Content uploaded by Deven Kammerichs-Berke
Author content
All content in this area was uploaded by Deven Kammerichs-Berke on Jul 01, 2022
Content may be subject to copyright.
The effect of shade tree species on bird
communities in central Kenyan coffee farms
DEVEN KAMMERICHS-BERKE
1
*, FANTER J. LANE
1
,
FRANK JUMA ONG’ONDO
1,2
, EDSON M. MLAMBA
2
, WILLIAM T. BEAN
1,3
,
JULIE A. JEDLICKA
4
, PETER NJOROGE
2
an d MATTHE W D. JOH N SON
1
1
California State Polytechnic University, Humboldt, Arcata, California, USA.
2
National Museums of Kenya, Nairobi, Kenya.
3
California Polytechnic State University, San Luis Obispo, California, USA.
4
Missouri Western State University, St. Joseph, Missouri, USA.
*Author for correspondence; email: dk1166@humboldt.edu
(Received 12 February 2021; revision accepted 01 November 2021)
Summary
Shade coffee is a well-studied cultivation strategy that creates habitat for tropical birds while also
maintaining agricultural yield. Although there is a general consensus that shade coffee is more
“bird-friendly” than a sun coffee monoculture, little work has investigated the effects of specific
shade tree species on insectivorous bird diversity. This study involved avian foraging observations,
mist-netting data, temperature loggers, and arthropod sampling to investigate bottom-up effects of
two shade tree taxa - native Cordia sp. and introduced Grevillea robusta - on insectivorous bird
communities in central Kenya. Results indicate that foliage-dwelling arthropod abundance, and the
richness and overall abundance of foraging birds were all higher on Cordia than on Grevillea.
Furthermore, multivariate analyses of the bird community indicate a significant difference in
community composition between the canopies of the two tree species, though the communities of
birds using the coffee understorey under these shade trees were similar. In addition, both shade
trees buffered temperatures in coffee, and temperatures under Cordia were marginally cooler than
under Grevillea. These results suggest that native Cordia trees on East African shade coffee farms
may be better at mitigating habitat loss and attracting insectivorous birds that could promote
ecosystem services. Identifying differences in prey abundance and preferences in bird foraging
behaviour not only fills basic gaps in our understanding of the ecology of East African coffee farms,
it also aids in developing region-specific information to optimize functional diversity, ecosystem
services, and the conservation of birds in agricultural landscapes.
Keywords: Ornithology, Coffee Ecology, Ecosystem Services
Introduction
Agricultural intensification is one of the greatest threats to biodiversity (Foley et al. 2005),
particularly because of its association with deforestation, which has a disproportionately negative
effect on biological communities (Donald 2004, Betts et al. 2017). In the tropics, where most of the
world’s biodiversity is concentrated (Brown 2014), an emphasis on agricultural habitats is vital for
Bird Conservation International (2022)1–19.©The Author(s), 2022. Published by Cambridge University Press on behalf of
BirdLife International. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence
(https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any
medium, provided the original work is properly cited.
doi:10.1017/S0959270921000502
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
successful conservation for a variety of ecological and socioeconomic reasons (Perfecto et al. 2009,
Perfecto and Vandermeer 2010, Renwick et al. 2014). Currently, agricultural landscapes cover
approximately 37% of the earth’s land surface, and agricultural production is projected to increase
100–110%by2050 to meet growing global crop demand (Tilman et al. 2011). Meeting this rising
agricultural demand will require identifying strategies to minimize the loss of biodiversity while
also maximizing agricultural yield (Vandermeer and Perfecto 1997, Fischer et al. 2014, Mehrabi
et al. 2018).
Coffee Coffea spp. grown beneath shade trees, called “shade coffee,” is a well-studied example of
integrating crop production with biodiversity conservation (Perfecto et al. 2009, Jha et al. 2014,
Perfecto and Vandermeer 2015), contrasting with a more industrial strategy, generally referred to
as “sun coffee,” which involves few to no shade trees to maximize short-term production (Jha et al.
2014). In east Africa (Douglas et al. 2013, Buechley et al. 2015), the Neotropics (Armbrecht and
Perfecto 2003, Philpott et al. 2008, Philpott and Bichier 2012) and India (Raman 2006), research
suggests that the shade strategy supports a high diversity of economically important taxa such as
birds (Johnson and Hackett 2016). In turn, insectivorous bird populations can play a key role in the
provisioning of natural pest control services in coffee through top-down effects on pest arthropods
(Perfecto et al. 2004, Kellermann et al. 2008, Philpott et al. 2009, Karp et al. 2014). Bird species
richness (Perfecto et al. 2004, Van Bael et al. 2008), density (Perfecto et al. 2004), abundance
(Jedlicka et al. 2011), and functional richness (Philpott et al. 2009) are all positively correlated with
the top-down control of pests, especially the coffee berry borer Hypothenemus hampei.
In all regions, the term shade coffee belies tremendous variation among and within farms that
contain shade trees (Moguel and Toledo 1999). Smaller-scale farms in Kenya tend to have a higher
diversity of native trees, either planted intentionally or as remnants of adjacent forests as a form of
rustic farm management (Moguel and Toledo 1999, Lengkeek et al. 2005, Kindt et al. 2006).
Conversely, a shade plantation strategy that utilizes one or only a few species of tree, called a
“shaded monoculture” (Moguel and Toledo 1999), is common in many regions, including among
the large plantations established during the colonial era in Kenya and now run usually by African
or international enterprises (Tignor 2015). Often, a few key tree species dominate shaded mono-
cultures within a region, such as Grevillea robusta in Kenya, Uganda, and Brazil (Baggio et al.
1997, Muchiri 2004, Kiyingi et al. 2016), several species of Inga in Mexico and Jamaica (Johnson
2000a, Romero-Alvarado et al. 2000), and Erythrina poeppigeana in Costa Rica (Perfecto and
Vandermeer 2015).
The selection of shade tree species has important implications for both farmers and the wildlife
that may use coffee farms. Farmers’ criteria for selecting shade tree species tend to revolve around
ecological or economic benefits provided by the trees, as well as aspects of tree phenology indirectly
related to microclimates, which can promote increased crop yield (Beer 1987, Soto-Pinto et al. 2000,
Pinard et al. 2014b). Shade tree products such as fruit and timber can also buffer the impact of coffee
income volatility, particularly for coffee farmers with small land holdings (Jassogne et al. 2012,
Davis et al. 2017), and recent evidence suggest shade trees may help farmers adapt to a warming
climate (Rahn et al. 2018, Schooler et al. 2020).
Understanding the ecology of specific shade tree species is also important because they can affect
coffee understorey pests by influencing the abundance and richness of natural bird predators that
can act as a top-down control on pest populations (Kellerman et al. 2008, Railsback and Johnson
2014) and by lowering understorey temperatures, which can slow pest reproduction (Jaramillo
et al. 2011). Johnson (2000a) found that Jamaican coffee plantations in which the native genus Inga
was dominant supported the highest abundances of both birds and non-pest arthropods. Similarly,
in central Kenyan plantations, Kammerichs-Berke (2020) found higher densities of non-pest
arthropods on native Cordia trees. This follows ecological theory regarding insect coevolution
with plants as summarised by Tallamy (2004). Insects adapt to evolutionarily novel plants slowly
(Southwood et al. 1982), and coevolution with particular host plants is a strong driving force for
species diversification and radiation for many insect taxa (Farrell and Mitter 1998, Becerra and
Venable 1999). Most herbivorous insects specialise on one or a few native plant groups with which
D. Kammerichs-Berke et al. 2
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
they have shared an evolutionary history (Erhlich and Raven 1964, Bernays and Graham 1988,
Forister et al. 2015), with specialisation being more pronounced at lower latitudes (Schemske et al.
2009). Thus, ecosystems dominated by non-native plants tend to exhibit lower insect diversity,
abundance, and biomass than systems dominated by native host plants (Burghardt et al. 2010, Litt
et al. 2014). This interaction has implications for the selection of shade tree species and their effects
on top-down impacts of insectivorous pest-eating birds in shade coffee farms (Narango et al. 2018).
