breed under homestead supplementation, m: male, f: female. Cha ter 4 Table 4.22: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BCS) of different groups of supplemented beef cattle kept in Ouagadougou (October 2014 - February 2016). Variable Season p Group Dry Pregnant Lactating Young Suckling Adult cow cows cow cattle calf male (m,f) (m,f) cattle Supplemented local zebu EDS-LDS-RS 0.202 0.094 0.501 0.094 0.219 0.053 LW EDS-LOS 0.587 0.207 1.000 0.170 0.277 0.908 EDS-RS 0.232 <0.05 0.180 0.057 0.317 <0.05 LDS-RS 0.063 0.564 0.564 0.349 0.157 <0.05 EDS-LDS-RS 0.943 0.153 0.116 0.317 0.598 ADG EDS-LOS 0.827 0.221 0.062 0.304 EDS-RS 1.000 1.000 0.590 0.898 LDS-RS 0.554 0.102 0.080 0.317 0.616 EDS-LDS-RS 0.361 0.730 0.607 0.385 0.472 0.152 BCS EDS-LOS 0.361 0.905 1.000 0.806 0.480 0.147 EDS-RS 0.436 0.564 0.180 0.317 0.414 LDS-RS 0.543 0.414 0.234 0.480 0.073 Supplemented Sahelian zebu EDS-LDS-RS 0.165 <0.05 LW EDS-LOS 0.221 <0.01 EDS-RS 0.121 <0.05 LDS-RS 0.221 0.370 EDS-LDS-RS 0.317 0.057 ADG EDS-LOS 0.447 EDS-RS 0.317 <0.05 LDS-RS 0.063 EDS-LDS-RS 0.135 <0.01 BCS EDS-LOS 0.157 <0.01 EDS-RS 1.000 <0.05 LDS-RS 0.157 0.220 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, m: male, f: female.

145 Cha ter 4 Table 4.23: Effects of season on live weight (LW), daily weight change (ADG), and body condition score (BGS) of different groups of non-supplemented local beef cattle kept in Ouagadougou (October 2014 - February 2016). Variable Season p Group Dry Pregnant Lactating Young Suckling Adult cow cows cow cattle calf male (m,f) (m,f) cattle Non supplemented local zebu EDS-LDS-RS <0.001 <0.05 0.322 0.605 <0.05 0.581 LW EDS-LOS <0.001 0.059 0.156 0.464 <0.05 0.289 EDS-RS <0.05 0.088 0.865 0.584 0.253 0.714 LDS-RS 0.181 0.126 0.286 0.180 <0.05 0.903 EDS-LDS-RS <0.001 0.054 <0.01 <0.05 0.343 <0.05 ADG EDS-LOS 0.52 0.077 <0.01 0.271 0.180 <0.05 EDS-RS <0.001 0.119 <0.05 <0.05 1.000 0.508 LDS-RS <0.001 0.068 <0.05 <0.05 0.221 0.057 EDS-LDS-RS <0.001 <0.05 0.061 <0.05 0.666 0.199 BCS EDS-LOS <0.001 <0.05 0.052 <0.05 0.673 0.106 EDS-RS 0.936 0.127 0.407 0.877 0.513 0.623 LDS-RS <0.05 0.702 0.053 0.190 0.386 0.240 p: p-value, RS: Rainy season, EDS: Early dry season, LOS: Late dry season, m: male, f: female.

The breed had a significant effect on the coverage of ME and GP requirements though homestead feeding in households supplementing their animals. The coverage of ME and GP requirements was significantly different for adult male and young cattle in both breeds with the highest coverages of ME and GP requirement observed in local zebu. The group effect on the coverage of ME and GP requirements varied considerably. In local zebu there was no significant difference between groups for the coverage of ME requirements, and the coverage of GP requirements was significantly different (p � 0.05) between dry and pregnant, male and young animals. In Sahelian zebu the coverage of ME requirement was similar amongst the different groups. The coverage of GP requirement was significantly different (p � 0.05) between young and male animals.

The coverage of ME and GP requirements through intake at the homestead depicted a wide range of variation from severe ME (or GP) deficit to substantial ME (or GP) supply (Figure 4.5; Figure 4.6). In supplemented local zebu the proportion of severe ME deficit cases was highest (with three quarters of the 43% overall cases observed in adult males), followed by substantial ME supply (three fifths of the 30% overall cases observed in adult males and one third in young cattle) (Table 4.27). Only one quarter of

146 Cha ter 4 the cases were shared amongst adequate ME supply, moderate ME deficit, and moderate ME supply. Almost half of the cattle experienced severe ME deficit and one third of cases were shared amongst adequate ME supply, moderate ME deficit, and moderate ME supply (though mainly male and young animals were found in these categories for Sahelian zebu). Almost the same trends that were observed for the coverage of ME requirements were observed for the coverage of CP requirements in both breeds with severe deficit and substantial supply accounting for more than three quarters of the cases (Table 4.28).

450 •• 400 • ••.... • • ...• •• • 350 ...·- ...... •• ...... ··· . •...... • ••. • • • ... • ••• 1 • :. I • • • 300 .... • •• ••••• ...... • ...... • -... • •••••••••••• .. ..--...... --- ...... ,,· �250 ...... �· •• •. ••• ...... • ...... ,. .• �t-··················"""'--- . • ...... ······ . -� .. 'iii 200 • • • • . J�····- . .: •• ·: • s: •. • ...... • • 1 •• . . . . I •• _.. ..e I •' .• 1 • • • • ••...... -· 1 • • • • • y = •30.641x D,6108 150 • • • •• • • • •• ·• a •·•····• • • . R2 = 0.61 •• •• • •• 11•t•••"· a•I•• • I• • •• • •• • • • • ••a.·i"··· ...... 100 . y = -0.1208x2 + 11.458x + 8.8586 • • 1...� : 2 •�f • R = 0.53 50 ... y = -0.0594x2 + 7.0341x + 18.882 , R2 = 0.66 0 0 10 20 30 40 50 60 70 80 Age (month)

• LBN • LBS • SBS ········· Log. (LBN) ••••••• Poly. (LBS) --Pot.(SBS)

Figure 4.4: Weight-age diagram for non-supplemented local zebu breed (LBN), supplemented local zebu breed (LBS) and supplemented Sahelian zebu breed (SBS), in Ouagadougou, Burkina Faso.

147 Cha ter 4 Table 4.24: Intake at the homestead (mean±S.D. for the 16 month study period) of feed dry matter, crude protein, phosphorus and metabolisable energy by cows, males and young tropical livestock units (TLU) of two different beef zebu cattle breed types in Ouagadougou (October 2014 - February 2016). Breed type Group n Live weight Dry matter Crude protein Phosphorus Metabolisable CP to ME (kg) (kg) (g) (g) energy (MJ) ratio(%) Pregnant cow 12 2196 ±60 4.9 6 ±3.7 521.5 6 ±427.6 23.5 6 ±22.3 44.0 6 ±33.4 12.4 ab ± 4.9 Lactating cow 6 180 b ±20 4.7 bed ±6.3 593.1 bed ±625.3 30.0 bed ±31.0 42.3 bed ±49.5 16.6 abe ± 7.3 Local breed Dry cow 17 194 a ±53 1.6 a ±2.6 201.6 a ±336.4 7.7 a ±14.8 15.1 a ±24.5 13.2 abe ± 2.0 c ed cd ed ac Adult male cattle 211 259 ±74 4.1 ±4.7 474.4 ±562.5 23.0 cd ±26.7 37.5 ±42.8 13.2 ± 4.7 d b b be Young cattle(m, f) 67 129 ±40 4.2 b ±4.6 462.2 ±493.8 25.1 b ±29.2 38.7 ±43.4 12.4 ± 3.2 ae ae ac ac Suckling calf (m, f) 6 35 e ±16 1.2 ±3.0 268.2 ±657.0 14 .4 ±35.4 12.8 ±31.4 29.9 C ± 0.0 p� 0.001 0.001 0.001 0.001 0.001 0.01 Pregnant cow Lactating cow 5 263 a ±12 3.3 a ±4.8 254.9 a ±303.0 11.9 a ±18.5 27.2 a ±38.1 14.4 a ±10.3 Sahelian Dry cow a a a a a a +>- breed Adult male cattle 51 280 ±78 3.9 ±3.5 387.2 ±356.2 27.8 ±23.8 38.4 ±33.0 10.1 ± 3.0 Young cattle(m, f) 2 71 b ± 9 2.7 a ±3.0 160.2 a ± 43.2 3.9 a ± 4.9 19.9 a ±17.9 15.1 a ±15.8 Suckling calf (m, f) 3 49 b ± 6 p� 0.001 n.s. n.s. 0.05 n.s. n.s. Pregnant cow Lactating cow 0.01 n.s. n.s. n.s. n.s. n.s. p� Dry cow Adult male cattle n.s. n.s. n.s. 0.01 n.s. 0.01 Young cattle(m, f) 0.05 n.s. n.s. n.s. n.s. n.s. Suckling calf (m, f) n.s. p: p-value, n: number of animals per group, n.s.: not significant. Different superscript letters indicate a significant difference (p�0.05) between groups of the same breed type (in the same column) as determined by Kruskal-Wallis test, m: male, f: female. Table 4.25: Energy and protein requirements (mean±S.D.) for maintenance and growth of different groups of beef cattle of two different breed types receiving homestead supplementation in Ouagadougou (October 2014 - February 2016). Performance Energy requirement Protein requirement 1 1 1 1 (MJ ME animar d· ) (g CP animar d· )) Breed type Group Growth (g d"1) Maintenance Growth Maintenance Growth Pregnant cow 2436 ± 79 27.2 6 ±5.5 5.5 be ± 4.6 209.4 6 ±42.3 92.2 6± 29.9 Lactating cow -213 8 ±437 23.5 b ±2.0 1.7 b ± 1.8 181.4 b ±15.5 37.7 a± 0.0 Local breed Dry cow 0 25.7 a ±2.1 0.0 198.4 a ±16.5 b b e Adult male cattle 322 ±440 30.9 C ±6.4 5.0 ± 9.4 238.4 C ±49.1 189.0 ±110.5 e be Young cattle (m, f) 523 ±161 18.4 d ±3.9 8.2 c± 9.7 142.1 d ±29.8 1 98. 6 ± 61 .1 Suckling calf (m, f) 268 b ±163 6.8 e ±2.4 3.0 be ± 5.3 52.3 e ±18.6 101.9 abc± 62.2 p� 0.01 0.001 0.01 0.001 0.01 Pregnant cow Lactating cow -125 8 ± 37 31.4 8 ±1.1 0.0 241.7 a± 8.3 Sahelian Dry cow b ...>. breed Adult male cattle 646 ±339 32.6 a±7.0 12.5 a ±14.0 251.4 a ±53.8 245.5 a ±128.9 b b b b b CD Young cattle (m, f) 208 ±108 11.8 a ±1 .1 7.1 a ± 3.7 90.7 a ± 8.4 79.0 ± 40.9 b b b Suckling calf (m, f) 131 C ± 0 8.3 ±0.9 3.0 a± 2.6 64.1 ± 6.6 50.1 ± 0.0 p� 0.001 0.01 n.s. 0.01 0.05 Pregnant cow p� Lactating cow n.s. 0.01 n.s. 0.01 Dry cow Adult male cattle 0.01 n.s. 0.001 n.s. 0.05 Young cattle (m, f) 0.05 0.05 n.s. 0.05 0.05 Suckling calf (m, f) n.s. n.s. n.s. n.s. n.s. p: p-value, n .s.: not significant. Different superscript letters indicate significant differences (p�O .05) between groups of the same breed type (in the same column) as determined by Kruskal-Wallis, m: male, f: female. Table 4.26: Energy and protein requirements (mean±S.D.) for maintenance and growth of different groups of non supplemented local zebu beef cattle in Ouagadougou (October 2014 - February 2016). Breed type Group Live weight Growth performance Energy requirement Protein requirement (kg) (g d"1) (MJ ME animal"1 d·1) (g CP animal"1 d·1) Maintenance Growth Maintenance Growth Pregnant cow 231 b ±46 114 be ±333 28.3 e ±4.2 5.9 be ±8.3 218.2 e ±32.5 112.8 ab ± 97.0 Lactating cow 221 b ±27 -283 a ±589 27.5 be ±2.5 2.8 a ±6.9 212.1 be ±19.2 143.3 ab ±109.4 Dry cow 179 a ±31 -27 ae ±338 23.4 a ±3.0 3.8 abe ±6.3 180.9 a ±23.5 116.6 a ± 70.9 Local breed Adult male cattle 210 a ±78 62 bed ±266 26.1 ab ±7.4 3.1 ab ±4.9 201.2 ab ±56.8 85.1 b ± 54.6 Young cattle (m,f) 88 e ±28 143 bd ±160 13.7 d ±3.3 4.9 ed ±4.9 105.7 d ±25.7 68.4 be ± 52.5 Suckling calf (m,f) 32 d ±14 268 e ± 46 6.3 e ±2.0 3.6 abd ±4.7 48.7 e ±15.8 1 01 .6 ae ± 17.7 p� 0.001 0.01 0.001 0.001 0.001 n.s. p: p-value, n.s.: not significant, Different superscript letters indicate significant differences (p�0.05) between groups (in the same column) as determined by Kruskal-Wallis, m: male, f: female.

(J1 0 :§' 0 w ,� 4 6 ctJ) Q) E 0 Q) L. ·sc:r � 3 .. 0 'o w ,� 6 Q) .:::t:.

.8C 2

0 Q) 0) ro � u0 Q,-'----,-----,----r-----,-----,----r-----,-----,--___.B LZ-Preg LZ-Lact LZ-Dry LZ-Male LZ-Young SZ-Lact SZ-Male SZ-Young Group

151 Cha ter 4 Table 4.27: Adequacy of coverage of metabolisable energy requirements (intake (MJ ME d-1): requirements (MJ ME d-1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Beef cattle Group n Ade. Mod. Mod. Sev. Sub. Total breed type % supply deficit Supply deficit supply Pregnant n 0 1 3 1 7 12 cows % o.o ab 8.3 b 25.0 b 8.3 a 58.3 b 100.0 Lactating n 0 1 0 3 2 6 cows % 0.0 8 16.7 8 0.0 8 50.0 8 33.3 8 100.0 Dry n 0 0 2 13 2 17 Local breed cows % o.o ab o.o ab 11.8 ab 76.5 a 11.8 b 100.0 Adult male n 17 7 29 101 57 211 C abc ab 8 be cattle % 8.1 3.3 13.7 47.9 27.0 100.0 Young n 17 2 5 15 28 67 c abc 8 8 b cattle (m,f) % 25.4 3.0 7.5 22.4 41.8 100.0 Total n 34 11 39 133 96 313 local breed % 10.9 b 3.5 b 12.5 ab 42.5 ab 30.7 a 100.0 Pregnant n cows % Lactating n 0 0 1 3 1 5 cows % 0.0 8 0.0 8 20.0 8 60.0 8 20.0 8 100.0 Sahelian Dry n breed cows % Adult male n 9 4 8 22 8 51 8 8 8 a a cattle % 17.6 7.8 15.7 43.1 15.7 100.0 Young n 0 1 0 1 0 2 cattle (m,f) % 0.0 8 50.0 8 0.0 8 50.0 8 0.0 8 100.0 Total n 9 5 9 26 9 58 b b b ab a Sahelian % 15.5 8.6 15.5 44.8 15.5 100.0 breed n 43 16 48 159 105 371 Total % 11.6 4.3 12.9 42.9 28.3 100.0 Different superscript letters indicate significant differences (p�0.05) between levels of coverage of metabolisable energy requirements within the same group (in the same row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

152 ::=:: 8 0 .._"O 9 c(J) Q) 7 0 E '-Q) ·s 0 O" 6 ..� :§ 0 9 Q) ::s:. 5

2C '=- (J) c 4 Q) E '-Q) ·s '-g 3 Q_ 0 0 Q) 2 Ol (U >

0 LZ-Preg LZ-Lact LZ-Dry LZ-Male LZ-Young SZ-Lact SZ-Male SZ-Young Group

Figure 4.6: Coverage of crude protein (CP) requirements (intake (g d-1):requirements (g d"1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016), LZ: Local zebu, SZ: Sahelian zebu, Preg: pregnant cow, Lact: Lactating cow, Dry: Dry cow, Male: Adult male cattle, Young: Young cattle.

153 Cha ter 4 Table 4.28: Adequacy of coverage of CP requirements (intake (g d-1): requirements (g d-1)) through homestead feeding in two different beef zebu cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016). Beef cattle Group n Ade. Mod. Mod. Sev. Sub. Total breed type % supply deficit supply deficit supply Pregnant n 1 0 1 1 9 12 cows % 8.3 ab o.o ab 8.3 ab 8.3 a 75.0 b 100.0 Lactating n 1 0 0 2 3 6 cows % 16.7 8 0.0 8 0.0 8 33.3 8 50.0 8 100.0 Dry n 0 0 0 13 4 17 Local breed cows % 0.0 8b 0.0 8b o.o ab 76.S a 23.S b 100.0 Adult male n 12 7 9 96 87 211 cattle % 5.7 8be 3.3 b 4.3 8e 45.5 8b 41.2 e 100.0 Young n 2 0 5 16 44 67 cattle ( m, f) % 3.0 8b 0.0 8b ?.S b 23.9 8 65.? b 100.0 Total n 16 7 15 128 147 313 local breed % 5.1 e 2.2 abe 4.8 be 40.9 8b 47.0 8 100.0 Pregnant n cows % Lactating n 1 0 0 2 2 5 8 8 8 8 8 cows % 20.0 0.0 0.0 40.0 40.0 100.0 Sahelian Dry n breed cows % Adult male n 10 2 6 20 13 51 cattle % 19.6 8 3.9 8 11.8 8 39.2 a 25.S a 100.0 Young n 0 0 0 2 0 2 cattle ( m, f) % 0.0 8 0.0 8 0.0 8 100.0 8 0.0 8 100.0 Total n 11 2 6 24 15 58 Sahelian % 19.0 e 3.4 8be 10.3 be 41.4 ab 25.9 8 100.0 breed n 27 9 21 152 162 371 Total % 7.3 2.4 5.7 41.0 43.7 100.0 Different superscript letters indicate significant differences (p:5:0.05) between levels of coverage of crude protein requirements within the same group (in the same row) as assessed by the Chi square test, m: male, f: female, Ade.: Adequate, Mod.: Moderate, Sev.: Severe, Sub.: Substantial.

