Introduction
Coffee is the second most traded commodity after oil and the main source of foreign currency earnings for over 80 tropical countries (Gole, Reference Gole2015). Coffee is the backbone of the economy of Ethiopia, accounting for 25–30% and 10% of the country’s total foreign currency earnings and government revenues, respectively (Worku, Reference Worku2023) and 50% of the smallholders’ yearly earnings of the total agricultural income (Minten et al., Reference Minten, Dereje, Engida and Kuma2019). It also provides employment opportunity and livelihood for about 25% of the country’s population (Worku, Reference Worku2023). More than 90% of the country’s total coffee production is cultivated by around 15 million smallholder farmers who possess less than 2 ha of land and produce coffee on less than 1 ha using the traditional cultivation methods (Minten et al., Reference Minten, Dereje, Engida and Kuma2019).
Ethiopia is the world’s fifth largest coffee producer and third largest arabica coffee producer (ICO, 2023). It is considered the centre of origin and genetic diversity of arabica coffee (Meyer, Reference Meyer1965) and possesses a suitable growing environment for coffee. However, the productivity of Ethiopian coffee (0.64-0.72 t/ha) (Ayele et al., Reference Ayele, Worku and Bekele2021) is much lower than its potential (1.39–2.14 t/ha) (JARC, 2017) and that of the other world top producers, such as Brazil (1.65 t/ha) and Colombia (1.04 t/ha). It also shows a cyclical decreasing trend over years (Ayele et al., Reference Ayele, Worku and Bekele2021). This is because of old coffee trees and poor agronomic management practices coupled with soil nutrient depletion and climate change (Worku, Reference Worku2023). It is reported that more than 78% of the Ethiopian coffee is old and unproductive and needs renovation and rehabilitation (USAID, 2017) and the soils of most coffee-growing regions in Ethiopia become more acidic, weathered and leached, and nitrogen and phosphorus are the most yield limiting nutrients (Endris et al., Reference Endris, Kebede and Yakob2008). Studies also show a gradual temperature increase over years with more erratic rainfall and drought incidence in the coffee-growing regions of Ethiopia that have led to yield fluctuations (Moat et al., Reference Moat, Williams, Baena, Wilkinson, Demissew and Challa2017a, Reference Moat, Williams, Baena, Wilkinson, Gole and Challa2017b; Gomm et al., Reference Gomm, Ayalew, Hylander, Zignol, Börjeson and Tack2024). This shows the importance of seeking good and efficient agricultural practices for the improvement of coffee productivity in Ethiopia under challenging environments.
Pruning and fertilisation are essential cultivation techniques in coffee plantation business for improving crop productivity and restoring the performance of the old, unproductive coffee trees. Pruning is important for improving productivity, productive life span and regular cropping of coffee plantations (Galearachchi et al., Reference Galearachchi, Wolf and Jank2021). This is because pruning changes physiology (e.g., CO2 assimilation, growth hormone production and flowering), growth (e.g., canopy diameter and root length) and yield of coffee (DaMatta et al., Reference DaMatta, Ronchi, Maestri and Barros2007; Baitelle et al., Reference Baitelle, Filho, Freitas, Miranda, Vieira and Vieira2019; Muliasari et al., Reference Muliasari, Dewi, Rochmah, Malala and Adinuran2021; Rohani et al., Reference Rohani, Prayogo, Suprayogo and Wicaksono2024) and soil chemistry, litter input and microclimate of the farm (Kurniawan et al., Reference Kurniawan, Nugroho, Nuklis, Wibowo, Anggraini, Balangga, Ardianti, Indraningsih, Aisyawati and Nugroho2024). Studies show a greater increase in plant growth and crop yield (Filho et al., Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016; Colodetti et al., Reference Colodetti, Rodrigues, Cavatte, dos Reis, Filho, Brinate, Martins, Christo, Júnior and Tomaz2020) and a lower rate of stem ageing following a multiple-stem pruning and training practice of arabica coffee trees (Colodetti et al., Reference Colodetti, Rodrigues, Cavatte, dos Reis, Filho, Brinate, Martins, Christo, Júnior and Tomaz2020). However, pruning can attain a sustainable coffee yield provided that the other management practices are properly applied. This is because growth and yield responses of coffee to pruning vary with agronomic management practices (e.g., mulching, compost addition and plant density), cultivar and location (Mauri et al., Reference Mauri, Bittenbender, Fleming and Gautz2003; Wolde et al., Reference Wolde, Tefera, Yared, Gezahagn and Tadesse2017; Yilma et al., Reference Yilma, Aman and Mekonnen2020). Recent studies reported a three-way interaction effect between pruning, shade and organic fertiliser on yield and yield-related traits (Adriani et al., Reference Adriani, Masrul, Chairani and Abubakar2020; Karim et al., Reference Karim, Hifnalisa and Manfarizah2021) and a two-way interaction effect between pruning and fertiliser on coffee yield and soil properties (Sudharta et al., Reference Sudharta, Hakim, Fadhilah, Fadzil, Prayogo, Kusuma and Suprayogo2022). This depicts the interplay impacts between pruning and cultivar, growing environments (e.g., location and shade) and agronomic management practices (e.g., fertilisation and plant density) on plant growth and crop yield of coffee plus soil properties. An interplay impact between pruning and fertiliser rate on yield and yield-related traits can also be existed as observed in Souza et al. (Reference Souza, de Carvalho, Teramoto, Silva and Guimarães2024) who studied physiology, anatomy, morphology and yield of fertigated coffee plants subjected to intensive pruning and different macronutrient rates.