In central Kenya, two of the most common trees on shaded coffee monocultures are Grevillea
robusta (hereafter Grevillea) and several species of Cordia, especially Cordia africana (collectively
hereafter Cordia). Grevillea is a deciduous tree introduced to Kenya from eastern Australia in the
19th century and is well-regarded amongst farmers because of its moderate to fast growth (as much
as 3m per year in some sites) and a tall branch system that provides a strong windbreak (Negash
1995). Cordia, on the other hand, is an evergreen native to East Africa that generally has a shorter
and wider branching canopy than Grevillea, as well as broader leaves (D. Kammerichs-Berke pers.
obs.) that provides high amounts of shade. Both species are often intentionally planted in evenly
spaced rows, and both tree species are also appealing as shade trees due to their nitrogen-fixing
abilities (Negash 1995, Lott et al. 2000). Despite the prominence of these two shade tree species,
ecological aspects of shade tree selection on East African coffee farms remains understudied.
Our study quantified the influence of these two tree species on the avian community in large
scale Kenyan coffee farms, with a special emphasis on insectivorous birds. We hypothesised that
native Cordia trees offer more potential for pest control services in Kenyan coffee farms than non-
native Grevillea because Cordia attracts more insectivorous birds that could act as a top-down
control on pest populations. Specifically, we tested the following predictions: (1) Non-pest foliage
arthropods are more abundant on Cordia than Grevillea,(2) greater numbers of insectivorous birds
forage in Cordia than in Grevillea,(3) insectivorous birds foraging in the shade layer also use the
coffee understorey (at the species level), and this pattern differs between Cordia and Grevillea, and
(4) insectivorous birds are more common in the coffee layer under or near Cordia than Grevillea.
Additionally, we measured understorey temperatures beneath Cordia and Grevillea shade trees to
shed light on potential bottom-up effects of shade trees on pests.
Methods
Study Area
This study was conducted on large-scale coffee plantations along an elevational gradient (1,567–
1,874 m) in Kiambu County, Kenya from 16 December 2018 to 19 January 2019. Both sun and
shade coffee farms occur along this elevational gradient (Jaramillo et al. 2013), with variation in
farming intensity, acreage, and habitat components. A variety of tree species are utilised within the
shade farms in this region, including acacias Acacia spp., broad-leaved croton Croton macrosta-
phylus, Meru oak Vitex keniensis, and Nandi flame Spathodea campanulate, though on large
plantations the two most commonly used species are Grevillea and Cordia (Johnson et al. unpubl.
data). Because of the focus on tree species selection, we only selected shade farms with low total tree
species diversity and a relatively even distribution of both Grevillea and Cordia, and full sun farms
were excluded from this study. Surveys were conducted on six sites (Figure 1a); each site was a
different coffee farm, except in one case a single farm was divided into two sites because it was large
(approximately 91 ha) and contained multiple fields (separated by dirt roads or paths) with
different characteristics (size and density of shade trees, density of coffee trees).
Survey methods
Arthropod, bird, vegetation, and temperature sampling was organised around individual shade
trees at each study site. To select trees, a four-quadrant grid fitted to the size of each farm was
overlaid on an aerial image of the site, recording the UTM coordinates for the centre of each
Effect of shade on Kenya bird communities 3
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
quadrant (Figure 1b). Then, in the field from the centroid of each quadrant, 3–4avian observation
points were selected, defined as locations with 3–4Cordia or Grevillea trees that could be visually
monitored simultaneously for avian foraging observations and also met the survey criteria:
23–40 cm diameter at breast height (dbh), at least 50 m from the site edge, and within 20 mof
Figure 1. (a) Map depicting the spatial arrangement of the six farms surveyed in Kiambu County,
Kenya from 16 December 2018–19 January 2019. (b) Site map depicting 4quadrants overlaid on
one of the coffee farms. Avian observation points were selected by going to the center of each
quadrant (green dots) and from there selecting 3–4points each with 3–4trees between 23–40 cm
diameter at breast height (dbh). All points were at least 50 m from the site edge (shown here in red)
and within 20 m of each other
D. Kammerichs-Berke et al. 4
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
each other. This dbh range was selected to minimise the confounding effects of tree size and
corresponds to the 25
th
and 75
th
percentiles of trees measured in a companion study of these farms
in 2017–2018 (Schooler et al. 2020, Kammerichs-Berke unpubl. data). An effort was made to
survey an equal number of Cordia and Grevillea trees at each site, though this was not always
possible due to their arrangement and availability. Of the 333 trees (184 Cordia and 149 Grevillea)
used in avian surveys, we sampled 146 (75 Cordia and 71 Grevillea) for arthropods, and 72 (36 Cor-
dia and 36 Grevillea) sampled with mist-nets. Basic vegetation data were recorded for all 333 trees.
Avian surveys were conducted at the avian observation points from 06h00–10h00 EAT, a time of
day when birds are most active (D. Kammerichs-Berke, pers. obs.). Two trained and experienced
field technicians conducted all surveys, and they generally alternated between sampling Cordia and
Grevillea trees throughout the morning. Due to the spatial design, one observer surveyed 71 more
trees in total than the other, but the difference in proportions of Cordia and Grevillea was not
significant (χ
2
=1.605,df=1,P=0.205). Once at an observation point, each observer simultan-
eously monitored the 3–4focal survey trees that were near the point, for a total of 10 minutes.
While this simultaneous design is unusual, we found that the number of birds present in or coming
to/from a given tree in a 10-minute period was low (see Results), and the habitat was open and
individual trees easily monitored, so this design optimised replication while maintaining precision.
For each survey, observers recorded species abundances, and the number of individuals actively
foraging in the trees. Foraging was defined as any of the stereotyped behaviours described in
Remsen and Robinson (1981). If there were more than 10 individuals of a species within a tree,
observers estimated flock size to the nearest five; for groups of a species fewer than 10, observers
counted individuals. Observers counted all birds seen in the trees within the 10-minute observation
period, including arriving birds.
Mist nets were used to quantify presence/absence and relative abundances of insectivorous birds
in the coffee layer. A team of field technicians set up 30-mm mesh nets in the coffee layer under
12 of the survey trees at each site, with nets deployed so that half were near Grevillea and half near
Cordia. Nets were placed no more than 5m from the base of a tree, parallel to the coffee crop rows.
Nets were opened 10 min before sunrise and were run for five hours for three mornings per site.
Birds were banded using bands supplied by the National Museums of Kenya. Recaptures from the
same day as initial banding were released directly at the net without re-processing, while recap-
tures from a previous day were processed and recorded. Recaptures from previous days were not
included in the analysis as a measure of abundance.
We sampled arthropod communities using the branch clipping method described in Johnson
(2000b) to sample arthropod communities. At each sampled tree, two branches were sampled,
selected from areas of the foliage profile most similar to those generally used by foliage-gleaning
birds (Johnson 2000a) during focal tree observations and within reach of extendable poles (i.e. outer
branches <5m high). Although an effort was made to sample two branches per tree, some trees
only had one sampleable branch, leading to an odd number of branches surveyed in total (147 Cor-
dia and 136 Grevillea, for a total of 283 branches across all farms). After a branch was selected, a
pole fitted with a fabric bag was extended to the height of the branch, the branch was enclosed
within the bag, and a drawstring pulled to cinch the bag over the branch as quickly as possible. A
pruning pole was used to clip the branch free. Once the branch was free, the bagged branch was
shaken to dislodge any arthropods. The clipped branches were checked for arthropods afterwards to
ensure that all insects were captured in the sample and weighed with a spring scale to obtain wet
biomass. The number of arthropods identified to order or class was recorded.
Tree species, height, and diameter at breast height (dbh) were measured at each surveyed shade
tree (n=333). Canopy coverage (via densiometer), crown length, width, and depth, and flowering
score were also measured for a third of shade trees (n=146). Tree height and crown depth were
calculated from angles to top and bottom of tree and the bottom of crown (excluding small branches
at the bottoms, where the bulk of the trees leaves end; measured with a clinometer) and distance to
the tree (measured with a rangefinder in m). Crown width was estimated as the average of the
crown diameter measured on two axes with a 50-m tape below the tree. Flowering was recorded on
Effect of shade on Kenya bird communities 5
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
a scale of 0–4, representing none, up to 25% of branches with flowers, up to 50%, 75%, and 100%,
respectively.
Coffee understorey data were measured in a square 10 x10 m plot directly adjacent to each
surveyed tree (n=146). The number of coffee shrubs (stems) in each quadrant of the 10 x10m plot
was recorded, the percentage coffee cover in each quadrant was visually estimated (to nearest
10%), and the coffee flowering (if any) was recorded using the same scale as the shade tree
measurements. Additionally, whether there was prominent flowering (>10 stems) and/or seed
prevalence in the understorey was recorded.