154 4.4 Discussion

In this 16 month duration study the effect of the feeding management, breed, animal age, sex and physiological status on production performance of pigs and beef cattle was assessed. Observed pigs belonged to two breeds, crossbred and local pigs. Likewise, beef cattle belonged to local zebu and Sahelian zebu and were kept under different feeding intensities.

4.4.1 Pig production

4.4.1.1 Production management

Both local and crossbred pig keeping households keep manly adult females and their offspring. The absence of adult male pigs questions the mating method used by farmers keeping both pig breed types. For local pigs uncontrolled mating certainly happens when animals are out of the farms and free ranging; scavenging was reported for EDS in several cases for pig feeding on harvested fields when available. In crossbred pigs where animals were enclosed year round, mating males were borrowed from or to other pig farms, a practice already reported for other West Africa pig farming systems (Nonfon, 2005). This practice, which has the advantage of not spending resources on keeping adult male pigs, is also risky as far as diseases transmission is concerned. Births mainly happened during LOS and RS in local pigs and crossbred pigs, respectively. For local pigs this may have been influenced by the natural increase in fertility during colder temperatures and fodder abundance during RS and EDS (post harvest season). Besides a year round production cycle, for local pigs there was an emphasis of sales during EDS. Thus, animals were sold prior to fodder scarcity. Rising prices of feed or feed shortage can be expected with increasing time since the last harvest at the end of rainy season or at the beginning of the EDS. For crossbred pigs, the production cycle was less perceptible and delayed, so that more animals were sold during LOS. A low fertility was observed both in crossbred and local reproductive sows, with an average litter size of less than 6 living piglets. This result was similar to the litter size of 6.5 piglets in local pigs and lower than the 8-11 piglets observed in crossbred and exotic breeds, respectively, in Bobo-Dioulasso, Burkina Faso (Kiendrebeogo et al., 2014) as well as in Cameroon (Fualefac et al., 2014; Kouamo et al., 2015). It remains unknown whether measures or actions to increase litter size (and also to reduce early mortality) can be of major impact for pig keeping households which keep both reproductive animals and fattening or growing animals since adequate feeding and health care 155 Cha ter 4 should simultaneously be implemented and a good market guarantied. It is well established that when a farmer is a seller of weaned piglets the litter size and low early mortality are central production parameters that have to be optimised throughout the production cycle. Though highly variable the average inter-farrowing interval of reproductive sows, which averaged 206±42 days for both breeds in this study, is similar to the 180±25 days observed in Cameroon (Kouamo et al., 2015). Reasons for this rather short inter-farrowing interval might a short suckling period and a high early mortality. Early weaning is common practice in intensive pig production systems of West Africa (Kouamo et al., 2015; Okello et al., 2015). Dry sows depicted on average very poor body condition (BSC less than 2), certainly as a consequence of severe ME and nutrient deficits triggered by improper feeding in view of the sows' pronounced nutrient turnover during suckling periods (Thaker and Bilkei, 2005). Though prolonged suckling is positively correlated with piglet growth, when coupled with insufficient ME and nutrient requirement coverage both negatively affect the following farrows and impair the sows' ability to properly nourish the piglets later on (Thaker and Bilkei, 2005). Overall only 3.3 and 2.8 piglets per year and adult sow were registered for crossbred and local pigs, respectively, as a consequence of keeping unproductive adult females coupled with a disproportionally high death of suckling local piglets. A mortality rate of 61% and 14% in local pigs and crossbred pigs was observed during the study period, thereby excluding postnatal deaths; suckling and young pigs were the most affected groups. The same trends have been reported for local and (exotic) crossbred pigs in Bobo-Dioulasso with mortality ranging from 0.8 to 27% in piglets of different breeds kept in intensive and (semi-) extensive pig farming systems, respectively (Kiendrebeogo et al., 2014), and with almost no deaths in adults animals. In Ndjamena, Chad, the mortality of local pigs was around 40% and mainly affected piglets and young pigs aged 5.7±4. 6 months (Mopate Logtene et al., 2014 ). The causes of the high mortality rate of young pigs in (semi-) extensive pig farming systems in sub-Saharan Africa might be the lack of sound exploitation of the available feeding potential, a high pathogen pressure during certain periods of the year, poor maternal qualities of local reproductive sows as well as poor farming hygiene and poor animal welfare. Though the most frequently reported reasons for deaths were diseases, no further investigation was carried out in the context of this study as far as reasons for death cases are concerned. Therefore the evidence of the spread of epidemic disease(s) highly affecting young local pigs has not been made. However, previous studies carried out in sub-Saharan Africa have identified poor health conditions characterised by digestive, respiratory and skin problems as

156 major health constraints of (peri-) urban pig farms in Ndjamena, Chad (Mopate Logtene et al., 2014), Dschang, Cameroon (Fualefac et al., 2014 ), Bobo-Dioulasso, Burkina Faso (Kiendrebeogo et al., 2014 ), and Mecha District, Ethiopia (Mekuriaw and Asmare, 2014). It seems that farmers keeping local pigs - and allowing scavenging - provide less health care to their animals as compared to farmers keeping crossbreds - and keeping them constantly at the homestead (Chapter 2). For both breeds, animals were mainly sold at the average age of biological maturation (10 to 11 months) or earlier. A similar selling age was reported by Kiendrebeogo et al. (2012a) from other cities of Burkina Faso. It may be due to market related factors such as buyers' preference for fatless young pigs (Mopate Logtene and Kabore-Zoungrana, 2013), economic reasons (e.g., farmers needing cash to afford some commodities including animal feed), increasing fodder costs since the last post-harvest season or a risk management strategy to avoid animal losses due to high mortality rates. Therefore it remains questionable whether further fattening could be a strategy to improve the productivity and the profitability of pig farming given the lower mortality rate of adult animals and their similar growth performance (linear growth curve) as compared to young animals.

4.4.1.2 Feeding management and intensity

Most feedstuffs used were protein-rich industrial by-products and only crossbred pigs were partly supplied with commercial (or compound) pig feeds, but also with roughage for adult animals during LOS, which led to a drop in the diets' nutrient concentration. Apart from commercial pig feed the other feed types have been commonly found in farming systems of West Africa cities such as Bobo-Dioulasso, Burkina Faso (Kiendrebeogo et al., 2014) and Ndjamena, Chad (Mopate Logtene et al., 2014 ). The feeding intensity was always significantly higher in crossbred pigs than in local pigs for reproductive (pregnant and lactating) sows and young and suckling animals. Crossbred pig farmers therefore more intensively feed those groups and it was common that DM supplies of multiple times the biological intake capacity were observed in reproductive animals. Whether this was related to important feed wastage due to improper housing or to true feed refusals was impossible to investigate. However it must be pointed out that requirements of ME and nutrients for lactation and pregnancy were not taken into consideration for the calculation of the total requirements. Furthermore, as observed in Iberian local pigs, scavenging can require up to half of the digested energy in extensive systems due to walking activity and to poor hygienic conditions (Rodrfguez-Estevez et

157 Cha ter 4 al., 2010). There was a high seasonal, breed and group variability of feeding intensities resulting from the variation in feed availability and farm feeding management. Likewise, previous studies on traditional and non-commercial smallholder livestock production systems in sub-Saharan Africa have reported high seasonal effects over feed offers resulting from the variation of feed availability rather than animals' actual nutritional requirements (Chapter 3) (Rodrfguez-Estevez et al., 2010). This implies that malnutrition during off-season (dry season) is a core problem in those production systems that shou Id be mitigated. In (peri-) urban local and cross bred pig farming of Ouagadougou, ME supply was frequently the more constraining factor for optimal growth in comparison to supply of DP. Therefore strategies toward shifting or reducing part of the high amount of protein feed offer and increasing energy feeds in the diet can improve the overall nutrition of pigs. Only then the unexploited growth potential can be reached. In addition, farmers should pay attention to feed containing anti-nutritional factors such as rice husks that should be removed from the diets.

4.4.1.3 Weight development

Most crossbred pig groups were twice as heavy as their local pig counterparts and the difference in ADG between both breeds was also within the same range. Likewise, mother sows in crossbred pigs were losing weight about twice as fast during lactation even though they received about 40% more DM per kg MW. Live weight lost due to lactation is positively correlated to piglet growth (Tantasuparuk et al., 2001; Thaker and Bilkei, 2005) and therefore indicates better maternal properties of the crossbred pigs. The overall productivity of pig production in and around Ouagadougou can be considered as low, with average weight gains of young animals being 110±116 and 70±69 g d-1 in crossbred and local pigs, respectively. For local pigs the ADG obtained in this study is higher than the 40 g d-1 reported for local young pigs (aged less than 12 month) in Benin (Nanton, 2005) and lower than the 87 to 91 g d-1 and 150 g d-1 observed in Ndjamena, Chad (Mopate Logtene et al., 2011) and Districts of Kenya (Carter et al., 2013), respectively. For crossbred pigs lower ADGs of 71 and 94 g d-1 were observed in Nigeria for male and female F1 crosses, respectively (Okeudo et al., 2007). In Uganda, ADGs of 700 g d-1 were reached with commercial and silage based feed supplied to young pigs aged between 199-209 days (Carter et al., 2017). The growth curves developed in the present study were showing relatively high regression coefficients (R2) and an almost linear trend line for both breeds, which may result from underfeeding leading to continual growth compensation towards reaching the

158 physiological appropriate weight. The poor average BCS observed in most groups reinforced that interpretation. The phenomenon was even more pronounced for crossbred pigs. However it should be emphasized that prolonged periods of undernutrition have more deleterious effects on sows' overall or lifetime productivity (Thaker and Bilkei, 2005) than on growing pigs which can easily compensate periods of undernutrition once they are supplied adequately afterwards even if they don't reach their maximum weight (Daza et al., 2003). Therefore (peri-) urban pig farmers should improve the nutrition of their sows.

4.4.1.4 Feed use efficiency of pig farms

The overall feed use efficiency (feed ME or nutrients, respectively, required per unit of output) can be considered as poor since for ME and DP supply levels the proportion of animals being adequately supplied was low. The situation was even worse for local pigs as far ME offer is concerned. For the sake of comparison, studies assessing the efficiency of West African pig farms are rare. However it is likely that the observations made in this study also hold true in other locations of the sub-region despite variability in production systems.

4.4.2 Beef cattle production

4.4.2.1 Production management

Unlike non-supplementing households which kept reproductive females, supplementing beef cattle households had a high share of adult male animals. Probably those animals were the male offspring from dairy cattle farms and non-supplementing beef cattle farms. The kept breeds were identical for both types of beef cattle farms, except that Sahelian zebu were mainly kept by supplementing households. Supplementing households had a much stronger seasonal production cycle in comparison to non supplementing households. Households supplementing their animals can therefore be classified as commercial farms that have adopted cattle fattening to increase their income, while non-supplementing beef cattle farms are less commercial or market oriented and rear their animals in a more traditional and extensive way. Two different cattle fattening systems have been described in West Africa: a rainy season fattening systems and a dry season fattening system (Dia Sow et al., 2004; Escot, 2011 ). For the rainy season fattening, farmers aim at taking advantage of the pastures during the rainy and post-harvest seasons when feed costs are low and fodder quality and quantity are good. For dry season fattening farmers rely mainly on stored and bought feeds from

159 Cha ter 4 markets, as already described for several West African cities or countries (Sanou et al., 2011; Ka bore et al., 2012; Ayantunde et al., 2014 ). Cattle fattening usually lasts for about 3 months and farmers usually target religious ceremonies when the price of animals is high (during LOS) and therefore the beef cattle market is attractive (Dia Sow et al., 2004; Escot, 2011 ).

4.4.2.2 Feeding management and intensity

Adult male animals were the most intensively managed group in supplementing households both for local and Sahelian zebu, while non-supplemented local zebu were extensively kept. The non negligeable contribution of roadside grasses, fallow vegetation and stubble material from public spaces, fallowed and harvested fields within and mainly around the city to the (peri-) urban feeding of ruminants has already been described in Mali (Amadou et al., 2015), Niger (Dan Gamma et al., 2017) and Burkina Faso (Schlecht et al., 2019). In Ouagadougou animals covered distances of 7.2-11.4 km daf1 and actively grazed for 4.3-8.8 h daf1 so that the contribution of pasture to daily OM intake of ruminants varied between 31-85% with an average OM intake on pasture of 31 g OM kg-1 MW daf1 for beef bulls (Schlecht et al., 2019). For supplementing households cattle homestead diets mainly consisted of protein rich concentrates from industrial by-products and roughages, such as diverse straws and hays, but also protein roughages which highly contributed to the fulfilment of fibre requirements at the homestead and compensated lack of grazing as observed for dairy cattle (Chapter 3). Industrial by-products mainly derived as residues from grain production and processing, were strongly lignified, and had a low nutritional value. As a way to avoid competition with human beings and similarly to observations already made in dairy cattle (Chapter 3), almost no grain was fed to beef cattle. This is because grains are preferably used in human nutrition. It also means that households supplementing their beef cattle get the ingredients for the animals' diet from various sources by harvesting or purchasing from fodder markets as reported for several West African cities (Sanou et al., 2011; Kabore et al., 2012; Ayantunde et al., 2014). The increasing feeding intensity from RS over EDS and LOS was pronounced for OM and fibre offers but was much less marked in terms of total ME and CP offered, since the diet quality decreased over the seasons, while feeding intensity increased. In terms of ME and CP offers, supplemented beef cattle were provided year round but with a very high seasonal variation, and also a high variation between different physiological groups. However, the feeding intensities were not always driven by the animals' actual requirements. Here again feed availability,

160 affordability (costs) and structural capacities for storage played a role, explaining also the variability in type, quality, and quantity of supplied feed (Buerkert and Schlecht, 2013; Amole and Ayantunde, 2016b).

4.4.2.3 Weight development

Sahelian zebu partly showed a high variation of LW and a lack in observation of middle aged animals. This may be due to the purchase of older rather than younger animals for fattening. For supplemented Sahelian beef cattle, adult males had the highest ADG of 1 1 646±339 g d- as compared to local zebu with ADG of 322±440 g d- . Meanwhile the 1 ADG in non-supplemented adult males was lowest (62±266 g d- ). The results for supplemented adult male Sahelian zebu are similar to the ADG of 788 g d-1 obtained for fattened local zebu (Gobra breed) in Senegal (Dia Sow et al., 2004). In Mali lower ADG of 63±126.1 g d-1 and 15±158.5 g d-1 were observed in non-supplemented and supplemented zebu cattle (Amadou et al., 2015), similarly to the present values for non supplemented local males and the ADG of 35.3±190 g d-1 observed in adult males of local zebu on extensive dairy farms in Ouagadougou, Burkina Faso (Chapter 3). The availability of quantity and quality fodder during the rainy season might explain similar LW, ADG and BCS obtained both in supplemented and non-supplemented beef cattle. During RS most farmers relied on the freely available feed (pasturing) and animals were in good condition. This is in line with observations made by authors in other West African cities, whereby grazing patterns differed with the types of husbandry units (Chapter 3; Amadou et al., 2015; Diogo et al., 2010) and are therefore related to the level of farm dependence on pasture and/or autonomous capacities for homestead feeding (supplementation). For beef cattle the growth curves developed in this study had relatively poor regression coefficients (R2). These imply a high variability of LW at a given age within the same breed and feeding intensity. The breed heterogeneity and the management system were the main factors affecting the variance in this case. Sahelian zebu are heavier than local zebu. Non-supplemented local zebu were heavier than supplemented ones up to 70 months of age. This might have been related to the rationale behind the choice of animals for fattening. Actually farmers usually buy cheaper animals and most probably animals in poor condition (Escot, 2011 ), which may explain the lower weight observed in supplemented local zebu beef cattle as compared to non-supplemented ones (that were mostly raised on-farm). In addition, for all adult groups, there was a growth depression during the LOS, or even a massive weight loss in non-supplemented

161 Cha ter 4 local beef zebu animals, whereas in supplemented local zebu only lactating cows lost weight during LOS. Therefore the animals' growth potential was only partially expressed, mainly in non-supplemented beef cattle, and therefore lagged behind the genetic potential as a consequence of low nutrient and ME supply relative to requirements. This is in line with results obtained in other West African countries where high input animals gained 97 g d-1 of LW while low input animals lost 68 g d-1 during the LOS in Niamey, Niger (Diogo et al., 2010) and losses of 22 and 91 g d-1 were observed in supplemented and non-supplemented cattle in Sikasso, Mali (Amadou et al., 2015) during the same season. Furthermore, similar weight losses have already been observed in central Mali (Schlecht et al., 1999).

4.4.2.4 Feed use efficiency in beef cattle production

Though both supplemented Sahelian zebu and local zebu showed a much better response to homestead feed offers in terms of weight gain than non-supplemented local zebu (which only responded to feed intake during grazing), a high variability in both feeding frequencies and excess feeding lead to inefficiencies. In supplementing households, local and Sahelian zebu of the same groups were supplied at the same intensity, showing that farmers were not taking into considerations the animals' actual ME and nutrient requirements. This means that farmers supply their animals indifferently with feed, irrespective of breed and they were mainly led by feed availability. Still a high proportion of animals were not adequately supplied and both extreme supply levels (severe deficit and substantial oversupply) of both ME and GP were frequently observed in the different groups of animals of both breeds. As already discussed in Chapter 3, this situation in not uncommon in West African beef cattle enterprises and has also been described for Bobo-Dioualasso (Dossa et al., 2015a), Niamey (Diogo et al., 2010) and Sikasso (Amadou et al., 2015). This also means that there is considerable room for improvement of resource use efficiency in beef cattle farms in Ouagadougou, both at farm and city level, despite the existing seasonal variability in feeding intensity from one farm to the other. Therefore, feeding frequency, feed quality and quantity are key factors for the improvement of resource use efficiency, production performance and profitability of beef cattle enterprises in (peri-) urban areas of Ouagadougou and other West African cities.