Despite some documents on interactive impacts of pruning and different agronomic practices and growing conditions on coffee growth and yield shown above, data on interactions between different pruning types and fertilisation rates are scare. This information is particularly limited for Ethiopian coffee. Considering the importance of pruning and fertilisation in growth and yield restoration of coffee plantations, particularly in a context of climate change, better knowledge on the interaction between pruning types and fertiliser rates is crucial to sustain high coffee yield and environmental safety. Thus, the objectives of the study were to examine the hypotheses that (1) the optimum fertiliser rate for plant growth and yield of coffee can vary with pruning type as it varies with location, growing conditions, and age and yield potential of coffee trees (EIAR, 2016) and (2) there will be a specific pruning–fertiliser rate combination that maximises growth and yield of arabica coffee and agronomic efficiency (AE) of fertiliser for a specific growing area of coffee and (3) thereby to find the best combination of pruning type and nitrogen, phosphorus and sulphur (NPS) mixed fertiliser rate for restoring yield and yield components of old and unproductive arabica coffee trees and improving AE of fertilisers in southwest Ethiopia.
Materials and methods
Description of the study area
The experiment was carried out at Green Coffee Agroindustry P.L.C. farm (7°19′ N, 36°1 E; 1774 m asl), which is located in Gimbo and Shisho-Inde Districts of the Kaffa Zone in southwest Ethiopia (Figure 1). The southwest Ethiopian highlands are the centre of origin and genetic diversity for arabica coffee and consist of montane rainforests, which are the genetic home of ‘wild’ arabica coffee populations. Some of these forests (Yayu, Kaffa, Sheka and Majang forests) are registered as UNESCO Man and Biosphere Reserves. The study area has a warm humid tropical rainy climate with mean annual rainfall ranging from 1710 mm at Bonga Meteorological Station (7°9′ N, 36°10′ E; 1725 m asl) to 1892 mm at Wushwush Meteorological Station (7°10′ N, 36°7′ E; 1950 m asl) and mean minimum and maximum temperatures of 11.3 and 27.4 °C, respectively (19.4 °C mean temperature). The study area experiences unimodal rainfall with a rainy season lasting from March/April to October. The coolest months of the area are in the middle of the main rainy season (July and August), while the hottest months are from February to May (Riechmann, Reference Riechmann2007). Nitisols is a dominant soil type in the study area. However, wetlands and poorly drained areas consist of Vertisols and Gleysols (Denboba, Reference Denboba2005). According to the soil analysis report of Horticulture-Coop Ethiopia Soil and Water Laboratory in 2016, the soil of the experimental site is characterised by a strong acidity (pH = 5.31–5.37), a high NH4 +-N content; a medium S content; the optimum contents of organic carbon, electric conductivity, K, Mn and Cu; and the low contents of P, Ca and B. The detail soil properties of the study area are presented in Table 1. For the strong acidic soils (pH = 5.0–5.5), availability of macro elements (N, P, K, S, Ca and Mg) to plants is very low, whereas that of micro elements (Fe, B, Mn, Cu and Zn) is very high (Jones, Reference Jones2001). However, coffee grows well on soils with pH between 4.37 and 6.78 (Riechmann, Reference Riechmann2007).
Figure 1. Geographical of the study area (Yakob et al., Reference Yakob, Gebremicheal, Aklilu and Melaku2015).
Table 1. Physicochemical soil properties of the study area (Kufa, Reference Kufa2011; Mogisao, Reference Mogisao2016)
N, nitrogen; C, carbon; P, phosphorus; K, potassium; Ca, calcium; Mg, magnesium.
Treatments and experimental design
The experiment consisted of three pruning types (stumping, heavy pruning and light pruning) and four rates of NPS mixed fertiliser (100, 140, 180 and 220 g per coffee tree per year, which is equivalent to 320, 448, 576 and 704 kg per ha per year, respectively with ca. 3200 trees per ha). Without fertiliser treatment was also included for the determination of AE of fertiliser rates. The fertiliser rates were formulated based on the full fertiliser rate recommended for high-yielding, old coffee trees or plantations growing without shade or with minimum shade on depleted soils in the study area, i.e., 63 g urea + 166 g NPSB-blended fertiliser per tree per year (EIAR, 2016). The mixed fertiliser used for the study was formed from 28.6% urea (46% N) (286 g/kg) and 71.4% (714 g/kg) commercially blended NPS fertiliser and consisted of 26.7% (267 g/kg) N, 27.1% (271 g/kg) P2O5 and 5.0% (50 g/kg) SO4. The commercially blended NPS fertiliser consisted of 19% (190 g/kg) N, 38% (380 g/kg) P2O5 and 7% (70 g/kg) SO4. The experiment was carried out in a split-plot design in four replicates for three consecutive years (2019–2021), where the pruning types and fertiliser rates were randomly assigned to the main plots and subplots, respectively. Each subplot consisted of 16 coffee trees (four rows each with four trees) with 1.7 and 1.8 m intra- and inter-row spacings, respectively and a 2-m pathway between blocks (replications). The middle two rows in each subplot were being considered the effective plot. The experiment was set up on a 16-years old coffee plantation (planted coffee with an average plant density of 3200 trees per ha), which was planted in 1999 under shade trees with a shade level of 30–40%. The variety of the plantation (i.e., 74112) has a compact canopy or growth habit and it is resistant to coffee berry disease (Colletotrichum kahawae). The dominant shade tree species in the plantation include Acacia albida, Albizia gummifera, Cordia africana, Spium ellipticum, Olea welwitschii and Schefflera abyssinica.
