Abstract
Uganda lies within the drier end of the natural distribution range ofCoffea canephora and contains unexplored genetic material that
could be drought-adapted and useful for developing climate-resilient
varieties. Using experimental treatments, (i) ample and (ii)
restricted-water, response of 148 genotypes were studied comprising
wild, feral and cultivated C. canephora. Biomass
allocation, standing leaf area and leaf area growth data were collected.
Linear mixed effect models and PCA were used to analyse effect of
drought on genotypes from different: (i) cultivation status, (ii)
genetic groups and (iii) locations. We assessed the relationship between
drought tolerance for relative growth rate in leaf area
(RGRA), total number of leaves (TNL),
total leaf area (TL) and total leaf dry weight
(TLDW) of genotypes at final harvest. Restricted-water
reduced RGRA across genetic groups (3.2 – 32.5%) and
locations (7.1 – 36.7%) but not cultivation status. For
TNL, TL and TLDW,
genotypes that performed well in ample-water performed worse under
restricted-water, indicating growth-tolerance trade-off. Drought
tolerance in RGRA and TNL were
negatively correlated with wetness index suggesting some degree of
adaptation to local climate. Findings indicate a growth-tolerance
trade-off within this tropical tree species and drought tolerance of
Uganda’s C. canephora is somewhat associated with local climate.
Keywords: Intraspecific variation, drought stress, growth
tolerance trade-off, local adaptation Coffea canephora.
INTRODUCTION
Water availability is a major factor limiting global coffee production
largely because of the drought sensitivity of Coffea species and
because a large fraction of the production is sustained by small-holder
farmers who usually lack resources to establish irrigation facilities
(DaMatta and Cochicho Ramalho, 2006; Wintgens, 2009; Craparo et
al. , 2015). Problems of water limitation in coffee production are
expected to be aggravated by climate change. This is because across the
coffee production belt, a temperature increase of 2.1° C has been
predicted by 2050 (Parry et al. , 2007; IPCC, 2014) and this
warming can directly result in increased vapour pressure deficits,
higher potential evapotranspiration and hence drought stress in plants.
Indirectly, the increase in global average temperatures is expected to
result in shifts in the annual precipitation with more frequent
occurrences of severe droughts (Schiermeier, 2008). The changes in
temperature and precipitation together may have strong negative effects
on coffee production (Bunn et al. , 2015), although Verhage et al.
(Verhage, Anten and Sentelhas, 2017) reported that the
CO2 fertilization effect arising from elevated
CO2 concentrations could offset the negative effects of
climate change on average coffee yields by a small net increase. The
global distribution and production of coffee is therefore likely to be
significantly affected by climate change (DaMatta and Cochicho Ramalho,
2006; Davis et al. , 2012; Jassogne et al. , 2013). There is
a need for finding or developing drought-tolerant genotypes, and one way
of working towards this is to explore the natural diversity in wild
coffee populations.
C. canephora Pierre ex A. Froehner is a tree native to African
tropical lowland forests stretching from Guinea in West Africa through
the Congo River Basin to Uganda in East Africa (Berthaud, 1986; Coste,
1992; Montagnon, Leroy and Yapo, 1992; Davis et al. , 2006).
Generally, these tropical forests are characterized by abundant rainfall
(precipitation > 2000 mm y-1) with a
short or no dry season, high atmospheric humidity and stable average
temperatures between 24 °C and 26 °C (Coste, 1992; DaMatta and Cochicho
Ramalho, 2006; Damatta et al. , 2018). However, even in these
moist tropical forests, there occur periodic water shortages due to dry
spells (Engelbrecht et al. , 2006). Furthermore, the natural
geographical distribution of C. canephora extends into the
somewhat drier areas (Masih et al. , 2014), e.g. in Uganda.
Tree growth (e.g. biomass or leaf area increment, referred to as
performance hereafter) is commonly observed to decrease with drought
intensity (Grime and Hunt, 1975; Chapin, 1980; Garnier and Poorter,
2007). Across tree species (at interspecific level), there tends to be a
negative correlation between growth under well-watered conditions and
drought tolerance which is defined as the extent to which plants can
maintain these growth rates under water-stressed conditions
(i.e ., drought tolerance in growth, the ratio of growth under
stressed and unstressed conditions) (Chapin, 1980; Garnier and Poorter,
2007; Ouédraogo et al. , 2013). Growth and survival under dry
conditions tend to be associated with traits such as low specific leaf
area (leaf area/mass ratio), fewer or smaller stomates, small stem
vessel diameter, high fractions of dry mass in roots, low leaf area to
root mass ratio and low leaf area to sapwood ratio which tend to reduce
growth rates under well-watered conditions (Lambers, Chapin and Pons,
2008).