4.5 Conclusions

Since currently the (peri-) urban local and crossbred pig farming has a low productivity due to poor prolificacy of sows, poor management conditions and high mortality rates 162 (mainly for local pigs), improvement measures have to be urgently taken to sustain pig production in and around Ouagadougou. Pig farmers should certainly as much as possible define clear farming purposes (fattening, selling of weaned piglets or both) and adapt farm management strategies purposively. The improvement of production performances can be realised through an optimised management that is cancelation of scavenging, adoption of proper housing, year round stall feeding that might reduce the risk of diseases, improve biosecurity and reduce the mortality of animals, plus appropriate weaning of piglets. Feeding management should be improved through appropriate supply of quality feed, thereby balancing the high variability in feed supply. The latter especially means to supply adequate energy feeds and reduce the supply of protein rich feeds that are more easily available. Moreover controlled mating and proper health care of pigs should be implemented, especially in local pigs. In the context of feed resources scarcity that goes along with climate change, agricultural land shrinkage and population growth, non-supplementing beef cattle farmers might need to adapt their farm management towards homestead feeding and therefore be more commercially oriented so as to be able to bear production or feeding cost in order to achieve performant beef cattle production in and around West African cities such as Ouagadougou. Fattening beef cattle tamers or households already supplementing their animals should also improve on their feed used efficiency by narrowing the range of supply levels around the adequate supply level whereby animals are fed based on their requirements. For both (peri-) urban pig and beef cattle production relevant advice and support from extension services and other food and agricultural driven organisations are welcome. However, since farmers' decision to adopt any technology depends not only on productivity or production factors but also on other external factors, sound approaches and efficient resource use strategies taking into consideration most if not all of those factors should be developed and progressively implemented by farmers. In this regards participatory prioritisation of improved resource use strategies has to be developed.

163 Cha ter 5

5. General discussion

164 Cha ter 5

5.1 General aspects

Like in most metropoles and megacities, the increasing (peri-) urban population growth in West African cities triggers an undeniably increasing demand for food of animal origin and vegetables (Graefe et al., 2008; Orsini et al., 2013). Urban and peri-urban farming is common in Ouagadougou, Burkina Faso, like in most West African cities, both for crops and livestock (Dossa et al., 2011; Abdulkadir et al., 2012). With respect to livestock many species are reared by farmers from various socio-cultural and economic backgrounds (Chapter 2) (Thys et al., 2005; Roessler et al., 2016). There are several reasons for keeping livestock, including household income generation and food security (Sch iere et al., 2006; Thys, 2006; Grace et al., 2015). Around the (peri-)urban development of livestock production, several businesses such as marketing of animal feed and related activities (Sanou et al., 2011; Kabore et al., 2012; Ayantunde et al., 2014), job opportunities and income generating activities along the livestock value chain have emerged (Thys, 2006). However, livestock keeping in the city poses some problems like pollution (both chemical and physical) and the risk of disease spread (Schiere et al., 2006; Thys, 2006; Grace et al., 2015). Considerable variations exist within and across cities driven by anthropologic and non-anthropologic factors (Dossa et al., 2011; Abdulkadir et al., 2012; Roessler et al., 2016). Challenges for (peri-)urban livestock production are the use of poorly performing or non-adapted breed for specific production systems, as well as poor animal managerial practices, tamers' limited knowledge about adequately feeding of animals, poor environmental and health conditions of livestock, and limited pro- (peri-) urban livestock policies (Schiere et al., 2006). Therefore classifying and identifying different livestock production systems that operate in and around West African cities and designing site specific improved resource use and management strategies in order to improve farms' production performance and profitability should be of major concern for stakeholders in the livestock sector (Notenbaert et al., 2009).

5.2 Trends in (peri-) urban livestock production systems and farm characteristics

In this study a systematic characterisation of Ouagadougou's (peri-)urban livestock farms permitted a basic classification and identification of four different livestock production systems operating in that city (Chapter 2). Several studies have used the same methodology to classify and identify livestock production systems in West African cities, namely Kano (Nigeria), Bobo-Dioulasso (Burkina Faso) and Sikasso (Mali) 165 Cha ter 5 (Dossa et al., 2011; Abdulkadir et al., 2012; Dossa et al., 2015a), as well as in Asia, namely Faisalabad (Pakistan) (Tariq et al., 2014) and Yunnan Province (China) (Riedel et al., 2012). The use of cross tabulation analysis and logistic regression guided our analysis so as to relate different production features to the identified livestock production systems (Chapter 2; Roessler et al., 2016) altogether providing evidence for or against specific hypotheses related to crop-livestock integration, specialisation and production intensification. By the identification of social, technical and natural patterns that are common to production systems with respect to resource use and use efficiency, sound recommendations have been developed to successfully support (peri-) urban livestock production development in cities of sub-Saharan Africa in general and in Ouagadougou, Burkina Faso, specifically (Chapters 3 and 4).

In West African cities several livestock species are kept by (peri-) urban livestock households that are also involved or not in crop production (Dossa et al., 2011 ; Abdulkadir et al., 2012). The four livestock production systems identified in this study also kept several species such as ruminants (large and small), pigs and poultry (Chapter 2; Roessler et al., 2016). The socio-economic and production variables that decimated the identified livestock production systems were ethnicity (Fulani or Massi), pig farming, cattle feeding management and the sale of milk. In West Africa, historically the Fulani ethnic group is composed of agro-pastoral communities that keep ruminants and poultry and have no interest in pig rearing (lro, 1994; Awa et al., 2004; Gautier et al., 2016) as other ethnic groups do (Kiendrebeogo et al., 2008; Fualefac et al., 2014; Mopate Logtene et al., 2014 ).

Crop-livestock integration is a way to improve resource use efficiency and (peri-) urban households in Ouagadougou are most frequently involved in both crop and livestock production, particularly when located in peri-urban areas. The availability of farm land is certainly the main reason for this observation. How these farmers will react to the city's expansion, shrinkage of agricultural land and shortage of feed resources is interesting for evaluating future prospects. Moving away from city edges, specialisation in one livestock species and intensification might be the only suitable and sustainable option for most of these farmers. At farm level, crop-livestock integration is translated into the use of crops or crop residues as input to livestock production and crop fields receive livestock outputs processes (manure, draught power) as input to crop production of the same household. At city level, crop-livestock integration consists in the use of crops and crop residues or industrial crop by-products in livestock feeding and the use (sale) of

166 Cha ter 5 manure by (to) crop farmers. Both processes however imply costs related to information gathering, collective decision-making, and operational and monitoring costs that are influenced by several factors such as external environmental attributes, resources used in crop-livestock integration, and participating actors and their arrangements (Martin et al., 2016; Martin et al., 2018). Different on-farm activities such as proper storage of crop residues and good manure handling enhance nutrient recycling and therefore resource use (Rufino et al., 2006). In the context of (peri-) urban livestock farming, studies focusing on the enhancement of crop-livestock integration are rather scarce. In Ouagadougou, cropping is carried out by the majority of livestock farmers regardless of the animal species kept (Chapter 2; Roessler et al., 2016). In addition single livestock species farms are not frequently encountered and keeping different livestock species can be seen as farmers' strategy to manage risk and uncertainty not only related to livestock production but also to the whole household including family members' health, education and household food security (Awa et al., 2004). It is likely that multi-species keeping still persists for many decades in (peri-) urban livestock production systems in sub-Saharan Africa. It might be interesting to ask oneself what makes a species the most important in such systems so that the farmer invests more resources in that species, to the detriment of others. In Fulani communities usually large animals (cattle, horse) belong to the household head (generally a man) while small ruminants and poultry are often owned by other household members including women (lro, 1994). It is likely that such socio-anthropological considerations also exist in (peri-) urban livestock production systems, especially since most household heads have a long-term experience since a young age. Dairy cattle farming is still strongly influenced by ethnicity (Chapter 2) and is mostly carried out by Fulani households that can be considered as traditional farmers with long-term experience as dairy producers (lro, 1994; Awa et al., 2004; Gautier et al., 2016). Yet there emerges a set of new dairy producers made up of non-Fulani ethnic groups; these are often called "new actors" and are often more open to innovations and adoption of new production technologies (Bonfoh et al., 2007). Given the specificities of rearing dairy animals in landless or land constrained (peri-)urban areas, the high urban demand for dairy products and policies towards enhancement of dairy production (Bonfoh et al., 2007; Duncan et al., 2013; Chagunda et al., 2015), this study compared resource use and use efficiency between both groups (Chapter 3).

Cattle production was not only limited to dairy production but was also extended to beef cattle production with different management intensities. And the shift from dairy to beef

167 Cha ter 5 cattle production and vice versa is foreseeable as cattle (beef and dairy) farmers partly keep the same breeds (Chapter 2; Roessler et al., 2016). In cities of sub-Saharan Africa, (peri-) urban pig farming as well as small ruminant production also show different levels of intensification and specialisation (Chapter 2; Roessler et al., 2016). The process of intensification of (peri-) urban cattle and sheep production has been evidenced to be less pronounced in small cities (with lower population) such as Sikasso, Mali (Amadou et al., 2015) as compared to large capital cities (with millions of dwellers) such as Bamako (Mali) (Bonfoh et al., 2007), Dakar (Senegal) (Fall et al., 2000; Diao and Ba, 2004), Ouagadougou and Bobo-Dioulasso (Burkina Faso) (Millogo et al., 2008; Roessler et al., 2019) and Niamey (Niger) (Diogo et al., 2010). Driving forces of livestock production intensification are highly related on one hand to market factors such as the demand of livestock and animal products, the (lack of) importation of similar products, the competiveness of local products and on the other hand to livestock production factors mainly related to resources availability (McDermott et al., 2010). Given the changing (peri-) urban environments with the underlying socio-political and economic dynamics, farmers should adopt improved but site-specific livestock production strategies because only then they will be able to take advantage of the emerging opportunities for improving their livelihoods (Notenbaert et al., 2009). Homestead feeding, use of exotic (cross-) breeds, proper housing, health care provision and market orientation observed in this study are specific features of the intensification of (peri-)urban livestock production systems.

5.3 Trend towards intensification of (peri-) urban dairy production

In Ouagadougou dairy cattle production systems depict different production strategies and levels of intensification (Chapter 2; Millogo et al., 2008; Gnanda et al., 2016). Diverse cattle breeds are used for dairy production ranging from local zebu, Sahelian zebu (Azawak, Gudali), cross breeds from exotic breeds used for upgrading the local zebu breed such as the Montbeliarde, the Tarentaise, the Brown Swiss, and the Holstein (CORAF/WECARD, 2013). Genetic improvement through cross breeding is mainly observed in intensive (OUA-3) dairy farms. Proper feeding or adequate nutritionnal management is a key factor that affects overall farm productivity as well as health, milk production and reproduction, all highly depend on feeding. Pasture use and homestead feeding are carried out by almost all (peri-) urban dairy farmers in Ouagadougou, but at different extent or intensity (Chapter 3). Pasturing in open (peri-) urban grasslands is of interest to all dairy farmers but more intensive dairy cattle

168 Cha ter 5 farmers (OUA-3) use pasture after strategic appraisal or assessment of its quality and they stop when they judge it is not worth sending animals on pasture. Key features used to make that decision are still to be studied. As evidenced, due to the decrease in pasture quality from RS to LOS extensive dairy farms (QUA 4) have to shift pasture lands from fallows to crop fields and valleys while grazing is progressively replaced by browsing as fodder become scarce (Chapter 3). Intensive dairy farms take more advantage of the freely available pasture by acquiring the necessary resources or capacities (cash, labour, housing, equipment ... ) for harvesting and storage of wild hay which they then use when their animals are off pasture. The contribution to the dry matter offer to ruminants by roadside grasses, fallow vegetation and stubble material from public spaces, fallowed and harvested fields within and particularly around the city is known to be important in West African countries such as Burkina Faso (Sanou et al., 2011; Kabore et al., 2012), Mali (Amadou et al., 2015) and Niger (Dan Gamma et al., 2017). A high competition for resources between intensive and extensive dairy cattle happens during RS and EDS when all farms use pasture though pasture quality is also better. Various feed types, qualities and quantities of feedstuffs are supplied to dairy cattle at homestead with high variation between and within dairy production systems and a remarkably high intra-seasonal and inter-seasonal variability in the same production system in terms of types, qualities and amounts of feedstuffs offered to all animal species and categories, and in particular to local cattle. As far as feed types are concerned, in addition to a higher variability of feed types in intensive dairy farms an interesting marker of intensification that differentiates both (peri-) urban dairy production systems is mainly the use of industrial and commercial dairy feed (protein and energy mix) in intensive dairy farms and accessory the use of cereal (maize and sorghum) silage produce on-farm, distillated spent grain from breweries, and the amount of hay that is harvested and stored for future use. Industrial by-products such as rice bran, cottonseed expeller, sesame chaff, maize grain residues produced in regional mills and oil factories are the main feed types used in OUA-4 dairy farms (Chapter 3). These feeds are also found in feed markets of other West African cities (Ayantunde et al., 2014; Soubeiga, 2015). With further city expansion it is likely that for sustainable dairy production OUA-4 tamers will have to make the necessary shift toward supplying their dairy cattle more intensively at homestead in periods of poor quality pasture in the near future as a consequence of overgrazing and hay harvesting intensity. Nowadays they are no more fully extensive dairy farmers given their considerable contribution to the city dairy market (Duteurtre,

169 Cha ter 5 2007) and because many of those farms already apply homestead supplementation though far below animal requirements (Chapter 3; Bonfoh et al., 2007). In the context of feed shortage long distance walking enhance the ME energy deficit which impairs lifetime productivity by widening the inter-calving interval, reducing milk yield and calf growth (De Ridder et al., 2015). In a dairy farm keeping adult males on-farm is not profitable and not efficient in the context of resource scarcity, and homestead supplementation. Given the high demand of milk and dairy products in Ouagadougou and other West African cities (Rosegrant et al., 2009), a sound analysis of the benefit cost analysis of applying (limited) long distance walking and better supplementation during dry season in extensive dairy farms could help designing proper dairy farm management that will allow animals to express their full (re-) production potential (Sidibe-Anago et al., 2008; De Ridder et al., 2015). As users of high productive dairy cattle breeds, intensive dairy farms should already be aware of the content (in terms of ME and nutrients) of feed they provide to animals and pay attention to metabolic diseases mainly due to unbalanced diets that reduce dairy cow's performance (Chapter 3; Roessler et al., 2019). At city level industrial feed suppliers, extension services, and NGO's should support farmers' training on intensive dairy cattle nutrition so that they can improve on-farm efficiency and therefore farm productivity and profitability.

5.4 Efficiency of (peri-) urban fattening and traditional beef cattle production

Cattle breeds used for beef production in Ouagadougou are primarily local zebus and secondary Sahelian zebu used by fatteners (Chapter 4). Actually sub-Saharan Africa accounts for a high number of cattle units (about 76,000,000) which produce limited amounts of milk and dairy products to meet the regional demand (FAOSTAT, 2017). This means that meat production is actually highly dominated by the traditional or extensive system both from the rural and (peri-) urban zone and that genetic improvement of breeds used for meat production only has gained low interest from extension services and decision makers in West African governments given the low number of projects that focused solely on improving meat production through breeding across the region (Salla, 2017). Traditional rearing of cattle is a low input and low output system where animals do not always express their full genetic potential (Otte and Chilonda, 2002) as observed in this study for non-supplementing households. Here again the poor productivity of pastures and vanishing grazing lands accompanying the growth of cities will make the situation worst for the extensive or non-supplementing beef cattle farmers (Chapter 3.4.2.3). Though fattening of beef cattle is rather a

170 Cha ter 5 speculative production strategy, fattened (i.e., supplemented) cattle show higher performances in comparison to non-supplemented animals. Apart from industrial compound feeds and silages, feedstuffs used for cattle fattening are similar to those used on dairy farms. Therefore the same observations made with respect to dairy cattle about high variability (of feed types, quality and amounts) and opportunistic feeding strategies hold true for fattening beef cattle farms. Likewise the improvement of the efficiency of fattening farms remains a key issue for a business which is already highly speculative (Gnanda et al., 2015; Blama et al., 2016) and depending on several market factors such as the price of the (lean) animals to be fattened, the costs of feeding and the market price of the fattened animals (Fikru, 2015; Dadi et al., 2017). Since a majority of fatteners purchase animals directly or indirectly from traditional (dairy or beef) cattle farms (Escot, 2011) it becomes obvious that the abandonment or change of operation by the latter will affect the production strategy of the opportunistic intensive specialised fattening segment and different future scenarios can be expected, from the disappearance of the cattle fattening system to the adoption of new strategies like buying animals further away on rural livestock markets that will therefore bear the risk of increasing meat prices.

5.5 Efficiency of (peri-) urban pig farming

Two distinct pig breeds have been identified on (peri-) urban pig farms in Ouagadougou, the local pig breed and crossbred pigs (Chapter 4 ). The latter result from cross breeding of local pigs with exotic breeds such as Large White and Landrace which have been introduced in the past through government or private attempts to improve the local breeds (Amills et al., 2013). The management of both breeds however can be considered as poor resulting in low productivity. Unbalanced feeding and high mortality (certainly due to poor hygienic conditions and major health care deficits) in local pigs are key weaknesses of Ouagadougou's (peri-) urban pig farming sector. Here again, the use of industrial commercial pig feed and fishmeal at low extent is limited to crossbreds which can be considered to be more intensively managed than local pigs. When available brewers' spent grain is heavily used for pigs of both breeds, together with grain residues - a phenomenon observed in most West African cities (Chapter 4; Mopate Logtene et al., 2014). By suppling local pigs mainly with protein rich feeds, metabolisable energy is the main limiting factor in the animals' diet. A seasonal variation in pig feeding patterns is observed as well but is seemingly less pronounced as observed in dairy and beef cattle. The traditional belief that pigs are omnivorous animals

171 Cha ter 5 that can easily feed themselves still prevails so that local pigs in many sub-Saharan countries are scavenging at least part of the year and are sometimes only enclosed during the cropping seasons. This is a way to avoid the destruction of crop fields by pigs which may lead to conflicts between pig owners and crop farmers. As far as feed supply is concerned, the consequence of scavenging on farmers' pig feeding know-how is that, since no or very little effort has to be made to feed the animals, farmers lack the necessary knowledge to supply pigs with appropriate feed quality and quantity with respect to reproduction and growth requirements (Ayizanga et al., 2018). Farmers' poor knowledge on feed nutrient content and pig feeding is further evidenced by the striking use of feedstuffs containing anti-nutritional factors (Chapter 4). This is at the same time an evidence for feed scarcity in the study area, which has already been described as one of the most challenging factors for pig farming in other West African cities (Kiendrebeogo et al., 2014; Mopate Logtene et al., 2014; Ayizanga et al., 2018). Attempts towards improved use strategies of brewers' spent grain and grain residues and the supply of energy-rich feeds in the diets of (peri-) urban pigs in Ouagadougou will certainly have a remarkable impact on pig farm performances. Besides feeding, better pig farm management also includes controlled breeding, a practice which for crossbreds is frequently carried out unlike for local pigs. In addition, crossbred pig keepers usually also invest more in housing (Kiendrebeogo et al., 2014). Crossbred pigs displayed better performance (mainly LW development and ADG) than local pigs though animals did not express their full growth potential. The high variability of pig performances might also mean that at farm level for both breeds there are both good and bad performers. For the latter it is not clear whether keeping heterogeneous groups of fattening pigs is a strategy of pig tamers to limit the excess deposit of fat in order to increase the chances to sell their animals at a better price or an involuntary consequence of poor production management. Resource use efficiency in (peri-) urban pig farms is rather poor, and pig breeds that are used by pig farmers are not exploited to their full potential in the different production systems and can perform better if properly managed. Thus the actual practices need to be improved towards resource efficient pig farming such as skillful allocation of feedstuffs, controlled mating (so as to avoid consanguinity), appropriate provision of healthcare and improved farm biosecurity.