Stumping, which is the activity of cutting the main vertical stems at 30–45 cm from the soil to rejuvenate the entire older coffee trees, was carried out in December 2015 and each coffee tree was trained to have two or three vigorous and healthy verticals by removing the weak suckers. Heavy and light pruning were carried out in December 2016 and 2017, respectively. December is the end of harvesting period in the study area. Heavy pruning, also known as rejuvenation pruning and usually done every few years, involves cutting back the coffee trees considerably to encourage new growth and renew older coffee trees that have become unproductive. It encompasses all the rehabilitation prunings, such as cutting of old verticals and exhausted trees at 100–120 cm above ground leaving one or two verticals per tree, toping (cutting) of the upper part of the tree, cutting of the exhausted branches at 30–40 cm away from the trunk, removing of the umbrella shaped parts of trees, removing of some primary branches growing between two or more verticals, and cyclical removal of verticals. Light pruning, also known as selective pruning, is often done annually to maintain the shape and health of the coffee trees and to promote the growth of new and productive branches. It involves the removal of exhausted (dead, diseased, or unproductive) verticals and branches, toping of trees, cutting of verticals with died central branches at waist height, removing of some primary branches growing between two verticals and cyclical removal of exhausted verticals. Each pruning type was followed by de-suckering and maintenance pruning throughout the experimental period. Maintenance pruning, which includes removing of dead, broken, or diseased branches, interlocked branches, primary branches that are dropped or close to the ground, alternative primary branches on central verticals, and secondary branches growing up or down, or close to each other within a 15-cm length in the centre of the tree as well as cutting back the primary branches at 1 m, was carried out after coffee harvest (in December) in every year. The fertiliser treatment was started to be applied when the coffee plots treated by different pruning types reached at their full fruit-bearing stage and the fertiliser was applied following the local fertiliser application practice in two splits, one in March and the other in July, for three consecutive years (2019, 2020 and 2021). The other common agronomic management practices of coffee, such as shade tree, weed, pest and disease management practices were carried out throughout the experimental period as per the agronomic activity calendar of the farm.
Measurements of the response variables
Number of verticals per tree, number of primary branches per tree, number of fruiting nodes per tree, canopy diameter, yield per ha and agronomic efficiency of fertiliser (AE) were considered as the response variables for the applied treatments. The number of verticals, primary branches and fruiting nodes per tree are considered as yield components per tree, and canopy diameter is considered as the overall growth parameter of the tree. AE is a measure of the impact of fertiliser on crop yield and it is the amount of additional yield harvested per kg of fertiliser applied.
The data of yield components were recorded during and after flowering periods (February and March, and May and June, respectively) from five randomly selected coffee trees in the central rows of each subplot, while canopy diameter was measured in the summer season (July–September) from the same tree samples. Yield was measured from the entire plot (16 coffee trees) in the harvesting period (September–November). In the study area, February–March and May–June, and July–September, respectively are the periods of the growing year where yield estimation is carried out and coffee plants can attend a larger canopy size. Number of primary branches, number of fruiting nodes, canopy diameter, yield and AE were measured for the three study years (2019, 2020 and 2021). However, due to its negligible change in 2019, the number of verticals was measured for two years (2020 and 2021). The number of verticals per tree was determined by counting new verticals (suckers) sprouted after fertiliser application at a plant height of 0.5–1.5 m and fostered for yield. The number of primary branches per tree was determined by counting the primary branches on the bigger vertical of each tree sample. After determining the number of primary branches, the number of fruiting nodes per tree was determined by counting fruiting nodes on the middle primary branches or on those primary branches found in between the upper and the bottom most primary branches of the tree samples. The canopy diameter of coffee trees was measured at 0.6-1.0 m above ground (immediately below the middle part of the tree) and from north to south and east to west, and then, the average of the two measurements was calculated. The data of each of these response variables were divided by the number of tree samples (n = 5) to obtain per tree data. The yield data were taken by harvesting red cherries from the entire plot (16 coffee trees) and converting to clean coffee (green beans at 11.5% moisture content) by using a red cherry to clean coffee conversion ratio (6:1) (Worku et al., Reference Worku, Astatkie and Boeckx2022). The yield of green beans per ha was calculated by multiplying the yield data per tree by the average number of trees per ha (i.e., 3200 trees). The yield data per tree was calculated by dividing the yield data per plot to the number of trees per plot (i.e., 16). AE was calculated as a ratio of the difference between with and without fertiliser in green bean yield to the amount of fertiliser applied:
$$AE = \left( {Y -{Y_0}} \right)/F$$
where Y is the green bean yield per ha with fertiliser, Y 0 is the green bean yield per ha without fertiliser and F is the amount of fertiliser applied per ha per year. AE is calculated for each rate of fertiliser treatment.