Lastly, Maxim iButton temperature loggers were deployed under 12 of the trees (six Cordia and
six Grevillea, one per species per site) and at six locations nearby under no shade trees that acted as
unshaded control samples (one sun location per site). Loggers were tied to the stem of coffee shrubs
within 3m of a shade tree (or at least 15 m from a shade tree for unshaded samples), 2m above
ground and not in direct sunlight. The loggers collected data once every half hour to capture the
warmest and coolest parts of the day, until the batteries died (approximately 43 days). Temperature
loggers were retrieved in April 2019, with 11 successfully located and retrieved (four Cordia, four
Grevillea, three unshaded control).
Statistical analysis
Multiple linear mixed-effects models were used to examine the effects of tree vegetation covariates
on arthropod abundance. A two-sample t-test showed mean branch weights of Cordia and
Grevillea were unequal (df =234.37,t=5.5236,P<0.001). As such, arthropod density was
used as the response variable, calculated as the number of individual arthropods per g of clipped and
inspected branch biomass 100. A Shapiro-Wilks normality test indicated the raw response
variable was not normally distributed, so arthropod density was log-transformed to improve
normality (W =0.9888,P=0.03613). Since arthropods were sampled from the same trees for
which full vegetation variables were measured, model selection for predicting arthropod biomass
included all vegetation variables. Because multiple branches were sampled from the same trees,
tree was treated as a random effect in the model.
Generalized linear mixed-effects models (GLMM) with a Poisson distribution were used to
examine the effects of vegetation variables on bird communities in the canopy of shade trees on
farms. Although data were collected for all bird species detected on the farms regardless of foraging
guild (Appendix S1in the online supplementary material), analysis of bird communities was
limited to insectivores, since that is the guild most relevant to farmers in terms of potential pest
control services. Species were classified as insectivorous based on major dietary preferences
(HBWA 2018). Three separate stepwise model selection analyses were conducted for the bird
community data, using species richness, total abundance of individuals, and abundance of foraging
individuals specifically as response variables, respectively. Rarefaction revealed that the bird
community was sampled adequately with the full sample size (n=333 trees; Appendix S2), but
not with the subset of trees that also included arthropod and detailed vegetation sampling (n=
146 trees), so predictive models for the bird community included only the vegetation data collected
at all trees (tree species, dbh, height). None of the final vegetation variables had a strong correlation
with each other (all r <0.75, VIF <5), so collinearity was not an issue. A Poisson distribution was
used to account for the zero-inflated nature of the detection data and helped meet the model
assumptions necessary for GLMMs. For each analysis, site was treated as a random effect to account
for unmeasured farm-level variation that may have influenced species richness or abundance
(e.g. elevation).
GLMMs with a Poisson distribution was also used to examine the effects of vegetation variables
on bird communities sampled by mist-nets in the crop layer. The number of captures per net and
number of species per net were used as indices of abundance and species richness of birds as the
response variables (see Smith et al. 2015), with tree species, height, canopy cover, dbh, coffee
flowering score, and average percentage understorey cover as predictor variables; site was again
D. Kammerichs-Berke et al. 6
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
used as a random effect. For both canopy and crop layer GLMM analyses, Akaike Information
Criterion corrected for small sample size (AICc) was used to establish model weights and select top
models and (Burnham and Anderson 2002).
Non-metric multidimensional scaling (NMDS) was used to ordinate Bray-Curtis dissimilarity
indices and to identify patterns in the bird community composition data. Because ordinations
cannot be constructed using zero values, the survey data was subsampled to only include trees that
had at least one detection of any species (n=139 trees). Bird community matrices were then
constructed for the canopy and understorey of each tree species from the foraging and banding
data, respectively. Bray-Curtis dissimilarity distances were calculated between each tree, which
were ordinated using a NMDS with no more than 1,000 random starts and 4dimensions (k =4).
Four dimensions were used because any scaling done with fewer dimensions failed to converge
after 1,000 starts. A pairwise Permutational Multivariate Analysis of Variance (PERMANOVA)
with a Bonferroni P-value correction was conducted to compare the community composition of
each analysis of canopy and understory, under the null hypothesis that there is no difference in
community composition between four vegetation levels (canopy and understory each of Cordia
and Grevillea). In all, 999 permutations were used for the PERMANOVA. A multivariate analogue
of Levene’s test was used to test for homogeneity of group variances (Anderson 2006). Simpson’s
indices of diversity and evenness were calculated to determine community diversity and evenness
for each vegetation level. Lastly, an analysis of variance (ANOVA) and Tukey’s honestly significant
difference (HSD) test were used to compare differences in daily maximum, minimum, and mean
daily temperatures between each tree species and the control.
Results
Overall, 2,386 individuals across 23 arthropod taxa groups were detected on native Cordia, while
682 individuals across 18 arthropod groups were detected on non-native Grevillea. The top
performing model predicting arthropod density included tree species and height (Table 1), with
Grevillea and tree height both negatively associated with arthropod density (Figure 2). The mean
Table 1. AICc results of the competing linear regression model set which included tree species, tree height,
and diameter at breast height (dbh) as predictors to arthropod biomass on coffee farms in Kiambu County,
Kenya, winter 2018–2019.
Response
Variable Model K
a
Log
e
(L)
b
AIC
cc
ΔAIC
cd
Wi
e
Arthropod
Biomass
Tree Species þHeight 5-392.29 794.81 0.00 0.65
Tree Species 4-393.90 795.96 1.15 0.35
Tree Species þHeight þAv. Crown Spread. 6-394.46 801.23 6.42 0.03
Tree Species þHeight þAv. Crown Spread
þCanopy Cover
7-398.79 812.00 17.19 0.00
All Vegetation 8-401.63 819.81 25.00 0.00
Height 4-406.21 820.58 25.76 0.00
Null 3-420.67 847.43 52.61 0.00
Av. Crown Spread 4-421.29 850.72 55.91 0.00
Canopy Cover 4-422.17 852.50 57.69 0.00
Dbh 4-423.71 855.58 60.77 0.00
a
Number of parameters
b
Log
e
(likelihood)
c
Akaike’s Information Criterion corrected for small sample size
d
Difference between AIC
c
and top model AIC
c
e
AIC
c
weight
Effect of shade on Kenya bird communities 7
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
density of arthropods per 100 g of clipped and inspected branch vegetation was over four times
higher on Cordia branches (17.07 2.10) than on Grevillea (3.39 0.39).
In total, 841 individuals of 19 insectivorous bird species were detected in the avian surveys: Batis
molitor, Terpisphone viridis, Melaniparus albiventris, Sylvietta whytii, Apalis flavida,Phyllosco-
pus trochilus, Ploceus baglafecht, two species of sylviid warblers (Family Sylviidae), two white-
eyes (Family Zosteropidae), two Old World flycatchers (Family Muscicapidae), and six species of
sunbirds (Family Nectariniidae; Table 2). Tree species and height were the top predictors of avian
species richness, total abundance, and abundance of foraging individuals (Table 3). Grevillea was
negatively associated with richness (β=-0.743 0.097,95%CI=-0.935,-0.554), total abundance
(β=-1.019 0.092,95%CI=-1.203,-0.835), and foraging abundance (β=-1.327 0.133,95%
CI =-1.595,-1.069). Tree height was positively associated with richness (β=0.038 0.009,95%CI
=0.019,0.057), total abundance (β=0.035 0.008,95%CI=0.018,0.053), and foraging
abundance (β=0.039 0.012,95%CI=0.015,0.063;Table 4). Relative to Grevillea,a10-min
survey of Cordia trees on average contained þ0.98 species, þ1.61 total birds, and þ1.1foraging
birds (Figure 3).
In total, 278 individuals of the same 19 insectivorous bird species were detected by mist-nets in
the understorey of shade farms. Average coffee flowering score, canopy cover, and understorey
cover were the top predictors of total relative abundance in the crop layer, whereas average
coffee flowering score and canopy cover were top predictors of species richness (Appendix S3).
