Fungal disease dynamics, genetic variation and biodiversity-yield relationships — a study along a gradient of coffee management in southwestern Ethiopia

Beyene Zewdie Hailu

Beyene Zewdie H

ailu Fun

gal disease dynam

ics, genetic variation

and biodiversity-yield relation

ships 

Doctoral Thesis in Plant Ecology at Stockholm University, Sweden 2020

Department of Ecology, Environment andPlant Sciences

ISBN 978-91-7911-352-0

Beyene Zewdie HailuEcologist interested in theepidemiology of diseases and theirnatural enemies in agroecosystems

This thesis includes the following papers:I. Zewdie, B., Tack, A. J. M., Adugna, G., Nemomissa, S. & Hylander,K. (2020). Patterns and drivers of fungal disease communities onArabica coffee along a management gradient. Basic and AppliedEcology 47, 95-106.

II. Zewdie, B., Tack, A. J. M., Ayalew, B., Adugna, G., Nemomissa, S.& Hylander, K. Temporal dynamics and biocontrol potential of ahyperparasite on coffee leaf rust in Arabica coffee’s native range.Submitted.

III. Zewdie, B., Bawin, Y. Tack, A. J. M., Nemomissa, S., Tesfaye, K.,Janssens, S. B., Ruttink, T., Van Glabeke, S., Roldán-Ruiz, I., Honnay,O. & Hylander, K. Genetic composition of Arabica coffee explainsfungal disease incidence in its area of origin. Manuscript.

IV. Zewdie, B., Tack, A. J. M., Ayalew, B., Wondafrash M.,Nemomissa, S. and Hylander, K. A steep decline in biodiversity withincreasing coffee yields across a gradient of management in Arabicacoffee’s native range. Manuscript.

Fungal disease dynamics, genetic variation andbiodiversity-yield relationships— a study along a gradient of coffee management in southwesternEthiopiaBeyene Zewdie Hailu

Academic dissertation for the Degree of Doctor of Philosophy in Plant Ecology at StockholmUniversity to be publicly defended on Wednesday 16 December 2020 at 14.00 in ViviTäckholmsalen (Q-salen), NPQ-huset, Svante Arrhenius väg 20.

AbstractIntensification of agricultural systems is a major threat to the associated biodiversity and could also affect the dynamicsof pests and pathogens. One such system that is currently under an intensification trajectory is the production of Arabicacoffee. In this thesis, I studied the relationships between fungal diseases and their natural enemies, the genetic variation incoffee, coffee yield and associated biodiversity along a coffee management gradient in southwestern Ethiopia.

The specific goals of this thesis were to investigate variation in fungal diseases on coffee and their natural enemiesalong a gradient of management (I, II), how genetic variation in coffee among sites relate to variation in incidence ofthe fungal diseases (III), and to investigate the trade-offs in biodiversity-yield relationships along the gradient of coffeemanagement (IV). To answer these questions, I selected 60 sites along a gradient of management that ranged from coffeenaturally growing in only little disturbed forests to intensively managed plantations. I used both observational studies andmolecular approaches.

In paper I, I examined if the severity of the four major fungal diseases on coffee varied along the gradient and assessedthe main drivers of variation in disease severity. I found that two of the fungal diseases were more severe in the intensivelymanaged coffee sites, while the other two were more severe in the less intensively managed sites. Altitude was the maindriver for the diseases, but related in a different way to the different diseases. In paper II, I examined the temporal dynamicsin coffee leaf rust-hyperparasite interaction, the biocontrol potential of the hyperparasite and environmental drivers for thetwo species for three consecutive years during the dry and wet seasons. I found that the rust was more common during thedry season and in managed sites while the hyperparasite was common during the wet season and in sites that were lessmanaged. My results also revealed that higher hyperparasite incidence during the wet season resulted in a lower growthrate of the rust during the subsequent dry season. In paper III, I investigated if genetic composition and diversity of coffeesites relate to the incidence of the fungal diseases assessed. I found that genetic composition of the coffee stands was linkedto the incidence of the four fungal diseases, but genetic diversity among the coffee sites did not relate to the incidence ofthe diseases. In paper IV, I examined biodiversity-yield trade-offs and shape of the relationships between biodiversity andyield along the gradient of management. I found a steep, concave shape initial decline in biodiversity values as coffee yieldincreased to a certain level, after which a further increase in yield did not have much effect on biodiversity values.

In conclusion, I found different drivers for the different diseases and for the parasite-hyperparasite interaction. It isdifficult to achieve a single management approach that can suit the different pathogen species investigated. High geneticdiversity among coffee sites did not reduce disease pressure. While the more complex, less managed sites provide highbiodiversity values, and could potentially serve as habitats for natural pest control and in situ conservation for coffee geneticdiversity, the yield gap compared to more intensively managed sites was very high. To optimize coffee management andconservation of biodiversity in these landscapes, there is a need to develop strategies whereby the smallholder farmerswho depend on coffee and the forest as the main source of livelihood can benefit through for example coffee certificationschemes that can pay premium prices for biodiversity-friendly coffee management.

Keywords: Armillaria root rot, biodiversity, biodiversity-yield trade-offs, coffee, Coffea arabica, coffee berry disease,coffee leaf rust, coffee wilt disease, coffee yield, genetic composition, genetic diversity, hyperparasite, managementintensity gradient, southwestern Ethiopia.

Stockholm 2020http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-186507

ISBN 978-91-7911-352-0ISBN 978-91-7911-353-7

Department of Ecology, Environment and PlantSciences

Stockholm University, 106 91 Stockholm

FUNGAL DISEASE DYNAMICS, GENETIC VARIATION ANDBIODIVERSITY-YIELD RELATIONSHIPS  

Beyene Zewdie Hailu

Fungal disease dynamics, genetic variation and biodiversity-yield relationships — a study along a gradient of coffee management in southwestern Ethiopia

Beyene Zewdie Hailu

©Beyene Zewdie Hailu, Stockholm University 2020 ISBN print 978-91-7911-352-0ISBN PDF 978-91-7911-353-7 Printed in Sweden by Universitetsservice US-AB, Stockholm 2020

To the memory of mylate mother AsnakuArarsa Gemeda andsister Aberu Zewdie       ***Haacaaluu Hundeessaaf            ***To my family: Birtukan,Naol and Maya 

List of papers This thesis is based on the following papers which are referred to in the text by their Roman numerals.

I. Zewdie, B., Tack, A. J. M., Adugna, G., Nemomissa, S. & Hylander, K. (2020). Patterns and drivers of fungal disease communities on Arabica coffee along a management gradient. Basic and Applied Ecology 47, 95-106. Doi:10.1016/j.baae.2020.05.002 © 2020 The Author(s).

II. Zewdie, B., Tack, A. J. M., Ayalew, B., Adugna, G., Nemomissa, S. & Hylander, K. Temporal dynamics and biocontrol potential of a hyperparasite on coffee leaf rust in Arabica coffee’s native range. Submitted.

III. Zewdie, B., Bawin, Y. Tack, A. J. M., Nemomissa, S., Tesfaye, K., Janssens, S. B., Ruttink, T., Van Glabeke, S., Roldán-Ruiz, I., Hon-nay, O. & Hylander, K. Genetic composition of Arabica coffee ex-plains fungal disease incidence in its area of origin. Manuscript.

IV. Zewdie, B., Tack, A. J. M., Ayalew, B., Wondafrash M., Ne-momissa, S. and Hylander, K. A steep decline in biodiversity with increasing coffee yields across a gradient of management in Arabica coffee’s native range. Manuscript.

Paper I is reprinted with permission from Elsevier under a Creative Com-mons licence (CC BY-NC-ND 4.0).

Contributions

I II III IV

Idea and design BZ, AT, KH BZ, AT, KH BZ, AT, KH, YB, TR, OH, IR

BZ, AT, KH

Data collection BZ BZ, BA BZ BZ, BA

Molecular work and bioinformatics

YB, BZ, SJ, SVG, TR

Analysis BZ, AT, KH BZ, AT, KH BZ, AT, KH BZ, AT, KH

Manuscript preparation BZ BZ BZ, YB BZ

Manuscript reviewing AT, KH, GA, SN

AT, KH, SN AT, KH, YB, SJ, TR, OH, SN, KT

AT, KH, SN

BZ = Beyene Zewdie AT = Ayco Tack BA = Biruk Ayalew GA = Girma Adugna IR = Isabel Roldán-Ruiz KH = Kristoffer Hylander KT = Kassahun Tesfaye

OH = Oliver Honnay SJ = Steven Janssens SN = Sileshi Nemomissa SvG = Sabine Van Glabeke TR = Tom Ruttink YB = Yves Bawin

Main supervisor Professor Kristoffer Hylander, Department of Ecology, Environment and Plant Sciences, Stockholm University, Sweden Co-supervisor Dr. Ayco Tack, Department of Ecology, Environment and Plant Sciences, Stockholm University, Sweden

Contents Abstract ........................................................................................................... 1

Introduction ..................................................................................................... 3 General background ................................................................................... 3 Coffee production and biodiversity conservation....................................... 6 Fungal diseases on coffee ........................................................................... 6

Aims of the study ............................................................................................ 9

Methods ........................................................................................................ 11 Study area ................................................................................................. 11 Coffee management systems .................................................................... 13 Study sites and surveying ......................................................................... 15 Environment and management variables ................................................. 15 Major fungal diseases on coffee along the gradient of management (Paper I) ............................................................................................................... 17 Temporal variation and biocontrol potential of a parasite–hyperparasite interaction on coffee (Paper II) ................................................................ 17 The relationship between genetic variation and the incidence of major fungal diseases on coffee (Paper III) ........................................................ 19 Trade-offs in biodiversity-yield relationships (Paper IV) ........................ 20

Results and discussion .................................................................................. 22 Fungal disease dynamics, drivers of the diseases, temporal variation, and biocontrol potential of the rust hyperparasite ........................................... 22 Biodiversity yield trade-offs ..................................................................... 26 Implications for management ................................................................... 28

Concluding remarks and future perspectives ................................................ 32

Svensk sammanfattning ................................................................................ 35

Acknowledgements ....................................................................................... 37

References ..................................................................................................... 40

1

Abstract Intensification of agricultural systems is a major threat to the associated bio-diversity and could also affect the dynamics of pests and pathogens. One such system that is currently under an intensification trajectory is the production of Arabica coffee. In this thesis, I studied the relationships between fungal dis-eases and their natural enemies, the genetic variation in coffee, coffee yield and associated biodiversity along a coffee management gradient in southwest-ern Ethiopia.

The specific goals of this thesis were to investigate variation in fungal dis-eases on coffee and their natural enemies along a gradient of management (I, II), how genetic variation in coffee among sites relate to variation in incidence of the fungal diseases (III), and to investigate the trade-offs in biodiversity-yield relationships along the gradient of coffee management (IV). To answer these questions, I selected 60 sites along a gradient of management that ranged from coffee naturally growing in only little disturbed forests to intensively managed plantations. I used both observational studies and molecular ap-proaches.

In paper I, I examined if the severity of the four major fungal diseases on coffee varied along the gradient and assessed the main drivers of variation in disease severity. I found that two of the fungal diseases were more severe in the intensively managed coffee sites, while the other two were more severe in the less intensively managed sites. Altitude was the main driver for the dis-eases, but related in a different way to the different diseases. In paper II, I examined the temporal dynamics in coffee leaf rust-hyperparasite interaction, the biocontrol potential of the hyperparasite and environmental drivers for the two species for three consecutive years during the dry and wet seasons. I found that the rust was more common during the dry season and in managed sites while the hyperparasite was common during the wet season and in sites that were less managed. My results also revealed that higher hyperparasite inci-dence during the wet season resulted in a lower growth rate of the rust during the subsequent dry season. In paper III, I investigated if genetic composition and diversity of coffee sites relate to the incidence of the fungal diseases as-sessed. I found that genetic composition of the coffee stands was linked to the incidence of the four fungal diseases, but genetic diversity among the coffee sites did not relate to the incidence of the diseases. In paper IV, I examined biodiversity-yield trade-offs and shape of the relationships between biodiver-sity and yield along the gradient of management. I found a steep, concave

2

shape initial decline in biodiversity values as coffee yield increased to a certain level, after which a further increase in yield did not have much effect on bio-diversity values.