172 Cha ter 5 5.6 Testing of initial hypotheses

Based on the insights gained by the survey and monitoring work of this study and the analysis of the gathered data, the working hypotheses postulated in the beginning of this study are tested for their plausibility and commented in Table 5.1.

Table 5.1: Study hypotheses (Chapt. 1) as verified by the study results (Chapt. 2, 3, 4). Hypothesis Verification • Intensification and specialisation • Partly confirmed: intensive segments of lead to high animal performances, (peri-) urban livestock production systems due to continuously high levels of use high inputs and show relatively high feed input through purchase and performances. The latter also depends on storage in intensive dairy, beef the proper mastering of factors such as cattle and pig farms, and to a feeding strategies and animal health care reduced variability of feed offer that can limit the production performances throughout the year or production when poor. With regard to feeds used and cycles. feeding intensity there is a high farm to farm, within farm and seasonal variability.

• Market orientation is positively • Rejected: intensive farms are more correlated to resources use market-oriented, however their resource efficiency. The more market use efficiency is poor and in some cases oriented a farm is, the higher is its dominated by important wastage of production and resource use resources while extensive farms lack efficiency. adequate resources to produce at a level that fully exploits the animals' (genetic) potential.

• Resource use efficiency in (peri-) • Rejected: the use of high performance urban livestock systems is related improved breeds is an indicator of to the breeds used, whereby the intensification but does not necessarily use of improved breeds enhances translate to high performances ( especially resource use efficiency. in pigs) and better resource use efficiency (in all species).

• In cattle, resource use efficiency is • Fully confirmed: intensive dairy cattle similar for dairy and beef cattle farms (OUA-3) and supplementing beef production systems, with cattle farms (fattening farms) depict similar comparable level of resource use resource use and use efficiency with efficiency in both production important farm to farm and farm-internal systems according to the level of seasonal variability of feed use, high intensification. inputs and often oversupply of energy and nutrients. Likewise, traditional Fulani dairy farms (OUA-4) and non-supplementing beef cattle farms highly depend on freely available (but shrinking) pasture resources which cannot meet the animals' requirements and result in poor performances.

173 Cha ter 5 5. 7 General conclusions and recommendations

The (peri-) urban livestock production systems in Ouagadougou are in a transitory phase whereby traditional and extensive farms are progressively replaced by more intensive, modern and commercial farms. Farmers are using improved breeds but they seem to lack the necessary knowledge to adequately feed and manage those animals for optimal livestock production. In addition, even local breeds also face poor production and inefficient management strategies that reinforce their traditional reputation of being poorly performing breeds. Such notion has been and is influencing farmers, decision makers, extension services and civil society organisations involved in the development of livestock production to invest in genetic improvement programs. The latter are mainly based on the importation of improved exotic breeds into West African countries. So far the outcomes have often been far below expectations or genetic improvement goals. Cropping is still widely practiced by most livestock farmers and there is a strong seasonal pattern that influences resource use such as feeding intensities, frequencies and feed type utilization. This leads to high variabilities in performances such as live weight changes and milk production. The very high intra-seasonal farm-to-farm and inter-seasonal within-farm variability in feeding strategies and intensities is contrasting with the rationale of an intensifying and specialising urban livestock sector. Feeding intensity highly depends on availability of feed resources rather than animal requirements. In this context, environmental conditions such as pasture productivity and pathogen pressure negatively impact on production performances. For quantified feed offers or supply levels, overfeeding (substantial supply) was as common as underfeeding (severe deficit) though related to livestock production systems but regardless of the animals' activity on pasture. Grazing in any case was insufficient to fulfill the animals' requirements on a year round basis. As a consequence poor health conditions such as ketosis prevail in some dairy farms and high animal mortality is still observed in some (peri-) urban livestock production systems. Both milk production and growth potential of animals are underexploited in most (peri-) urban livestock production systems resulting in poor resource use efficiency. Balancing diets by increasing the energy offer and reducing the offer of proteins will in most cases contribute to improve the overall energy and nutrient use efficiency. In the so-called traditional livestock production systems, animals should be fed more intensively to improve their overall productivity and reduce the share of nutrients required for maintenance by increasing the proportion of nutrients used for production performances.

174 Cha ter 5 The following recommendations are made for a successful development of (peri-) urban livestock production in Ouagadougou as well as in other West African cities:

Create awareness about good quality fodder production and storage, Champion a better feed and nutrient distribution through storage of the total amount of annually available nutrients across seasons, Promote of a switch of the production focus from grain to fodder plants and silage production, Promote the use of dual purpose crops that are optimised for good quality crop residues and result in better feed quality supplied to livestock, Disseminate knowledge from government, feed processing industries and civil society organisations about appropriate animal feeding based on animals' requirements and production potential, Raise awareness on animal health and welfare issues and their management for improved production.

175 References References

Abdu, S., & Gashaw, A. 2010. Production system dynamism and parasitic interaction of swine in and around Holetta, Ethiopia. Ethiopian Veterinary Journal 14, 71-81. Abdulkadir, A., Dossa, L. H., Lompo, D. J.-P., Abdu, N., & van Keulen, H. 2012. Characterisation of urban and peri-urban agroecosystems in three West African cities. International Journal of Agricultural Sustainability 10, 289-314. Adebayo, A. A., & Oseni, S. 0. 2016. Evaluation of the milk yield performance of Jersey cows under tropical climatic conditions in Nigeria. Livestock Research for Rural Development 28. http://www.lrrd.org/lrrd28/1O/adeb28173.html. Accesssed 2018. Adesehinwa, A. 0. K., Oluwole, 0. 0., Saka J. 0., & Olaseinde, T. 2010. Growth performance, blood parameters and return on investment of growing weaner pigs of Nigerian indigenous pig (NIP), Exotic (Large White x Landrace) and hybrid (F1 Crossbred of NIP x Exotic) origins under intensive management system. In: Conference on International Research on Food Security, Natural Resource Management and Rural Development. Tropentag, September 14-16, Zurich, Swizerland. Adetunji, M. 0., & Adeyemo, K. E. 2012. Economic efficiency of pig production in Oyo State, Nigeria: A stochastic production frontier approach. American Journal of Experimental Agriculture 2, 382-394. Agbokounou, A. M., Ahounou, G. S., Karim, I. Y. A., Mensah, G. A., Koutinhouin, B., & Hornick, J.-L. 2016. Ethnologie et potentialites du pore local d' Afrique. Journal of Animal and Plant Sciences 29, 4665-4677. Agyemang, K., Yapi-Gnaore, C. V., Ouologuem, B., Barje, P. P., & Adediran, S. A. 2006. Characterisation of inland-valleys for smallhoder dairy production in West Africa. Agronomie Africaine 18, 105-116. Ahunu, B. K., Aboagye, G. S., & Arthur, P. F. 1994. Cattle crossbreeding research in southern Ghana. In: Smith C., Gibson, J.S., Benkel, B., Chesnais, J., Fairfull, W., Gibson, J.P., Kennedy, B.W. and Burnside E.B. (Eds), Proceedings of the 5th World Congress on Genetics Applied to Livestock Production held at the University of Guelph, Ontario, Canada, 7-12. Organizing Committee, 5th World Congress on Genetics Applied to Livestock Production. Guelph, Ontaraio, Canada, pp. 464-467. Ahunu, B. K., & Acquaah, E. K. 1987. Production indices for selection in crossbred dairy cattle at Legon. Legon Agricultural Research Bulletin 2, 27-31. Ajala, M. K., & Adesehinwa, A. 0. K. 2007. Roles and efficiency of participants in pig marketing in the northern part of Nigeria. Journal of Central European Agriculture 8, 311-326. Ajala, M. K., & Adesehinwa, A. 0. K. 2008. Analysis of pig marketing in Zango Kataf local government area of Kaduna State, Nigeria. Tropicultura 26, 229-239. Ajala, M. K., Adesehinwa, A. 0. K., & Mohammed, A. K. 2007. Characteristics of smallholder pig production in Southern Kaduna area of Kaduna State, Nigeria. American-Eurasian Journal of Agricultural and Environmental Sciences 2, 182- 188. Akpo, L. E., & Grouzis, M. 2000. Valeur pastorale des herbages en region soudanienne, le cas des parcours saheliens du Nord-Senegal. Tropicultura 18, 1-8. Amadou, H., Dossa, L. H., & Schlecht, E. 2012. A comparison between urban livestock production strategies in Burkina Faso, Mali and Nigeria in West Africa. Tropical Animal Health and Production 2050, 1631-1642.

176 References Amadou, H., Konate, L., Nantoume, H., Buerkert, A., & Schlecht, E. 2015. Nutrient use efficiency in peri-urban dairy cattle and sheep holdings in southern Mali. Nutrient Cycling in Agroecosystems 102, 65-78. Amills, M., Ramfrez, 0., Galman-Omitogun, 0., & Clop, A. 2013. Domestic pigs in Africa. Nutrient Cycling in Agroecosystems 30, 73-82. Amole, T. A., & Ayantunde, A. A. 2016a. Climate-smart livestock interventions in West Africa: A review. CCAFS Working Paper N°178, CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Copenhagen, Denmark. Amole, T. A., & Ayantunde, A. A. 2016b. Assessment of existing and potential feed resources for improving livestock productivity in Niger. International Journal of Agricultural Research 11, 40-55. Arelovich, H. M., Abney, C. S., Vizcarra, J. A., & Galyean, M. L. 2008. Effects of dietary neutral detergent fiber on intakes of dry matter and net energy by dairy and beef Cattle: Analysis of published data. The Professional Animal Scientist 24, 375-383. Atherstone, C., Galiwango, R. G., Grace, D., Alonso S., Dhand, N. K., Ward, M. P., & Mor, S. M. 2018. Analysis of pig trading networks and practices in Uganda. Tropical Animal Health and Production 51, 137-147. Awa, D. N., Njoya, A., Mopate Logtene, Y., Ndomadji, J., Onana, J. A., Awa, A. A., Ngo Tama, A. C., Djoumessi, M., Loko, B., Bechir, A. B., et al. 2004. Contraintes, opportunites et evolution des systemes d' elevage en zone semi-aride des savanes d'Afrique Centrale. Cahiers Agricultures 13, 331-340. Awosanya, E. J., Olugasa, B., Ogundipe, G., & Grohn, Y. T. 2015. Sero-prevalence and risk factors associated with african swine fever on pig farms in Southwest Nigeria. BMC Veterinary Research 11, 1-11. Ayantunde, A. A., & Amole, T. A. 2016. Improving livestock productivity: Assessment of feed resources and livestock management practices in Sudan-Savanna zones of West Africa. African Journal of Agricultural Research 11, 422-440. Ayantunde, A. A., Blummel, M., Grings, A., & Duncan, A. J. 2014. Price and quality of livestock feeds in suburban markets of West Africa' s Sahel : Case study from Bamako, Mali. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 67, 13-21. Ayanwuyi, E., Adedeji, T. A., & Olatilewa, M. 2012. Contribution of dairy farming to household welfare in Edu local government area of Kwara State, Nigeria. African Journal of Agricultural Research 2, 26-38. Ayizanga, R. A., Kayang, B. B., Adomako, K., & Asamoah, L. 2018. Rural pig production systems and breeding preferences of pig farmers in Northern Ghana. Ghanaian Journal of Animal Science 9, 49-57. Ayssiwede, S. B., Dieng, A., Houinato, M. R. B., Chryssostome, C. A. A. M., lssay, Hornick, J. L., & Missohou, A. 2013. Elevage des poulets traditionnels ou indigenes au Senegal et en Afrique Subsaharienne: Etat des lieux et contraintes. Annales de Medecine Veterinaire 157, 103-119. Ba Diao, M. 2005. Situation et conditions de developpement de la production laitiere intensive dans les Niayes au Senegal. PhD Thesis. Cheikh Anta Diop University, Dakar, Senegal. Ba Diao, M., Dieng, A., & Seek, M. M. 2006. Pratiques alimentaires et productivite des femelles laitieres en zone periurbaine de Dakar. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 59, 43-49.

177 References Bailey, K. W., Jones, C. M., & Heinrichs, A. J. 2005. Economic returns to Holstein and Jersey herds under multiple component pricing. Journal of Dairy Science 88, 2269- 2280. Baltenweck, I., Staal, S. J., Ibrahim, M. N. M., Herrero, M. T., Holmann, F. J., Jabbar, M.A., Manyong, V. M, Patil, B. R., Thornton, P. K., Williams, et al. 2003. Crop- livestock intensification and interactions across three continents. Main report, ILRI, System-wide Livestock Program (SLP), project on transregional analysis of crop- livestock systems, Nairobi, Kenya. Baroin, C., & Boutrais, J. 2008. Betail et societe enAfrique. Journal desAfricanistes 78, 9-52. Bayemi, P. H., Bryant, M. J., Perera, B. M.A. 0., Mbanya, J. N., Cavestany, D., & Webb, E. C. 2005. Milk production in Cameroon : A review. Livestock Research for Rural Development 17. http://www.lrrd.org/lrrd17/6/baye17060.htm. Accessed 2016. Bechir, A. B. 2010. Productivite, dynamique des parcours et pratiques d'elevage bovin en zone soudanienne du Tchad. PhD Thesis. Polytechnic University of Bobo- Dioulasso, Burkina Faso. Bechir, A. B., & Kabore-Zoungrana, C. 2012. Browse trees of Chad savannas: Present state of populations and pasture use. Cameroon Journal of Experimental Biology 8, 35-46. Bellwood-Howard, I., Haring, V., Karg, H., Roessler, R., Schlesinger, J., & Shakya, M. 2015. Characteristics of urban and peri-urban agriculture in WestAfrica: Results of an exploratory survey conducted in Tamale (Ghana) and Ouagadougou (Burkina Faso). IWMI Working Paper N°163, International Water Management Institute, Colombo, Sri Lanka. Bequette, B. J., Backwell, F. R. C., & Crompton, L. A. 1998. Current concepts of amino acid and protein metabolism in the mammary gland of the lactating ruminant. Journal of Dairy Science 81, 2540-2559. Bjerre-Harp0th, V., Friggens, N. C., Thorup, V. M., Larsen, T., Damgaard, B. M., lngvartsen, K. L., & Moyes K. M. 2012. Metabolic and production profiles of dairy cows in response to decreased nutrient density to increase physiological imbalance at different stages of lactation. Journal of Dairy Science 95, 2362-2380. Blama, Y., Kadoumai, S., & Rigolot C. 2016. Analyse technico-economique des unites d'embouche bovine en zone semi-aride du Cameroun. Revue internationale des sciences et technologie 12. https://www.researchgate.net/publication/317338532 _A nalyse_technico economique_des_unites_d'embouche_bovine_en_zone_se mi-aride_ du_Cameroun. Accessed 2018. Bonfoh, B., Fokou, G., Ould Taleb, M., Fane, A., Woirin, D., Laimaibao, N., & Zinsstag, J. 2007. Dynamiques des systemes de production laitiere, risques et transformations socio economiques au Mali. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 60, 67-76. Boro, P., Naha, B. C., Chandra, P., Ambadas, M., Kumar, N., Kumari, A., & Channa, G. P. 2016. Genetic and non-genetic factors affecting milk composition in dairy cows. International Journal ofAdvanced Biological Research 6, 170-17 4. Botoni, H. E. 2003. Interactions Elevage-Environnement. Dynamique des paysages et evolution des pratiques pastorales dans les fronts pionniers du Sud-Quest du Burkina Faso. PhD Thesis. Paul Valery University, Monpellier Ill, France. Boudet, G. 1975a. lnventaire et cartographie des paturages enAfrique de l'Ouest. In: lnventaire et cartographie des paturages tropicaux africains. Actes du Colloque de Bamako, du 3-8 mars 1975, Bamako, Mali, pp. 57-77.