Average coffee flowering score was negatively associated with total abundance (β=-0.688 0.184,
95%CI=-1.061,-0.333), whereas canopy cover was positively associated with abundance
(β=0.013 0.003,95%CI=0.006,0.019), as was understorey cover (β=0.006 0.003,95%
CI =0.0008,0.013). Average coffee flowering score was negatively associated with species richness
Figure 2. Arthropod density (arthropods per 100 g clipped and inspected vegetation) on Cordia and
Grevillea trees on coffee farms in Kiambu County, Kenya, winter 2018-2019.Cordia had signifi-
cantly higher arthropod density than Grevillea (P=0.0002), and shorter trees had higher biomass
regardless of tree species (P=0.0167). Enlarged dots represent the mean arthropod density for each
tree species and mean height.
D. Kammerichs-Berke et al. 8
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
(β=-0.899 0.241,95%CI=-1.393,-0.440), while canopy cover was positively associated with
richness (β=0.007 0.003,95%CI=-0.0001,0.0147; Appendix S4). Shade tree species was not
strongly associated with bird abundance or richness sampled by mist-nets in the understorey.
All 19 focal species were detected in the canopy of Cordia,18 in the understorey of Cordia,12 in
the canopy of Grevillea, and 17 in the understorey of Grevillea, with 10 species detected in all four
vegetation levels (Table 2). In the ordination, a stress level of 0.141 was obtained at convergence,
indicating good ordination goodness-of-fit. The canopy community of Cordia was marginally
more even than in Grevillea, and the understory community of Grevillea was the most even and
most diverse of all four vegetation levels (Appendix S5). Pairwise PERMANOVA indicated that the
bird community composition in the Grevillea canopy was significantly different from the Cordia
canopy (r
2
=0.086,F=6.437,p
adj
=0.006,df=1), the Cordia understory (r
2
=0.103,F=7.857,p
adj
=0.006,df=1), and the Grevillea understory (r
2
=0.100,F=7.185,p
adj
=0.006,df=1). The
community composition did not differ significantly between any other pair of vegetation layers
(Table 5,Figure 4). Variance was also shown to be unequal between most groups (F =21.596,P<
0.001,df=3), with only Cordia understorey and Grevillea understorey communities having equal
variance. However, pairwise PERMANOVAs are resilient to heterogeneity of variance in balanced
designs such as this one (Anderson and Walsh 2013), so the results of the pairwise PERMANOVA
should not be a result of inequal variances.
Shade trees buffered temperatures in coffee, and this affect was similar under Cordia and
Grevillea. The maximum daily temperature was 3.2–3.5
o
C lower under shade trees than in the
unshaded control, and this affect was significant for both Cordia and Grevillea (P<0.01; Appendix
S6). Likewise, the minimum daily temperature was warmer under shade than in the unshaded
Table 2. Detected abundances of each focal insectivorous bird species for each vegetation level on coffee
farms in Kiambu County, Kenya, winter 2018–2019. Birds were detected at the canopy level using 10-minute
focal tree observations and at the understory level using mist nets.
Common Name Latin Name
Vegetation Level
Canopy-
Cordia
Understory-
Cordia
Canopy-
Grevillea
Understory-
Grevillea
Chinspot Batis Batis molitor 24 0 2
African Paradise-Flycatcher Terpsiphone viridis 37 211
White-bellied Tit Melaniparus
albiventris
19 0 3
Red-faced Crombec Sylvietta whytii 17 1 9
Yellow-breasted Apalis Apalis flavida 18 5 5
Willow Warbler Phylloscopus trochilus 17 7 1 7
Eurasian Blackcap Sylvia atricapilla 46 0 7
Garden Warbler Sylvia borin 11 0 1
Pale White-Eye Zosterops flavilateralis 73 7 2
Kikuyu White-Eye Zosterops kikuyuensis 9471319
Pale Flycatcher Agricola pallidus 74 0 2
White-eyed Slaty-Flycatcher Melaenornis fischeri 42 0 5
Collared Sunbird Hedydipna collaris 14 0 0
Green-headed Sunbird Cyanomitra verticalis 11 1 0
Amethyst Sunbird Chalcomitra
amethystina
10 1 1
Scarlet-chested Sunbird Chalcomitra
senegalensis
41 1 9
Bronze Sunbird Nectarinia kilimensis 12 20 2 15
Variable Sunbird Cinnyris venustus 16 12 3 13
Baglafecht Weaver Ploceus baglafecht 212 1 12
Effect of shade on Kenya bird communities 9
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
control, and this was significant for Cordia (þ1.2
o
C, P<0.01) but not Grevillea-Control: þ0.8
o
C,
P=0.15). Mean daily temperatures were similar among both shade tree species and in the unshaded
control sites, though mean temperatures were marginally cooler under Cordia than Grevillea
(-0.5
o
C difference, 95%CI=-0.043,1.096,P=0.08).
Table 3. AICc results of the competing general linear model set which included tree species, tree height, and
diameter at breast height (dbh) as predictors to insectivorous bird species richness, abundance, and foraging
on coffee farms in Kiambu County, Kenya, winter 2018-2019.
Response Variable Model K
a
Log
e
(L)
b
AIC
cc
Delta AIC
cd
Wi
e
Richness Species þHeight þ(1|Site) 4-620.70 1249.51 0.00 0.69
Species þHeight þdbh þ(1|Site) 5-620.48 1251.15 1.64 0.30
Species þdbh þ(1|Site) 4-624.85 1257.82 8.31 0.01
Species þ(1|Site) 3-628.83 1263.73 14.22 0.00
dbh þ(1|Site) 3-643.92 1293.91 44.40 0.00
Height þdbh þ(1|Site) 4-643.03 1294.17 44.66 0.00
1þ(1|Site) 2-651.96 1307.95 58.44 0.00
Height þ(1|Site) 3-651.85 1309.77 60.26 0.00
Abundance Species þHeight þ(1|Site) 4-825.37 1658.86 0.00 0.68
Species þHeight þdbh þ(1|Site) 5-825.15 1660.48 1.62 0.30
Species þdbh þ(1|Site) 4-829.42 1666.96 8.10 0.01
Species þ(1|Site) 3-833.29 1672.64 13.78 0.00
Height þdbh þ(1|Site) 4-875.74 1759.60 100.74 0.00
dbh þ(1|Site) 3-875.74 1772.28 113.42 0.00
Height þ(1|Site) 3-892.08 1790.23 131.37 0.00
1þ(1|Site) 2-893.32 1790.67 131.80 0.00
Foraging Species þHeight þ(1|Site) 4-614.42 1236.95 0.00 0.69
Species þHeight þdbh þ(1|Site) 5-614.36 1238.89 1.94 0.26
Species þdbh þ(1|Site) 4-617.27 1242.65 5.70 0.04
Species þ(1|Site) 3-619.43 1244.92 7.97 0.01
Height þdbh þ(1|Site) 4-658.99 1326.10 89.15 0.00
dbh þ(1|Site) 3-666.94 1339.96 103.00 0.00
Height þ(1|Site) 3-671.36 1348.79 111.84 0.00
1þ(1|Site) 2-673.43 1350.90 113.95 0.00
a
Number of parameters
b
Log
e
(likelihood)
c
Akaike’s Information Criterion corrected for small sample size
d
Difference between AIC
c
and top model AIC
c
e
AIC
c
weight
Table 4. Results from top models for insectivorous bird species richness, abundance, and foraging on coffee
farms in Kiambu County, Kenya, winter 2018-2019.
Response Covariate βSE CI (95%)
Richness Intercept 0.327 0.147 0.036,0.616
Species (Grevillea) -0.743 0.097 -0.935,-0.554
Height 0.038 0.009 0.019,0.057
Abundance Intercept 0.700 0.183 0.316,1.084
Species (Grevillea) -1.019 0.092 -1.203,-0.835
Height 0.035 0.008 0.018,0.053
Foraging Intercept 0.096 0.232 -0.381,0.572
Species (Grevillea) -1.327 0.133 -1.595,-1.069
Height 0.039 0.012 0.015,0.063
D. Kammerichs-Berke et al. 10
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
Discussion
Shade coffee is important for the conservation of birds globally, but there is a need to better
understand the effects of particular shade tree species on bird communities and the implications for
shade tree use for conservation and ecosystem services (Narango et al. 2018,2019). As predicted by
ecological theory relating native vegetation to species diversity (Tallamy 2004), native Cordia trees
in Kenyan shade coffee farms hosted not only a higher density of arthropods than did non-native
Grevillea (Figure 4), but Cordia also had higher abundance of insectivorous birds and specifically
more foraging individuals than Grevillea (Figure 5). Cordia also had greater bird species richness
than did Grevillea. All 19 focal species were detected in Cordia, and the most abundant species
(Willow Warbler Phylloscopus trochilus) accounted for 18% of all individual detections. In con-
trast, only 12 of the focal insectivorous bird species were detected in Grevillea, and one species
(Kikuyu White-eye Zosterops kikuyuensis) accounted for 34% of all detections.