In conclusion, I found different drivers for the different diseases and for the parasite-hyperparasite interaction. It is difficult to achieve a single man-agement approach that can suit the different pathogen species investigated. High genetic diversity among coffee sites did not reduce disease pressure. While the more complex, less managed sites provide high biodiversity values, and could potentially serve as habitats for natural pest control and in situ con-servation for coffee genetic diversity, the yield gap compared to more inten-sively managed sites was very high. To optimize coffee management and con-servation of biodiversity in these landscapes, there is a need to develop strat-egies whereby the smallholder farmers who depend on coffee and the forest as the main source of livelihood can benefit through for example coffee certi-fication schemes that can pay premium prices for biodiversity-friendly coffee management. Keywords: Armillaria root rot, biodiversity, biodiversity-yield trade-offs, coffee, Coffea arabica, coffee berry disease, coffee leaf rust, coffee wilt dis-ease, coffee yield, genetic composition, genetic diversity, hyperparasite, man-agement intensity gradient, southwestern Ethiopia

3

Introduction

General background Tropical forests are characterized by structural complexity and floristic diver-sity that serves as a habitat for large proportion of the world’s biodiversity (Gardner et al., 2010; Mittermeier et al., 1998; Myers et al., 2000). While gen-erally there is a biodiversity loss globally, the major threats to biodiversity occur in these areas due to anthropogenic influences through deforestation, forest degradation and intensification of agriculture to meet the food require-ments of the growing population (Bradshaw et al., 2009; Ellis et al., 2010; Tilman et al., 2002). Forest simplification for agriculture could increase pro-duction but involves negative consequences on the maintenance of ecosystem services and functions in the long term (Foley et al., 2011) which again affects the society (Cardinale et al., 2012). Ecosystem services associated with di-verse systems include crop pollination by insects (Berecha et al., 2015, 2014; Samnegård et al., 2014), natural pest control (Power, 2010) or soil nutrient improvement (Perfecto and Vandermeer, 2015). The complexity in forest sys-tems could also be at a disadvantage to people when for example forests serve as habitats for crop raiding animals (Ango et al., 2014; Lemessa et al., 2013), herbivores (Zhang et al., 2007) or plant diseases (Avelino et al., 2018). In most parts of the tropics the livelihoods of inhabitants close to natural forests de-pends on what they obtain from these forests. Nevertheless, productivity in natural forests is generally low (Perfecto et al., 2005; Schmitt et al., 2010), pushing them to somehow manage these habitats to improve their livelihood. Since more than 90% of the tropical biodiversity occurs in human modified landscapes (Chazdon et al., 2009), management approaches that can reconcile biodiversity conservation and enhance livelihood of the community are ur-gently needed for effective management of human impacts on forests (Gard-ner et al., 2009; Harvey et al., 2008; Perfecto and Vandermeer, 2008).

One such strategy is different types of agroforestry practices that is increas-ingly recognized as a promising approach for combining goals of enhancing the livelihoods for resource poor farmers and conservation of biodiversity (Bhagwat et al., 2008; Schroth et al., 2004). Agroforestry involves intentional management of crops with diverse shade trees (Schroth et al., 2004). Coffee and cocoa are the main agroforestry crops while vanilla, rubber tree and oil palm among others are also grown under shade of trees (Schroth et al., 2004).

4

While mostly agroforestry systems involve a land use change from a previ-ously forested land through thinning and selective felling of trees and keeping the preferred ones, it could also involve a potential reforestation of an open land (e.g., Martin et al., 2020). Agroforestry systems are important in that larger proportion of the biodiversity is conserved compared to under more drastic land transformations (Bhagwat et al., 2008; De Beenhouwer et al., 2013; Harvey and González Villalobos, 2007; Hylander et al., 2013), while it should also be noted that most of forest specialist species could be lost in ag-roforests (Harvey and González Villalobos, 2007; Senbeta and Denich, 2006). Agroforestry systems could also serve as corridors or a conductive matrix for dispersal of fauna and flora across landscapes (Bhagwat et al., 2008; Perfecto and Vandermeer, 2008). Agroforestry systems are expected to vary in terms of their floristic complexity, for example based on management intensity (Geeraert et al., 2019; Hundera et al., 2013; Jose, 2012). Thus there could also be a trade-off between yield and biodiversity values (Aerts et al., 2011; Per-fecto et al., 2005). Yet, shade adapted crops such as coffee need some level of shade for optimum production (Soto-Pinto et al., 2000; Vaast et al., 2006) as they suffer from over-bearing and dieback if subjected to the open sun (Beer et al., 1998; DaMatta et al., 2008). Moreover, coffee and cocoa benefit from the shade as it balances the microclimate (De Beenhouwer et al., 2016; Lin, 2007), which is essential for moisture control (Zhang et al., 2007). Therefore, we need to manage agroforestry systems in ways that could optimize both bi-odiversity and long-term benefits for smallholder farmers without endanger-ing natural habitats.

Due to the homogeneity of production in intensified systems and lack of natural pest control, pests and disease damages could be expected to be higher in more simplified systems than in agroforestry and natural agroecosystems (Bianchi et al., 2006). In intensive agriculture, pests and diseases can be con-trolled by different practices like crop rotation with non-host crops, fallow periods, tillage and burning practices and pesticide application (Avelino et al., 2011; Schroth et al., 2000). On the other hand, this could be a problem in agroforestry systems, where perennial crops dominate. In simplified agroeco-systems, crops might be vulnerable to pests and diseases due to the loss of some beneficial biodiversity as compared to natural ecosystems. In natural ecosystems, plants and diseases infecting them could co-evolve together over a longer period of time (Avelino et al., 2011), and this process could result in lower disease risks (Bianchi et al., 2006; Cheatham et al., 2009; Tilman et al., 2002). Natural ecosystems are also habitats for several beneficial organisms for pest and disease control, but the interaction of the different organisms with

5

the environment and with each other is less clearly understood (Avelino et al., 2011; Cheatham et al., 2009).

Moreover, a management that can help to reduce one disease can have a negative effect on the other (Avelino et al., 2018, 2011; Staver et al., 2001). For example, shade is suggested to reduce coffee berry disease intensity in Cameroon through decreasing splash dispersal (Bedimo et al., 2008). On the other hand, shade buffers the temperature and provides humid microclimate close to the coffee shrubs; a condition that some disease causal organisms like Hemileia vastatrix require for successful germination and infection processes (Avelino et al., 2006, 2004; Staver et al., 2001). However, there is a contro-versy over the effect of shade management on coffee leaf rust development. Some authors reported higher incidence of coffee leaf rust in shaded coffee systems (Staver et al., 2001), while others showed a decline in coffee leaf rust incidence with shade complexity (Chala et al., 2010; Soto-Pinto et al., 2002). There could be several mechanisms in which shade could reduce coffee leaf rust. Shade creates moist microclimate conducive for potential biocontrol organisms (Staver et al., 2001); shade decreases yield and reduces stress on coffee leaves which reduces rust infection (Avelino et al., 2004) and shade trees might reduce wind speed and probably block different wind-borne path-ogens such as coffee leaf rust (Beer et al., 1998; Schroth et al., 2000; Staver et al., 2001).

While most plants are attacked by several pests and diseases at the same time or simultaneously, the majority of studies on pests and diseases focus on the dynamics of a single disease at a time (but see Avelino et al., 2018) and decisions for management of the diseases based on studies on a single disease could be misleading. In a landscape where the host occurs as a native plant in the wild and in managed agricultural areas, it is important to study the dynam-ics of several co-occurring diseases on a host to understand their combined effects. Such landscapes are also important to understand the interaction be-tween diseases and their natural enemies, how genetic variation in the host plant relates to the disease pressures, and to design ecological approaches for disease management. Moreover, smallholder farmers want to get enough rev-enues from their small plots of land to improve their livelihood. Studies on the trade-offs between biodiversity and yield are important to address the yield reductions associated with biodiversity-friendly production systems. Such studies are necessary to make appropriate conservation decisions that can en-hance both the livelihood of the community and biodiversity.

6

Coffee production and biodiversity conservation Across the world, coffee grows in diverse settings that range from complex semi-natural forests or rustic coffee to ‘sun coffee’ where coffee grows with no overhead shade (Moguel and Toledo, 1999; Perfecto et al., 2005). Natu-rally, coffee is a shade adapted crop that needs some level of shade for opti-mum performance, long-term productivity and resilience to climatic effects such as extreme temperatures (DaMatta et al., 2008; Lin, 2007). Traditional coffee production has been highlighted to have a positive value for biodiver-sity as the crop grows under a diverse shade tree canopy (Perfecto et al., 2005, 2003, 1996; Philpott et al., 2008; Philpott and Dietsch, 2003). Coffee is one of the major cash crops for the tropical world and it is mainly cultivated by smallholder farmers. Arabica coffee, Coffee arabica is native to the understo-rey of the Afromontane forests of southwestern Ethiopia (Anthony et al., 2002; Tesfaye et al., 2014). In Ethiopia, coffee is exclusively shade grown but the level of shade varies under the different management types. The shade requirement of coffee has a strong positive impact on the biodiversity conser-vation in the forest patches as it actually limits further deforestation and con-version to annual crop agriculture (Gove et al., 2008; Hylander et al., 2013). In my study area, coffee grows across a gradient of management that ranges from natural forests, where only ripe coffee berries and other spices can be collected by farmers, to intensively managed commercial plantations. In Ethi-opia, the largest proportion of production of coffee is from smallholder farm-ers, who on average grow coffee on less than 1 ha of land, each under different intensities of management and with various combinations of shade trees. On the other hand, coffee from the forest and commercial plantations each ac-counts only for ca. 5% of the nation’s production (Labouisse et al., 2008). Coffee significantly contributes to the livelihoods of an estimated 15 million people and thus it is an integral part of Ethiopia’s economy (Petit, 2007).

Fungal diseases on coffee Coffee production across the world is challenged by several insect pests and diseases (Avelino et al., 2018). The major ones include coffee leaf rust, caused by Hemileia vastatrix (Avelino et al., 2015; Hindorf and Omondi, 2011), cof-fee berry borer, Hypothenemus hampei (Jaramillo et al., 2011; Vega et al., 2015), coffee berry disease, caused by Colletotrichum kahawae Waller and Bridge (Waller et al., 1993), and coffee wilt disease, caused by Giberella

7

xylarioides (Fusarium xylarioides) (Girma et al., 2009, 2001). Among the fun-gal diseases, coffee leaf rust is by far the most important coffee disease glob-ally with a worldwide distribution (McCook, 2006), while coffee berry disease and coffee wilt disease are limited to the African continent (Avelino et al., 2018). Coffee leaf rust is specific to coffee leaves and is characterized by or-ange powdery spores on the lower side of the coffee leaves (Fig. 1a). The urediospores of the rust are distributed by wind and coffee workers. Heavy leaf defoliation due to severe infestation by the rust could lead to secondary yield losses (Cerda et al., 2017).

In the Ethiopian coffee landscape, where most of the farmers grow land-races, coffee berry disease is the most important fungal disease. The causal pathogen can infect flower buds, fruits and maturing bark (Hindorf and Omondi 2011; Bedimo et al., 2008). The disease symptoms are visible when young developing coffee berries start to develop black, sunken lesions and heavy infections lead to mummified beans (Fig. 1b) and can result in a total loss in yield. The disease is most common in high altitude areas (Hindorf and Omondi, 2011) and it is aggravated by extended showers of rain (Bedimo et al., 2010). Conidia of C. kahawae are sticky and the spores need droplets of water such as rain splashes or contact to liberate and disperse them. Coffee wilt disease is another fungal pathogen that can do great damage to the coffee shrubs as the fungus kills the infected shrubs within a few weeks. It is a vas-cular disease that blocks the xylem vessels so that the shrubs are deprived of water (Girma et al., 2009; Hakiza et al., 2009). Symptoms of the wilt disease are observable on dying or fresh dead stems as blue-black discolouration un-der the bark of the stems (Fig. 1c). Towards the end of rainy season, the fungal fruiting bodies (ascospores) are also visible as clusters on the cracked stem (Fig. 1d, Girma et al., 2001). Another pathogen that is sporadic in occurrence but damaging is the root rot caused by Armillaria mellea. A coffee shrub in-fected with Armillaria root rot shows symptoms of wilting and a big crack at the collar region, which extends higher up the stem and with unique black rhizomorph growth in the cracked stem (Fig. 1e). In the field, a shrub that died as a result of coffee wilt disease and Armillaria root rot are easy to differenti-ate as coffee shrubs that died as a result of coffee wilt disease will remain firm in the soil when pushed while shrubs which died as a result of Armillaria root rot can easily be felled when pushed aside. Occasionally, the fungal fruiting bodies (mushrooms) of Armillaria develop in the cracked stem (Fig. 1f). Both coffee wilt disease and Armillaria root rot spread through contacts and man-agement practices like hoeing and weeding, and movement of the infected

8

coffee shrubs between different sites can disperse them over wider areas. Spe-cifically, coffee wilt disease is transmitted mainly through wounds to the stem and coffee shrub damage by machete during weeding and pruning of healthy coffee shrubs with the same scissors after infected ones can spread the disease.

Fig. 1 Major fungal diseases of coffee along the gradient of management in south-western Ethiopia. Top left (coffee leaf rust), top middle (coffee berry disease), top right (coffee wilt disease), bottom left (perithecia that bear the ascospores of coffee wilt disease), bottom middle (Armillaria root rot with black rhizomorph in the cracked stem), and bottom right (fruiting bodies of Armillaria root rot sprouting from the cracked coffee stem).

9

Aims of the study The main aim of this thesis was to understand the variability of ecological conditions in sites that differ in the intensity of human management and how this variability has affected the distribution of plant diseases, interactions be-tween pathogen species and the general biodiversity patterns and relationships to yield in a unique landscape in southwestern Ethiopia. I focus on three main aims.