Data analysis
The analysis of variance (ANOVA) was done based on a split-plot design nested in year using the GLM procedure of SAS ver.9.4. The effect of the main plot factor (pruning type) on number of verticals, number of primary branches, number of fruiting nodes and canopy diameter was tested by using the error term of the main plot × replication (block), while that of the subplot factor (fertiliser rate) and the interaction of the main plot and subplot factors were tested by the pooled error. The effect of the main plot factor and its interaction with year on yield were tested by the error term of the main plot × replication (block × year), while that of the subplot factor and its interaction with the main plot and year were tested by the pooled error. Before deciding the analysis of the yield data over years together (a combined analysis of variance) or separately (a separate analysis of variance), a homogeneity test of the residual variances across years was carried out using a mean square error ratio and a chi-square test. As the residual variance was similar across years, the combined analysis of variance was used (Moore and Dixon, Reference Moore and Dixon2015). The validity of model assumptions (normal distribution and constant variance assumptions on the error terms) was also verified for each response variable by examining the residuals as described in Montgomery (Reference Montgomery2020). When these assumptions are violated, an appropriate transformation was applied to the response variables; however, the means reported are back-transformed to the original scale. When the main effect is significant (P < 0.5), the means of the significant factor were compared using least significant difference (LSD) at the 5% level of significance and when the interaction effect is significant (P < 0.5), multiple means comparison was conducted to compare the combination of the factors using least squares means (lsmeans) statement of SAS at the 5% level of significance. A multiple samples t-test was used to test the difference in yield between years. To determine the significance of correlation between each yield component and yield, Pearson’s correlation coefficient was calculated and tested whether P = 0 or not using the Corr procedure of SAS.
Results
Analysis of variance
The main effect of pruning type significantly (P < 0.05) affected the number of primary branches and fruiting nodes per tree, while that of fertiliser rate significantly (P < 0.05) influenced the number of verticals per tree, the number of fruiting nodes per tree, canopy diameter, yield per ha and AE. The interaction effect of pruning type and fertiliser rate was significant (P < 0.05) on the number of fruiting nodes per tree, canopy diameter, yield per ha and AE (Tables 2 and 3). The year effect and the three-way interaction effect of pruning, fertiliser and year were also significant (P < 0.05) on coffee yield (Table 3).
Table 2. ANOVA P values that show the main and interaction effects of pruning type and fertiliser rate on vegetative growth response variables of coffee and agronomic efficiency of fertiliser
Table 3. ANOVA P values that show the main and interaction effects of growing year, pruning type and fertiliser rate on coffee yield
Main effects
As shown in Table 4, the numbers of primary branches per tree and fruiting nodes per tree were significantly higher for stumping and heavy pruning than for light pruning. However, stumping and heavy pruning did not significantly vary for both variables although their mean values for heavy pruning were greater than for stumping. The pruning types also showed a similar order in their mean yield value: heavy pruning (1.17 t/ha) > stumping (1.07 t/ha) > light pruning (1.04 t/ha). The number of verticals per tree was much higher for 100 and 220 g NPS per tree per year than the remaining fertiliser rates (Table 5). However, the number of fruiting nodes per tree significantly reduced as the fertiliser rate increased from 100 g NPS per tree per year to 180 g NPS per tree per year while there was not much variation among these fertiliser rates in canopy diameter. The canopy diameter was considerably higher for 220 g NPS per tree per year than for the other fertiliser rates. Yield per ha and AE were significantly higher for 100 g NPS per tree per year than for the other rates, other than 220 g NPS per tree per year in case of yield. Yield of the 220 g NPS per tree per year and AE of the 140 g NPS per tree per year was much higher than that of 180 g NPS per tree per year, and 180 and 220 g NPS per tree per year, respectively. In general, similar to fruiting nodes, both yield and AE reduced as the NPS fertiliser rate increased from 100 to 180 g per tree per year (Table 5). Coffee yield in 2019 was significantly higher than the yield in 2020 and 2021. However, it did not vary significantly between the production years of 2020 and 2021 (Table 6).
Table 4. Mean values of number of primary branches per tree and number of fruiting nodes per tree of coffee under different pruning types
LSD, least significant difference.
Means followed by the same letter in the same column are not significantly different at P < 0.05.
Table 5. Mean values of number of verticals per tree, number of fruiting nodes per tree, canopy diameter and yield per ha of coffee and agronomic efficiency of fertiliser under different fertiliser application rates
LSD, least significant difference.
Means followed by the same letter in the same column are not significantly different at P < 0.05.
Table 6. Mean values of coffee yield in three consecutive growing years
LSD, least significant difference.
Means followed by the same letter in the same column are not significantly different at P < 0.05.