Optimal foraging theory predicts that animals distributed in patchy environments should select
the most profitable patches to forage in and decide when to leave the patch they are using, given
Table 5. Pairwise PERMANOVA results for insectivore community similarities between each pair of
vegetation levels on coffee farms in Kiambu County, Kenya, winter 2018-2019.
Pairs Df Sum of Squares F R
2
P
adj
Canopy-Cordia / Understory-Cordia 10.530 3.269 0.043 0.054
Canopy-Cordia / Canopy-Grevillea 10.795 6.437 0.086 0.006**
Canopy-Cordia / Understory-Grevillea 10.366 2.233 0.031 0.300
Understory-Cordia / Canopy-Grevillea 11.046 7.857 0.103 0.006**
Understory-Cordia / Understory-Grevillea 10.170 0.981 0.014 1.000
Canopy-Grevillea / Understory-Grevillea 10.957 7.185 0.100 0.006**
** Statistically significant (p
adj
<0.05)
Figure 3. Mean number (X
¯/10-minute survey 1SE) of total individuals, foraging individuals,
and bird species richness per 10-minute survey of Cordia and Grevillea shade trees on coffee farms
in Kiambu County, Kenya, winter 2018-2019.
Effect of shade on Kenya bird communities 11
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
that the intake rates will vary among patches (Pyke 1984). Based on the functional response of
animals to prey density (Holling 1965), feeding insectivorous birds should distribute among
feeding patches according to their supply of insects, the so-called “habitat matching” rule
(Fretwell 1972, Fagen 1987, Johnson and Sherry 2001). Because most insect taxa specialise on
one or few native host plants, it is expected that herbivorous insects should be more common on
native than exotic plants (Burghardt et al. 2010, Litt et al. 2014), and correspondingly insect-eating
birds should forage more on natives than exotics (Narango et al. 2018). Although this study
involved only a single pair of native and non-native tree species, the results are consistent with
ecological theory of higher abundances of arthropods on native plants, which in turn would support
more insectivorous birds that can forage on pest arthropods in the crop layer (Narango et al. 2018).
This is relevant to farm managers because many of the ecosystem services that birds provide in
agricultural landscapes result from their dietary preferences and foraging behaviour (Wenny et al.
2011). Insectivorous birds are more likely than other foraging guilds to provide beneficial top-
down control of pest species (Kellermann et al. 2008, Philpott et al. 2008, Johnson et al. 2010), and
are generally also at higher conservation risk due to their stronger associations with forest habitats
(Bennun et al. 1996, Sekercioglu et al. 2002, HBWA 2018).
The notion that shade trees could attract insectivorous birds helpful for control of pests on coffee
shrubs rests on the assumption that birds using the shade trees also forage in the associated
understory, but this has rarely been examined explicitly (but see Smith et al. 2012). Because the
preferred vegetation profiles for foraging vary among bird species, some natural variation between
canopy and crop level bird communities is expected. Nonetheless, the bird communities were
nearly identical between the Cordia canopy and the crop layer (94.7% species overlap), whereas
they were much less so between Grevillea canopy and understorey (64.7% species overlap), with
several species detected in the Grevillea understorey but not its canopy. The crop layer under both
Cordia and Grevillea trees more closely resembled the canopy-level communities in Cordia trees,
Figure 4. Non-metric multi-dimensional scaling (NMDS) plot of insectivorous bird community
similarities between each vegetation level on coffee farms in Kiambu County, Kenya, winter 2018-
2019. Canopy-Grevillea differs significantly from Canopy-Cordia (adj-p =0.006), Understory-
Cordia (adj-p =0.006), and Understorey-Grevillea (adj-p =0.006). Ellipses represent 95%CI
around the centroids of each community.
D. Kammerichs-Berke et al. 12
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
suggesting that Grevillea had comparatively less influence on the crop-level bird communities. The
resemblance between the crop layer, regardless of shade tree species, and the Cordia canopy
suggests that Cordia attracts birds to the canopy with its high abundance of non pest arthropods,
and birds then move down and spread out to forage throughout the crop layer. These results,
coupled with the relatively low arthropod abundance on in the crop layer (Milligan 2014, Smith
et al. 2018), suggest Cordia attracts greater numbers of insect-eating birds to the canopy that
subsequently spill over into the crop layer, increasing the potential for birds to predate on pest
species such as coffee berry borer, white coffee stem-borer Xylotrechus quadripes, and scale insects
(Superfamily Coccoidea). Our data did not suggest that native trees increased the abundance of pest
species, because none of the species observed during the arthropod sampling were known coffee
pests. In the Neotropics, avian predators of coffee berry borer and other coffee insects are mainly
small-billed, small bodied, foliage gleaning insectivores, such as Parulid warblers (Karp and Daily
2014, Sherry et al. 2016). Diet data are not yet available for the birds inhabiting East African coffee,
but based on morphology, white-eyes (Zosterops spp.) may be a likely candidate for pest control.
Notably, there were considerably more Z. kikuyuensis in the crop layer below Cordia than
Grevillea, even though Z. kikuyuensis comprised the majority of individuals detected in the
canopy of Grevillea. While more Z. kikuyuensis were detected in the canopy of Grevillea than
Cordia, most of the individuals were observed collecting nesting material such as spiderweb and
tree fibre mand were rarely seen actively foraging. Of course, insectivorous birds could remove
more pest-eating insects than the pests themselves, and this intra-guild predation could result in a
net negative effect of birds on coffee pests (Mu
¨ller and Brodeur 2002, Perfecto et al. 2014).
Cordia may be preferred by farmers for other reasons besides their attractiveness to insect-
eating birds. Cordia are generally wide-canopied trees, which, while sometimes taking up more
space on the farm, provide the coffee crop with greater amounts of shade that may help adapt to
expected climate warming (Kammerichs-Berke 2020). Coffee berry borer reproductive rates are
associated with warming temperatures (Jaramillo et al. 2009,2011), and data indicate that coffee
under the canopy of both Cordia and Grevillea trees had a more restricted temperature range than
in the sun, with marginally cooler mean temperatures under Cordia than Grevillea. These buffered
temperatures could affect the productivity of pests that would proliferate under warmer temper-
atures (Jaramillo et al. 2009) and help adapt to expected climate warming (Schooler et al. 2020).
Grevillea robusta proliferated as a shade tree in central Kenya in the latter half of the 20
th
century largely due to the growth of the Greenbelt Movement. With the mission of community
empowerment and conservation, the Green Belt Movement planted millions of trees throughout
Kenya, particularly in agricultural areas such as the Kiambu region (Chikwendu 2008). Grevillea
was chosen largely because it grows quickly (up to 3m per year; SelecTree 2020) and yields high,
immediate material benefits such as firewood. However, in recent decades the Greenbelt Move-
ment has shifted its stance to encourage the use of native species, including Cordia, in environ-
mentally sensitive areas (Murithi et al. 2009). Cordia, while slower growing, may yield greater
environmental conservation benefits as well as similar material benefits in the long term
(Alemayehu et al. 2016). Cordia has various uses as medicine, food, firewood, fodder, and mulch
(Alemayehu et al. 2016), and is considered an attractive species for beekeeping and honey pro-
duction (Fichtl and Adi 1994). Cordia also provides a greater windbreak than Grevillea, offering
better crop protection during rainy season storm events (J. Murithi pers. comm.).
Research priorities
The clear next step is to test if species detected in the crop understorey are in fact removing insects
from coffee plants. Insectivorous birds have been confirmed to help control coffee pests in the
Neotropics (Kellermann et al. 2008, Johnson et al. 2010, Karp et al. 2013, Sherry et al. 2016), but
this phenomenon has been much less studied in East Africa. Exclosure experiments in Tanzanian
coffee farms confirmed a significant increase in herbivory rates on bushes from which birds and
bats were excluded (Classen et al. 2014), and a sentinel pest removal experiment in Nyeri County,
Effect of shade on Kenya bird communities 13
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
Kenya, documented greater insect removal rates in shade versus sun farms (Milligan et al. 2016).