First, I wanted to explore the dynamics of the four major fungal diseases on coffee, how the diseases respond to the different environmental and man-agement drivers and the co-occurrence patterns of pathogen species along the gradient of coffee management (I). Coffee leaf rust damage is generally low in the native coffee forests of southwestern Ethiopia and one probable reason is the co-occurrence of the rust and hyperparasite. Given that coffee leaf rust is a polycyclic disease, I wanted to explore the temporal dynamics of the rust and the hyperparasite over several years during contrasting seasons, their main environmental drivers and the biocontrol potential of the hyperparasite on the rust (II). Moreover, my study landscape includes coffee managed with differ-ent intensities and one might expect genetic variation in coffee populations along the gradient of management. I wanted to explore variation in genetic composition and genetic diversity of coffee in the different sites and whether that variation was related to the severity of the major fungal diseases on coffee or not (III).

Second, I aimed to explore the relationship between biodiversity and coffee yield along the gradient of management and assessed if I could find a proxy for yield that could represent yield in biodiversity-yield studies (IV). My study sites included a wide management gradient ranging from near-natural forest with extreme floristic diversity where coffee was not much managed but har-vested from a few coffee shrubs with berries to commercial plantations where only a few shade trees are maintained, but high yielding and coffee berry dis-ease resistant cultivars are planted and intensively managed. Thus, a trade-off in biodiversity-yield relationship could be inevitable, but I wanted to know the shape of the relationship. Understanding the shape of biodiversity-yield rela-tionship could enable to explore if a win-win situation could be reached where smallholder farmers can increase their yield without much effect on biodiver-sity or if they need to be compensated for the decline in yield in favor of bio-diversity.

10

Third, I aimed to derive management implications for the smallholder farmers for sustainable coffee production and biodiversity conservation (I-IV). How do my studies on different diseases help in the management of the disease? How can we optimize the potential top-down control in the system? What management interventions are needed to make wise use of the genetic variation in coffee for the management of fungal diseases? How do my studies help in harmonizing livelihood and biodiversity conservation in coffee’s na-tive range?

11

Methods

Study area All studies included in this thesis were conducted in Gera and Gomma districts of Jimma zone, southwestern Ethiopia (7°37’ - 7°56’ N; 36°13’ - 36°39’ E, Fig. 2a and b). The altitude of the study sites ranges between 1500 and 2200 m a.s.l. The area is characterized by a mosaic, heterogeneous landscape com-posed of large continuous forest areas, small forest patches, open areas with annual agriculture and communal grazing lands, home gardens and residential areas (Lemessa et al., 2013). In the eastern part of the landscape, coffee culti-vation started earlier (McCann, 1995) and the forests are much fragmented into small patches, while in the western part of the landscape large intact forest patches still exist and the inhabitants might have depended only on ripe coffee from the forest systems and cultivation is believed to have started more re-cently (Fig. 2). The floristic composition and density of tree varies much across the landscape. The area is characterized by a unimodal rainfall pattern with the main rainy season between June and September and the dry season between November and February, with occasional showers of rain in the re-maining months. The annual precipitation of the area varies between 1480 and 2150 mm per year, with mean daily minimum and maximum temperatures of 12 °C and 28 °C, respectively (Ethiopia National Meteorological Service Agency, unpubl. document). Coffee has both social and cultural values for the community. It is the major source of income for the smallholder farmers and their livelihood is strongly linked to the crop (Ango et al., 2014).

12

Fig. 2 Study area and plot design. a) Map of Ethiopia where the study site is indicated with a black rectangle. b) Aerial view of the landscape where the study sites are rep-resented with the yellow dots and red triangles showing the 60 plots; the red triangles represent sites from commercial plantations; the dark green area shows little managed forests, the scattered green area shows forest patches and the light green areas show open areas for grazing, annual crop agriculture and home gardens; and the two blue stars show Gomma and Gera districts. c) Shows the layout of the plots at each site where 1-16 indicate the coffee shrubs from which I assessed coffee diseases, took leaf samples for genetic analysis and quantified coffee yield.

13

Coffee management systems In my study area, coffee grows across a wide gradient of management inten-sity (Fig. 3). In the little disturbed forest areas, coffee grows wild under dense canopies of indigenous trees, shrubs, and lianas and these forests are also res-ervoirs for coffee genetic diversity (Gole et al., 2008). The coffee shrubs are from naturally dispersed seeds, and the density of coffee can be very low and scattered or very high when seeds remain in the soil and germinate in bulk. Farmers pick ripe coffee berries and native spices like korarima, Aframomum corrorima, and long pepper, Piper capense. There is not much management in the forest systems, but on forest edges close to their residential areas farmers sometimes cut lianas, small shrubs and partially remove weeds to facilitate movement and harvesting of coffee berries. Coffee shrubs growing under in-tense shade are much thinner and with few productive shoots, often bearing very few berries. The dominant shade trees in the little disturbed forest sys-tems in this landscape include the climax vegetation of the moist Afromontane species such as Syzygium guineense, Schefflera abyssinica, Prunus africana, Pouteria adolfi-friederici, Ficus sur and Olea welwitschii (Friis et al., 2010).

The smallholder farmers coffee plots are in the intermediate level of man-agement. Since coffee management has a long history in this region, farmers cultivate their own landraces, which might have originated from the closest forest. Farmers regularly manage canopy of trees through selective thinning of the shrub and liana layer to open space to plant seedlings and allow light to build a strong stem with large number of fruiting shoots (Aerts et al., 2011; Schmitt et al., 2010; Senbeta and Denich, 2006). Dead coffee shrubs are reg-ularly replaced either by self regenerating seedlings or cultivars in few cases. Slashing of the herbal vegetation is common to avoid competition and usually practiced at least once to facilitate berry harvesting. In the small coffee forest patches, the regular management of shade trees has caused a shift in dominant shade tree species from those dominating in the forest trees, to other species, that still are native in the landscape, but have more preferred shade properties such as Albizia spp., Acacia abyssinica, Millettia ferruginea, Cordia africana or fast growing early successional species such as Croton macrostachyus (Aerts et al., 2011; Hundera et al., 2013). At the smallholder level, coffee pro-duction in the area is de facto organic, since herbicides, pesticides and ferti-lizers are not commonly used.

14

Fig. 3 Coffee management intensity gradient. Left to right shows management inten-sity gradient from little managed (left), intermediate management (middle) and highly managed (right). The top row shows canopy cover photos and the middle row shows coffee shrubs structure photos taken from sites along the gradient of management; the bottom row shows schematic diagram of management gradient of coffee.

The most intensively managed sites, i.e., plantation coffee systems and few

of the smallholder sites are characterized by low diversity and density of shade trees. In the plantation coffee systems, only shade trees with some desirable shade attributes such as wide spreading branching habit, less defoliation dur-ing the dry season, fixing nitrogen etc. are maintained or planted with pre-ferred spacing. Coffee cultivars that are resistant to coffee berry disease and high yielding are exclusively grown in the plantation systems. Management practices like fertilization, pruning, weeding or herbicide application and thin-ning of the canopy are practiced at regular basis while pesticides are not com-monly applied for control of pests and diseases (Labouisse et al., 2008).

Most often, a terminology in which coffee production in southwestern Ethiopia is categorized based on management intensity is used in the literature, i.e. forest coffee, semi-forest coffee, semi-plantation coffee, garden coffee and plantation coffee. Nevertheless, there are no clear boundaries for this classifi-cation. Our plots are selected from a landscape with a clear continuous gradi-ent in attributes such as coffee structure and tree layer composition, and we

15

therefore chose to regard management intensity as a continuous gradient. I developed a continuous variable, coffee structure index, that can capture the management intensity gradient of the sites (see the details below).

Study sites and surveying The data for papers I, II and IV are entirely based on field assessments while data for paper III includes field assessments and laboratory analysis of coffee leaves for genetic composition and diversity using genotyping-by-sequencing (GBS) approach. For the field surveys and samplings, I selected 60 sites that varied in terms of their shade tree density, diversity and canopy layers and intensity of management across the landscape (Fig. 2b). At each site, I made a plot of 50 × 50 m and further divided it into 10 × 10 m grids. At the inter-sections of each grid cell, I selected and marked 16 coffee shrubs from the central 30 × 30 m area (Fig. 2c). I recorded the GPS readings at the center and four corners of each plot and revisited the sites several times for all the studies. The field samplings were conducted between March 2017 and August 2019.

Environment and management variables To understand drivers of the different fungal diseases, I measured several en-vironmental and management related variables in each site. These include: i) altitude, ii) slope, iii) canopy cover, iv) number of large shade trees (>20 cm in diameter at breast height, DBH), v) coffee density as a count of coffee shrubs >1.5 m height, vi) proportion of non-coffee shrubs to coffee, vii) coffee dominance as a proportion of coffee to other woody shrubs, and viii) coffee structure index.

Coffee structure index was created from several measurements on the 16 coffee shrubs at each site to reflect the characteristic growth pattern of coffee shrubs subjected to the different intensity of management. First, from each of the 16 coffee shrubs per site, I measured i) the number of primary and second-ary orthotropic (vegetative or vertical leading shoots) and plagiotropic (hori-zontal fruiting shoots), ii) stem diameter at knee height, iii) average of two perpendicular diameters of the ground projection of canopy of coffee shrub, iv) proportion of the coffee height with plagiotropic branches, and v) height of the coffee shrub. Then, I ran a cluster analysis with K-means clustering algorithm with three groups. Cluster 1 contained coffee shrubs with narrow

16

canopy diameter, low number of bearing branches usually towards the tip of main stem which is typical characteristic for coffee in the least managed part of the gradient; cluster 2 contained coffee shrubs with many coarse primary stems due to the free growth as a result of a more open canopy which is a typical characteristic of coffee that has been under intermediate level manage-ment for a long period of time; and the typical coffee shrub in cluster 3 was characterized by many fruiting shoots often occurring more proportionally across the height of the shrub, which is common in intensively managed coffee plantations. Finally, for each site I calculated an index value to reflect the management intensity gradient by i) multiplying the number of shrubs in: clus-ter 1 by 1, cluster 2 by 2, and cluster 3 by 3; and ii) summing these values and dividing it by the total number of shrubs in a plot (n = 16, with a few excep-tions). I hereafter refer to this site level index as ‘Coffee Structure Index’ in which the smaller values indicate the only little managed forest sites and higher values denote sites with more intensive management (Fig. 4).

Fig. 4 Schematic diagram showing coffee shrub structure along gradient of manage-ment. Typical coffee shrub structure in the only little managed forest (a), with inter-mediate management (b) and intensively managed coffee systems (c).

17

Major fungal diseases on coffee along the gradient of management (Paper I) Fungal diseases are the major problem in coffee production in Ethiopia. To investigate the severity of the fungal diseases along the gradient of manage-ment, I assessed the four major fungal diseases of coffee (coffee leaf rust, cof-fee berry disease, coffee wilt disease and Armillaria root rot) at each of the 60 sites. Coffee leaf rust was surveyed during the dry season when the disease was most severe while the other three diseases were surveyed during the wet season. Coffee berry disease infection is peak during the rainy season when the berries are at developmental stages before maturity. Coffee wilt disease and Armillaria root rot produce clear disease symptoms during the rainy sea-son facilitating an easy diagnosis of these diseases in the field. For coffee leaf rust, I counted the total number of leaves and leaves infected with the rust on three branches from the bottom, middle and upper parts of the shrubs on the 16 coffee shrubs per site. For coffee berry disease, I selected three representa-tive fruiting branches per shrub and counted the total number of berries and berries showing coffee berry disease symptoms. I also counted the total num-ber of fruit bearing branches showing coffee berry disease infection on each of the 16 coffee shrubs per site. I counted the total number of coffee shrubs and the number of coffee shrubs showing coffee wilt disease and Armillaria root rot symptoms within the 50 × 50 m plot. I used incidence and/or severity of the four fungal diseases as response variables and environmental and man-agement variables (altitude, canopy cover, number of larger shade trees, cof-fee density, proportion of non-coffee shrubs to coffee, slope, and coffee struc-ture index) as explanatory variables to evaluate if the variation in severity of the fungal diseases differed along the gradient of management and to identify the major drivers for the variation in the disease severity.

Temporal variation and biocontrol potential of a parasite–hyperparasite interaction on coffee (Paper II) To investigate the temporal dynamics in the rust-hyperparasite interaction, and potential hyperparasitism, I assessed coffee leaf rust and its hyperparasite (Fig. 5) for three consecutive years, twice a year (during the dry and wet seasons), following the same procedures as for paper I. I counted the total number of leaves on three selected braches per shrub, number of leaves infected with coffee leaf rust, and number of rusted leaves infected with the hyperparasite

18

on the 16 coffee shrubs per site. I also estimated the proportion of leaf area covered by the rust and proportion of the rusted part of the leaves covered by the hyperparasite. To evaluate the temporal variation in the incidence of coffee leaf rust, I modeled incidence of coffee leaf rust as a function of year, season and their interaction. Similarly, to evaluate the temporal variation in the hy-perparasite incidence, I modeled incidence of the hyperparasite as a function of year, season and their interaction. To understand the drivers for the rust incidence, I modeled the incidence of the rust as a function of environmental and management variables (altitude, canopy cover, number of larger shade trees, coffee density and coffee structure index), in separate model for each season during the three years. To understand the effect of the hyperparasite on the rust, I included hyperparasite incidence or index to the reduced model and re-run the model. Similarly, to identify drivers for the hyperparasite incidence, I modeled incidence of the hyperparasite as a function of the five environmen-tal and management variables. To understand if the hyperparasite is affected by the incidence of the rust, I included the rust incidence in the model. Fur-thermore, to evaluate the biocontrol potential of the hyperparasite on the rust, I modeled the rust growth rate (calculated as rust incidence during season t+1 divided by rust incidence during the preceding season (t)) as a function of hyperparasite incidence during season t. I included the preceding season’s rust incidence as a covariate.

