Lessons from Ethiopian coffee landscapes for global conservation in a post-wild world.

The reality for conservation of biodiversity across our planet is that all ecosystems are modified by humans in some way or another. Thus, biodiversity conservation needs to be implemented in multifunctional landscapes. In this paper we use a fascinating coffee-dominated landscape in southwest Ethiopia as our lens to derive general lessons for biodiversity conservation in a post-wild world. Considering a hierarchy of scales from genes to multi-species interactions and social-ecological system contexts, we focus on (i) threats to the genetic diversity of crop wild relatives, (ii) the mechanisms behind trade-offs between biodiversity and agricultural yields, (iii) underexplored species interactions suppressing pest and disease levels, (iv) how the interactions of climate change and land-use change sometimes provide opportunities for restoration, and finally, (v) how to work closely with stakeholders to identify scenarios for sustainable development. The story on how the ecology and evolution of coffee within its indigenous distribution shape biodiversity conservation from genes to social-ecological systems can inspire us to view other landscapes with fresh eyes. The ubiquitous presence of human-nature interactions demands proactive, creative solutions to foster biodiversity conservation not only in remote protected areas but across entire landscapes inhabited by people.

Genetic composition and diversity of Arabica coffee in the crop's centre of origin and its impact on four major fungal diseases.

Conventional wisdom states that genetic variation reduces disease levels in plant populations. Nevertheless, crop species have been subject to a gradual loss of genetic variation through selection for specific traits during breeding, thereby increasing their vulnerability to biotic stresses such as pathogens. We explored how genetic variation in Arabica coffee sites in southwestern Ethiopia was related to the incidence of four major fungal diseases. Sixty sites were selected along a gradient of management intensity, ranging from nearly wild to intensively managed coffee stands. We used genotyping-by-sequencing of pooled leaf samples (pool-GBS) derived from 16 individual coffee shrubs in each of the 60 sites to assess the variation in genetic composition (multivariate: reference allele frequency) and genetic diversity (univariate: mean expected heterozygosity) between sites. We found that genetic composition had a clear spatial pattern and that genetic diversity was higher in less managed sites. The incidence of the four fungal diseases was related to the genetic composition of the coffee stands, but in a specific way for each disease. In contrast, genetic diversity was only related to the within-site variation of coffee berry disease, but not to the mean incidence of any of the four diseases across sites. Given that fungal diseases are major challenges of Arabica coffee in its native range, our findings that genetic composition of coffee sites impacted the major fungal diseases may serve as baseline information to study the molecular basis of disease resistance in coffee. Overall, our study illustrates the need to consider both host genetic composition and genetic diversity when investigating the genetic basis for variation in disease levels.

Multi-scale mosaics in top-down pest control by ants from natural coffee forests to plantations.

While top-down control plays an important role in shaping both natural and agricultural food webs, we lack insights into how top-down control effects vary across spatial scales. We used a multi-scale survey of top-down control of coffee pests and diseases by arboreal ants to examine if colony location creates a small-scale mosaic in top-down control around trees and if the strength of that control varies between sites at the landscape scale. We investigated pest and disease levels on coffee shrubs at different distances from shade trees with and without a Crematogaster spp. ant colony in 59 sites along a coffee management intensity gradient in southwestern Ethiopia. Within sites, ants significantly suppressed herbivory and coffee leaf rust at distances less than 10 m from nesting trees. Top-down control varied between sites, with stronger top-down control of free-feeding herbivory near ant colonies at sites with lower management intensity and stronger top-down control of a skeletonizer at sites with higher canopy cover. We conclude that the strength of top-down control by ants is highly heterogeneous across spatial scales, as a consequence of the biology of the predator at the small scale and herbivore density or changes in herbivore-ant interactions at the landscape scale.