Interaction effects
Table 7 presents the mean values of number of fruiting nodes per tree, canopy diameter and yield of coffee and AE of fertiliser under different pruning types and fertiliser rates, and Table 8 presents the mean values of coffee yield under different pruning types and fertiliser rates in each study year. Heavy pruning and 100 g NPS per tree per year produced a significantly higher number of fruiting nodes per tree (ca. 543), coffee yield (1.27 t/ha) and AE (2.24 kg/kg) than did most combinations of pruning types and fertiliser rates. Stumping and 220 g NPS per tree per year also produced a much higher number of fruiting nodes per tree (ca. 534), canopy diameter (3.36 m) and coffee yield (1.23 t/ha) than did the other combinations of pruning types and fertiliser rates. Stumping and 100 g NPS per tree per year also provided a higher AE (2.05 kg/kg) compared to the remaining pruning and fertiliser rate combinations. However, light pruning and 220 g NPS per tree per year produced the lowest number of fruiting nodes per tree (212.6) and stumping and 180 g NPS per tree per year produced the lowest coffee yield (0.88 t/ha) and AE (0.56 kg/kg) (Table 7).
Table 7. Mean values of number of fruiting nodes per tree, canopy diameter and yield of coffee and agronomic efficiency of fertiliser under different pruning types and fertiliser rates
NPS, nitrogen, phosphorus and sulphur.
Means followed by the same letter in the same column are not significantly different at P < 0.05.
Table 8. Mean values of coffee yield under different pruning types and fertiliser rates in three consecutive years
NPS, nitrogen, phosphorus and sulphur.
Means followed by the same letter in the same column are not significantly different at P < 0.05.
As indicated in Table 8, the combination of stumping and 220 g NPS per tree per year provided the highest coffee yield in 2019 (1.59 t/ha); however, it did not significantly differ from stumping and 100 g NPS per tree per year and heavy pruning and 180 g NPS per tree per year in the same year (1.42 and 1.32 t/ha, respectively), and heavy pruning and 140 g NPS per tree per year in 2020 (1.46 t/ha). On the other hand, the coffee yield obtained from the combination of stumping and 180 g NPS per tree per year in 2020 (0.78 t/ha) was much lower than those obtained from most of the combinations of pruning types and fertiliser rates across the study years (2019, 2020 and 2021).
Correlations between yield components and yield
The results of the correlation analysis between yield components and yield are presented in Table 9. The number of verticals per tree and the number of fruiting nodes per tree significantly (P < 0.05) correlated with yield. However, the number of primary branches per tree and canopy diameter did not correlate significantly with yield.
Table 9. Correlation coefficients and P values between yield components and yield of arabica coffee
Discussion
Treatment effects on yield components
Pruning, fertilisation and their interaction significantly affected yield-related traits of arabica coffee in its centre of origin in southwest Ethiopia. This indicates that these agricultural practices improve plant growth and crop productivity of arabica coffee. Specifically, pruning significantly affected number of primary branches and fruiting nodes per tree with the higher values being observed for stumping and heavy pruning than light pruning. This is in line with Yilma et al. (Reference Yilma, Aman and Mekonnen2020) who reported a considerable difference among three different pruning systems (single-stem capped, multiple-stem capped and multiple-stem uncapped) in new branch regrowth of arabica coffee in southwest Ethiopia and Rahman et al. (Reference Rahman, Malek, Hossain and Islam2024) who observed a significant difference between pruning treatments (pruning at 10, 20, 30, 40 and 50 cm from branch apex and without pruning) in branch length, number of nodes, number of nodes with fruits and internode length of robusta coffee in Bangladesh. A recent study carried out to determine the effectiveness of shoot pruning and 50% pruning on growth of arabica coffee seedlings also observed a significant difference between pruning types with the higher average plant height, number of leaves and number of branches being observed for shoot pruning than 50% pruning (Wisdawati et al., Reference Wisdawati, Yusuf, Tambaru and Pasareang2023).
The result of this study also shows that the removal of the entire plant canopy at 30-45 cm above ground (stumping) or all verticals at 100–120 cm above ground leaving one or two verticals per tree (heavy pruning) and training each tree to have 2 or 3 verticals is better than the selective removal of exhausted verticals and branches (light pruning) to rehabilitate productivity of old or exhausted arabica coffee plantations. This suggests that stumping and heavy pruning, especially when the former is combined with a higher fertiliser rate and the latter with a lower fertiliser rate, lead to a good vegetative recovery that in turn leads to a higher crop yield in arabica coffee plantation. As indicated in Filho et al. (Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016), these pruning types can also favour the source-to-sink relations in coffee plants, which is due to the favourable changes in the photosynthetic apparatus of coffee. This is because a higher number of stems per plant with better leafiness, photosynthetic capacity and fruiting nodes promotes a greater production of photo-assimilates (reserves) and flowers. Valadares et al. (Reference Valadares, Rosa, Martinez, Venegas and de Lima2013) reports that denser crops provide a better use of solar radiation and cycling of nutrients due to higher leaf surface and root density, which provide higher production rates with the use of a larger number of stems per area. In addition, pruning helps considerably in nutrient cycling, soil conservation, soil organic matter maintenance, among others, it provides the vegetative parts taken from plants grown in the culture medium (Filho et al., Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016).