178 References Boudet, G. 1975b. Problemes poses par !'estimation de la capacite de charge d'un paturage naturel tropical. In: lnventaire et cartographie des paturages tropicaux africains. Actes du colloque de Bamako du 3-8 mars 1975, Bamako, Mali, pp. 265- 267. Bouissou, M.-F.,& Boissy, A. 2005. Le comportement social des bovins et ses consequences en elevage. INRA Productions Animales 18, 87-99. Boukary, A. R., Saegerman, C., Abatih, E., Fretin, D., De Deken, R., Harouna, H. A., Yenikoye, A., & Thys, E. 2013. Seroprevalence and potential risk factors for Brucella Spp. Infection in traditional cattle, sheep and goats reared in urban, periurban and rural areas of Niger. PLOS ONE 8, 1-12. Boussini, H., Traore, A., Tamboura, H. H.,& Bessin R. 2012. Prevalence de la tuberculose et de la brucellose dans les elevages bovins laitiers intra-urbains et periurbains de la ville d' Ouagadougou au Burkina Faso. Revue Scientifique et Technique de l'Office International des Epizooties 31, 943-951. Bruggeman, H. 1994. Pastoral women and livestock management: Examples from Northern Uganda and Central Chad. International Institute for Environment and Development (IIED) Issue Paper N°50, London, United Kingdom. Bruinenberg, M. H., Zorn,R. L. G., & Valk, H. 2002. Energy evaluation of fresh grass in the diets of lactating dairy cows. Netherlands Journal of Agricultural Science 50, 67-81. Brunner, N., Groeger, S., Raposo, J., Bruckmaier, R. M., & Gross, J. J. 2018. Prevalence of subclinical ketosis and production diseases in dairy cows in Central and South America, Africa, Asia, Australia and New Zealand, and Eastern Europe. Translational Animal Science 3, 84-92. Buerkert, A.,& Schlecht, E. 2009. Performance of three GPS collars to monitor goats' grazing itineraries on mountain pastures. Computers and Electronics in Agriculture 65, 85-92. Buerkert, A.,& Schlecht, E. 2013. Agricultural innovations in small-scale farming systems of Sudano-Sahelian West Africa: Some prerequisites for success. Secheresse 24, 322-329. Camara, A., Dieng, A.,& Mergeai, G. 2015. Amelioration des systemes de production mixtes en Afrique Soudano-Sahelienne. Role de l'espece Stylosanthes hamata (L.) Taub. (synthese bibliographique). Biotechnology Agronomy Society and Environment 19, 281-289. Carter, N. A, Dewey, C. E., Mutua, F., de Lange, C. F. M.,& Grace, D. 2013. Average daily gain of local pigs on rural and peri-urban smallholder farms in two districts of Western Kenya. Tropical Animal Health and Production 45, 1533-1538. Carter, N. A., Dewey, C. E., Grace, D., Lukuyu, B., Smith, E.,&. de Lange C. F. M. 2017. Average daily gain and the impact of starting body weight of individual nursery and finisher Ugandan pigs fed a commercial diet, a forage-based diet, or a silage-based diet. Journal of Swine Health and Production 25, 121-128. Cesar, J. 2005. Production fourragere en zone tropicale. L'evaluation des ressources fourrageres naturelles. CIRDES/CIRAD, production animale en Afrique de l'Ouest, Fiche N°17, Bobo-Dioulasso, Burkina Faso. Chagunda, M. G. G., Mwangwela, A., Mumba, C., Dos Anjos, F., Kawonga, B. S., Hopkins, R., & Chiwona-Kartun, L. 2015. Assessing and managing intensification in smallholder dairy systems for food and nutrition security in sub-Saharan Africa. Regional Environmental Change 16, 2257-2267. Chaibou, M., lllia, A. S.,& Marichatou, H. 2011. Pratiques de gestion et performances de production dans les elevages bovins laitiers urbains et peri-urbains de Niamey. Revue des Bioressources 1, 1-12.

179 References Charray, J., Coulomb, J., &Mathon, J. C. 1977. Le Croisement Jersiais X N'Dama en Cote-d' lvoire: Analyse des performances des animaux demi-sang produits et eleves au centre de recherches zootechniques de Minankro. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 30, 67-80. Chase, L. E., Higgs, R. J., & Van Amburgh, M. E. 2012. Feeding low crude protein rations to dairy cows - What have we learned ? In: 23rd Ruminant Nutrition Symposium. University of Florida, Gainesville, Florida, pp. 32-42. Close, W., & Menke, K. 1986. Selected topics in animal nutrition. In: A manual prepared for the 3rd Hohenheim course on animal nutrition in the tropics and semi-tropics. Deutsche Stiftung fur lnternationale Entwicklung, Feldafin. Coffie, I., Annor, S. Y., Kagya-Agyemang, J. K., & Bonsu, F. R. K. 2015. Effect of breed and non-genetic factors on milk yield of dual-purpose cattle in Ashanti Region, Ghana. Livestock Research for Rural Development 27. http://www.lrrd.org/lrrd 27/7/coff27134.htm. Accessed 2019. Cohen, J. 1988. Statistical power analysis for the behavioral sciences. 2nd editiion. Hillsdale, NJ. Erlbaum. Cohen, M. J., & Garrett, J. L. 2009. The food price crisis and urban food (in)security. IIED and UNFPRA, London, UK. http://www.iied.org/pubs/display.php?o=1057411ED. Accessed 2017. Commune de Ouagadougou. 2011. Annuaire statistique 2010. Ouagadougou, Burkina Faso. CORAF/WECARD 2013. Etat de la gestion et de la conservation des ressources genetiques animales en Afrique de l'Ouest (Burkina Faso, Benin, Mali), Project 03.GRN.16. http://www.coraf.org/ressourcesgenetiques/wp-content/uploads/2014/ 11 /Status-of-ANR-WA 1.pdf. Accessed 2019. Corniaux, C. 2015. L'industrie laitiere en Afrique de l'Ouest : Histoire, strategies et perspectives. Milk way for development/CIRAD/PPZS, Dakar Senegal. Corniaux, C., Duteurtre, G., Dieye, P. N., & Poccard-Chapuis, R. 2005. Les minilaiteries comme modele d' organisation des filieres laitieres en Afrique de l'Ouest : Succes et limites. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 58, 237-243. Coulibaly, A. 2006. Country pasture/forage resource profiles Mali. Food and Agricultural Organisation, Rome, Italy. Gour, J.M. 2001. The Sahel in West Africa: Countries in transition to a full market economy. Global Environmental Change 11, 31-47. D' Orgeval, R. 1997. Le developpement de la production porcine en Afrique. PhD Thesis. INA PG, Paris, France. Dadi, G., Urge, M., & Teklebrhan, T. 2017. Assessment of commercial beef cattle fattening practices and performance in East Shoa zone. International Journal of Agricultural Science and Food Technology 3, 67-76. Daget, P., & Poissonet, J. 1971. Une methode d'analyse phytologique des prairies : Criteres d' application. Annales agronomiques 22, 5-41. Daget, P., & Poissonet, J. 1972. Un procede d'estimation de la valeur pastorale des paturages. Fourrages 49, 31-39. Dan Gomma, A., Chaibou, I., Banoin, M., & Schlecht, E. 2017. Commercialisation et valeur nutritive des fourrages dans les centres urbains au Niger : Cas des villes de Maradi et de Niamey. International Journal of Innovation and Applied Studies 21, 508-521.

180 References Darfour-Oduro, K. A., Sottie, E.T., Hagan, B. A., & Okantah, S. A. 2010. Milk yield and lactation length of Ghana Sanga and its crosses with the Friesian raised under agropastoral system. Tropical Animal Health and Production 42, 349-356. Daza, A., Rodriguez, I., Ovejero, I., and Lopez-Bote, J. 2003. Effect on pig performance of feed restriction. Spanish Journal of Agricultural Research 1, 3-8. De Ridder, N., Sanogo, 0.M., Rufino, M. C., Van Keulen, H., & Giller, K. E. 2015. Milk: The new white gold? Milk production options for smallholder farmers in SouthernMali. Animal 9, 1221-1229. De Zeeuw, Van veenhuizen, H., R., & Dubbeling, M. 2011. The role of urban agriculture in building resilient cities in developing countries.The Journal of Agricultural Science 149, 153- 163. DEFRA 1998. Condition scoring of pigs. In: Department of Environment, F.a.R.A.(Ed.), PB 3480. Defra Publications, London, United Kingdom. DEFRA. 2001. Condition scoring of dairy cows. In: Department of Environment, F.a.R.A.(Ed.), PB 3480. Defra Publications, London, United Kingdom. Dewa Kassa, K. A., Nenonene, A. Y., Tchainiley, L., & Koba, K. 2018. Pratiques de la production et d'exploitation des fourrages dans la region des Plateaux auTogo. International Journal of Biological and Chemical Sciences 12, 1415-1422. Dia Sow, F., Aminata, B., &Maloney, M. 2004. L'embouche paysanne, un exemple d'adaptation de l'elevage traditionnel a la nouvelle situation agricole dans le bassin arachidier du Senegal. Cahiers Agricultures 13, 211-219. Diao, M., & Ba, T. 2004. Situation et contraintes des systemes urbains et periurbains de production horticole et animale dans la region de Dakar. Cahiers Agricultures 13, 39-49. Diarra, L., Hiernaux, P., & Leeuw, P. N. 1995. Foraging behavior of cattle grazing semi- arid rangelands in the Sahel ofMali. In: Powell, J.M., Fernandez-Rivera, S. Williams, T.O., Renard C (Ed.), Livestock and Sustainable Nutrient Cycling inMixed Farming Systems of sub-Saharan Africa Volume II : Technical Papers. Addis Ababa, Ethiopia, pp. 101-116. Dieye, P. N., Duteurtre, G., Guzon, J., & Dia, D. 2007. Livestock, liberalisation and trade negotiations in West Africa. Outlook on Agriculture 36, 93-99. Dieye, P. N., Duteurtre, G., Sissokho, M.M., Sall, M., & Dia, D. 2003. La production I aitiere periurbaine au sud du Senegal. Saisonnalite de l'offre et performances economiques. Tropicultura 21, 142-148. Diogo, R. V. C., Buerkert, A., & Schlecht, E. 2010. Resource use efficiency in urban and peri urban sheep, goat and cattle enterprises. Animal 4, 1725-1738. Diop, A., lckowicz, A., Diene, M., & Nzimulinda, J. 2009. Production laitiere dans la zone sylvopastorale du Senegal: Etude des facteurs de variation et modes de gestion par les populations locales. Revue d' Elevage et deMedecine Veterinaire des PaysTropicaux 62, 39-47. Dossa, L. H., Abdulkadir, A., Amadou, H., Sangare, S., & Schlecht, E. 2011. Exploring the diversity of urban and peri-urban agricultural systems in Sudano-Sahelian West Africa : An attempt towards a regional typology. Landscape and Urban Planning 102, 197-206. Dossa, L. H., Sangare, M., Buerkert, A., & Schlecht, E. 2015a. Intra-urban and peri-urban differences in cattle farming systems of Burkina Faso. Land Use Policy 48, 401- 411 . Dossa, L. H., Sangare, M., Buerkert, A., & Schlecht, E. 2015b. Production objectives and breeding practices of urban goat and sheep keepers in West Africa: Regional analysis and implications for the development of supportive breeding programs. Springerplus 4, 1-12.

181 References Duku, S., van der Zijpp, A. J., & Howard, P. 2010. Small ruminant feed systems: Perceptions and practices in the transitional zone of Ghana. Journal of Ethnobiology and Ethnomedicine 6, 11. Dumont, B., Meuret, M., Boissy, A., & Petit, M. 2001. Le paturage vu par l'animal: Mecanismes comportementaux et applications en elevage. Fourrages 166, 213- 238. Duncan, A. J., Teufel, N., Mekonnen, K., Singh, V. K., Bitew, A., & Gebremedhin, B. 2013. Dairy intensification in developing countries: Effects of market quality on farm-level feeding and breeding practices. Animal 7, 2054-2062. Duteurtre, G. 2007. Commerce et developpement de l'elevage laitier en Afrique de l'Ouest : une synthese. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 60, 209-223. Duteurtre, G., & Corniaux, C. 2013. Etude relative a la formulation du programme d'actions detaille de developpement de la filiere lait en zone UEMOA. GIRAD, Montpellier, France. Duteurtre, G., & Koussou, M. 0. 2001. l'observatoire de la filiere lait au Tchad, un outil de concertation entre acteurs. In: Communication au 2eme Comite scientifique du PRASAC, Observatoire du Developpement du PRASAC / LRVZ de Farcha/ CIRAD-EMVT, Maroua. Escot, F. 2011. Le marche urbain de l'embouche bovine a Bamako: Rapport d'etude qualitative et quantitative, MISELI, Bamako, Mali. Fall, S. T., Fall, A. S., Cisse, I., Badiane, A., Fall, C. A., & Diao, M. B. 2000. Integration horticulture - elevage dans les systemes agricoles urbains de la zone des Niayes (Senegal). Bulletin de l'APAD 19, 1-17. FAO 2005. Bukina Faso: Livestock sector brief. Food and Agricultural Organisation, Rome, Italy .http://www.fao.org/ag/againfo/resources/en/pubs _sap.html. Accessed 2016. FAO 2007. Analyse de la filiere betail- viande au Burkina Faso. FAO service d'appui a !'assistance aux politiques (EASYPol), Ouagadougou, Burkina Faso. http://www. fao.org/docs/up/easypol/885/analyse_filiere_b%E9tail-viande_105fr.pdf. Accessed 2019. FAO 2011. Mapping Supply and demand for Animal-Source Foods to 2030. by Robinson T.P.& Pozzi F., Animal Production and Health Working Paper N° 2, Rome, Italy. http://www. fao.org/ docrep/014/i2425e/i2425e00.pdf. Accessed 2017. FAQ 2012. Pig sector Kenya. FAQ Animal Production and Health, livestock country, Reviews N° 3, Rome, Italy. http://www.fao.org/3/a-i2566e.pdf. Accessed 2017. FAO 2019. Food-based dietary guidelines. http://www.fao.org/nutrition/education/food- dietary guidelines/background/en/. Accessed 2019. FAO, AU-IBAR, & ILRI 2017. Strategie regionale du controle de la peste porcine africaine en Afrique. Accra, Ghana. http://www.fao.org/3/a-i6053f.pdf. Accessed 2018. FAOSTAT 2017. Food and Agriculture Organization of the United Nations. FAO Statistics Database. http://www.fao.org/faostat/en/#data/TA. Accessed 2018. FAOSTAT 2019. Food and Agriculture Organization of the United Nations. FAO Statistics Database. http://www.fao.org/faostat/en/#data/QL. Accessed 2019. Feedipedia 2017. Animal feed resources information system. Joint feed database project by INRA, GIRAD, AFZ and FAO. https://www.feedipedia.org/. Accessed 2018. Fikru, S. 2015. Assessment of Cattle Fattening and Marketing Practice in Harshin District of Somali Regional State, Ethiopia. Advances in Dairy Research 3, 137. Le Floc'h, E. 2007. Methodological guide for the study and monitoring of flora and vegetation. TC Number. ROSEL T/OSS Collection (Ed.), Technical Contribution N° 1, Obsevatory Network for Long Term Ecological Monitoring. ROSELT/OSS, Tunis, pp. 1-155.

182 References Fogwe, Z. N. 2015. The Development of a potential local dairy for an emerging peri-urban population of Cameroon. World Journal of Business and Management 1, 43-57. Fualefac, D. H., Raphae, K. J., Bime, M.-J., Ndebi, G., Yemele F., Manjeli, Y., Zoli, P. A., Teguia, A., & Tchoumboue, J. 2014. Socioeconomic and technical characteristics of pig farming in the urban and peri- urban zone of Dschang - West region of Cameroon. Discourse Journal Of Agriculture and Food Sciences 2, 11-20. Gantner, V., Bobic, T., Potocnik, K., & Gregic, M. 2018. Metabolic disorders in dairy simmentals - prevalence risk and effect on subsequent daily milk traits. Mljekarstvo 68, 77-84. Gaston, A., & Lamarque, G. 1994. Les paturages saheliens de l'Afrique de l'Ouest. Extraits des atlas "Elevage et potentialites pastorales saheliennes." In: Gaston, A. & Lamarque, G. (Ed.), CTA and CIRAD-EMVT. Wageningen, Netherlands, pp. 1- 140. Gautier, D., Seignobos, C., Njoya, A., Reounodji, F., & Ankogui-Mpoko, G.-F. 2016. Agriculteurs et eleveurs des savanes d'Afrique centrale: De la coexistence a !'integration territoriale. Espace Geographique 34, 223-236. Gidiglo, K. F. 2014. Milk production and marketing in Ghana: The case of Accra plains. Journal of Biology, Agriculture and Healthcare 4, 60-65. Gnanda, I. B., Bougouma-Yameogo, V. M. C., Wereme/N'Diaye, A., Ouedraogo, T., Kabore, A., Lodoun, B., & Sinon, B. 2015. L'embouche bovine dans les elevages du Plateau Central du Burkina Faso: Resultats economiques d'une demarche de validation d'un referentiel technico-economique sur la speculation. International Journal of Biological and Chemical Sciences 9, 2648-2662. Gnanda, I. B., Bado, D., Reusse, S., & Pare, L. 2016. Caracteristiques et typologie des exploitations laitieres bovines de la zone periurbaine de Ouagadougou, au Burkina Faso. Rencontres Recherches Ruminants 23, 165-168. Gomma, A. D., Chaibou, I., Banoin, M., & Schlecht, E. 2017. Commercialisation et valeur nutritive des fourrages dans les centres urbains au Niger : Gas des villes de Maradi et de Niamey. International Journal of Innovation and Applied Studies 21, 508-521. Groupe Agreco. 2006. Analyse de la problematique de !'importation de la poudre de lait au Burkina Faso et de son effet sur le developpement de la filiere lait. Goupe Agreco, Sainte Foy, Quebec, Canada. Grace, D., Lindahl, J., Correa, M., & Kakkar, M. 2015. Urban livestock keeping. In: De Zeeuw, H. & Drechsel, P. (Ed.), Cities and agriculture - developing resilient urban food systems. London, UK, Oxfam, pp. 255-284. Grace, D., Randolph, T. F., Omore, A., Schelling, E., & Bonfoh, B. 2007. Place of food safety in evolving pro-poor dairy policy in East and West Africa. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 60, 153-162. Graefe, S., Schlecht, E., & Buerkert, A. 2008. Opportunities and challenges of urban and peri urban agriculture in Niamey, Niger. Outlook on Agriculture 37, 47-58. Hamadou, S., Kamuanga, M., Abdoulaye, A. T., & Lowenberg-Deboer, J. 2005. Facteurs affectant !'adoption des cultures fourrageres dans les elevages laitiers periurbains de Bobo Dioulasso (Burkina Faso). Tropicultura 23, 29-35. Hamadou, S., & Sanon, Y. 2006. Synthese bibliographique sur les filieres laitieres au Burkina Faso. Reseau de Recherche et d' Echanges sur les Politiques Laitieres (REPOL), Document de Travail N°3, Dakar, Senegal. Hansen, P. J., & Arechiga, C. F. 1999. Strategies for managing reproduction in the heat stressed dairy cow. Journal of Dairy Science 82, 36-50.

183 References Hector, K. L., & Nakagawa, S. 2012. Quantitative analysis of compensatory and catch-up growth in diverse taxa. Journal of Animal Ecology 81, 583-593.