Fig. 5 Coffee leaf rust caused by Hemileia vastatrix (yellow spores) and the hyper-parasite Lecanicillium lecanii (white spores).

19

The relationship between genetic variation and the incidence of major fungal diseases on coffee (Paper III) To explore how the genetic composition and diversity of coffee relate to the incidence of fungal disease, I used the same data on fungal diseases but with one additional year data on coffee leaf rust and coffee berry disease. From the same coffee shrubs that I used for assessing the fungal diseases, I collected leaf samples for DNA extraction. A piece of silica gel dried leaf material from each of the 16 coffee shrubs per site was pooled together to form a population (coffee stand). The molecular analysis involved many steps, but in short DNA was extracted from the ground pooled leaf samples using an optimized cetyltrimethylammonium bromide (CTAB) for GBS libraries preparation. GBS library replicates were constructed for each pool and equimolar quanti-ties of each GBS libraries were pooled, bead-purified and sent for sequencing. Then, GBS read data were evaluated and further processed. Next, reads were mapped onto the reference genome sequence of Coffea canephora Pierre ex A.Froehner. C. canephora is one of the two parents of the allotetraploid spe-cies C. arabica (2n = 4x = 44), the other parent being C. eugenoides (Tesfaye et al., 2014). Afterwards, single nucleotide polymorphic variants were called from the reads. Genetic differentiation between pools was quantified as FST

following (Nei and Chesser, 1983) and the expected heterozygosity on each variant position was calculated as HE = 2*RAF*(1-RAF), where HE refers to the expected heterozygosity and RAF refers to reference allele frequency. Mean expected heterozygosity was calculated for the coffee stands as a robust measure of genetic diversity. The full methods for the molecular work are de-tailed in the methods section in paper III.

To explore if the genetic composition of coffee stands varied across the landscape, I performed an indirect ordination with principal component anal-ysis (PCA) on the site-by-allele frequency matrix of the Hellinger transformed RAF values. To determine which environmental variables were related to the genetic composition of the coffee stands, I fitted a direct ordination using re-dundancy analysis (RDA) on the RAF values with the five environmental var-iables (listed above) as constraining variables. To explore if there was a spatial contribution to the genetic composition of coffee stands, I created Moran’s Eigenvector Maps (MEMs). I used mean expected heterozygosity as a re-sponse variable and environmental variables as predictors to investigate if ge-netic diversity of coffee varied along the management gradient. To assess the relationship between genetic composition of the coffee stands and the inci-

20

dence of the major fungal diseases, I used incidence of the diseases as a re-sponse variable and the first three PCA axis scores of the genetic composition as explanatory variables along with other environmental variables to fit a gen-eralised linear mixed effects model. Similarly, to explore if the incidence of the fungal diseases was related to the genetic diversity of coffee among the sites, I fitted a linear model with incidence of the diseases as a response vari-able and the first three PCA axis scores of the genetic composition as explan-atory variables along with other environmental variables. I assessed coffee leaf rust and berry disease at coffee shrub level and I was motivated to explore if the among shrub variation in the incidence of the two diseases (standard deviation of the incidence of the diseases) was related to the genetic diversity among coffee sites. For this, I fitted a linear model with standard deviation of the incidence of the two diseases as response variable and the mean expected heterozygosity as explanatory variable.

Trade-offs in biodiversity-yield relationships (Paper IV) To explore the trade-offs in biodiversity and yield as well as the shape of the relationship, I assessed several biodiversity components and estimated clean coffee yield. For the biodiversity components I assessed woody plants (trees, shrubs, lianas), non-woody vascular plants (herbaceous vegetation and epi-phytic plants) and non-vascular plants (bryophytes i.e., mosses and liverworts) at each of the 60 coffee sites. I also measured the yield of coffee for three successive years in the same sites. I also quantified coffee density, coffee dom-inance and coffee structure index in these sites to explore if they could be used as a good proxy for coffee yield. To evaluate the relationship between yield and biodiversity, I used biodiversity values (species richness of woody plants, non-woody vascular plants, bryophytes and the total) as response variables and average coffee yield as explanatory variable. I fitted generalized additive model (GAM) to explore the shape of the relationship. I did the same for the species composition of the different biodiversity values, but with a slightly different approach. First, I ran canonical redundancy analysis (RDA) on the species composition of woody plants, non-woody plants, bryophytes and the total species list) with average coffee yield as a constraining variable. Second, I used the RDA axis scores for each of the four ordination of species compo-sition groups as response variable and average coffee yield as explanatory var-iable to fit GAM models. To explore if the biodiversity-yield relationship was consistent during the different years, I fitted a similar species richness and

21

composition model for each year separately. Finally, to explore a proxy for yield that can capture the biodiversity-yield relationships, I chose three coffee management related variables (coffee structure index, coffee density and cof-fee dominance) and performed Pearson’s product-moment correlation test be-tween these variables and average coffee yield. I chose the variable that had higher correlation coefficient with average yield as a proxy for yield. I fitted the same model for the species richness and composition as above, but with the yield proxy (coffee structure index) as explanatory variable instead of the average coffee yield. This yield proxy could be used in place of average coffee yield to establish biodiversity-yield relationships when yield data from several years is lacking.

22

Results and discussion

Fungal disease dynamics, drivers of the diseases, temporal variation, and biocontrol potential of the rust hyperparasite The severity of the four fungal diseases varied across the landscape, with dif-ferent drivers for the different diseases. Coffee leaf rust and Armillaria root rot were severe in the intensively managed part of the gradient while coffee berry disease and coffee wilt disease were severe in the least managed part of the gradient (I). Altitude, as a proxy for climate (e.g. temperature and relative humidity) was the main driver for the variation in the severity of the fungal diseases (I, II, III), but the diseases responded differently (Fig. 6). For exam-ple, coffee leaf rust was severe towards the low altitude sites, while the other three diseases and the rust-hyperparasite were more common at the other end of the altitudinal gradient (I, II, III). At higher altitudes, low night tempera-tures are suggested to increase the latent period of the rust and thereby limiting its epidemic development (Chala et al., 2010; Waller, 1982). On the other hand, the relatively higher rainfall conditions at higher altitudes modify the microclimate in favor of coffee berry disease, since the morning rise and the afternoon fall in temperature are less abrupt compared to at lower altitudes providing sufficiently long time for invasion (Nutman, 1970).

Coffee wilt disease was previously reported to be a major problem in in-tensively managed plantations at lower altitudes and in garden coffee systems (Girma et al., 2001). However, with active forest management and movement of planting materials and farm tools between different coffee farms, there is a high chance for the wide and fast spread of the disease to more natural systems (Getachew et al., 2013, 2012). Armillaria root rot severity at higher altitudes could be related to the preference of the causal agent for relatively cool soil conditions (Gezahgne et al., 2004; Otieno et al., 2003), which is expected at higher altitude sites. At site level, the severity of the four diseases did not co-vary even though three of the diseases were severe at higher altitudes. The diseases might have specific niche requirements besides altitude. However, at coffee shrub level (within sites), shrubs with high level of coffee berry disease also had a higher rust severity and vise versa (I). This could suggest that weak-ened host defenses could predispose it for infection with another pathogen (Singer, 2010). Coffee leaf rust was reported to be severe on stressed plants for example due to high fruit load (Avelino et al., 2004; Eskes, 1983) and

23

infection by diseases such as the American leaf spot caused by Mycena citri-color (Allinne et al., 2016; Avelino et al., 2018).

Fig. 6 Relationship between a) coffee leaf rust, b) coffee berry disease, c) coffee wilt disease and d) Armillaria root rot and altitude. Each dot represents mean severity of the diseases at each of the 60 sites.

The incidence of coffee leaf rust strongly varied between years, while the hyperparasite showed less clear variation between years (II). This could be due to the variation in environmental drivers of the rust during the different seasons. The hyperparasite might be even more strict in its environmental re-quirements as it also depends on the rust prevalence as the only energy source in this system. The two interacting species also had higher incidences during contrasting seasons. The rust was more common during the dry season while the hyperparasite was common during the wet season (II). The rust develop-ment starts with the showers of rain at the start of rainy season, germination and infection processes continue during the wet season without much visible

24

disease symptoms unless the infection process started very early. Thus, to-wards the start of the dry season the leaves could show severe rust attack (Chala et al., 2010; Daba et al., 2019; Garedew et al., 2019). Heavily infected leaves drop-off the coffee shrub, subsequently reducing disease inoculum car-ryover to the next wet season. Lack of moisture during the dry season also prevents further germination and infection process and result in lower rust during the next wet season (Avelino et al., 2004; Garedew et al., 2019; Waller, 1982). However, this could slightly be different towards shade tree rich, higher altitude sites where infection is often less during the dry season and infected leaves could remain attached. This rust spores may continue infection due to availability of moist conditions at higher altitudes to result in slightly higher rust incidence during the wet season. The hyperparasite prefers moist condi-tions (Eskes et al., 1991; Staver et al., 2001), and the lower infection level at higher altitudes that does not lead to leaf drop could be at an advantage for the hyperparasite to coexist. On the other hand, due to shortage of moisture the hyperparasite might fail to establish on the rust spores at the lower altitudes during the dry season when the rust is more severe.

Hyperparasite incidence during the wet season was negatively related to the rust growth rate during the subsequent dry season implying a top-down control of the rust by the hyperparasite (II, Fig. 7). This time lag could be related to the hyperparasite development as coffee leaf rust is the only resource for the hyperparasite in this landscape and it develops on the rust that has al-ready well established. As the dry season starts, the rust might need only little moisture such as morning dew to continue infection (Hindorf and Omondi, 2011), while moisture could perhaps be limiting for the hyperparasite (Eskes et al., 1991). Unlike my study system, in the coffee systems in Mexico, the hyperparasite has another primary host, the green coffee scale Coccus viridis on which it can build a high population as an inoculum source to infect coffee leaf rust (Vandermeer et al., 2009). Interestingly, Vandermeer et al. (2009) found a minor effect of the hyperparasite on the rust in the same year but Jack-son et al. (2012) found a one-year time lag for a strong top-down control by the hyperparasite where the hyperparasite has to develop a population that can effectively control on the rust during the next season before the rust develops a local epidemics.

Apart from the environmental and management variables (I, II), variation in the incidence of the fungal diseases was also related to the genetic compo-sition of the coffee stands, but the relationships were unique for the different diseases (III). Specifically, the relationship between the severity of coffee berry disease and genetic composition of coffee was stronger. This could be

25

attributed to the known variation in resistance of coffee shrubs to coffee berry disease across the landscape (Labouisse et al., 2008). However, the genetic diversity among the coffee sites did not relate to the severity of the fungal diseases across the landscape. I found a higher among shrub variation in coffee leaf rust severity compared to among sites (I), but this among coffee shrub variation did not relate to the genetic diversity of coffee shrubs (III). In a re-cent study, Daba et al. (2019) showed that cultivars resistant to coffee berry disease did not show resistance to coffee leaf rust. Thus, the among shrub var-iation in coffee leaf rust could mainly be driven by environmental or manage-ment related variations at coffee shrub level. For example, coffee shrubs that are not under direct overhead shade could get more direct rain and spores are washed resulting in less rust compared to shrubs under shade where the shade can break the heavy droplets of rain and spores not get washed away.

Fig. 7 The relationship between coffee leaf rust growth rate and hyperparasite index during the preceding season. Black dots represent the log-transformed growth rates of coffee leaf rust for each of the 60 sites. Red solid line is the fitted regression line and the grey shaded region shows the 95% confidence interval around the fitted line. The blue broken line indicates zero growth.

On the other hand, I found a higher variation in coffee berry disease sever-ity among sites compared to within sites (I), which could be attributed to var-iation in coffee shrubs to coffee berry disease due to some farmers and plan-tation coffee growing coffee berry disease resistant cultivars (Benti et al.,

26

2020; Labouisse et al., 2008). Alternatively, coffee berry disease spores are splash-borne and do not disperse over wider distances unless other vectors like humans are involved (Bedimo et al., 2008; Hindorf and Omondi, 2011), unlike the coffee leaf rust spores which are air-borne and could disperse over a rela-tively wider area. Among coffee shrub variation in coffee berry disease was related to the genetic diversity of coffee (III) which could apparently be re-lated to the fact that most smallholder farmers grow susceptible and resistant cultivars in combination resulting in higher among shrub variation in coffee berry disease incidence.

