Leucaena interspecific hybrid ‘KX4-Hawaii’ as a source of agricultural biomass in a water-scarce small island developing state

Establishment of Leucaena ‘KX4-Hawaii’

Forty, 30-cm-long rooted air layers of Leucaena cultivar ‘KX4-Hawaii’ were imported into Barbados from the University of Hawaii in 2015. These were acclimated in the Biology Gardens at the Cave Hill Campus in plastic nursery pots (33 cm diameter, 25 cm high) filled with a 1:1 ratio of loamy topsoil and construction sand with daily irrigation. The surviving air layers (25) were planted at the Barbados Agriculture and Manufacturing Company’s (BAMC) research station at Groves, Saint George, Barbados to establish stock trees. These were spaced 2 m apart and 15 g of triple super phosphate fertilizer was placed in each planting hole.

After the stock trees had been established for 2 years, 240 air layers were created by removing 3 cm wide strips of bark from branches 4 to 6 cm in diameter, and cheese cloth containing 125 cm³ of moist Sphagnum peat moss was applied, before wrapping the wounded portion with plastic wrap. Up to three air layers were initiated per branch and the branches were cut after 5 weeks to produce 40 cm long air layers. These were acclimated as was done with the imported ‘KX4-Hawaii’ air layers and 183 air layers (76.25%) survived. Further air layers were created (cut after 8 weeks) and these were planted directly in the field without acclimatization in pots to test the feasibility of skipping the nursery stage of propagation. Of the 39 air layers directly planted, 33 survived (84.6%).

Experimental field plot and planting

A field experiment was established at the Barbados Ministry of Agriculture’s Graeme Hall complex, in the parish of Christ Church, (13.074379°N 59.573093°W), in 2018. This site is located 11 m above sea level in Barbados’ black soil association, characterized by fine textured clays (Ahmad, 2011). A randomized block design (with five blocks arranged linearly along a single hedgerow on one continuous ridge) was used to evaluate the plant spacing (50, 75 and 100 cm) between ‘KX4-Hawaii’ trees planted as a field border hedgerow. The ridge was 1.2 m wide. There were three plots per block (the position of each plot within each block was randomized), each plot was 9 m long, and plots were spaced 2 m apart for a total block length of 33 m. Each 50, 75, and 100 cm spacing plot contained 18, 12 and 9 plants, respectively, for a total of 39 plots per block. Due to seasonal variations such as the dry season in Barbados between December and May (Mohan, Clarke & Chadee, 2020), a 4-month pruning interval was chosen so that adequate yields could still be acquired during dry periods when Leucaena biomass production drops (Proverbs, 1984). After each pruning a 0.15% Roundup Ultra herbicide solution (a.i. glyphosate, Bayer AG, Germany) was spot sprayed to control weeds based on previously reported excellent weed control and minimum Leucaena mortality using this technique in Barbados (Proverbs, 1984).

Experimental blocks 1–4 were planted using the 183 acclimated air layers. As more land was available at the study site than initially discussed with the Agriculture Ministry, a fifth block (originally unplanned) was planted using the 39 air layers directly planted in the field to utilize the additional land. Supplying was done 2 weeks after planting. The trees were established for 7 (blocks 1–4) or 9 months (block 5), with daily drip irrigation for the first month. Then they were pruned to 30 cm to standardize the plants to the same height in September 2018 for blocks 1 to 4 and in January 2019 for block 5.

Biomass data collection

There were four prunings (spanning 16 months) until the experiment was prematurely concluded due to COVID-19 lockdowns in Barbados in 2020 (Table 1). Block 5 was established later than the other blocks and data were not collected from this block until May 2019 (13 months after planting). Apart from the first interval (ending January 2019) the amount of precipitation received was low.

The plant height and canopy width of two plants in the center of each plot were recorded and averaged, and biomass samples were taken from a single plant in the center of each plot (resources to process biomass samples from two plants per plot were not available). Data were collected from plants in the center of each plot to reduce any edge effects on the results. Plant height and canopy width (perpendicular across hedgerow) were measured with a meter ruler. Light interception was measured using a LI-190R point quantum sensor and one meter LI 191R line quantum sensor (LI COR, Inc., Lincoln, NE, USA) between 10:00 and 12:00 am. The line sensor was centered on the ground below the plant in each plot to be used for biomass sampling and measurements were only taken if the overhead light was at least 1,000 μmol s⁻¹ m⁻².