Herrero, M., Havlik, P., Mcintire, J.-M., Palazzo, A., & Valin, H. 2014. African livestock futures: Realizing the potential of livestock for food security, poverty reduction and the environment in sub-Saharan Africa. Office of the special representative of the UN secretary general for food security and nutrition and the United Nations System Influenza Coordination (UNSIC), Geneva, Switzerland. Heuer, C., Schukken, Y. H., & Dobbelaar, P. 1999. Postpartum body condition score and results from the first test day milk as predictors of disease, fertility, yield, and culling in commercial dairy herds. Journal of Dairy Science 82, 295-304. IBM Corp. 2011. IBM SPSS Statistics for Windows, Version 20.0. lckowicz, A., & Mbaye, M. 2001. Forets soudaniennes et alimentation des bovins au Senegal : Potentiel et limites. Bois et Forets des Tropiques 270, 47-61. IEMVT 1975. Growth and milk production of local and imported breeds, Bambui, Cameroon. Report of activities, lnstitut d'Elevage et de Medecine Veterinaire des Pays Tropicaux, France. I NSD 2017. lnstitut national de la statistique et de la demographie. Annuaire statistique 2016. Ouagadougou, Burkina Faso. INSD 2014. lnstitut national de la statistique et de la demographie. Annuaire statistique 2013. Ouagadougou, Burkina Faso. lro, I. 1994. From nomadism to sedentarism [microform]: An analysis of development constraints and public policy issues in the socioeconomic transformation of the pastoral Fulani of Nigeria. PhD thesis, Howard University, Washington, D. C, USA.

Kabore, A., Traore, A., Gnanda, B. I., Ouedraogo, P., Tamboura, H. H., & Belem, G. A. M. 2012. Diversity and characteristics of green natural fodders sold in the city of Ouagadougou in Burkina Faso, West Africa. International Research Journal of Agricultural Science and Soil Science 2, 377-385. Kabuga, J. D., & Agyemang, K. 1984. Performance of Canadian Holstein-Friesian cattle in the humid forest zone of Ghana. I. Milk production. Tropical Animal Health and Production 16, 85-94. Kadurumba, C., Kadurumba, 0. E., & Ahiwe, E. U. 2014. Measurement and sources of economic efficiency in pig production in Ebonyi State, Nigeria: A stochastic frontier function approach. Advances in Agriculture Sciences and Engineering Research. 4, 1522-1527. Kagone, H. 2001. Profil fourrager, Burkina Faso. Food and Agriculture Organization, Ouagadougou, Burkina Faso. Kahan, D. 2013. Market-oriented farming: An overview. Food and Agriculture Organization, Rome, Italy.

Karasov, W. H., & Douglas, A. E. 2013. Comparative digestive physiology. Comprehensive Physiology 3, 741-783. Kassa, K. S., Ahounou, S., Dayo, G.-K., Salifou, C., lssifou, M. T., Dotche, I., Gandonou, P. S., Yapi-Gnaore, V., Koutinhouin, B., Mensah, G. A., et al. 2016. Performances de production laitiere des races bovines de l'Afrique de l'Ouest. International Journal of Biological and Chemical Sciences 10, 2316-2330. Kiema, A., Sawadogo, I., Ouedraogo, T., & Nianogo, A. 2012. Strategies d' exploitation du fourrage par les eleveurs de la zone sahelienne du Burkina Faso. International Journal of Biological and Chemical Sciences 6, 1492-1505.

184 References Kiendrebeogo, T., Hamadou, S., Mopate Logtene, Y., & Kabore-Zoungrana, C.-Y. 2008. Typologie des elevages porcins urbains et periurbains de Bobo -Dioulasso (Burkina Faso). Revue Africaine de Sante et de Productions Animales 6, 205-212. Kiendrebeogo, T., Mopate Logtene, Y., & Kabore-Zoungrana, C.-Y. 2014. The typology of the pig breeding in Burkina Faso : Cases of the towns of Bobo-Dioulasso and Gaoua in soudanian area ; Kaya and Dori in Sahelian area. International Journal of Agronomy and Agricultural Research 4, 119-136. Kiendrebeogo, T., Mopate Logtene, Y., Kondombo, S. R., & Kabore-Zoungrana, C.-Y. 2012a. Approvisionnement en pores vifs et viande porcine de la ville de Bobo- Dioulasso (Burkina Faso). Journal of Agriculture and Environment for International Development 106, 105-122. Kiendrebeogo, T., Y. Mopate Logtene, Kondombo, S. R., & Kabore-Zoungrana, C. -Y. 2012b. Characterisation and importance of pig breeds in the pork industry of the Bobo-Dioulasso zone (Burkina Faso, West Africa). International Journal of Biological and Chemical Sciences 6, 1535-1547. Kouame-Sina, S. M., Makita, K., Costard, S., Grace, D., Dadie, A., Dje, M., & Bonfoh, B. 2012. Hazard identification and exposure assessment for bacterial risk assessment of informally marketed milk in Abidjan, Cote d' lvoire. Food and Nutrition Bulletin 33, 223-234. Kouamo, J., Tankou Tassemo, W. F., Zoli, P. A., Bah, G. S., & Ngo Ongla, A. C. 2015. Assessment of reproductive and growth performances of pig breeds in the peri- urban area of Douala (Equatorial Zone). Open Veterinary Journal 5, 64-70. Kouassi, A. F., Adou,Y. C. Y., lpou lpou, J., & Kamanzi, K. 2010. Diversite floristique des zones cotieres paturees de la Cote d' lvoire : Gas du cordon littoral Port-Bouet- Grand-Bassam (Abidjan). Sciences and Nature 7, 69-86. Kouassi, A. F., Koffi, K. J., Beranger, N. K. S., & lpou lpou, J. 2014. Potentiel de production fourragere d' une zone paturee menacee de destruction : Gas du cordon littoral Port- Bouet Grand-Bassam. Journal of Applied Biosciences 82, 7403-7410. Koutinhouin, G. B., Youssao, A. K. I., Toleba, S. S., Kpodekon, T. M., Ahounou, G. S., Bonou, A. G., & Bessanvi, J. 2009. Effet du mode d'elevage sur la prolificite des truies de race locale du Benin et la viabilite de leurs porcelets, de la naissance au sevrage. International Journal of Biological and Chemical Sciences 3, 819-829. Lanyasunya, T. P., Musa, H. H., Yang, Z. P., Mekki, D. M., & Mukisira, E. A. 2005. Effects of poor nutrition on reproduction of dairy stock on smallholder farms in the tropics. Pakistan Journal of Nutrition 4, 117-122. Lekule, F. P., & Kyvsgaard, N. C. 2003. Improving pig husbandry in tropical resource- poor comm unities and its potential to reduce risk of porcine cysticercosis. Acta Tropica 87, 111- 117. Lowry, 0. H., & Lopez, J. A. 1946. The determination of inorganic phosphate in the presence of labile phosphate esters. Journal of Biological Chemistry 162, 421-428. Machebe, N. S ., Onyekuru, N. A., & Ekweogu, N. 2007. Socio-economic factors affecting pig production in Enugu state Nigeria. Journal of Agriculture, Forestry and the Social Sciences 7, 41-49. Marichatou, H., Kore, H., Motcho, K. H., & Vias, G. 2005. Synthese bibliographique sur les filieres laitieres au Niger. Reseau de Recherche et d' Echanges sur les Politiques Laitieres (REPOL), Document de Travail N°3, Dakar, Senegal. Marshall, K., Tebug S., Juga J., Tapio M., & Missohou, A. 2016. Better dairy cattle breeds and better management can improve the livelihoods of the rural poor in Senegal. ILRI Research Brief 65. https://core.ac.uk/download/pdf/132682937. pdf. Accessed 2018.

185 References Martin, G., Moraine, M., Ryschawy, J., Magne, M.-A., Asai, M., Sarthou, J.-P., Duru, M., & Therond, 0. 2016. Crop-livestock integration beyond the farm level: A review. Agronomy for Sustainable Development, Springer Verlag/EDP Sciences/IN RA 36, 53. Martin, G.,. Hoshide, A. K, de Wit, J., Asai, M., Moraine, M., & Ryschawy, J. 2018. Critical factors for crop-livestock integration beyond the farm level: A cross-analysis of worldwide case studies. Land Use Policy 73, 184-194. Mbah, D. A., Mbanya, J., & Messine, 0. 1987. Performance of Holsteins, Jerseys and their zebu crosses in Cameroon:Preliminary results. Science and Technology Review, Agronomic, and Animal Sciences Series 3, 115-126. Mbah, D. A., Tawah, L. C., & Messine, 0. 1991. Performance of dairy cattle in Adamawa Cameroon. Cattle Research Network Newsletter 3, 10-12. 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, 95-109. McDowell, R. E., Amble, V. N., Jain, J.P.,Perez, C. M. B.,Prince, H. R., Rojas, H. C., Bhatnagar, D. S., Taneja, V. K., Basu, S. B., Rao, et al. 1985. Crossbreeding in tropical areas with emphasis on milk, health, and fitness. Journal of Dairy Science 68, 2418-2435. Mekuriaw, Y., & Asmare, B. 2014. Assessment of pig production and constraints in Mecha District, Amhara Region, Northwestern Ethiopia. Advances in Agriculture. 2014, 1- 5. Messine, 0., Schwalbach, L. J. M., Mbah, D. A., & Ebangi, A. L. 2007. Non-genetic factors affecting gestation length and postpartum intervals in Gudali zebu cattle of the Adamawa highlands of Cameroon. Tropicultura 25, 129-133. Metzger, R., Centres, J.-M., Thomas, L., & Lambert, J.-C. 1995. L'approvisionnement des villes africaines en lait et produits laitiers. Food and Agricultural Organisation, Rome, Italy. http://www.fao.org/docrep/v4870fN4870F00.htm#Contents. Accessed 2017. Millogo, V. 2010. Milk production of hand-milked dairy cattle in Burkina Faso. PhD Thesis, Swedish University of Agricultural Sciences, Uppsala, Sweden. Millogo, V., Ouedraogo, G. A., Agenas, S. & Svennersten-Sjaunja, K. 2008. Survey on dairy cattle milk production and milk quality problems in peri-urban areas in Burkina Faso. African Journal of Agricultural Research 3, 215-224. Mingoas, K. J.P., Zoli,P. A., Tchoumboue, J., Ebene, N. J., & Toukala, J.P. 2014. Socio economic characteristics and husbandry practices of cattle breeders in the Vina division. International Journal of LivestockProduction 5, 36-46. Ministere des Ressources animales. 2012. Statistiques du secteur de l'elevage, annuaire 2011. Ouagadougou, Burkina Faso. Ministere des Ressources animales. 2015. Annuaires des statistiques de l'elevage 2014. Ouagadougou, Burkina Faso. Mopate Logtene, Y., & Kabore-Zoungrana, C.-Y. 2013. Production, commerce et consommation de la viande porcine dans quelques villes d' Afrique de l'Ouest et du Centre. Revue Scientifique du Tchad 2013, 34-44. Mopate Logtene, Y., Kabore-Zoungrana C.-Y., & Facho, B. 2011. Structure des troupeaux et performances des elevages porcins de la zone de N'Djamena au Tchad. International Journal of Biological and Chemical Sciences 5, 321-330. Mopate Logtene, Y., Vounparet, Z., Issa Youssouf, A., & Kabore-Zoungrana, C.-Y. 2014. Production practices and constraints of pig farms in N' Djamena area, Chad. International Journal of Livestock Production 5, 196-203.

186 References Moreki, J. C., & Mphinyane, H. G. 2011. Opportunities and challenges of pig production in Botswana. Livestock Research for Rural Development 23. http://www.lrrd.org/ lrrd23/4 /more23087.htm. Accessed 2019. Moustier, P., & Danso, G. 2006. Local economic development and marketing of urban produced food. In: Van Veenhuizen, R. (Ed.), Cities farming for the future: Urban agriculture for green and productive cities. RUAF /IDRC/IIRR, Leusden, Netherland, pp.171-207. Mrode, R. A. 1988. Lactation performance of the White Fulani cattle in Southern Nigeria. Tropical Animal Health and Production 20, 149-154. Muhangi, D. 2014. African swine fever: An epidemiological overview. British Journal of Virology 1, 42-47. Muhanguzi, D., Lutwama, V., & Mwiine, F. 2012. Factors that influence pig production in Central Uganda - Case study of Nangabo Sub-County, Wakiso district. Veterinary World 5, 346-351. N'goran, K. E., Yapi Gnaore, C. V., Fantodji, T. A., & N'goran, A. 2008. Caracterisation phenotypique et performances des vaches laitieres de trois regions de Cote d'Ivoire. Archivos de Zootecnia 57, 415-426. Nare, R. W. A., Savadogo, P. W., Gnankambary, Z., Nacro, H. B., & Sedogo, M. P. 2015. Analyzing risks related to the use of pesticides in vegetable gardens in Burkina Faso. Agriculture, Forestry and Fisheries 4, 165-172. Ndebi, G., Kamajou, J., & Ongla, J. 2009. Analyse des contraintes au developpement de la production porcine au Cameroun. Tropicultura 27, 70-76. Ndebi, G., & Ongla, J. 2006. Fonctionnement des systemes de distribution du pore au Cameroun. Tropicultura 24, 73-81. Ngodigha, E. M., & Etokeren, E. 2009. Milk yield traits of Holstein Friesian x Bunaji crossbred cows with different Holstein Friesan inheritance. Journal of Animal and Veterinary Advances 8, 1145-1148. Njwe, R. M., Kwinji, L. N., Gabche, A. L., & Tambi, E. N. 2001. Contributions of Heifers Project International (HPI) to small-scale dairy development in Cameroon. In: Rangnekar, D. & Thorpe, W. (Ed.), Proceedings of a South-South workshop held at National Dairy Development Board (NDDB), 13-16 March 2001, Anand, India, pp. 414-430. Nonfon, W. 2005. La filiere de production du pore local au Benin : L' amelioration de sa productivite par l'alimentation. PhD Thesis. University of Liege, Belgium. Notenbaert, A., Herrero, M., Kruska, R., You, L., Wood, S., Thornton, P., & Omolo, A. 2009. Classifying livestock production systems for targeting agricultural research and development in a rapidly changing world. ILRI Discussion Paper N°19, International Livestock Research Institute, Nairobi, Kenya. Nouala, F. S., Akinbamijo, 0. 0., Bosso, N. A., & Agyemang, K. 2003. The comparative performance of N'Dama and N'Dama Crossbred cows under two supplementation levels in The Gambia. Livestock Research for Rural Development 15. http://www.lrrd.org/lrrd15/10/noua151 O.htm. Accessed 2018. Oba, M., & Wertz-Lutz, A. E. 2011. Ruminant nutrition symposium: Acidosis: New insights into the persistent problem. Journal of Animal Science 89, 1090-1091. Okantah, S., Oddoye, E., Obese, F., Gyawu, P., & Asante, Y. 1998. Characterisation of peri urban dairy production in Ghana. 2. Health and other Constraints in the production system. Ghana Journal of Agricultural Science 31, 197-202.

187 References Oke, U. K., lbe, S. N., Ologbose, F. I., & Amaefula, K. U. 2006. Effect of breed of sire on growth performance of exotic crossbred pigs in a humid tropical environment. Journal of Animal and Veterinary Advances 5, 744-748. Okello, E., Amonya, C., Okwee-Acai, J., Erume, J., & De Greve, H. 2015. Analysis of performance, management practices and challenges to intensive pig farming in peri-urban Kampala. International Journal of Livestock Production 6, 1-7. Okeudo, N. J., Aladi, N. 0., Okoli, I. C., & Akanno, E. C. 2007. Comparative study of the growth and carcass characteristics of the Nigerian indigenous and Large White pigs. Asian Journal of Animal Sciences 1, 57-66. Omore, A., Staal, S. J., Osafo, E. L. K., Kurwijilla, L., Barton, D., Mdoe, N., Nurah, G., & Aning G. 2004. Market mechanisms, efficiency, processing and public health risks in peri-urban dairy product markets : Synthesis of findings from Ghana and Tanzania. LPP Project R7321, ILRI, Nairobi, Kenya. Onana, J., & Devineau, J. L. 2002. Afzelia africana Smith ex Persoon dans le Nord- Cameroun . Etat actuel des peuplements et utilisation pastorale. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 55, 39-45. Orsini, F., Kahane, R., Nono-Womdim, R., & Gianquinto, G. 2013. Urban agriculture in the developing world: A review. Agronomy for Sustainable Development 33, 695- 720. Osei, S. A., Effah-Baah, K., & Karikari, P. 1991. The reproductive performance of Friesian cattle bred in the hot humid forest zone of Ghana. Animal Genetic Resources 68, 490-497. Otte, M., & Chilonda, P. 2002. Cattle and small ruminant production systems in sub- Saharan Africa. Food and Agriculture Organization of the United Nations, Rome, Italy, pp. 1-105. Oudet, M. 2005. La revolution blanche est-elle possible au Burkina Faso, et plus largement en Afrique de l'Ouest? MISEREOR, Aachen, Germany. Ouedraogo-Kone, S., Kabore-Zoungrana, C.-Y., & Ledin, I. 2006. Behaviour of goats, sheep and cattle on natural pasture in the sub-humid zone of West Africa. Livestock Science 105, 244-252. Ouedraogo, B., Bale, B., Zoundi, J. S., & Sawadogo, L. 2015. Caracteristiques de l'aviculture villageoise et influence des techniques d' amelioration sur ses performances zootechniques dans la province du Sourou, region Nord-Quest Burkinabe. International Journal of Biological and Chemical Sciences 9, 1528- 1543. Ouedraogo, B., Zoundi, J. S., & Sawadogo, L. 2017. Caracteristiques de l'elevage avicole en zone sahelienne du Burkina Faso. Revue lvoirienne des Sciences et Technologie 30, 263- 280. Oxfam. 2002. Milking the CAP: How Europe's dairy regime is devastating livelihoods in the developing world. Oxfam, London, United Kingdom. Petit, S., & Mallet, B. 2001. L' emondage d' arbres fourragers : Detail d' une pratique pastorale. Bois et Forets des Tropiques 270, 37-45. Porphyre, V. 2009. Enjeux et contraintes des filieres porcines en Afrique de l'Ouest. Grain de Sel 46-47, 26-27. Rege, J. E. 0., Aboagye, G. S., Akah, S., & Ahunu, B. K. 1994. Crossbreeding Jersey with Ghana Shorthorn and Sokoto Gudali cattle in a tropical environment: Additive and heterotic effects for milk production, reproduction and calf growth traits. Animal Production 59, 21-29. Rekwot, P., Abubakar, Y., & Jegede, J. 2005. Swine production characteristics and management systems of smallholder piggeries in Kaduna and Benue States of North Central Nigeria. Nigerian Veterinary Journal 24, 34-40.