Biodiversity yield trade-offs

In the 60 sites across the landscape, I found 407 plant species belonging to woody plants (71), non-woody plants (243) and bryophytes (93). Woody plant species richness decreased with an increase in coffee yield with a concave non-linear relationship until ca. 750 kg ha-1 where it leveled off (Fig. 8). The species richness of non-woody plants, bryophytes and the total did not have a relationship with average coffee yield. However, the species composition of all the species groups changed across the yield gradient, with a similar pattern to the species richness of woody plants. The trade-off in biodiversity-yield relationship was less clear when related to the coffee yield from the three years separately, compared to the average yield of the three years. My finding indi-cates that we need to be cautious when establishing biodiversity-yield rela-tionships using yield data from a single year. This is even more important in crops like coffee that show biennial nature in yield. In such instances, re-searchers often search for simple proxies for yield that can be used to establish reliable biodiversity-yield relationships in place of yield. I found Coffee Struc-ture Index to be the best proxy for coffee yield as the relationship between biodiversity and this variable was similar to what I found when the average coffee yield was used.

I found a strong trade-off between woody species richness and coffee yield. Naturally, coffee and cocoa are shade adapted crops that have optimum per-formance under shade (Beer et al., 1998; DaMatta et al., 2008). Nevertheless, when shade is very dense, often corresponding to sites with high biodiversity, yields are very low. Such strong biodiversity and yield trade-offs have been reported for coffee (Jha et al., 2014; Perfecto et al., 2005; Soto-Pinto et al., 2000) and cocoa (Beer et al., 1998; Somarriba et al., 2013; Steffan-Dewenter

27

et al., 2007). On the other hand, when shade levels are low, many forest de-pendent species do not thrive. Some studies, however, show that it is possible to get good yield without much effect on biodiversity (Clough et al., 2011; Jezeer et al., 2017). The strength of the biodiversity-yield trade-off perhaps depends on how wide a gradient in complexity has been under investigation. In this regard, my study included near-natural forest where the floristic diver-sity is very high while coffee is harvested from a few coffee shrubs with ber-ries to intensively managed commercial plantation coffee where cultivars with high yield are grown, resulting in the steep decline of biodiversity with in-creasing yield. Therefore, it is unlikely to get a win-win situation for biodiver-sity and yield in this landscape.

Fig. 8 The relationship between woody species richness and average coffee yield. Each dot represents an average of three years coffee yield (kg ha-1) for each of the 60 sites. Regression slopes with 95% confidence interval from a GAM-model are shown with solid lines (p < 0.001, R2 = 0.30).

The trade-off between biodiversity and yield was stronger for woody spe-cies richness than for non-woody plants and bryophytes. The low yield of cof-fee in the relatively less-disturbed forests is the major driver for the small-holder farmers to slightly reduce the canopy of trees, specifically the small trees, shrubs and lianas (Aerts et al., 2011; Schmitt et al., 2010; Senbeta and

28

Denich, 2006). Such practices help to direct more light to the coffee shrubs so that they develop more strong branches that can support better yield (Aerts et al., 2011). Reduction of trees and shrubs which are in the same height with coffee might reduce competition for space and light but results in structural simplification of the forest (Aerts et al., 2011; Hundera et al., 2013b; Tadesse et al., 2014). Due to the open canopies, sun adapted lower canopy herbaceous vegetation and bryophytes which for example prefer coffee as their host could still flourish (Hylander and Nemomissa, 2009, 2008). This could be the reason why we did not get a negative relationship between species richness of non-woody plants and bryophytes along the yield gradient. However, the species composition of these groups also changed fast along the yield gradient and often the species composition is more important for biodiversity conservation than species richness per se (Aggemyr et al., 2018) as species richness only provides taxonomic diversity.

Implications for management Smallholder coffee farmers are often challenged by crop yield losses due to different pests and diseases. Nevertheless, most studies focus on the dynamics of a certain disease at a time but such single species studies are not enough to make good management decisions. In this thesis, I studied the dynamics of multiple diseases on a single host and their environmental drivers (I), parasite-hyperparasite interaction and potential biological control (II), the relationship between genetic variation in coffee and disease incidences (III), and the trade-offs in biodiversity-yield relationships (IV) along a gradient of coffee man-agement in southwestern Ethiopia. Even though it is impossible to get a single management strategy that can fit the different goals (disease management, bi-odiversity, productivity), based on my findings in the different papers, I pro-vide some insights on management which are specific to the different diseases.

Coffee leaf rust was more severe in the managed sites at lower altitudes (I, II). This could be attributed to the better distributions of the rust spores by coffee workers and suitability of microclimate such as temperature at lower altitudes. Coffee leaf rust also decreased with canopy cover (I). This indicates that coffee shrubs in the lower altitude areas could benefit from a slight in-crease in shade tree density. Shade might buffer the microclimate thereby re-ducing the temperature, which could increase the latent period of the coffee leaf rust pathogen and result in lower rust severity. Increased shade cover might also result in a moist microclimate that is conducive for the potential

29

top-down control of coffee leaf rust by the hyperparasite (II). In intensively managed coffee systems, stress on coffee shrubs due to fruit load is suggested to predispose the shrubs for coffee leaf rust infection (Avelino et al., 2006, 2004). Increased shade results in lower coffee yield, and subsequently reduc-ing stress on coffee shrubs for the rust infection. However, shade is suggested to increase coffee leaf rust (Avelino et al., 2004; Staver et al., 2001), probably because other studies compare the effect of shaded and unshaded systems (e.g. López-Bravo et al., 2012). Whereas, in my study system, coffee is exclusively grown under shade and moisture levels under the low shade areas might be optimum for the rust to thrive even at the lower altitudes. Thus, improvement of shade at the lower altitude areas does not seem to have a positive effect on the rust development in this landscape. I also found that variation in coffee leaf rust incidence had a relationship with genetic composition of coffee (III). This might be helpful for further investigation of valuable genes that might be useful for breeding of the crop for disease resistance.

I found a higher incidence of coffee berry disease in sites with less inten-sive management (I). Association of coffee berry disease with little managed sites might rather be due to the genotype of coffee grown besides altitude ef-fect than the management per se (III). Yet, management can aggravate the problem if the disease prevails in the sites due to the nature of the disease inoculums. Coffee berry disease spores are sticky, and need moisture to liber-ate the spores for dispersal (Bedimo et al., 2008, 2010; Hindorf and Omondi, 2011). In the presence of other vectors like coffee workers or animals, the disease can be dispersed easily within the field. Farmers living close to Gera forest said coffee shrubs which are far away in the forest that are less fre-quently visited by humans and grazing domestic animals are not much af-fected by the coffee berry disease. Similarly, some farmers perceive that slash-ing of weeds during the berry development season can result in high berry disease infection (Zewdie, personal communication). Both conditions are more related to the dispersal of the disease inoculum due to contacts with dif-ferent coffee shrubs, and thus it is advisable to reduce management during the peak times of dispersal of the disease, not least in the natural forest systems. Coffee berry disease is the major challenge for the community as it sometimes causes total yield loss (I). The coffee populations in this landscape show var-iation in susceptibility to the disease (III), perhaps due to the known variation in the resistance of coffee shrubs to coffee berry disease (Labouisse et al., 2008). Thus, highly susceptible landraces could be replaced by resistant coffee cultivars so that farmers obtain some revenues from their small plots of land. However, gradual replacement of the smallholders landraces will reduce the

30

genetic diversity of the crop. As a complementary to the genetic diversity in the forest systems, ex situ conservation of the genetic material of all the land-races across the landscape is very important.

Coffee wilt disease and Armillaria root rot showed lower incidences gen-erally (I) but these two diseases completely kill infected coffee shrubs and could cause considerable yield losses. I found that coffee wilt disease was more common in the little managed forest sites while Armillaria root rot was more common in the more intensively managed sites (I). At its early stage of outbreak, coffee wilt disease was considered as a less important disease in the little management forest and semi-forest coffee systems of Ethiopia (Pieters and van der Graaff, 1980). However, with the current increasing trend of forest management, the disease could be a major cause of death for coffee shrubs in the forest systems and a potential threat to the genetic diversity of coffee in the forest systems. Thus, to save coffee genetic diversity it would be advisable to discourage coffee management in the more natural forest systems. The in-cidence of both coffee wilt disease and Armillaria root rot was related to the genetic composition of coffee stands (III). This highlights the need to look for coffee genotypes that have potential resistance to the these diseases. Coffee wilt disease and Armillaria root rot are distributed by cultural practices like weeding, hoeing and movement of infected coffee shrubs between sites. This indicates that smallholder farmers might need training on the disease symp-toms, dissemination mechanisms and management techniques that can reduce the damage from these diseases. It is also believed that the forest coffee sys-tems have a lot of microorganisms that act as natural pest control organisms in the rhizosphere (Mulaw et al., 2010; Muleta et al., 2007), of which Tricho-derma spp. interact with both coffee wilt disease and Armillaria root rot (Mu-law et al., 2010; Waller et al., 2007). Thus, in the absence of management that can aggravate dissemination of the diseases, there could be a natural balance through potential top-down control in the system.

Another big problem with this landscape is the low yield of coffee in areas with lower management (IV). Since the little managed forest sites have higher biodiversity values of both the plant species (IV) and coffee genetic diversity, it would be valuable to avoid coffee management in these areas. This could be possible through planning for different goals in different parts of the land-scape, whereby the near-natural forest sites can be saved for biodiversity con-servation (of both plants species and coffee genetic diversity) and the other parts of the landscape can be more intensified for a better yield (Phalan et al., 2011). However, coffee production in this landscape is often traditional, inputs like fertilizers and pesticides are not used to improve yield and yield of coffee

31

is generally low even in the smallholder farmers sites with intermediate man-agement. Therefore, coffee certification programs that pay premium prices for the biodiversity-friendly coffee production approaches might help to improve the livelihood of smallholder farmers who depend on sites with low yield in the less intensified coffee sites.

32

Concluding remarks and future perspectives

The main aim of my study was to assess the dynamics of fungal diseases, their drivers and natural enemies, the relationship between genetic variation in cof-fee and disease intensity and biodiversity-yield relationships along the gradi-ent of coffee management in southwestern Ethiopia. I found out that diseases could be driven by the same environmental conditions but might have specific niches. Since crop plants are often affected by several pests and diseases sim-ultaneous, studies involving multiple diseases on a host are valuable in that they will give more insights to make better decision on pests and disease man-agement. It is interesting to note that the rust and the hyperparasite had a dif-ferent specific environmental requirements to optimize the environment for the rust and the hyperparasite co-occurrence. Prevalence of a conducive envi-ronment for the hyperparasite is essential for a potential top-down control of the rust by the hyperparasite. The finding that the different fungal diseases responded differently to the genetic composition of coffee is very important in light of the future improvement of the corp. This study will pave the way for further research that could potentially disentangle which specific genes are responsible for the variation in the diseases and potential use of those genes in coffee breeding. It is often known that biodiversity could trade-off with yield but it was not studied in detail for the coffee systems in Ethiopia. It is interesting to note that there is a strong trade-off between biodiversity and coffee yield in this landscape and that a win-win situation is unlikely. This is very important for smallholder farmers who depend on the forest for their live-lihoods, and coffee certification and conservation organizations to wisely manage the conflict between biodiversity conservation and livelihood en-hancement.

I believe I have shed some light on the main aims of my study with the findings of the four papers included in this thesis. However, I am also left with a lot of questions that are not answered. I will list one main question for the first three paper that I think are novel and could be developed to a new project or are already in progress.

In the first paper of my study, I aimed to understand the environmental and management drivers for the different fungal diseases. I found that altitude, as a proxy for the climate, was the main driver for the fungal diseases. I was eager to understand how specific microclimate variables like temperature and relative humidity relate to the development of the different fungal diseases.

33

For this, I set out data loggers to measure the soil and canopy temperatures and relative humidity of the air under the coffee canopy on several coffee shrubs at each of the sixty sites. This new experiment involved several seasons of data on the microclimate variables and the fungal disease assessments. Cur-rently, a new PhD student Biruk Ayalew is involved as the main investigator of this project.

In the second paper of my study, I investigated the interaction between coffee leaf rust and its hyperparasite and environmental drivers for the two species for three years during the dry and wet seasons. I found that the two species have specific environmental requirements and that the hyperparasite has the potential to control the rust growth rate. My study involved several seasons of the rust and hyperparasite assessment, but it is a one-time assess-ment each dry and wet season. However, I think that it could be much useful to closely follow up on the development of the two interacting species with both manipulative greenhouse and field experiments to understand more about the potential top-down control and specific environmental requirements of the two species. This idea could be developed into a new project in the future.