While multispecies comparisons are useful to understand ecological
strategies and community composition, questions regarding natural
selection and applications for breeding require additional intraspecific
comparisons across wild accessions of a species. When an environmental
stress gradient such as water availability acts as a selective force,
one may expect tolerance of a genotype to this stress factor to be
related to the climate in the site of origin (Alberto et al. ,
2013). Analysing such patterns is important as it may provide insights
into natural selection but may also provide basic information to assess
the adaptive potential to climate change and, for crops, identify
drought-tolerant genotypes (Alberto et al. , 2013; Rungwattanaet al. , 2018). However, very few studies have compared wild
accessions from different climates for tropical trees such as coffee.
Rungwattana et al. (2018) compared wild accessions of rubber
(Hevea brasiliensis ) from different locations across a rainfall
gradient in the Amazon forest and found no correlation between any of
the traits investigated and either temperature or rainfall at the site
of origin. In C. canephora’ s congener, C. arabica ,
comparisons between nine accessions from different Ethiopian forests
showed that accessions from drier areas were more plastic in leaf gas
exchange traits in response to changes in water availability than those
from wetter areas (Beining, 2007) but another study with a similar set
of accessions found no correlations between water availability as an
experimental factor and leaf gas exchange traits (Kufa and Burkhardt,
2011).
Uganda has been reported to have substantial C . canephoradiversity (Musoli et al. , 2009; Ngugi and Aluka, 2019; Kiwuka,
2020; Kiwuka et al. , 2021) which could be explored to identify
functional diversity in regards to drought stress. But to our knowledge,
intraspecific comparisons of drought-related traits in C .canephora have been limited to cultivated material e.g. in
(DaMatta et al. , 2003; Pinheiro and Var, 2004; Dias et
al. , 2007; Silva et al. , 2013; King’oro, 2014; Menezes-Silvaet al. , 2015). While the aforementioned studies give important
insights into the morphological and physiological drivers of drought
tolerance, exploration of the variation in drought tolerance across wild
populations and potential correlations with climate need to be done.
Furthermore, none of the studies on tropical trees has explored the
extent to which drought tolerance is associated with genetic diversity,
a link that would provide helpful information to interpret drought
adaptation. Finally, as far as we know, drought tolerance in coffee has
also not been explored along a cultivation status trajectory, i.e.
comparing wild, feral (second generation or higher of formerly
cultivated material and abandoned for over 50 years) and cultivated
genotypes. It is therefore unknown whether the cultivation of C.
canephora has been selected for or against drought tolerance.
This study was set out to determine: (i) the effect of drought on
vegetative growth (biomass and leaf area increment) of C.
canephora genotypes, collected across a climatic gradient in Uganda and
categorised by (a) cultivation status, (b) genetic groups as
characterised by Kiwuka et al., (2021), (c) and location, indicating the
different climatic envelopes (for the years 1950 -2000), (ii) the
relationship between performance under restricted and ample-water
conditions, (iii) the relationship between drought tolerance of
genotypes and wetness index (WI) at their native location. WI, the ratio
of mean annual precipitation to mean annual potential evapotranspiration
(PET) is a reasonable proxy for local climate wetness, whereby high WI
indicates wetter climates and vice versa (note that we do not use the
original but confusing term, aridity index, from Zomer et al., (2008)).
We hypothesized that, since Uganda’s wild C. canephorapopulations occur in different climatic envelopes, genotypes from dry
(lower WI) locations characterised by high temperatures, low
precipitation, and high PET and will have comparatively higher growth
and performance under restricted-water conditions than genotypes from
locations with low to moderate temperatures, high precipitation, higher
WI and low PET (wet location). Additionally, we expect a trade-off
between drought tolerance and performance, whereby the mechanisms that
underlie drought tolerance in material from dry locations are associated
with slow growth and the inability to exploit favourable conditions
(McGill et al. , 2006; Lambers, Chapin and Pons, 2008; Sade,
Gebremedhin and Moshelion, 2012; Amissah et al. , 2018).
MATERIALS AND METHODS