However, confirmation of Kenyan birds as pest predators awaits examination of their diets and
additional experimental exclosure studies. In our study area, fecal samples were collected from
birds captured in mist-nets, and on-going molecular analysis should reveal diet compositions of
insectivorous birds (Jedlicka et al. unpubl. data).
Despite the economic, cultural, and ecological significance of coffee in Africa, its role in conser-
vation on the continent is poorly understood, especially compared to the abundance of coffee-
related ecological research done in the western hemisphere. With a combined worth of US$
70 billion, the coffee industry plays a significant role in the global economy (Osorio 2002). Coffee
is a major export of several tropical and subtropical countries in Central and South America, Asia,
and Africa, and the industry supports approximately 125 million people worldwide (Osorio 2002,
FAO 2016). With roughly 20% of the world’s 10 million ha of harvested area, Africa is one of the
world’s leading producers of coffee. Coffee is a major cash crop in Kenya, third only to tea and
horticulture produce in export earnings. Approximately 110,000 ha of land are harvested for coffee,
and the industry supports about 5million people within these areas (KALRO 2015). Few studies on
birds and coffee have been conducted in East Africa, but among them they show conflicting results
(Pinard et al. 2014a, Buechley et al. 2015, Smith et al. 2015, Milligan et al. 2016). These various
results arise from the first few studies of birds in East African coffee farms, and they have followed
basic survey designs completed much earlier and replicated many times in the Neotropics, from
which broad observable patterns have now emerged (Philpott et al. 2008). It is therefore vital to
continue examining birds and other wildlife in coffee systems in East Africa to gain a more
complete understanding of the agroecosystems in this region.
Supplementary Materials
To view supplementary material for this article, please visit http://doi.org/10.1017/
S0959270921000502.
Acknowledgements
We are grateful to James Murithi of Sasini Ltd and Duncan Kanyi of Theta Country Farm for
granting us permission to work on their coffee farms. We thank Mark Colwell and Erik Jules, whose
input, particularly with the multivariate analyses, was invaluable. Kristina Wolf, Manny Hernan-
dez, and Christopher Watson helped with data collectionin the field. This project was funded by the
National Science Foundation (IRES award #1657973 and #1657836).
References
Alemayehu, G., Asfaw, Z. and Kelbessa, E.
(2016)Cordia africana (Boraginaceae) in
Ethiopia: A review on its taxonomy, distri-
bution, ethnobotany and conservation sta-
tus. Internatn. J. Bot. Stud. 1:38–46.
Anderson, M. J. (2006) Distance-based tests
for homogeneity of multivariate disper-
sions. Biometrics 62:245–253.
Anderson, M. J. and Walsh, D. C. I. (2013)
PERMANOVA, ANOSIM, and the Mantel
test in the face of heterogeneous disper-
sions: What null hypothesis are you test-
ing? Ecol. Monogr. 83:557–574.
Armbrecht, I. and Perfecto, I. (2003) Litter-
twig dwelling ant species richness and pre-
dation potential within a forest fragment
and neighboring coffee plantations of con-
trasting habitat quality in Mexico. Agric.
Ecosyst. Environ. 97:107–115.
Baggio, A. J., Caramori, P. H., Filho, A. and
Montoya, L. (1997) Productivity of South-
ern Brazilian coffee plantations shaded by
different stockings of Grevillea robusta.
Agroforest. Syst. 37:111–120.
Becerra, J. X., and Venable, D. L. (1999)Macro-
evolution of insect-plant associations: The
D. Kammerichs-Berke et al. 14
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
relevance of host biogeography to host affili-
ation. Proc. Natl. Ac. Sci. 96:12626–12631.
Beer, J. (1987) Advantages, disadvantages and
desirable characteristics of shade trees for
coffee, cocoa and tea. Agroforest. Syst. 5:
3–13.
Bernays, E. M. and Graham, M. (1988) On the
evolution of host specificity in phytopha-
gous arthropods. Ecology 69:886–892.
Bennun, L., Dranzoa, C. and Pomeroy, D.
(1996) The forest birds of Kenya and
Uganda. J. East Afr, Nat. Hist. 85:23–48.
Betts, M. G., Wolf, C., Ripple, W. J., Phalan, B.,
Millers, K. A., Duarte, A., Butchart, S. H. M.
and Levi, T. (2017) Global forest loss dispro-
portionately erodes biodiversity in intact
landscapes. Nature 547:441–444.
Brown, J. H. (2014) Why are there so many
species in the tropics? J. Biogeogr. 41:8–22.
Buechley, E. R., S¸ ekerciog˘lu, C¸ . H., Atickem,
A., Gebremichael, G., Ndungu, J. K., Maha-
mued, B. A., Beyene, T., Mekonnen, T. and
Lens, L. (2015) Importance of Ethiopian
shade coffee farms for forest bird conserva-
tion. Biol. Conserv. 188:50–60.
Burghardt, K. T., Tallamy, D. W. Philips, C. and
Shropshire, K. J. (2010) Non-native plants
reduce abundance, richness, and host spe-
cialization in lepidopteran communities.
Ecosphere 1: art11.
Burnham, K. P. and Anderson, D. R. (2002)
Model selection and multimodel inference:
A practical information-theoretic approach.
New York, New York: Springer-Verlag.
Classen, A., Peters, M. K., Ferger, S. W.,
Helbig-Bonitz, M., Schmack, J. M., Maas-
sen, G., Schleuning, M., Kalko, E. K. V.,
Bo
¨hnig-Gaese, K. and Steffan-Dewenter, I.
(2014) Complementary ecosystem, services
provided by pest predators and pollinators
increase quantity and quality of coffee
yields. Proc. Roy. Soc. 281:20133148.
Chikwendu, E. (2008) Wangari Maathai:
Nobel laureate, environmentalist, and
engagement with the Kenya state.
Pp. 29–44 in C. J. Korieh and P. E. Okeke-
Ihejirika, eds. Gendering global transform-
ations: Gender, culture, race, and identity.
Abingdon-on-Thames, UK: Routledge.
Davis, H., Rice, R., Rockwood, L., Wood, T. and
Marra, P. (2017) The economic potential of
fruit trees as shade in blue mountain coffee
agroecosystems of the Yallahs River water-
shed, Jamaica W.I. Agroforest. Syst. 93:
581–589.
Donald, P. F. (2004) Biodiversity impacts of
some agricultural commodity production
systems. Conserv. Biol. 18:17–38.
Douglas, D. J. T., Nalwanga, D., Katebaka, R.,
Atkinson, P. W., Pomeroy, D. E., Nkuutu, D.
and Vickery, J. A. (2013) The importance of
native trees for forest bird conservation in
tropical farmland. Anim. Conserv. 17:
256–264.
Eagen, R. (1987) A generalized habitat match-
ing rule. Evol. Ecol. 1:5–10.
Ehrlich, P. L, and Raven, P. H. (1964) Butter-
flies and plants: A study in coevolution.
Evolution 18:586–608.
FAO (2016) FAOSTAT. Food and Agriculture
Organization of the United Nations.
<http://www.fao.org/faostat/en/#home>.
Accessed 6Sep 2018.
Farrell, B. D. and Mitter, C. (1998) The timing
of insect/plant diversification: might Tetra-
opes (Coleoptera: Cerambycidae) and
Asclepias (Asclepiadaceae) have co-
evolved? Biol. J. Linn. Soc. 63:553–577.
Fichtl, R. and Adi, A. (1994)Honeybee flora of
Ethiopia. Weikersheim, Germany: Margaf
Verlag.
Fischer, J., Abson, D. J., Butsic, V., Chappell,
M. J., Ekroos, J., Hanspach, J., Kuemmerle,
T., Smith, H. G. and von Wehrden, H.
(2014) Land sparing versus land sharing:
Moving forward. Conserv. Lett. 7:149–157.
Foley, J. A., DeFries, R., Asner, G. P., Barford,
C., Bonan, G., Carpenter, S. R., Chapin, F. S.,
Coe, M. T., Daily, G. C., Gibbs, H. K., Helk-
owski, J. H., Holloway, T., Howard, E. A.,
Kucharik, C. J., Monfreda, C., Patz, J. A.,
Prentice, I. C., Ramankutty, N. and Snyder
P. K. (2005) Global consequences of land
use. Science 309:570–574.