Fertiliser application in this study also significantly affected the number of verticals and fruiting nodes per tree and canopy diameter with the higher number of verticals and fruiting nodes being observed for 100 kg NPS per tree per year and number of verticals and canopy diameter for 220 kg NPS per tree per year. The former rate is less than half of the rate recommended by Jimma Agricultural Research Centre (JARC) for high-yielding, old coffee trees growing with minimum shade or without shade on depleted soils (i.e., 229 g = 63 g urea + 166 g NPSB-blended fertiliser per tree per year) in the study area while the latter rate is equivalent to the recommended rate (EIAR, 2016). However, the number of fruiting nodes decreased when the fertiliser rate increased from 100 to 180 kg NPS per tree per year. The larger number of verticals and canopy diameter, but a lower number of fruiting nodes observed for 220 kg NPS per tree per year shows that arabica coffee grows magnificently with long internodes when we apply fertiliser at higher rates, particularly N fertiliser, which enhances the vegetative growth of plants. On the other hand, the larger number of verticals and fruiting nodes, but a lower canopy diameter observed for 100 kg NPS per tree per year shows that arabica coffee grows slowly with short internodes when we apply fertiliser at lower rates. This agrees with a previous study (Cai et al., Reference Cai, Cai, Yao and Qi2007) reporting a significantly higher relative growth rate of plant height and lateral branch length at two growth peaks for a high fertilisation group (60 g NPK compound fertiliser at 1:1:1 ratio) than for a low fertilisation group. This finding also demonstrates that fertiliser application at higher rate regardless of pruning enhances the growths of verticals and canopies of arabica coffee, but at lower rate enhances the growths of verticals and fruiting nodes of arabica coffee. As observed in the present study, both conditions can lead to a higher crop yield due to a positive correlation between bean yield and each of these plant growth variables. However, the higher fertiliser rates (e.g., 180 and 220 g NPS per tree per year) showed a much lower fertiliser AE than the lower rates (e.g., 100 and 140 g NPS per tree per year).
The two-way interaction of pruning type and fertiliser rate significantly affected the number of fruiting nodes per tree and canopy diameter. This agrees with Rohani et al. (Reference Rohani, Prayogo, Suprayogo and Wicaksono2024) who observed enhancement of various coffee parameters, such as stem diameter, shoot length and chlorophyll content as a result of tree management and chicken manure: NPK fertiliser application. Due to the fact that stumping and 220 kg NPS per tree per year have demonstrated a higher canopy diameter with a higher number of fruiting nodes per tree than others, it is suggested that stumping combined with a higher fertiliser application rate leads to the largest plant canopy with a greater number of leaves and reproductive structures in arabica coffee plantation. Coffee plantations in this condition could also produce a higher crop yield although the present study did not show a significant relationship between canopy diameter and bean yield. A greater number of reproductive structures of arabica coffee that may lead to a higher crop yield can also be produced by heavy pruning combined with a lower fertiliser application rate. This is due to the fact that heavy pruning coupled with a lower fertiliser application rate (100 g NPS per tree per year) produced a higher number of fruiting nodes per tree, which was also significantly correlated with bean yield. Cerda et al. (Reference Cerda, Avelino, Gary, Tixier, Lechevallier and Allinne2017) also identified fruiting nodes and dead productive branches as the most important and useful predictors of yield and yield loss. Yet, unlike Cerda et al. and our expectation, the correlations of canopy diameter and number of primary branches per tree with bean yield were weak. Furthermore, as recent studies (Sudharta et al., Reference Sudharta, Hakim, Fadhilah, Fadzil, Prayogo, Kusuma and Suprayogo2022; Rohani et al., Reference Rohani, Prayogo, Suprayogo and Wicaksono2024) reported on the effect of the combined treatments of coffee tree management (pruning and bending) and chicken manure and NPK fertiliser application on bulk density, organic carbon content, and total and available nitrogen contents of soils, the combined treatments of pruning and fertiliser tested in this study can also affect soil properties. In Rohani et al. (Reference Rohani, Prayogo, Suprayogo and Wicaksono2024), the management approach of Bending + Chicken manure: NPK fertiliser increased stem diameter, shoot length and chlorophyll content of coffee and organic carbon content of soil and reduced bulk density of soil. This indicates the importance of including physiological and growth responses of coffee and soil and microclimate responses of agro-ecosystem when we study a combined effect of plant canopy and nutrition management practices of coffee to realise its impact on sustainable coffee production and environmental health in a climate change scenario.
According to a review by Filho et al. (Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016), a greater investment in area and volume of the plant canopy favours the soil exploration capacity of the plant in search for water and nutrients, due to the joint development of the root system in relation to shoots, and plants with more stems and vigorous canopies due to the removal of weak branches present a better development of the root system, contributing to their growth and production. Moreover, the formation of microclimates inside the canopy of vigorous coffee plants is capable of mitigating climate variables, such as radiation and temperature, which can promote a lower environmental pressure over physiological processes. Thus, the improvement of physiological characteristics of arabica coffee conducted with plant canopy management through pruning will improve crop yield over harvests.