188 References Riedel, S., Schiborra, A., Huelsebusch, C., Huanming, M., & Schlecht, E. 2012. Opportunities and challenges for smallholder pig production systems in a mountainous region of Xishuangbanna, Yunnan Province, China. Tropical Animal Health and Production 44, 1971- 1980. Rodrfguez-Estevez, V., Sanchez-rodrfguez, M., Garcfa, A.,&. G6mez-castro, A. G. 2010. Feed conversion rate and estimated energy balance of free grazing Iberian pigs. Livestock Science 132,152-156. Roessler, R., Mpouam, S. E., Muchemwa, T., & Schlecht, E. 2016. Emerging development pathways of urban livestock production in rapidly growing West Africa cities. Sustainability 8, 1199. Roessler, R., Mpouam, S. E, & Schlecht, E. 2019. Genetic and nongenetic factors affecting on farm performance of peri-urban dairy cattle in West Africa. Journal of Dairy Science 102, 2353-2364. Rosegrant, M. W., Fernandez, M., Sinha, A., Alder, J., Ahammad, H., de Fraiture, C., Eickhour, B., Fonseca, J., Huang, J., Koyama, 0., et al. 2009. Looking into the future for agriculture and AKST (Agricultural Knowledge Science and Technology). In: McIntyre, B. D., Herren, H. R., Wakhungu, J., and Watson, R. T. (Ed.), Agriculture at a Crossroads, global report. Washington, DC, USA: Island Press, pp. 307-376. Rouamba J.-P. 2016. Revue des filieres betail/viande & lait et des politiques qui les influencent au burkina faso. In: Niang M., Salla A., & Bedane B. (Ed.). Food and Agriculture Organization of the United Nations and the Economic Community of West African States, pp. 1-57. Rufino, M. C., Rowe, E. C., Delve, R. J., & Giller, K. E. 2006. Nitrogen cycling efficiencies through resource-poor African crop-livestock systems. Agriculture, Ecosystems and Environment 112, 261-282. Salla, A. 2017. Review of the livestock/meat and milk value chains and policy influencing them in West Africa. In: Olanrewaju, B. S.& Berhanu, B. (Ed.). Food and Agriculture Organization of the United Nations and the Economic Community of West African States, pp. 1-98. Sanon, H. 0., Savadogo, M., Tamboura, H. H.,& Kanwe, B. A. 2014. Caracterisation des systemes de production et des ressources fourrageres dans un terroir test de la zone soudanienne du Burkina Faso. Vertigo - la Revue electronique en Science de l'environnement 14. http://journals.openedition.org/vertigo/15171. Accessed 2018. Sanou, K. F., Nacro, S., Ouedraogo, M., Ouedraogo S., & Kabore-Zoungrana, C.-Y. 2011. La commercialisation de fourrages en zone urbaine de Bobo-Dioulasso (Burkina Faso) : Pratiques marchandes et rentabilite economique. Cahiers Agricultures 20, 487-493.

Savadogo, M., Zemmelink, G., Van Keulen, H., & Nianogo, A. J. 1999. Contribution of crop residues to ruminant feeding in different agroecological zones of Burkina Faso. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 52, 255- 262. Savadogo, P., Sawadogo, L., & Tiveau, D. 2007. Effects of grazing intensity and prescribed fire on soil physical and hydrological properties and pasture yield in the savanna woodlands of Burkina Faso. Agriculture, Ecosystems and Environment 118, 80-92. Sawadogo, L., Tiveau, D.,& Nygard, R. 2005. Influence of selective tree cutting, livestock and prescribed fire on herbaceous biomass in the savannah woodlands of Burkina Faso, West Africa. Agriculture, Ecosystems and Environment 105, 335-345. Schiborra, A., Gierus, M., Wan, H. W., Glindemann, T., Wang, C. J., Susenbeth, A.,& Taube, F. 2010. Dietary selection of sheep grazing the semi-arid grasslands of Inner Mongolia, China,

189 References at different grazing intensities. Journal of Animal Physiology and Animal Nutrition 94, 446- 454. Schiere, H., Thys, E., Matthys, F., Rischkowsky, B., & Schiere, J. 2006. Livestock keeping in urbanized areas, does history repeat itself? In: Van Veenhuizen, R. (Ed.), Cities farming for the future: Urban agriculture for green and productive cities. RUAF /IDRC/IIRR, Leusden, Netherland, pp. 349-379. Schlecht, E., Sangare, M., & Becker, K. 1999. Supplementation of zebu cattle grazing sahelian pasture. II. Development of body mass and empty body composition. Journal of Agricultural Science Cambridge 133, 83-95. Schlecht, E., Dickhoefer, U., Gumpertsberger, E., & Buerkert, A. 2009. Grazing itineraries and forage selection of goats in the Al Jabal al Akhdar mountain range of Northern Oman. Journal of Arid Environments 73, 355-363. Schlecht, E., Plagemann, J., Mpouam, S. E., Sanon, H. 0., Sangare, M., & Roessler, R. 2019. Input and output of nutrients and energy in urban and peri- urban livestock holdings of Ouagadougou, Burkina Faso. Nutrient Cycling in Agroecosystems. https://doi.org/10.1007/s10705-019-09996-x. Accessed 2019. Schneider, K., Gugerty, M. K., & Plotnick, R. 2010. Poultry market in West Africa: Burkina Faso. EPAR Brief N° 84, Washington. Sidibe-Anago, A. G., Ouedraogo, G. A., & Ledin, I. 2006. Effect of partly replacing cottonseed cake with Mucuna spp. (var. Ghana) hay on feed intake and digestibility, milk yield and milk composition of zebu cows. Tropical Animal Health and Production 38, 563-570. Sidibe-Anago, A. G., Ouedraogo, G. A., & Ledin, I. 2008. Effect of season and supplementation during late pregnancy and early lactation on the performance of zebu cows and calves. African Journal of Agricultural Research 3, 640-646. Sidibe, M., Boly, H., Lakouetene, T., Leroy, P., & Bosma, R.H. 2004. Characteristics of peri urban dairy herds of Bobo- Dioulasso (Burkina Faso). Tropical AnimalHealth and Production 36, 95-100. Somda, J., Kamuanga, M., Munstermann, S., Keita, K., &Mendes, A. 2004. Characteristics of the smallholder dairying farmers in West African countries: Economic viability and paths for improvement. Socio- economic research Working Paper N°2, ITC (International Trypanotolerance Centre), Banjul, The Gambia. Soubeiga, L. M. 2015. Analyse des aliments a betails vend us dans la ville de Ouagadougou: Structure et fonctionnement des marches. Unpublished report, Ecole Nationale de l'Elevage et de la Sante Animale (ENESA), Ouagadougou, Burkina Faso. Sow Dia, F., Somda, J., & Kamuanga, M. 2007. Dynamique des filieres laitieres en zone sahelienne : Gas de l'offre et de la demande du lait en zone agropastorale centre du Senegal. Revue d' Elevage et deMedecine Veterinaire des Pays Tropicaux 60, 77-88. Staples, C. R. 2007. Nutrient and feeding strategies to enable cows to cope with heat stress conditions. In: Proceedings of the 22nd Annual Southwest Nutrition andManagement Conference, Tempe, AZ, USA, pp. 93-108. Stoop, W. M., Bovenhuis, H., Heck, J.M. L., & Van Arendonk, J. A. M. 2009. Effect of lactation stage and energy status on milk fat composition ofHolstein-Friesian cows. Journal of Dairy Science 92, 1469-1478. SWAC-OECD/ECOWAS. 2008. Livestock and regional market in the Sahel and West Africa: Potentials and challenges. Sahel and West Africa Club/OECD, Paris, France.

190 References Tamboura, T., Bibe, B., Babile, R., & Petit, J.-P. 1982. Resultats experimentaux sur le croisement entre races locales et races laitieres ameliorees au Mali. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 35, 401-412. Tantasuparuk, W., Dalin, A., Lundeheim, N., Kunavongkrit, A., & Einarsson, S. 2001. Body weight loss during lactation and its influence on weaning-to-service interval and ovulation rate in Landrace and Yorkshire sows in the tropical environment of Thailand. Animal Reproduction Science 65, 273-281. Tariq, M., Hagmann, J., Dossa, L. H., Younas, M., & Schlecht, E. 2014. Structural characterisation of dairy production systems in Faisalabad, Pakistan as basis for their efficient resource management. Pakistan Journal of Agricultural Sciences 51, 1169-1180. Thaker, M. Y. C., & Bilkei, G. 2005. Lactation weight loss influences subsequent reproductive performance of sows. Animal Reproduction Science 88, 309-318. Thornton, P. K. 2010. Livestock production: Recent trends, future prospects. Philosophical Transactions of the Royal Society of London 365, 2853-2867. Thys, E., Oueadraogo, M., Speybroeck, N., & Geerts, S. 2005. Socio-economic determinants of urban household livestock keeping in semi-aridWestern Africa. Journal of Arid Environments 63, 475-496. Thys, E. 2006. Role of urban and peri-urban livestock production in poverty alleviation and food security in Africa. Academie Royale des Sciences d'Outre-Mer (Ed.), Memoire ln-8° Nouvelle Serie, Tome 26, fasc.1, Bruxelles, Belgique. Tialla, D., Kone, P., Kadja, M. C., Kamga-Waladjo, A., Dieng, C. B., Ndoye, N., Kouame, K. G. G., Bakou, S., & Akakpo, A. J. 2014. Prevalence de la brucellose bovine et comportements a risque associes a cette zoonose dans la zone periurbaine de Dakar au Senegal. Revue d' Elevage et de Medecine Veterinaire des Pays Tropicaux 67, 67-72. Tiffen, M. 2004. Population pressure, migration and urbanisation : Impacts on crop- livestock systems development in West Africa. In: Williams, T. 0., Tarawali, S. A., Hiernaux, P., and Fernandez-Rivera, S. (Ed.), Sustainable crop- livestock production for improved livelihoods and natural resource management in West Africa. International Livestock Research Institute, Nairobi, Kenya andand CTA, Wageningen, The Netherlands, UN-Habitat, pp. 3-27. Tindano, K., Moula, N., Traore, A., Leroy, P., & Antoine-Moussiaux, N. 2015. Characteristics and typology of sheep herding systems in the suburban area of Ouagadougou (Burkina Faso). Archives Animal Breeding 58, 415-423. Toure, A., Moula, N., Kouriba, A., Traore, B., Tindano, K., Leroy, P., & Antoine- Moussiaux, N. 2015. Dairy farms typology and management of animal genetic resources in the peri-urban zone of Bamako (Mali). Journal of Agriculture and Rural Development in the Tropics and Subtropics 116, 37-47. Traore, A., Tamboura, H. H., Bayala, B., Rouamba, D. W., Yameogo, N., & Sanou, M. 2004. Prevalence globale des pathologies majeures liees a la production laitiere bovine en systeme d' elevage intra- urbain a Hamdallaye (Ouagadougou). Biotechnology Agronomy Society and Environment 8, 3-8. Turkson, P. K. 2008. A comparison of the delivery of veterinary services to small-scale and medium to large-scale poultry keepers in peri-urban Ghana. Revue Scientifique et Technique de l'Office International des Epizooties 27, 719-730. Ulbrich, M., Hoffmann, M., & Drochner, W. 2004. Futterung und Tiergesundheit. Erste Auflage. Eugen Ulmer, Stuttgart, Germany 416p.

191 References Vall, E., & Bayala, I. 2004. Pilotage de !'amelioration des bovins. Note d'etat corporel des zebus soudaniens. CIRDES/CIRAD, production animale en Afrique de l'Ouest, Fiche N°12, Bobo Dioulasso, Burnia Faso. Vall, E., & Diallo, M. A. 2009. Savoirs techniques locaux et pratiques : La conduite des troupeaux aux paturages (Quest du Burkina Faso). Natures Sciences Societes 17, 122-135. VDLUFA 2012. Handbuch der Landwirtschaftlichen Versuchs- und Untersuchungs methodik (VDLUFA-Methodenbuch), Band Ill. Die chemische Untersuchung von Futtermitteln, 3. Auflage. Verband Deutscher Landwirtschaftlicher Untersuchungs- und Forschungsanstalten (ed.), VDLUFA- Verlag, Darmstadt, Germany West, J. W. 1999. Nutritional strategies for managing the heat-stressed dairy cow. Journal of animal science 77, 21-35.. World-Climate, 2016. Weather rainfall and temperature data. http://www.worldclimate. corn/. Accessed 2016. Wymann, M. N. 2005. Calf mortality and parasitism in periurban livestock production in Mali. PhD Thesis. University of Basel, Switzerland. Yameogo, G., Kiema, A., Yelemou, B., & Ouedraogo, L. 2013. Caracteristiques des ressources fourrageres herbacees des paturages naturels du terroir de Vipalogo (Burkina Faso). International Journal of Biological and Chemical Sciences 7, 2078- 2091. Younoussi, Z., & Piche, V. 2005. Migration et emploi urbain: Le cas de Ouagadougou au Burkina Faso. African Population Studies 20, 69-87. Zampaligre, N. 2012. The role of ligneous vegetation for livestock nutrition in the sub- Sahelian and Sudanian zones of West Africa: Potential effects of climate change. PhD Thesis. Kassel University, Kassel, Germany. Zezza, A., Federighi, G., Kalilou, A. A., & Hiernaux, P. 2016. Milking the data: Measuring milk off-take in extensive livestock systems. Experimental evidence from Niger. Food Policy 59, 174-186. Zoli, P. A., Shey-Njila, 0., Assana, E., Nguekam, J.-P., Dorny, P., Brandt, J., & Geerts, S. 2003. Regional status, epidemiology and impact of Taenia solium cysticercosis in Western and Central Africa. Acta Tropica 87, 35-42.

192 Appendix Appendix

Table A.1: Potential classification variables for two-step cluster analysis(% for nominal, mean±S.D. for continuous variables).

Potential variable Frequency (n=157) Selected for catPCA2 Selected for two-step cluster analysis3 Location Urban: 26.8 Yes Yes Peri-urban: 73.2 Age of household head (years) 51.9±12.01 Yes No Education of household head No education: 54.8 Yes Yes Primary school: 21.7 Secondary school: 14.7 Tertiary school: 8.9 Household member (n) 9.8±5.88 Yes No Ethnicity of household head Mossi: 71.3 Yes Yes Fulani: 24.2 Other: 4.4 Migrated Not migrated: 63.3 Yes Yes Migrated: 36.7 Livestock main occupation No: 86.6 Yes No Yes: 13.4 Crop production No: 29.3 Yes No Yes: 70.7 Farm time (years) 14.5±11.35 Yes No Labour use for livestock activities Family labour only: 61.8 Yes Yes Hired labour only: 8.3 Family and hired labour: 29.9 Cows and heifers (n) 8.2±12.32 No No Appendix Table A.1 continued Potential variable Frequency (n=157) Selected for catPCA2 Selected for two-step cluster analysis3 Cattle (n) Not considered No No Goats (n) 7.5±10.60 Yes No Sheep (n) 11.8±14.41 Yes No Pigs (n) 9.3±18.84 Yes Yes Local chicken, guinea fowl (n) 25.4±27.69 Yes No Mating controlled (natural, Al) No: 47.1 Yes Yes Yes: 52.9 Animals confined (stall, shelter, No: 45.2 Yes No fence) Yes: 54.8 Transhumance of cattle No: 62.4 Yes Yes Yes: 37.5 Milk sales No: 69.4 Yes Yes Yes: 30.6 :t> Egg sales No: 71.3 Yes No N Yes: 28.7 Average daily milk production in 7.9±26.17 No No 2013 (kg/households) Average daily milk sales in 2013 6.7±25.87 Yes No (kg/households) Livestock sales in 2013 (TLU1) 3.2±3.84 Yes Yes Manure use Manure thrown away: 7.0 No No Used on-farm: 58.0 Manure sold: 17.8 Manure given away: 9.6 Multiple uses: 7.6 Appendix Table A.1 continued Potential variable Frequency (n=157) Selected for catPCA' Selected for two-step cluster analysis3 Feeding method cattle Stall-fed: 5.6 Yes Yes Pasture: 67.3 Tethered: 2 .8 Stall-fed and pasture: 17 .8 Stall-fed and tethered: 2.8 Pasture and tethered: 3.7 Feeding method goats Stall-fed: 6.8 Yes Yes Pasture: 46.6 Tethered: 4.1 Stall-fed and pasture: 5.5 Stall-fed and tethered: 1.4 Pasture and tethered: 35.6 Feeding method sheep Stall-fed: 9.3 No No :t=, Pasture: 67.4 vJ Tethered: 1.2 Stall-fed and pasture: 12 .8 Stall-fed and tethered: 0.0 Pasture and tethered: 9.3 Feeding method pigs Stall: 57.4 Yes Yes Free range: 3.7 Stall and free range: 38.9 1 Tropical livestock unit (equivalent to 250 kg body weight), conversion factors used: 0.8 TLU for cattle, 0.1 TLU for sheep and goats, 0.2 TLU for pigs, 0.01 TLU for local poultry; 2 catPCA: categorical principal component analysis. The number of potential variables was gradually reduced. Spearman correlation tests, 112 statistics and Cramer's V were used to measure the association between variables, resulting in 24 variables that were selected for catPCA; 3 Through catPCA, the number of variables for the two-step cluster analysis was further reduced to 13, using a loading score �0.5 on one of the two components as selection criterion, S.D.: Standard deviation. A endix Table A.2: Proximate composition of feedstuffs offered to dairy, beef cattle and pigs in the (peri-) urban area of Ouagadougou (October 2014 - February 2016). OM OM CP p NDF ADF Feed group 1 Feed type n (g/kg FM) {g/kg OM) {g/kg OM} {mg/kg OM} (g/kg OM) (g/kg OM) Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Protein feed Maize grain residues 24 952 154.1 932 23.3 131 24.2 8.9 6.2 291 71.6 119 46.9 Protein feed Sesame chaff 2 973 3.3 842 9.3 205 13.8 4.1 0.2 281 0.0 361 0.0 Cereal grains Meal leftovers 2 548 468.1 925 59.4 185 93.5 2.4 0.5 123 51.6 34 3.8 Commercial feed Commercial dairy feed 9 938 15.2 904 40.3 216 21.7 14.1 8.1 471 125.5 263 112.4 Commercial feed Commercial pig feed 3 952 8.5 913 12.1 170 50.3 11.7 4.7 186 26.0 80 24.7 Other Fruit 1 976 4.7 952 - 48 - 1.6 - 197 - 146 Protein feeds Brewers grain (moist) 15 241 125.7 917 30.5 216 86.1 7.1 6.3 486 62.4 237 48.3 Protein feeds Cotton seed cake 15 958 27.7 951 14.3 240 34.4 7.0 4.0 562 51.8 397 66.9 Protein feeds Fish meal 1 964 - 753 - 620 - 19.0 - 278 - 31 Protein roughage Bean hulls 2 966 3.7 915 36.8 60 6.5 1.5 0.9 607 12.8 456 15.1 � Protein roughage Green grass 5 294 170.4 873 37.4 71 32.6 3.8 2.0 669 73.4 457 58.6 Protein roughage Green maize 4 306 41.5 936 17.2 69 6.2 3.5 1.6 650 112.2 345 55.9 Protein roughage Sorghum silage 3 296 39.1 901 17.3 91 18.8 4.1 2.1 670 29.0 400 23.5 Protein roughage Cereal straw 3 977 21.2 941 14.9 28 12.4 1.3 0.9 701 72.0 424 60.5 Straws and hays Grass hay 12 919 47.1 928 21.7 33 30.8 1.5 1.8 759 29.4 484 43.7 Straws and hays Maize straw 2 965 18.4 886 41.6 42 17.9 1.3 1.3 708 102.9 451 64.9 Straws and hays Millet straw 3 931 0.2 951 4.9 22 18.9 5.2 7.0 681 - 405 9.5 Straws and hays Rice husks 3 952 1.8 811 4.8 53 9.2 10 9.3 641 67.2 442 62.1 Straws and hays Rice straw 1 971 - 820 - 52 - 0.8 - 680 - 393 Straws and hays Sorghum straw 2 885 59.5 946 2.0 38 22.2 1.2 1.2 711 6.0 409 14.2 n: number of samples, OM: Dry matter, OM: Organic matter, CP: Crude protein, P: Phosphorus, NDF: Neutral detergent fiber, ADF: Acid detergent fiber Appendix Table A.3: Applied metabolisable energy (ME) values and nitrogen digestibility for feeding analysis. Feed name ME ruminants ME growing pigs Nitrogen digestibility pigs (MJ/kg OM) (MJ/kg OM) % Bean hulls 11.5 8.6 Brewers grain 10.0 10.0 77 Cereal straw 6.8 2.0 Commercial dairy feed 13.2 Commercial pig feed 12.3 76 Cotton seed cake 11.9 Fish meal 13.6 16.5 87 Fruit 13.3 13.5 80 Grass hay 6.9 48 Green grass 8.7 8.5 61 Green maize 9.6 Green sorghum 8.8 Maize grain 13.6 12.9 Maize grain residues 12.0 11.3 62 Maize silage 10.8 Maize straw 7.6 Meal leftovers 10.5 12.5 85 Millet straw 7.2 Molasses 9.6 13.0 Rice chaff 3.0 Rice husks 1.8 0.0 0 Rice straw 5.8 Sesame chaff 13.5 15.7 Sorghum silage 7.5 Sorghum straw 7.3 Tree fruit (wild) 12.7