In the third paper of my study, I investigated the genetic variation in coffee populations from the 60 sites and how that relates to the incidence of fungal diseases assessed from the same sites. I found that the incidence of fungal diseases of coffee is related to the genetic composition of coffee stands. How-ever, in this project, we characterized the neutral genetic variation and did not target any region of the genome related to disease resistance. This finding in-itiated a new project to characterize the genetic variation at the region of the genome that was related to disease resistance characteristics in Arabica coffee in other studies. For this project, we collaborate with researchers from KU Leuven and ILVO in Belgium. We aim to understand:

i. Which sequence variation is present in genomic regions of Arabica coffee related to disease resistance characteristics? and

ii. How is this variation structured along the gradient of coffee management?

In paper IV, I explored the relationship between biodiversity and coffee yield along the gradient of management. I found a strong conflict between biodiversity and coffee yield, and it is unlikely that a win-win solution could be achieved, especially in the little managed forest sites. Future research could be related to policy or economic analysis of the feasibility of different coffee

34

certification programs to understand the possibilities of harmonizing liveli-hood and biodiversity conservation in Arabica coffee’s native region.

************************************************************* I would like to thank Kristoffer Hylander and Ayco Tack for valuable comments and linguistic edits on this summary.

35

Svensk sammanfattning

Intensifiering av jordbrukssystem är ett stort hot mot dess biologiska mångfald och kan också påverka dynamiken hos skadedjur och patogener. Ett sådant system som för närvarande blir mer och mer intensifierat är produktion av Arabica-kaffe. I denna avhandling studerade jag sambandet mellan svampsjukdomar och deras naturliga fiender, den genetiska variationen av kaffe, variationen i kaffeskörd och associerad biologisk mångfald längs en gradient i kaffeodlingsintensitet i sydvästra Etiopien.

De specifika målen för denna avhandling var att undersöka variationen i svampsjukdomar på kaffe och deras naturliga fiender längs en gradient av kaffeodlinglingsintensitet (I, II), hur genetisk variation hos kaffe mellan platser relaterar till variation i förekomst av svampsjukdomarna (III), och att undersöka avvägningarna i förhållandet mellan biologisk mångfald och avkastning längs en gradient av kaffeodlingsintensitet (IV). För att svara på dessa frågor valde jag 60 platser längs en gradient av odlingssystem som sträckte sig från kaffe som naturligt växte i bara lite störda skogar till intensivt skötta plantager. Jag använde både observationsstudier och molekylära metoder.

I kapitel I undersökte jag om skadenivån från de fyra viktigaste svampsjukdomarna på kaffe varierade längs gradienten och vilka som var de viktigaste drivkrafterna för denna variation. Jag fann att två av svampsjukdomarna var allvarligare på de intensivt skötta odlingarna, medan de andra två var mer problematiska på de mindre intensivt skötta platserna. Altitud var den främsta förklarande variablen för variationen hos sjukdomarna, men relationen mellan altitud och skadenivå var olika för de olika sjukdomarna. I kapitel II undersökte jag den tidsmässiga dynamiken för interaktionen mellan sjukdomen kafferost och dess hyperparasit, potentialen för hyperparasiten att kontrollera rostsvampen samt vilka miljövariabler som påverkar de två arterna. Jag gjorde det under tre år i rad både på torrtiden och rengtiden. Jag fann att kafferost var vanligare under torrtiden och på mer intensivt skötta platser medan hyperparasiten var vanligare under regntiden och på mer skogsliknande kaffeodlingar. Mina resultat avslöjade också att mer hyperparasitangrepp under regntiden resulterade i en lägre tillväxttakt för kafferost fram till den efterföljande torrtiden. I kapitel III undersökte jag om den genetiska sammansättningen och mångfalden hos kaffebuskarna på olika

36

platser relaterar till variation i förekomst av svampsjukdomarna. Jag fann att den genetiska sammansättning hos kaffebuskarna på olika platser var kopplad till förekomsten av de fyra svampsjukdomarna, medan variationen av den genetiska mångfalden inte påverkade förekomsten av sjukdomarna. I kapitel IV undersökte jag avvägningar mellan bevarandet av biologisk mångfald och storleken på kaffeavkastningen. Jag fann att den biologisk mångfalden minskade kraftigt i början av kurvan av ökande skördar, varefter en ytterligare ökning av avkastningen inte hade så mycket effekt på den biologiska mångfalden.

Sammanfattningsvis hittade jag olika drivkrafter för olika sjukdomar och för parasit-hyperparasit-interaktionen. Därför är det svårt att föreslå ett enda odlingssystem som skulle minimera problemen från alla arterna. Hög genetisk mångfald minskade inte sjukdomstrycket. De mer komplexa, lite skötta platserna har höga biologiska mångfaldsvärden och kan potentiellt fungera som livsmiljöer för naturlig skadedjursbekämpning och för bevarande av kaffets genetiska mångfald. Dock var avkastningsklyftan jämfört med de mer intensivt skötta platserna mycket hög. För att optimera kaffeodling och bevarande av biologisk mångfald i dessa landskap, finns det ett behov av att utveckla strategier där småbönder som är beroende av kaffe och skogen som den främsta försörjningskällan kan dra nytta av till exempel certifieringssystem som kan betala högre priser för en mer biodiversitetsvänlig kaffeodling.

37

Acknowledgements

First of all, I thank the Almighty God for providing me with strength during the interesting as well as challenging Ph.D journey. I received various support and encouragement from several individuals and organizations during this study. I extend my heartfelt gratitude to all of them. I ask for an excuse be-forehand in case I forget someone.

I will start with my main supervisor Kristoffer Hylander. I could not have words to thank you enough for the extraordinary care, guidance and support you provided right from the start to the end of my Ph.D study. Thank you for carefully following my academic development and for improving my writing skills through your critical, insightful comments and suggestions. Thanks for your inspiring words, even when things are moving slowly. When I felt I am lost, you have always found a way to help me. Apart from science, you also care a lot for me and my family, and I am very grateful for that. Thanks also to you and your wife Eva for all the care and efferts to make me feel at home in Sweden.

Ayco, I must say that I am blessed to have two supervisors who are always there to help me when I needed them. Thank you very much for the critical discussions from the start that helped me to smoothly start the Ph.D study. Thanks also for helping me when I did not feel comfortable with what I am doing. I am always amazed by your insights and the speed in which you give a supper high quality and inspiring comments and suggestion on manuscripts. Thanks also for letting your door always open for questions and ready to help when needed.

My study was financed by the Swedish Research Council (Grant to Kristoffer Hylander), and I am very grateful for that.

Thanks to all fellow Ph.D. students and Postdocs from Kristoffer group Caro-line, Ditte, Irena, Jacqueline, Sonia, and Maria J for the discussions during our weekly meetings. Thanks to Maria for helping with the translation of my sum-mary to the Swedish language. Thanks to Caroline for the nice introduction to ImageJ and for sharing experience during the thesis printing process.

38

Thank you all in Ayco’s microfica group: Pil, Masha, Laura, Alvaro, Biruk, Anna, Etsuko, Niklas, Oliver, Ahmed, Ryan, Hannah, Jessie for our discus-sions during weekly meetings and for the traditional test presentations to help each other. Thanks to Pil and Masha for always being available for help during the start of my study. Thanks to Giada for helping with compiling coffee dis-ease pictures in a nice way that we used for the preparation of T-shirt as an outreach for the coffee farmers. Thank you Biruk for taking care of my plots and the encouragement during writing.

Thank you very much Elsa for being a nice office buddy and answering all my questions, especially for providing help when I am stuck with R.

Thanks to Peter and Sarahi for reviewing my thesis and providing valuable insights on the thesis.

Thank you to all Ekofika groups for the inspiring discussion and insightful comments during presentations. Thanks to all the seniors in the group: Ove, Peter, Johan, and Tanja and PhD sudents I did not mention above: Beate, Ma-tilda, Alicia, Xuyue, David, Juanita and Alejandro. I am very grateful to DEEP for providing a nice working environment for my Ph.D study. I would also like to thank Erica for her kind support with admin-istrative works which I didn’t know much about how to handle and Lars for fixing problems with my computer.

Thanks to Prof. Sileshi Nemomissa for handling the administrative and lo-gistic issues for the fieldwork in Ethiopia and facilitating the material permits from the Ethiopian biodiversity institute. I am indebted to Oromia Forest and Wildlife Enterprise and Coffee and Tea Authority and their respective offices at the district level for providing the permit to work in Gomma and Gera coffee forests. I am also grateful to the Ethiopian Biodiversity Institute for providing export permits for coffee leaves and bryophytes to Sweden for molecular anal-ysis and identification, respectively.

I am also grateful to coffee farmers for allowing me to work in their plots and coffee plantations (Horizon, Abana, and Tracon) for the permit to work in their coffee sites.

I extend my appreciation to our collaborators from KU Leuven (Honnay’s Lab) and ILVO (Tom and Isabel). Specifically, I thank Yves Bawin for his well-organized introduction to the molecular techniques we used for genetic analysis of our coffee samples.

39

It was also interesting to be joined by several masters students at different times for their fieldwork in Ethiopia: Abebe, Oliver, Amelie, Irena, and Tamiru. It has been fun with more people in the group, and I hope you have enjoyed the fieldwork and your stay in Gomma and Gera.

I am also very greatful to my colleagues from Jimma University College of Agriculture and Veterinary medicine for their support and incouragement dur-ing the fieldwork for this study.

I am also grateful for the support I got from my field assistants: Tijani Ibrahim, Raya A/Oli, and Shabu Jemal. Hundi keessanuu galatooma! Melaku Wonda-frash, it has been great to work with you. I have learned some plant species from you during the biodiversity assessment. Thanks to Teklu, Cheru, Ash-enafi, and Kassahun for fun and the nice services, you guys made our field-work enjoyable.

Thank you my friends Tola Gemechu and Abraham Kumsa for sharing your experiences as a PhD student at Stockholm University and on matters related to living with family in Sweden and much more. Maatii keessan waliinuu ga-lata guddaa narraa qabdu. Eebbifamaa!

I extend my heartfelt gratitude to my siblings: Dereje and his family, Meseret and her family, Astu, Fetene, Tsedu, Shimu, and Mule for their kind support and encouragement during my study. You have all played some role for me to succeed, and thank you very much! I am also very grateful to my father-in-law Mr. Solomon Habtamu and mother-in-law Mrs. Kebebush Tilahun for their support and encouragement during my study.

Finally, my special thanks go to my family: my wife Birtukan Solomon, and our lovely kids Naol and Maya. Inat I am deeply grateful for your care, com-mitment, and encouragement to help the family all throughout my study time. This Ph.D study would have been much difficult without your keen support and encouragement. Naol, thank you very much for taking me back to my childhood times, maybe I got the chance to play with you those I didn’t! Maya, you came to this world when I was much occupied with finalizing my thesis and couldn’t give you much time, but your smiles when I come home are charming and refreshing. I got a lot of energy from it and helped me to forget the busy days. Thank you very much Inat, Naol, and Maya for making my life meaningful. I love you so much!

40

References

Aerts, R., Hundera, K., Berecha, G., Gijbels, P., Baeten, M., Van Mechelen,

M., Hermy, M., Muys, B., Honnay, O., 2011. Semi-forest coffee cul-tivation and the conservation of Ethiopian Afromontane rainforest fragments. Forest Ecology and Management 261, 1034–1041. https://doi.org/10.1016/j.foreco.2010.12.025

Aggemyr, E., Auffret, A.G., Jädergård, L., Cousins, S.A.O., 2018. Species richness and composition differ in response to landscape and bioge-ography. Landscape Ecology 33, 2273–2284. https://doi.org/10.1007/s10980-018-0742-9

Allinne, C., Savary, S., Avelino, J., 2016. Delicate balance between pest and disease injuries, yield performance, and other ecosystem services in the complex coffee-based systems of Costa Rica. Agriculture, Eco-systems & Environment 222, 1–12.

Ango, T.G., Börjeson, L., Senbeta, F., Hylander, K., 2014. Balancing eco-system services and disservices: Smallholder farmers’ use and man-agement of forest and trees in an agricultural landscape in south-western Ethiopia. Ecology and Society 19. https://doi.org/10.5751/ES-06279-190130

Anthony, F., Combes, M.C., Astorga, C., Bertrand, B., Graziosi, G., Lash-ermes, P., 2002. The origin of cultivated Coffea arabica L. varieties revealed by AFLP and SSR markers. Theoretical and Applied Ge-netics 104, 894–900.

Avelino, J., Allinne, C., Cerda, R., Willocquet, L., Savary, S., 2018. Multi-ple-disease system in coffee: From crop loss assessment to sustaina-ble management. Annual Review of Phytopathology 56, 611–635.

Avelino, J., Cristancho, M., Georgiou, S., Imbach, P., Aguilar, L., Borne-mann, G., Läderach, P., Anzueto, F., Hruska, A.J., Morales, C., 2015. The coffee rust crises in Colombia and Central America (2008–2013): impacts, plausible causes and proposed solutions. Food Security 7, 303–321.

Avelino, J., ten Hoopen, G.M., DeClerck, F.A.J., 2011. Ecological mecha-nisms for pest and disease control in coffee and cacao

agroecosystems of the neotropics, in: Ecosystem Services from Agriculture

and Agroforestry. Routledge, pp. 125–152.

41

Avelino, J., Willocquet, L., Savary, S., 2004. Effects of crop management patterns on coffee rust epidemics. Plant Pathology 53, 541–547.