Treatment effect on yield and agronomic efficiency
The significant effects of fertilisation and its interaction with pruning were observed for bean yield and AE of fertiliser in the centre of origin and diversity of arabica coffee in southwest Ethiopia. The interaction among pruning type, fertiliser rate and harvest year was also observed for bean yield. Sudharta et al. (Reference Sudharta, Hakim, Fadhilah, Fadzil, Prayogo, Kusuma and Suprayogo2022) also reported the interaction effects between pruning of the coffee stems and adding of fertiliser (manure and NPK fertiliser) on bean yield and soil properties. Similarly, Adriani et al. (Reference Adriani, Masrul, Chairani and Abubakar2020) and Filho et al. (Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016), respectively, reported an interaction effect between pruning, shading and fertilising and between pruning and harvest year on coffee bean yield.
In the present study, stumping and heavy pruning combined with 220 and 100 kg NPS per tree per year produced much higher bean yields (1.23 and 1.27 t/ha, respectively) than most of the other treatments. Heavy pruning combined with 100 kg NPS per tree per year, which was statistically similar with stumping and 100 kg NPS per tree per year, also showed the highest AE of fertiliser. However, stumping and 220 g NPS per tree per year was one of the treatments that demonstrated the lowest AE of fertiliser. Heavy pruning can also be more costly than stumping due to its higher skilled manpower requirement and shorter pruning cycle that may not be compensated by its continual cropping, and less deaths and growths of coffee plants and weeds, respectively compared to stumping. However, this needs to be confirmed by a partial budget analysis of the different pruning types and fertiliser levels under study. In general, the results of this study show a harmony between bean yield of coffee and AE of fertiliser for stumping and heavy pruning, each combined with a lower fertiliser rate (100 kg NPS fertiliser per tree per year) and a controversy between bean yield of coffee and AE of fertiliser for each pruning type combined with the higher fertiliser rates (140, 180 and 220 kg NPS per tree per year). The 100 kg NPS fertiliser per tree per year is less than half of the recommended rate by JARC for coffee trees that are similar with those used in this study (i.e., 229 g = 63 g urea + 166 g NPSB-blended fertiliser per tree per year) (EIAR, 2016). This shows the importance of using a lower fertiliser application rate following coffee pruning to maximize fertiliser use efficiency and to reduce production cost and environmental pollution related to fertiliser application than the JARC’s recommended rates. It also suggests the need to revise the previous fertiliser dose recommendations by JARC for different coffee trees growing under different conditions in Ethiopia, for example, young, old or high-yielding coffee trees growing with or without shade under different plant density or soil fertility conditions (EIAR, 2016).
Regarding the three-way interaction between harvest year, pruning and fertilisation, higher bean yields were observed for stumping combined with 220 and 100 g NPS per tree per year in 2019 (1.59 and 1.42 t/ha, respectively), and for heavy pruning combined with 140 and 180 g NPS per tree per year in 2020 (1.46 t/ha) and 2019 (1.32 t/ha), respectively. However, light pruning combined with any NPS application rate did not significantly change the bean yield in any study year. Moreover, AE of fertiliser did not differ for light pruning combined with any NPS application rate excluding 100 g and it was much lower compared to that observed for stumping and heavy pruning, each combined with 100 g NPS per tree per year. This result depicts that light pruning and fertilisation of an exhausted or an old arabica coffee plantation at different rates do not bring changes in crop productivity, but does stumping or heavy pruning. This could be because a light pruning cannot renew the whole canopy or the verticals of coffee plants while does stumping or heavy pruning. Owing to this finding, to restore the productivity of an exhausted or an old coffee plantation, one can recommend to apply stumping or heavy pruning in combination with 100 and 220 g NPS application per tree per year, and 100 and 140 g NPS application per tree per year, respectively. However, considering both coffee productivity and AE of fertiliser is very important for the economic viability of the agronomic treatments to be applied. In the present study, the higher coffee bean yields with a higher AE of fertiliser were obtained from stumping and heavy pruning and 100 g NPS application per tree per year. But, it is worth noting that the productive life span of the stumped or heavily pruned coffee is much less than that of planted coffee. As per the study report from Central America, the re-sprouted suckers from the stumped coffee are only productive for little more than half of the productive lifespan of the planted coffee, a fact that it directly affects yield, labour cost and farm profitability (Cambou et al., Reference Cambou, Thaler, Clément-Vidal, Barthès, Charbonnier, Meersche, Vega, Avelino, Davrieux, Labouisse, Filho, Deleporte, Brunet, Lehner and Roupsard2021).