A5 A endix Table A.4: Requirements (Req) of metabolisable energy (ME) and crude protein (GP) or digestible protein (DP), respectively, for maintenance (m), protein and fat accretion (pt; i.e. growth) and lactation (I) of the different animal types distinguished in the study, as well as their physiological limit for dry matter intake (DMI). Cattle Req ME_m Req ME_pf Req CP_m Req CP _pf DMI limit (kJ kg-1 MW*d) (MJ kg-1 1'. BW) (g kg-1 MW*d) (g kg-1 1'. BW) (kg OM 100 kg-1 BW*d) European crossbred (dairy) Dry cow 480 34.0 3.7 380 7.46 Pregnant cow 480 34.0 3.7 380 8.94 Lactating cowa 480 34.0 3.7 380 11.56 Young female 480 42.8 3.7 400 3.27 Adult male 480 31.7 3.7 380 9.76 Young male 480 31.7 3.7 400 3.46 Local zebu (dairy) Dry cow 480 34.0 3.7 380 5.68 Pregnant cow 480 34.0 3.7 380 5.36 a )> Lactating cow 480 34.0 3.7 380 6.63 Young female 480 42.8 3.7 400 2.46 Adult male 480 31.7 3.7 380 6.19 Young male 480 31.7 3.7 400 2.41 Sahelian zebu (dairy) Dry cow 480 34.0 3.7 380 5.49 Pregnant cow 480 34.0 3.7 380 6.95 Lactating cowa 480 34.0 3.7 380 9.50 Young female 480 42.8 3.7 400 3.67 Adult male 480 31.7 3.7 380 9.55 Young male 480 31.7 3.7 400 3.39 Appendix

Table A.4 continued Cattle Req ME_m Req ME_pf Req CP_m Req CP _pf DMI limit (kJ kg-1 MW*d) (MJ kg-1 1'. BW) (g kg-1 MW*d) (g kg-1 1'. BW) (kg OM 100 kg-1 BW*d) Local zebu (Beef) Dry cow 480 34.0 3.7 380 4.94 Pregnant cow 480 34.0 3.7 380 4.33 Lactating cowa 480 34.0 3.7 380 4.77 Young female 480 42.8 3.7 400 3.42 Adult male 480 31.7 3.7 380 6.97 Young male 480 31.7 3.7 400 4.18 Pigs Crossbred Dry sow 449 25.3 2.6 300 2.01 Pregnant sow 440 15.0 2.6 300 2.16 Lactating sowb 440 26.0 2.6 300 3.45 -.J Young female 567 15.2 2.6 300 0.86 Adult male 423 29.3 2.6 300 2.82 Young male 579 14.7 2.6 300 0.79 Local Dry sow 440 26.0 2.6 300 2.00 Pregnant sow 440 15.0 2.6 300 1.17 Lactating so� 440 26.0 2.6 300 1.81 Young female 567 15.1 2.6 300 0.77 Adult male 480 19.4 2.6 300 1.72 Young male 590 13.8 2.6 300 0.62 BW is body weight, MW is metabolic weight (BW0.75), and 1'. is delta/difference (i.e. weight change). All values are taken from/calculated after Ulbrich et al. (2004) a A milk offtake by the suckling calf of 250 Lin a period of 8 months is assumed; Req ME_I= 5.3 MJ kg-1 fat corrected milk (FCM, 4% fat), Req CP_I= 85 g/kg FCM b For milk production, the daily requirements are as follows: Req ME_I= 17.13 MJ/d, Req CP_I= 208.3 g/d in crossbred sows; Req ME_I= 7.75 MJ/d, Req CP _I= 94.2 g/d in local sows. Appendix Table A.5: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to pregnant (a) and lactating (b) cows in two different beef cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

a) Breed type Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (2) 204.7±13.3 183.9±11.8 23.0±5.8 1.39±0.08 116.5±10.6 76.6±12.7 2001.6±237.0 Local breed EDS (6) 104.8±47.6 98.0±44.0 13.2±6.5 0.63±0.38 59.9±25.3 33.0±13.2 805.2±531.0 LOS (8) 119.0±21.3 111.9±19.2 11.0±5.9 0.51±0.31 77.9±10.0 45.3±4.24 964.0±229.4 p� n.s. n.s. n.s. n.s. n.s. 0.05 n.s. RS Sahelian breed EDS LOS p� �

b) � Breed type Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (3) 18.6± 20.0 17.4± 18.7 3.9±3.8 0.22±0.39 7.4± 7.4 3.2± 3.2 198.6±219.3 Local breed EDS (1) 17.0 15.8 2.0 0.20 4.7 1.8 202.8 LOS (2) 18.9± 10.3 17.9± 9.9 2.1±0.6 0.99±0.54 114.6±95.1 66.3±64.2 1620.8±562.3 p� RS (2) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 Sahelian breed EDS (2) 94.5±119.9 89.1±113.0 7.3±6.0 0.34±0.47 58.7±76.0 34.6±44.7 761.2±913.2 LOS (1) 69.7 66.0 5.7 0.26 36.3 20.9 643.3 p� E._� p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight. Appendix Table A.6: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to dry cows (a) and adult males (b) in two different beef cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

a) Breed type Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (2) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 Local breed EDS (7) 19.4±33.1 18.2±31.0 2.4±4.1 0.09±0.15 11.0±18.8 6.7±11.4 195.4±333. 7 LOS (8) 24.2±44.9 23.1±42.9 2.7±5.1 0.07±0.13 16.1±29.8 10.8±20.1 221.2±410.1 p� n.s. n.s. n.s. n.s. n.s. n.s. n.s. RS Sahelian breed EDS LOS p� p�

b) Breed type Season (n) OM OM CP p NDF ADF ME kJ/k MW RS (33) 20.9±48.8 19.1±43.9 2.7± 5.7 0.22±0.39 9.7±27.7 5.7±18.3 219.3±481.1 Local breed EDS (102) 56.7±62.6 52.9±58.7 6.1± 6.1 0.37±0.40 28.5±36.2 15.7±20.8 533.4±560.5 LOS (74) 89.4±81.8 83.9±77.1 0.8±10.1 0.41±0.37 55.0±55.5 35.0±37.5 815. 9±736.8 p� 0.001 0.001 0.01 0.05 0.001 0.001 0.001 RS (14) 24.0±14.7 22.6±13.8 2.8± 1.7 0.49±0.38 6.2±3.9 2.2± 1.5 287.0±175.5 Sahelian breed EDS (20) 51.7±49.2 48.6±46.3 5.5± 4.0 0.51±0.41 23.1±29.5 12.0±17.5 529.7±426.7 LOS (17) 88.0±61.6 83.1±58.3 7.7± 5.6 0.54±0.35 40.7±38.5 22.7±24.1 857.1 ±560.2 p� 0.01 0.01 0.01 n.s. 0.01 0.01 0.01 p� n.s. n.s. n.s. 0.01 n.s. n.s. n.s. p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight. Appendix Table A.7: Seasonal variation in the daily amount of proximate diet components (g/kg MW, mean±S.D.) offered to young cattle (a) and suckling calves (b) in two different beef cattle breed types in Ouagadougou, Burkina Faso (October 2014 - February 2016).

a) Breed type Season (n) OM OM CP p NDF ADF ME kJ/ka MW RS (11) 135.5±87.0 121.7±78.2 15.4±10.7 0.92±0.60 77.0±49.5 50.5±32.8 1322.7±852.9 Local breed EDS (40) 91.6±57.9 84.7±54.0 12.7± 7.4 0.65±0.42 48.9±32.8 27.0±18.2 826.3±534.9 LOS (15) 86.4±54.8 78.9±50.4 9.3± 5.9 0.45±0.29 53.4±34.1 32.4±20.4 756.6±479.0 p s; 0.05 n.s. 0.05 0.01 0.05 0.05 0.05 RS Sahelian breed EDS (2) 42.9±47.0 40.5±44.2 2.5± 0.7 0.06±0.08 28.1±32.9 16.5±19.0 316.6±284.4 LOS p s; s;

b) Breed type Season (n) OM OM p NDF ADF ME 0 CP kJ/ka MW RS (1) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 Local breed EDS (1) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 LOS (4) 29.0±58.0 27.1±54.2 6.4±12.8 0.34±0.69 11.8±23.7 5.2±10.4 305.8±611.6 p s; RS (2) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 Sahelian breed EDS LOS (1) 0.0 0.0 0.0 0.00 0.0 0.0 0.0 p s; E_s; p: p-value, n: number of measurements, n.s.: not significant, OM: dry matter, OM: organic matter, CP: crude protein, P: phosphorus, NDF: neutral detergent fiber, ADF: acid detergent fiber, ME: metabolisable energy, RS: Rainy Season, EDS: Early Dry Season, LOS: Late Dry Season, MW: metabolic body weight. Related publications Related publications

Publications related to the present study: Roessler, R., Mpouam, S.E., Muchemwa, T., Schlecht, E. 2015. Classification on urban and peri-Urban livestock farm types in Ouagadougou and Tamale. In: Conference on International Research on Food Security, Natural Resource Management and Rural Development. Tropentag, September 16-18, 2015, Berlin, Germany. Mpouam, S.E., Muchemwa, T., Roessler, R., Schlecht E. 2015. Manure use in two West Africa cities: Tamale (Ghana) and Ouagadougou (Burkina Faso). Poster presented at the 3rd UrbanFoodPlus Summer School, October 4-9, 2015, Bochum, Germany. Muchemwa, T., Mpouam, S.E., Roessler, R., Schlecht, E. 2015. Body weight changes and milking performance in urban and peri-urban livestock systems: Tamale and Ouagadougou. Poster presented at the 3rd UrbanFoodPlus Summer School, Oct. 4-9, 2015, Bochum, Germany. Roessler, R., Mpouam, S.E., Muchemwa, T., Schlecht, E. 2016. Emerging development pathways of urban livestock production in rapidly growing West African cities. Sustainability 8, 1199. Plagemann, J., Mpouam, S.E., Roessler, R., Schlecht, E. 2017. Performance and management of dairy cows, beef cattle and pigs in peri-/urban agriculture in Ouagadougou, Burkina Faso. In: Conference on International Research on Food Security, Natural Resource Management and Rural Development. Tropentag, September 20-22, 2017, Bonn, Germany. Plagemann, J., Mpouam, S.E., Roessler, R., Schlecht, E. 2017. Feeding dairy cows, beef cattle and pigs in peri-/urban agriculture in Ouagadougou, Burkina Faso. In: E. Tielkes (Ed.). Future Agriculture: Socio-Ecological Transitions and Bio-Cultural Shifts, Cuvillier, Gottingen, Germany, pp. 367. Roessler, R., Mpouam, S.E,. and Schlecht, E. 2018. Identification of appropriate livestock genotypes to improve production performances in small household farms in Ouagadougou (Burkina Faso). In: Proceedings of the World Congress on Genetics Applied to Livestock Production, Volume Genetic gain - Strategies for Local Breeds 1, 588.

A11 Related publications Roessler, R., Mpouam, S.E., Schlecht, E. 2019. Genetic and non-genetic factors affecting on-farm performances of peri-urban dairy cattle in West Africa. Journal of Dairy Science 102 (3), 2353-2364. Schlecht, E., Plagemann, J., Mpouam, S.E., Sanon, H.O., Sangare, M., Roessler R., 2019. Nutrient and energy flows in urban and peri-urban livestock systems of Ouagadougou, Burkina Faso. Nutrient Cycling in Agroecosystems. https://doi.org/10.1007/s10705-019-x. Mpouam, S.E., Roessler, R., Schlecht, E. 2019. Performances and efficiency of (peri-) urban pig breeds under different production management in Ouagadougou, Burkina- Faso. In: Conference on Filling Gaps and Removing Traps for Sustainable resources management. Tropentag, September 18-20, 2019, Kassel, Germany.

A12 Acknowledgements Acknowledgements

Having come to the end of the long and challenging journey of a PhD Thesis, I would like to express my gratitude to all those who positively helped, contributed and influenced this achievement.

First of all my profound and sincere gratitude goes to my supervisor Prof. Dr. Eva Schlecht for her trustfulness in my person, her valuable supervision, relevant scientific orientations throughout the progress of my PhD thesis and moreover her indefectible availableness despite her busy agenda.

To Prof. Dr. Andreas Burkert as one of the founding fathers of UrbanFoodPlus and for valuable advises during our work as well as his always willingness to take along heavy samples with him from West Africa to Germany for laboratory analyses.

To Prof Dr. Luc Hippolyte Dossa not only for accepting to become a member of my examination board but also for valuable scientific assistance right away from the start of this PhD research.

To Prof. Dr. Detlev Moller for instantly accepting to become a member of my examination board given his extensive knowledge on farm management.

To Dr. Regina Roessler who scientifically supported and advised us throughout this research process for being our relay and connection in Germany when in the field to renew and ship our broken materials and new ones.

To Dr. Hadja Oumou Sanon (lnstitut de !'Environnement et de Recherches Agricoles, Ouagadougou, Burkina Faso) and Dr. Mamadou Sangare (Centre International de Recherche-Developpement sur l'Elevage en zone Subhumide, Bobo-Dioulasso, Burkina Faso) for their continuous and tremendous assistance and collaboration during our field work in Ouagadougou, Burkina Faso.

To Sankara Ousmane and Hamidou Bagayan for their assistance during baseline survey data collection, Mr Julian Plagemann and Mr Felix Stiegler for data cleansing.

To Cecile Sarambe and Louise Marie Kabore (MSc students, Ouagadougou, Burkina Faso) for their assistance during animal observations on pasture and pasture quality assessment.

A13 Acknowledgements To my assistant Bokoum Hassane who relentlessly dedicated his time, effort and knowledge of (peri-) urban livestock farmers to the achievement of field work from the start to the end.

To all (peri-) urban livestock keepers in Ouagadougou, Burkina Faso for their unconditional willingness to participate in the different steps of this PhD research work from the baseline survey and throughout the sixteen months on farm monitoring data collection period where they always welcomed us with the legendary hospitality and human warmth of West Africa farmers.

To my colleagues of the UrbanFoodPlus Dr. Delphine M. Abusi, Dr. Juliane Erbach, Dr. Takemore Chagomoka, Dr. Eileen B. Nchanji, Dr. Edmund K. Akoto-Danso, Tichaona Muchemwa, Barbara Loehde, for their collaboration.

We highly acknowledge the funding from the German Federal Ministry of Education and Research (BMBF, Bonn and Berlin, Germany) and the German Federal Ministry for Economic Cooperation and Development (BMZ, Bonn and Berlin, Germany) without which this research carried out under the initiative GlobE - Research for the Global Food Supply, grant number 031A242-A through the UrbanFoodPlus Project (http://www.urbanfoodplus.org/ index.php?id=5) wouldn't have been made possible.

A14 Eidesstattliche Erklarung - Affidavit Eidesstattliche Erklarung

"Hiermit versichere ich, dass ich die vorliegende Dissertation selbstandig und ohne unerlaubte Hilfe angefertigt und keine anderen als die in der Dissertation angegebenen Hilfsmittel benutzt habe. Alie Stellen, die aus veroffentlichten oder unveroffentlichten Schriften entnommen sind, habe ich als solche kenntlich gemacht. Kein Teil dieser Arbeit ist in einem anderen Promotionsverfahren verwendet warden."

Affidavit

"I assure that this dissertation was written independently and without non-permissible help and that I used no sources other than those specified in the dissertation. All quotations that have been extracted from published or unpublished sources have been marked as such. No part of this work has been used in other PhD processes."

Serge Eugene Mpouam 23.06.2019

A15