Avelino, J., Zelaya, H., Merlo, A., Pineda, A., Ordoñez, M., Savary, S., 2006. The intensity of a coffee rust epidemic is dependent on pro-duction situations. Ecological Modelling 197, 431–447.

Bedimo, J.A.M., Njiayouom, I., Bieysse, D., Nkeng, M.N., Cilas, C., Notte-ghem, J.-L., 2008. Effect of shade on Arabica coffee berry disease development: toward an agroforestry system to reduce disease im-pact. Phytopathology 98, 1320–1325.

Bedimo, M.J.A., Bieysse, D., Nyassé, S., Nottéghem, J.L., Cilas, C., 2010. Role of rainfall in the development of coffee berry disease in Coffea arabica caused by Colletotrichum kahawae, in Cameroon. Plant Pa-thology 59, 324–329.

Beer, J., Muschler, R., Kass, D., Somarriba, E., 1998. Shade management in coffee and cacao plantations. Agroforestry Systems 38, 139–164. https://doi.org/10.1023/A:1005956528316

Benti, T., Gebre, E., Tesfaye, K., Berecha, G., Lashermes, P., Kyallo, M., Kouadio Yao, N., 2020. Genetic diversity among commercial Ara-bica coffee (Coffea arabica L.) varieties in Ethiopia using simple se-quence repeat markers. Journal of Crop Improvement 1–22.

Berecha, G., Aerts, R., Muys, B., Honnay, O., 2015. Fragmentation and management of Ethiopian moist evergreen forest drive composi-tional shifts of insect communities visiting wild Arabica coffee flow-ers. Environmental Management 55, 373–382.

Berecha, G., Aerts, R., Vandepitte, K., Van Glabeke, S., Muys, B., Roldán-Ruiz, I., Honnay, O., 2014. Effects of forest management on mating patterns, pollen flow and intergenerational transfer of genetic diver-sity in wild Arabica coffee (Coffea arabica L.) from Afromontane rainforests. Biological Journal of the Linnean Society 112, 76–88.

Bhagwat, S.A., Willis, K.J., Birks, H.J.B., Whittaker, R.J., 2008. Agrofor-estry: a refuge for tropical biodiversity? Trends in Ecology & Evolu-tion 23, 261–267. https://doi.org/10.1016/j.tree.2008.01.005

Bianchi, F.J.J.A., Booij, C.J.H., Tscharntke, T., 2006. Sustainable pest regu-lation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest control. Proceedings of the Royal Soci-ety B: Biological Sciences 273, 1715–1727. https://doi.org/10.1098/rspb.2006.3530

42

Bradshaw, C.J., Sodhi, N.S., Brook, B.W., 2009. Tropical turmoil: a biodi-versity tragedy in progress. Frontiers in Ecology and the Environ-ment 7, 79–87. https://doi.org/10.1890/070193

Cardinale, B.J., Duffy, J.E., Gonzalez, A., Hooper, D.U., Perrings, C., Ve-nail, P., Narwani, A., Mace, G.M., Tilman, D., Wardle, D.A., Kinzig, A.P., Daily, G.C., Loreau, M., Grace, J.B., Larigauderie, A., Srivastava, D.S., Naeem, S., 2012. Biodiversity loss and its impact on humanity. Nature 486, 59–67. https://doi.org/10.1038/na-ture11148

Cerda, R., Avelino, J., Gary, C., Tixier, P., Lechevallier, E., Allinne, C., 2017. Primary and secondary yield losses caused by pests and dis-eases: Assessment and modeling in coffee. PLoS ONE 12, e0169133. https://doi.org/10.1371/journal.pone.0169133

Chala, J., Chemeda, F., Girma, A., Holger, H., 2010. Coffee leaf rust epi-demics (Hemileia vastatrix) in montane coffee (Coffea arabica L.) forests in southwestern Ethiopia. East African journal of sciences 4, 86–95.

Chazdon, R.L., Harvey, C.A., Komar, O., Griffith, D.M., Ferguson, B.G., Martínez‐Ramos, M., Morales, H., Nigh, R., Soto‐Pinto, L., Breugel, M.V., Philpott, S.M., 2009. Beyond reserves: A research agenda for conserving biodiversity in human-modified tropical landscapes. Bio-tropica 41, 142–153. https://doi.org/10.1111/j.1744-7429.2008.00471.x

Cheatham, M.R., Rouse, M.N., Esker, P.D., Ignacio, S., Pradel, W., Ray-mundo, R., Sparks, A.H., Forbes, G.A., Gordon, T.R., Garrett, K.A., 2009. Beyond yield: plant disease in the context of ecosystem ser-vices. Phytopathology 99, 1228–1236.

Clough, Y., Barkmann, J., Juhrbandt, J., Kessler, M., Wanger, T.C., An-shary, A., Buchori, D., Cicuzza, D., Darras, K., Putra, D.D., Erasmi, S., Pitopang, R., Schmidt, C., Schulze, C.H., Seidel, D., Steffan-Dewenter, I., Stenchly, K., Vidal, S., Weist, M., Wielgoss, A.C., Tscharntke, T., 2011. Combining high biodiversity with high yields in tropical agroforests. Proceedings of the National Academy of Sci-ences 108, 8311–8316. https://doi.org/10.1073/pnas.1016799108

Daba, G., Helsen, K., Berecha, G., Lievens, B., Debela, A., Honnay, O., 2019. Seasonal and altitudinal differences in coffee leaf rust epidem-ics on coffee berry disease-resistant varieties in Southwest Ethiopia. Tropical Plant Pathology 1–7.

43

DaMatta, F.M., Ronchi, C.P., Maestri, M., Barros, R.S., 2008. Ecophysiol-ogy of coffee growth and production. Brazilian Journal of Plant Physiology 19, 485–510. https://doi.org/10.1590/S1677-04202007000400014

De Beenhouwer, M., Aerts, R., Honnay, O., 2013. A global meta-analysis of the biodiversity and ecosystem service benefits of coffee and cacao agroforestry. Agriculture, Ecosystems & Environment 175, 1–7.

De Beenhouwer, M., Geeraert, L., Mertens, J., Van Geel, M., Aerts, R., Vanderhaegen, K., Honnay, O., 2016. Biodiversity and carbon stor-age co-benefits of coffee agroforestry across a gradient of increasing management intensity in the SW Ethiopian highlands. Agriculture, Ecosystems & Environment 222, 193–199.

Ellis, E.C., Goldewijk, K.K., Siebert, S., Lightman, D., Ramankutty, N., 2010. Anthropogenic transformation of the biomes, 1700 to 2000. Global Ecology and Biogeography 19, 589–606. https://doi.org/10.1111/j.1466-8238.2010.00540.x

Eskes, A.B., 1983. Incomplete resistance to coffee leaf rust, in: Lamberti, F., Waller, J.M., Van der Graaff, N.A. (Eds.), Durable resistance in crops, NATO advanced science institutes series. Springer New York, Boston, MA, pp. 291–315. https://doi.org/10.1007/978-1-4615-9305-8_26

Eskes, A.B., Mendes, M.D.L., Robbs, C.F., 1991. Laboratory and field stud-ies on parasitism of Hemileia vastatrix with Verticillium lecanii and V. leptobactrum. Café, Cacao, Thé (Francia). v. 35(4) p. 275-282.

Foley, J.A., Ramankutty, N., Brauman, K.A., Cassidy, E.S., Gerber, J.S., Johnston, M., Mueller, N.D., O’Connell, C., Ray, D.K., West, P.C., Balzer, C., Bennett, E.M., Carpenter, S.R., Hill, J., Monfreda, C., Polasky, S., Rockström, J., Sheehan, J., Siebert, S., Tilman, D., Zaks, D.P.M., 2011. Solutions for a cultivated planet. Nature 478, 337–342. https://doi.org/10.1038/nature10452

Friis, I., Sebsebe, D., van Breugel, P., 2010. Atlas of the potential vegetation of Ethiopia. Det Kongelige Danske Videnskabernes Selskab.

Gardner, T.A., Barlow, J., Chazdon, R., Ewers, R.M., Harvey, C.A., Peres, C.A., Sodhi, N.S., 2009. Prospects for tropical forest biodiversity in a human-modified world. Ecology Letters 12, 561–582. https://doi.org/10.1111/j.1461-0248.2009.01294.x

Gardner, T.A., Barlow, J., Sodhi, N.S., Peres, C.A., 2010. A multi-region as-sessment of tropical forest biodiversity in a human-modified world.

44

Biological Conservation 143, 2293–2300. https://doi.org/10.1016/j.biocon.2010.05.017

Garedew, W., Lemessa, F., Pinard, F., 2019. Landscape context and plot fea-tures influence the epidemics of coffee leaf rust (Hemileia vastatrix) in southwest Ethiopia. Archives of Phytopathology and Plant Pro-tection 52, 71–89. https://doi.org/10.1080/03235408.2019.1580177

Geeraert, L., Hulsmans, E., Helsen, K., Berecha, G., Aerts, R., Honnay, O., 2019. Rapid diversity and structure degradation over time through continued coffee cultivation in remnant Ethiopian Afromontane for-ests. Biological Conservation 236, 8–16. https://doi.org/10.1016/j.bi-ocon.2019.05.014

Getachew, S., Adugna, G., Lemessa, F., Hindorf, H., 2013. Population struc-ture of Gibberella xylarioides Heim and Saccas in Ethiopian forest coffee (Coffea arabica L.) systems. African Journal of Biotechnol-ogy 12 (33), 5157-5163.

Getachew, S., Adugna, G., Lemessa, F., Hindorf, H., 2012. Coffee wilt dis-ease (Gibberella xylarioides Heim and Saccas) in forest coffee sys-tems of southwest and southeast Ethiopia. Plant Pathology Journal 11, 10–17.

Gezahgne, A., Coetzee, M.P.A., Wingfield, B.D., Wingfield, M.J., Roux, J., 2004. Identification of the Armillaria root rot pathogen in Ethiopian plantations. Forest Pathology 34, 133–145.

Girma, A., Hulluka, M., Hindorf, H., 2001. Incidence of tracheomycosis, Gibberella xylarioides (Fusarium xylarioides), on Arabica coffee in Ethiopia. Journal of Plant Diseases and Protection 136–142.

Girma, A., Million, A., Hindorf, H., Arega, Z., Teferi, D., Jefuka, C., 2009. Coffee wilt disease in Ethiopia. Coffee Wilt Disease 50–68.

Gole, T.W., Borsch, T., Denich, M., Teketay, D., 2008. Floristic composi-tion and environmental factors characterizing coffee forests in south-west Ethiopia. Forest Ecology and Management 255, 2138–2150.

Gove, A.D., Hylander, K., Nemomisa, S., Shimelis, A., 2008. Ethiopian cof-fee cultivation—Implications for bird conservation and environmen-tal certification. Conservation Letters 1, 208–216.

Hakiza, G.J., Kyetere, D.T., Musoli, P., Wetala, P., Njuki, J., Kucel, P., Aluka, P., Kangire, A., Ogwang, J., 2009. Coffee wilt disease in Uganda. Coffee wilt disease 28–49.

Harvey, C.A., González Villalobos, J.A., 2007. Agroforestry systems con-serve species-rich but modified assemblages of tropical birds and

45

bats. Biodiversity Conservation 16, 2257–2292. https://doi.org/10.1007/s10531-007-9194-2

Harvey, C.A., Komar, O., Chazdon, R., Ferguson, B.G., Finegan, B., Grif-fith, D.M., Martínez‐Ramos, M., Morales, H., Nigh, R., Soto‐Pinto, L., Breugel, M.V., Wishnie, M., 2008. Integrating agricultural land-scapes with biodiversity conservation in the Mesoamerican hotspot. Conservation Biology 22, 8–15. https://doi.org/10.1111/j.1523-1739.2007.00863.x

Hindorf, H., Omondi, C.O., 2011. A review of three major fungal diseases of Coffea arabica L. in the rainforests of Ethiopia and progress in breeding for resistance in Kenya. Journal of Advanced Research 2, 109–120.

Hundera, K., Aerts, R., De Beenhouwer, M., Van Overtveld, K., Helsen, K., Muys, B., Honnay, O., 2013a. Both forest fragmentation and coffee cultivation negatively affect epiphytic orchid diversity in Ethiopian moist evergreen Afromontane forests. Biological Conservation 159, 285–291.

Hundera, K., Aerts, R., Fontaine, A., Van Mechelen, M., Gijbels, P., Hon-nay, O., Muys, B., 2013b. Effects of coffee management intensity on composition, structure, and regeneration status of Ethiopian moist evergreen afromontane forests. Environmental Management 51, 801–809.

Hylander, K., Nemomissa, S., 2009. Complementary roles of home gardens and exotic tree plantations as alternative habitats for plants of the Ethiopian montane rainforest. Conservation Biology 23, 400–409. https://doi.org/10.1111/j.1523-1739.2008.01097.x

Hylander, K., Nemomissa, S., 2008. Home garden coffee as a repository of epiphyte biodiversity in Ethiopia. Frontiers in Ecology and the Envi-ronment 6, 524–528.