Forister, M. L., Novotny, V., Panorska, A. K.,
Baje, L., Basset, Y., Butterill, P. T., Cizek, L.,
Coley, P. D., Dem, F., Diniz, I. R., Pavel, D.,
Fox, M., Glassmire, A. E., Hazen, R., Hrcek,
J., Jahner, P., Kaman, O., Kozubowski, T. J.,
Kursar, T. A., Lewis, O. T., Lill, J., Marquis,
R. J. Miller, S. E., Morais, H. C., Murakami,
M., Nickel, H., Pardikes, N. A., Ricklefs,
R. E., Singer, M. S., Smilanich, A. M., Stire-
man, J. O., Villamarin-Cortez, S., Vodka, S.,
Effect of shade on Kenya bird communities 15
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
Volf, M., Wagner, D. L., Walla, T., Weiblen,
G. D. and Dyer, L. A. (2015) The global
distribution of diet breadth in insect herbi-
vores. Proc. Natl. Ac. Sci. 112:442–447.
Fretwell, S. D. (1972)Populations in a sea-
sonal environment. Princeton, USA:
Princeton University Press.
Handbook of Birds of the World Alive
[HBWA] (2018) Species home page.
<https://www.hbw.com/species>. Accessed
6Sep 2018.
Holling, C. (1965) The functional response of
predators to prey density and its role in
mimicry and population regulation.
Memoirs Entomol. Soc. Canada 97 (S45):
5–60.
Jaramillo, J., Chabi-Olaye, A., Kamonjo, C.,
Jaramillo, A., Vega, F. E., Poehling, H. M.
and Borgemeister, C. (2009) Thermal toler-
ance of the coffee berry borer Hypothene-
mus hampei: predictions of climate change
impact on a tropical insect pest. PLoS ONE
4:e6487.
Jaramillo, J., Muchugu, E., Vega, F. E., Davis,
A., Borgemeister, C. and Chabi-Olaye, A.
(2011) Some like it hot: the influence and
implications of climate change on coffee
berry borer (Hypothenemus hampei) and
coffee production in East Africa. PLoS
ONE 6:e24528.
Jaramillo, J., Setamou, M., Muchugu, E.,
Chabi-Olaye, A., Jaramillo, A., Mukabana,
J., Maina, J., Gathara, S. and Borgemeister,
C. (2013) Climate change or urbanization?
Impacts on a traditional coffee production
system in East Africa over the last 80 years.
PLoS ONE 8:e51815.
Jassogne, L., van Astern, P. J. A., Wanyama, I.
and Baret, P. (2012) Perceptions and outlook
on intercropping coffee with banana as an
opportunity for smallholder coffee farmers
in Uganda. Internatn. J. Agric. Sustain. 11:
144–158.
Jedlicka, J. A., Greenberg, R. and Letourneau,
D. K. (2011) Avian conservation practices
strengthen ecosystem services in California
vineyards. PLoS ONE 6:e27347.
Jha,S.,Bacon,C.M.,Philpott,S.M.,Ernesto
Me´ ndez, V., La
¨derach,P.andRice,R.A.
(2014) Shade coffee: Update on a disappearing
refuge for biodiversity. BioScience 64:
416–428.
Johnson, M. D. (2000a) Effects of shade-tree
species and crop structure on the winter
arthropod and bird communities in a Jamai-
can shade coffee plantation. Biotropica 32:
133–145.
Johnson, M. D. (2000b) Evaluation of an
arthropod sampling technique for measur-
ing food availability for forest insectivorous
birds. J. Field Ornithol. 71:88–109.
Johnson, M. D. and Sherry, T. W. (2001)
Effects of food availability on the distribu-
tion of migratory warblers among habitats
in Jamaica. J. Anim. Ecol. 70:546–560.
Johnson, M. D., Kellermann, J. L. and Stercho,
A. M. (2010) Pest reduction services by
birds in shade and sun coffee in Jamaica.
Anim. Conserv. 13:140–147.
Johnson, M. D. and Hackett, S. C. (2016) Why
birds matter economically: Values, markets,
and policies. Page 29 in Why birds matter:
Avian ecological function and ecosystem
services. Chicago, Illinois: University of
Chicago Press.
KALRO (2015)Annual report 2014-2015.
Nairobi, Kenya: Kenya Agricultural and
Livestock Research Organization.
Kammerichs-Berke, D. (2020) The effect of
shade tree species on bird communities in
central Kenyan coffee farms. Master’s the-
sis, Humboldt State University, Arcata,
California.
Karp, D. S., Mendenhall, C. D., Sandi, R. F.,
Chaumont, N., Ehrlich, P. R., Hadly, E. A.
and Daily, G. C. (2013) Forest bolsters bird
abundance, pest control and coffee yield.
Ecol. Lett. 16:1339–1347.
Karp, D. S. and Daily, G. C. (2014) Cascading
effects of insectivorous birds and bats in
tropical coffee plantations. Ecology 95:
1065–1074.
Karp, D.S., Judson, S., Daily, G. C. and Hadly,
E. A. (2014) Molecular diagnosis of bird-
mediated pest consumption in tropical
farmland. SpringerPlus:3:630.
Kellermann, J. L., Johnson, M. D., Stercho,
A. M. and Hackett, S. C. (2008) Ecological
and economic services provided by birds on
Jamaican Blue Mountain coffee farms. Con-
serv. Biol. 22:1177–1185.
Kindt, R., Van Damme, P. and Simmons, A. J.
(2006) Tree diversity in western Kenya:
using profiles characterize richness and
D. Kammerichs-Berke et al. 16
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
evenness. Biodivers. Conserv. 15:
1253–1270.
Kiyingi, I., Ocama, D., Mujuni, D. and
Nyombi, K. (2016) A bioeconomic analysis
of the carbon sequestration potential of
agroforestry systems: A case study Grevil-
lea robusta in Western Uganda. Uganda
J. Agric. Sci. 17:219–229.
Lengkeek, A. G., Kindt, R., van der Maesen,
L. J. G., Simons, A. J. and van Oijen, D. C. C.
(2005) Tree density and germplasm source
in agroforestry ecosystems in Meru, Mount
Kenya. Genet. Resour. Crop Evol. 52:
709–721.
Litt, A. R., Cord, E. E., Fulbright, T. E. and
Schuster, G. L. (2014) Effects of invasive
plants on arthropods. Conserv. Biol. 28:
1532–1549.
Lott, J. E., Howard, S. B., Ong, C. and Black,
C. R. (2000) Long-term productivity of a
Grevillea robusta-based overstorey agrofor-
estry system in semi-arid Kenya: II. Crop
growth and system performance. For. Ecol.
Manage. 139:187–201.
Mehrabi, Z., Ellis, E. C. and Ramankutty, N.
(2018) The challenge of feeding the world
while conserving half the planet. Nat. Sus-
tain. 1:409–412.
Milligan, M. C. (2014)Quantifying pest con-
trol services by birds and ants in Kenyan
coffee farms. Master’s thesis, Humboldt
State University, Arcata, California.
Milligan, M. C., Johnson, M. D., Garfinkel,
M., Smith, C. J. and Njoroge, P. (2016)
Quantifying pest control services by birds
and ants in Kenyan coffee farms. Biol. Con-
serv. 194:58–65.
Moguel, P. and Toledo, V. M. (1999) Biodiver-
sity conservation in traditional coffee sys-
tems. Conserv. Biol. 13:11–21.
Muchiri, M. N. (2004) Grevillea robusta in
agroforestry systems in Kenya. J. Trop.
For. Sci. 16:396–401.
Mu
¨ller, C.B. and Brodeur, J. (2002) Intraguild
predation in biological control and conser-
vation biology. Biol. Control 25:216–223.
Murithi, C. W., Ong, C. K., Craigon, J., Mati,
B. M., Ngumi, V. W. and Black, C. R. (2009)
Gas exchange and water use efficiency of
trees and maize in agroforestry systems in
semi-arid Kenya. Agric. Ecosyt. Environ,
129:297–307.
Narango, D. L., Tallamy, D. W. and Marra, P. P.
(2018) Nonnative plants reduce population
growth of an insectivorous bird. Proc. Natl.
Ac. Sci. 201809259.
Narango, D. L., Tallamy, D. W., Snyder, K. J.
and Rice, R. A. (2019) Canopy tree prefer-
ence by insectivorous birds in shade-coffee
farms: Implications for migratory bird con-
servation. Biotropica 51:387–398.