Studying the means of the bean yields along the harvest years show that a higher bean yield (1.22 t ha−1) was obtained in the first harvest (2019) than in the second and third harvests (2020 and 2021) (1.06 and 1.03 t/ha, respectively). However, Filho et al. (Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016) observed the higher crop yields in the first and second harvests than in the third, fourth and fifth harvests. The superior yield in the first harvest than in the other harvests might be due to the biennial bearing nature of arabica coffee in which a high bearing year will affect the bean yield in the following years. This is because in a year next to a high bearing year a non-structural carbohydrate reserve seldom mobilises for fruit production that would rely mainly on the recent assimilates (Bustan et al., Reference Bustan, Avni, Lavee, Zipori, Yeselson, Schaffer, Riov and Dag2011) and plant growth in coffee is low in the years with high fruit demands for assimilates (Charbonnier et al., Reference Charbonnier, Roupsard, Maire, Guillemot, Casanoves, Lacointe, Vaast, Allinne, Audebert, Cambou, Clément-Vidal, Defrenet, Duursma, Jarri, Jourdan, Khac, Leandro, Medlyn, Saint-André, Thaler, Meersche, Aguilar, Lehner and Dreyer2017). However, though the difference between the yields of the first harvest and those of the second and third harvests was statistically significant, the difference between the consecutive harvests in value term (0.16 and 0.03 t/ha, respectively) was not too large. This may exhibit the role of pruning and fertilisation for narrowing down the yield gaps between good and bad harvests due to the biennial bearing nature of arabica coffee.
The correlation between yield components and yield
A strong positive correlation between yield and number of verticals and fruiting nodes per tree shows a higher contribution of these yield components for crop yield of arabica coffee under the current study condition. This implies that crop yield of arabica coffee significantly increases as number of verticals and fruiting nodes per tree considerably increase, which in turn implies that agronomic practices that increase these yield-related traits (e.g., stumping and heavy pruning and fertilisation) also increase crop yield. The higher number of verticals will have a higher number of primary branches with a larger number of fruiting nodes that will produce flowers. This is because the higher number of stems per plant promotes greater production of photo-assimilates (reserves) and hence greater issuance of flowers (Filho et al., Reference Filho, Volpi, Ferrão, Ferrão, Mauri, da Fonseca, Tristão and Júnior2016). As per our observation in the study area of the present study, about 82–90% of the flowers in arabica coffee set fruits and as per Espindula et al. (Reference Espindula, Tavella, Schmidt, Rocha, Dias, Bravin and Partelli2021), the increase in the number of stems per plant promotes a quadratic crop yield response of individual plants of robusta coffee. Cerda et al. (Reference Cerda, Avelino, Gary, Tixier, Lechevallier and Allinne2017) also identified fruiting nodes and productive branches as the most important and useful predictors of yield. Thus, coffee yield increases as the number of verticals and fruiting nodes increase. Supporting this finding, Cerda et al. (Reference Cerda, Avelino, Gary, Tixier, Lechevallier and Allinne2017) realised that the number of fruits depends directly upon the number of vegetative nodes initiated during the previous year. However, the contribution of the number of fruiting nodes to crop yield depends on the number of fruits set per node, which ranges from 80-90% and varies between flowering times. This is because different factors cause a poor fruit set in coffee: (i) the development of atrophied flowers, attributed to prolonged drought or excessive rainfall during the critical stages of flower bud growth, and (ii) incomplete pollination or fertilisation of flowers owing to heavy rains, low temperatures or a shortage of pollinators during the blossoming times (Kumar, Reference Kumar1982).
As shown in Table 9, though they were not statistically significant, the associations between bean yield and number of primary branches and canopy diameter were positive. In coffee plants, primary (lateral) branches determine a canopy diameter of the plant and they are the main sources of fruiting nodes and photo-assimilates (leaves). This tells us that crop yield by default relates to the number of primary (lateral) branches and canopy diameter of coffee plants. In line with this, an earlier study on biometrical genetic in arabica coffee have shown that the selection efficiency for higher yield is substantially increased by considering various growth parameters and yield components, such as stem girth, canopy radius, percentage of bearing primaries, percentage of bearing nodes and number of berries per node (Walyaro, Reference Walyaro1983).
Conclusion
The current study elucidated the interactive effect of pruning and fertilisation on growth and yield of arabica coffee and its importance for the rehabilitation of old, unproductive coffee plantations. Particularly, stumping and heavy pruning combined with 100 g NPS application per tree per year can be used to restore vegetation and productivity of old or exhausted arabica coffee plantations and to maximise AE in the study area and other areas with similar growing conditions, followed by stumping and 220 g. The results of the study also suggest a revision of previous fertiliser doses recommended for different coffee trees growing under different conditions in the study area (i.e., 141–284 g NPSB per tree per year), which are much higher than the dose that provided a higher AE with a comparable yield for all coffee trees treated by the three pruning types in this study (i.e., 100 g NPS per tree per year). Also, it is vital for the future study on coffee canopy and nutrition management practices to consider partial budget analysis and physiological and growth aspects of coffee and soil and microclimate of the agro-ecosystem, as this study did not consider these issues.
Acknowledgements
The authors would like to acknowledge Green Coffee Agroindustry P.L.C. for its permission and financial support to conduct the research and the staff of Woshi farm for their technical support in field works of the experiment.
Author contributions
T.M. conceived and designed the study, conducted data gathering, performed statistical analyses and wrote the draft article, M.W. designed the study, visualise the results and wrote and revised the article.
Funding statement
Green Coffee Agroindustry P.L.C. supported the field work of the research.
Competing interests
The authors have no conflict of interest to declare.
Ethical standards
Not applicable.
Open access