Hylander, K., Nemomissa, S., Delrue, J., Enkosa, W., 2013. Effects of coffee management on deforestation rates and forest integrity. Conserva-tion Biology 27, 1031–1040.

Jackson, D., Skillman, J., Vandermeer, J., 2012. Indirect biological control of the coffee leaf rust, Hemileia vastatrix, by the entomogenous fun-gus Lecanicillium lecanii in a complex coffee agroecosystem. Bio-logical Control 61, 89–97.

Jaramillo, J., Muchugu, E., Vega, F.E., Davis, A., Borgemeister, C., Chabi-Olaye, A., 2011. Some like it hot: the influence and implications of climate change on coffee berry borer (Hypothenemus hampei) and

46

coffee production in East Africa. PLoS ONE 6, e24528. https://doi.org/10.1371/journal.pone.0024528

Jezeer, R.E., Verweij, P.A., Santos, M.J., Boot, R.G.A., 2017. Shaded coffee and cocoa – Double dividend for biodiversity and small-scale farm-ers. Ecological Economics 140, 136–145. https://doi.org/10.1016/j.ecolecon.2017.04.019

Jha, S., Bacon, C.M., Philpott, S.M., Mendes, V.E., Läderach, P., Rice, R.A., 2014. Shade coffee: update on a disappearing refuge for biodiver-sity. BioScience 64, 416–428.

Jose, S., 2012. Agroforestry for conserving and enhancing biodiversity. Ag-roforest Systems 85, 1–8. https://doi.org/10.1007/s10457-012-9517-5

Labouisse, J.-P., Bellachew, B., Kotecha, S., Bertrand, B., 2008. Current sta-tus of coffee (Coffea arabica L.) genetic resources in Ethiopia: im-plications for conservation. Genetic Resources and Crop Evolution 55, 1079. https://doi.org/DOI: 10.1007/s10722-008-9361-7

Lemessa, D., Hylander, K., Hambäck, P., 2013. Composition of crops and land-use types in relation to crop raiding pattern at different dis-tances from forests. Agriculture, Ecosystems & Environment 167, 71–78. https://doi.org/10.1016/j.agee.2012.12.014

Lin, B.B., 2007. Agroforestry management as an adaptive strategy against potential microclimate extremes in coffee agriculture. Agricultural and Forest Meteorology 144, 85–94. https://doi.org/10.1016/j.agrformet.2006.12.009

López-Bravo, D.F., Virginio-Filho, E. de M., Avelino, J., 2012. Shade is conducive to coffee rust as compared to full sun exposure under standardized fruit load conditions. Crop Protection 38, 21–29.

Martin, D.A., Osen, K., Grass, I., Hölscher, D., Tscharntke, T., Wurz, A., Kreft, H., 2020. Land-use history determines ecosystem services and conservation value in tropical agroforestry. Conservation Letters, e12740. https://doi.org/10.1111/conl.12740

McCann, J.C., 1995. People of the plow: an agricultural history of Ethiopia, 1800–1990. University of Wisconsin Press.

McCook, S., 2006. Global rust belt: Hemileia vastatrix and the ecological in-tegration of world coffee production since 1850. Journal of Global History 1, 177–195.

Mittermeier, R.A., Myers, N., Thomsen, J.B., Fonseca, G.A.B.D., Olivieri, S., 1998. Biodiversity hotspots and major tropical wilderness areas:

47

approaches to setting conservation priorities. Conservation Biology 12, 516–520. https://doi.org/10.1046/j.1523-1739.1998.012003516.x

Moguel, P., Toledo, V.M., 1999. Biodiversity conservation in traditional cof-fee systems of Mexico. Conservation Biology 13, 11–21.

Mulaw, T., Kubicek, C., Druzhinina, I., 2010. The Rhizosphere of Coffea Arabica in Its Native Highland Forests of Ethiopia Provides a Niche for a Distinguished Diversity of Trichoderma. Diversity 2, 527–549. https://doi.org/10.3390/d2040527

Muleta, D., Assefa, F., Granhall, U., 2007. In vitro antagonism of rhizobac-teria isolated from Coffea arabica L. against emerging fungal coffee pathogens. Engineering in Life Sciences 7, 577–586. https://doi.org/10.1002/elsc.200700004

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. https://doi.org/10.1038/35002501

Nei, M., Chesser, R.K., 1983. Estimation of fixation indices and gene diver-sities. Annals of Human Genetics 47, 253–259. https://doi.org/10.1111/j.1469-1809.1983.tb00993.x

Nutman, F.J., 1970. Coffee berry disease. PANS Pest Articles & News Sum-maries 16, 277–286.

Otieno, W., Sierra, A.P., Termorshuizen, A., 2003. Characterization of Ar-millaria isolates from tea (Camellia sinensis) in Kenya. Mycologia 95, 160–175.

Perfecto, I., Mas, A., Dietsch, T., Vandermeer, J., 2003. Conservation of bio-diversity in coffee agroecosystems: a tri-taxa comparison in southern Mexico. Biodiversity and Conservation 12, 1239–1252. https://doi.org/10.1023/A:1023039921916

Perfecto, I., Rice, R.A., Greenberg, R., Van der Voort, M.E., 1996. Shade coffee: a disappearing refuge for biodiversity: shade coffee planta-tions can contain as much biodiversity as forest habitats. BioScience 46, 598–608.

Perfecto, I., Vandermeer, J., 2015. Coffee agroecology: A new approach to understanding agricultural biodiversity, ecosystem services and sus-tainable development. Routledge.

Perfecto, I., Vandermeer, J., 2008. Biodiversity conservation in tropical agroecosystems. Annals of the New York Academy of Sciences 1134, 173–200. https://doi.org/10.1196/annals.1439.011

Perfecto, I., Vandermeer, J., Mas, A., Pinto, L.S., 2005. Biodiversity, yield, and shade coffee certification. Ecological Economics 54, 435–446.

48

Petit, N., 2007. Ethiopia’s coffee sector: A bitter or better future? Journal of Agrarian Change 7, 225–263.

Phalan, B., Onial, M., Balmford, A., Green, R.E., 2011. Reconciling food production and biodiversity conservation: Land sharing and land sparing compared. Science 333, 1289–1291. https://doi.org/10.1126/science.1208742

Philpott, S., Bichier, P., Rice, R., Greenberg, R., 2008. Philpott SM, Bichier P, Rice RA, Greenberg R. Biodiversity conservation, yield, and al-ternative products in coffee agroecosystems in Sumatra, Indonesia. Biodiversity and Conservation 17, 1805–1820. https://doi.org/10.1007/s10531-007-9267-2

Philpott, S.M., Dietsch, T., 2003. Coffee and conservation: a global context and the value of farmer involvement. Conservation Biology 17, 1844–1846.

Pieters, R., van der Graaff, N.A., 1980. Resistance to Gibberella xylarioides in Coffea arabica: evaluation of screening methods and evidence for the horizontal nature of the resistance. Netherlands Journal of Plant Pathology 86, 37–43. https://doi.org/10.1007/BF02650392

Power, A.G., 2010. Ecosystem services and agriculture: tradeoffs and syner-gies. Philosophical Transactions of the Royal Society B: Biological Sciences 365, 2959–2971.

Samnegård, U., Hambäck, P.A., Nemomissa, S., Hylander, K., 2014. Domi-nance of the semi-wild honeybee as coffee pollinator across a gradi-ent of shade-tree structure in Ethiopia. Journal of Tropical Ecology 30, 401–408.

Schmitt, C.B., Senbeta, F., Denich, M., Preisinger, H., Boehmer, H.J., 2010. Wild coffee management and plant diversity in the montane rainfor-est of southwestern Ethiopia. African Journal of Ecology 48, 78–86.

Schroth, G., Izac, A.-M.N., Vasconcelos, H.L., Gascon, C., da Fonseca, G.A., Harvey, C.A., 2004. Agroforestry and biodiversity conserva-tion in tropical landscapes. Island Press.

Schroth, G., Krauss, U., Gasparotto, L., Duarte Aguilar, J.A., Vohland, K., 2000. Pests and diseases in agroforestry systems of the humid trop-ics. Agroforestry Systems 50, 199–241. https://doi.org/10.1023/A:1006468103914

Senbeta, F., Denich, M., 2006. Effects of wild coffee management on spe-cies diversity in the Afromontane rainforests of Ethiopia. Forest Ecology and Management 232, 68–74.

49

Singer, M., 2010. Pathogen-pathogen interaction: a syndemic model of com-plex biosocial processes in disease. Virulence 1, 10–18.

Somarriba, E., Cerda, R., Orozco, L., Cifuentes, M., Dávila, H., Espin, T., Mavisoy, H., Ávila, G., Alvarado, E., Poveda, V., Astorga, C., Say, E., Deheuvels, O., 2013. Carbon stocks and cocoa yields in agrofor-estry systems of Central America. Agriculture, Ecosystems & Envi-ronment 173, 46–57. https://doi.org/10.1016/j.agee.2013.04.013

Soto-Pinto, L., Perfecto, I., Caballero-Nieto, J., 2002. Shade over coffee: its effects on berry borer, leaf rust and spontaneous herbs in Chiapas, Mexico. Agroforestry Systems 55, 37–45.

Soto-Pinto, L., Perfecto, I., Castillo-Hernandez, J., Caballero-Nieto, J., 2000. Shade effect on coffee production at the northern Tzeltal zone of the state of Chiapas, Mexico. Agriculture, Ecosystems & Environment 80, 61–69. https://doi.org/10.1016/S0167-8809(00)00134-1

Staver, C.P., Guharay, F., Monterroso, D., Muschler, R.G., 2001. Designing pest-suppressive multistrata perennial crop systems: shade-grown coffee in Central America. Agroforestry Systems 53, 151–170. https://doi.org/10.1023/A:1013372403359

Steffan-Dewenter, I., Kessler, M., Barkmann, J., Bos, M.M., Buchori, D., Erasmi, S., Faust, H., Gerold, G., Glenk, K., Gradstein, S.R., Guhardja, E., Harteveld, M., Hertel, D., Hohn, P., Kappas, M., Kohler, S., Leuschner, C., Maertens, M., Marggraf, R., Migge-Kleian, S., Mogea, J., Pitopang, R., Schaefer, M., Schwarze, S., Sporn, S.G., Steingrebe, A., Tjitrosoedirdjo, S.S., Tjitrosoemito, S., Twele, A., Weber, R., Woltmann, L., Zeller, M., Tscharntke, T., 2007. Tradeoffs between income, biodiversity, and ecosystem func-tioning during tropical rainforest conversion and agroforestry inten-sification. Proceedings of the National Academy of Sciences 104, 4973–4978. https://doi.org/10.1073/pnas.0608409104

Tadesse, G., Zavaleta, E., Shennan, C., 2014. Coffee landscapes as refugia for native woody biodiversity as forest loss continues in southwest Ethiopia. Biological Conservation 169, 384–391.

Tesfaye, K., Govers, K., Bekele, E., Borsch, T., 2014. ISSR fingerprinting of Coffea arabica throughout Ethiopia reveals high variability in wild populations and distinguishes them from landraces. Plant Systemat-ics and Evolution 300, 881–897.

Tilman, D., Cassman, K.G., Matson, P.A., Naylor, R., Polasky, S., 2002. Ag-ricultural sustainability and intensive production practices. Nature 418, 671–677. https://doi.org/10.1038/nature01014

50

Vaast, P., Bertrand, B., Perriot, J.-J., Guyot, B., Génard, M., 2006. Fruit thin-ning and shade improve bean characteristics and beverage quality of coffee (Coffea arabica L.) under optimal conditions. Journal of the Science of Food and Agriculture 86, 197–204. https://doi.org/10.1002/jsfa.2338

Vandermeer, J., Perfecto, I., Liere, H., 2009. Evidence for hyperparasitism of coffee rust (Hemileia vastatrix) by the entomogenous fungus, Lecanicillium lecanii, through a complex ecological web. Plant Pa-thology 58, 636–641.

Vega, F.E., Infante, F., Johnson, A.J., 2015. Chapter 11 - The genus Hypoth-enemus, with emphasis on H. hampei, the coffee berry borer, in: Vega, F.E., Hofstetter, R.W. (Eds.), Bark Beetles. Academic Press, San Diego, pp. 427–494. https://doi.org/10.1016/B978-0-12-417156-5.00011-3

Waller, J.M., 1982. Coffee rust—epidemiology and control. Crop Protection 1, 385–404.

Waller, J.M., Bigger, M., Hillocks, R.J., 2007. Coffee pests, diseases and their management. CABI.

Waller, J.M., Bridge, P.D., Black, R., Hakiza, G., 1993. Characterization of the coffee berry disease pathogen, Colletotrichum kahawae sp. nov. Mycological Research 97, 989–994.

Zhang, W., Ricketts, T.H., Kremen, C., Carney, K., Swinton, S.M., 2007. Ecosystem services and dis-services to agriculture. Ecological Eco-nomics 64, 253–260. https://doi.org/10.1016/j.ecolecon.2007.02.024

1

2

3