Negash, L. (1995)Native trees of Ethiopia:
Biology, uses and propagation techniques.
Umea, Sweden: SLU Reprocentralen.
Osorio, N. (2002)The global coffee crisis: a
threat to sustainable development. Johan-
nesburg, South Africa: International Coffee
Organization.
Perfecto, I., and Vandermeer, J. (2010) The
agroecological matrix as alternative to the
land- sparing/agriculture intensification
model. Proc. Natl. Ac. Sci. 107:5786–5791.
Perfecto, I., Vandermeer, J. and Philpott, S. M.
(2014) Complex ecological interactions in
the coffee agroecosystem. Annu. Rev. Ecol.
Evol. Syst. 45:137–158.
Perfecto, I., and Vandermeer, J. (2015)Coffee
agroecology: A new approach to under-
standing agricultural biodiversity, ecosys-
tem services and sustainable development.
Abingdon, UK: Routledge.
Perfecto, I., Vandermeer, J., Bautista, G. L.,
Ibarra Nunn
˜ez, G., Greenberg, R., Bichier,
P. and Langridge, S. (2004) Greater preda-
tion in shaded coffee farms: The role of resi-
dent Neotropical birds. Ecology 85:
2677–2681.
Perfecto, I., Vandermeer, J. and Wright, A.
(2009)Nature’s matrix: Linking agricul-
ture, conservation and food sovereignty.
Abingdon, UK: Routledge.
Philpott, S. M., Arendt, W. J., Armbrecht, I.,
Bichier, P., Diestch, T. V., Gordon, C.,
Greenberg, R., Perfecto, I., Reynoso-Santos,
R., Soto-Pinto, L., Tejeda-Cruz, C.,
Williams-Linera, G., Valenzuela, J. and
Zolotoff, J. M. (2008) Biodiversity loss in
Latin American coffee landscapes: Review
of the evidence on ants, birds, and trees.
Conserv. Biol. 22:1093–1105.
Philpott, S. M. and Bichier, P. (2012) Effects of
shade tree removal on birds in coffee agro-
ecosystems in Chiapas, Mexico. Agric. Eco-
syst. Environ. 149:171–180.
Effect of shade on Kenya bird communities 17
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
Philpott, S. M., Soong, O., Lowenstein, J. H.,
Pulido, A. L., Lopez, D. T., Flynn, D. F. and.
DeClerck, F. (2009) Functional richness and
ecosystem services: bird predation on
arthropods in tropical agroecosystems. Ecol.
Applic. 19:1858–1867.
Pinard, F., Joetzjer, E., Kindt, R. and Kehlen-
beck, K. (2014a) Are coffee agroforestry
systems suitable for circa situm conserva-
tion of native trees? A case study from Cen-
tral Kenya. Biodivers. Conserv. 23:467–495.
Pinard, F., Boffa, J. M. and Rwakagara, E.
(2014b) Scattered shade trees improve
low-input smallholder Arabica coffee prod-
uctivity in the northern Lake Kivu region of
Rwanda. Agroforest. Syst. 88:707–718.
Pyke, G. H. (1984) Optimal foraging theory: A
critical review. Annu. Rev. Ecol. Syst. 15:
523–575.
Rahn, E., Vaast, P., La
¨derach, P., van Asten, P.,
Jassogne, L. and Ghazoul, G. (2018) Explor-
ing adaptation strategies of coffee produc-
tion to climate change using a process-based
model. Ecol. Model. 371:76–89.
Railsback, S. F. and Johnson, M. D. (2014)
Effects of land use on bird populations and
pest control service on coffee farms. Proc.
Natl. Ac. Sci. 1320957111.
Raman, T. R. S. (2006) Effects of habitat struc-
ture and adjacent habitats on birds in trop-
ical rainforest fragments and shaded
plantations in the Western Ghats, India.
Biodivers. Conserv. 15:1577–1607.
Remsen, J. V. and Robinson, S. K. (1981)A
classification scheme for foraging behavior
of birds in terrestrial habitats. Stud. Avian
Biol. 13:144–160.
Renwick, A. R., Vickery, J. A., Potts, S. G., Bol-
wig, S., Nalwanga, D., Pomeroy, D. E., Mush-
abe, D. and Atkinson, P. W. (2014)Achieving
production and conservation simultaneously
in tropical agricultural landscapes. Agric.
Ecosyst. Environ. 192:130–134.
Romero-Alvarado, Y., Soto-Pinto, L., Garcı´a-
Barrios, L. and Barrera-Gayta´n, J. F. (2000)
Coffee yields and soil nutrients under the
shades of Inga sp. vs. multiple species in
Chiapas, Mexico. Agroforest. Syst. 54:
215–224.
Schemske, D. W., Mittelbach, G. G., Cornell,
H. V., Sobel, J. M. and Roy, K. (2009)Is
there a latitudinal gradient in the
importance of biotic interactions? Annu.
Rev. Ecol. Syst. 40:245–269.
Schooler, S. L., Johnson, M. D., Njoroge, P. and
Bean, W. T. (2020) Shade trees preserve
avian insectivore biodiversity on coffee
farms in a warming climate. Ecol. Evol. 10:
1–13.
Sekercioglu, C. H., Ehrlich, P. R., Daily, G. C.,
Aygen, D., Goehring, D. and Sandi, R. F.
(2002) Proc. Natl. Acad. Sci. 99:263–267.
SelecTree (2020) “Grevillea robusta Tree
Record.” CalPoly SelecTree Database
(1995-2020). <https://selectree.calpoly.
edu/tree-detail/grevillea-robusta>.
Accessed 31 July 2020.
Sherry, T.W., Johnson, M. D., Williams, K. A.,
Kaban, J. D., McAvoy, C. K., Hallauer,
A. M., Rainey, S. and Xu, S. (2016) Dietary
opportunism, resource partitioning, and
consumption of coffee berry borers by five
species of migratory wood warblers
(Parulidae) wintering in Jamaican shade cof-
fee plantations. J. Field Ornithol. 87:
273–292.
Smith, C. J., Johnson, M. D., Campos, B. R. and
Bishop, C. M. (2012) Variation in aggres-
sion of Black-throated Blue Warblers win-
tering in Jamaica. Condor 114:831–839.
Smith, C., Barton, D., Johnson, M. D., Wendt,
C., Milligan, M. C., Njoroge, P. and Gichuki,
P. (2015) Bird communities in sun and shade
coffee farms in Kenya. Global Ecol. Con-
serv. 4:479–490.
Smith, C., M. Milligan, C., Johnson, M. D. and
Njoroge, P. (2018) Bird community
response to landscape and foliage arthropod
variables in sun coffee of central Kenyan
highlands. Global Ecol. Conserv. 13:e00378.
Soto-Pinto, L., Perfecto, I., Castillo-
Hernandez, J. and Caballero-Nieto, J.
(2000) Shade effect on coffee production at
the northern Tzeltal zone of the state of
Chiapas, Mexico. Agric. Ecosyst. Environ.
80:61–69.
Southwood,T. R. E., Moran, V. C. and Kennedy,
C. E. J. (1982) The richness, abundance, and
biomass of the arthropod communities on
trees. J. Anim. Ecol. 51:635–649.
Tilman, D., Balzer, C., Hill, J. and Befort, B. L.
(2011) Global food demand and the
sustainable intensification of agriculture.
Proc. Natl. Ac. Sci. 108:20260–20264.
D. Kammerichs-Berke et al. 18
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
Tignor, R. L. (2015)Colonial transformation
of Kenya: The Kambu, Kikuyu, and Maasai
from 1900-1939. Princeton, New Jersey:
Princeton University Press.
Van Bael, S. A., Philpott, S. M., Greenberg, R.,
Bichier, P., Barber, N. A., Mooney, K. A. and
Gruner, D. S. (2008) Birds as predators in
tropical agroforestry systems. Ecology 89:
928–934.
Vandermeer, J. and Perfecto, I. (1997)
The agroecosystem: a need for the conser-
vation biologist’s lens. Conserv. Biol. 11:
591–592.
Wenny, D. G., DeVault, T. L., Johnson, M. D.,
Kelly, D., Sekercioglu, C. H., Tomback, D. F.
and Whelan, C. J. (2011) The need to quan-
tify ecosystem services provided by birds.
The Auk 128:1–14.
Effect of shade on Kenya bird communities 19
https://doi.org/10.1017/S0959270921000502 Published online by Cambridge University Press