Plots were pruned to 30 cm high and biomass samples were taken approximately every 4 months. All pruned biomass from a single plant (due to limited resources sampling two plants per block would not be possible) from the center of each plot was used to determine the green biomass (leaves and green tender stems ≤ 1 cm in diameter) and woody biomass (more lignified stems and branches > 1 cm in diameter) yield per plant. The fresh weight of each sample was recorded, subsamples were weighed and oven-dried at 105 °C until constant weight, and the dry weight of the subsamples were used to determine the dry mass of the samples.

Data analysis

Block 2 was not analyzed due to very shallow soil depths in that area of the hedgerow that resulted in plots 4 and 5, and half of plot 6, performing extremely poorly. The ‘jamovi’ 2.5.3 software with the GAMLj: General analyses for linear models 3.2.7 module was used for mixed model, correlation, and regression analyses (Gallucci, 2019). A log10 transformation/back transformation was used during mixed model analysis if the residuals were not normally distributed (canopy light interception and the biomass yield variables), and polynomial contrasts were applied during mixed modelling for trend analysis. The Tukey honest significant difference post hoc test was applied to significant (p ≤ 0.05) mixed model results for pairwise mean comparisons. The strength of correlations was defined based on Swinscow (1997).

‘KX4-Hawaii’ growth

During 2019, seasonal variations were observed for some growth parameters Plant height, and canopy light interception were at their highest on the January 2019, and canopy width was highest in September 2019, though not significantly different to January 2019 (Table 2). Plant height, canopy width and canopy light interception were at their lowest in May 2019 during the dry season. Although differences between the mean plant height achieved were significant throughout the sampling period, there were no significant differences in canopy width and light interception between the January 2019 and September 2019 sampling dates. There was also no significant difference in canopy light interception between the May 2019 and January 2020 sampling dates, although significantly wider canopies were recorded in January 2020 than in May 2019.

The 75 cm spacing treatment consistently resulted in greater plant heights and canopy light interception (Table 3). The 75 cm spacing treatment was significantly different to the 100 cm spacing treatment in these variables, although it was not significantly different to the 50 cm spacing treatment. Similarly, wider canopy widths were observed with a 75 cm spacing interval. At the 50, 75, and 100 cm spacing intervals the mean canopy widths were 222 ± 10.1, 255 ± 10.1, and 220 ± 10.1 cm, respectively, but the linear mixed model results were not significant (F = 4.77, d.f. = 2, p = 0.060). There was a quadratic response of plant height (p = 0.013), canopy width (p = 0.023), and light interception (p < 0.001) as plant spacing increased. There was also a linear response of canopy light interception to plant spacing (p < 0.001).

‘KX4-Hawaii’ biomass yields

The yield per plant at each sampling date for each spacing treatment can be found in Table 4. These yields were equivalent to 27.94 Mg/ha/year green biomass and 47.19 Mg/ha/year woody biomass for the 50 cm spacing treatment, 23.38 Mg/ha/year green biomass and 44.19 Mg/ha/year woody biomass for the 75 cm spacing treatment, and 13. 88 Mg/ha/year green biomass and 24.63 Mg/ha/year woody biomass for the 100 cm spacing treatment.

Green biomass comprised a lower fraction of the total biomass produced by ‘KX4-Hawaii’ (Table 5) and its yields were significantly different to the woody biomass yields at each sampling date apart from May 2019. Yet, overall biomass yields recorded in May 2019 did not differ significantly to those recorded in January 2020. However, the woody biomass yields were significantly different between January and September 2019.

The 75 cm spacing treatment resulted in consistently higher yields per plant, although this was not significant. There were no significant effects of the spacing×date interaction on ‘KX4-Hawaii’ biomass yields and spacing only significantly affected the biomass yields per meter of hedgerow (F = 6.78, d.f. = 2, p = 0.016). The 50, 75 and 100 cm spacing treatments resulted in a mean total biomass production of 2.70 ± 0.43, 2.54 ± 0.40, and 1.43 ± 0.23 kg per meter of hedgerow respectively. Biomass yields per meter of hedgerow decreased with higher spacings and polynomial contrasts indicated a significant linear response of biomass yield per meter to spacing (p = 0.009). Notably, the 50 and 75 cm spacing treatments did not significantly differ in biomass yields per meter (p = 0.021), but both resulted in significantly different biomass yields per meter to the 100 cm treatment.

There was a quadratic response (contrast) of the green/woody biomass ratio to plant spacing (p = 0.048) but the mixed model results were not significant (F = 2.69, d.f. = 2, p = 0.085). However, there was a significant effect of date on the green/woody biomass ratio (F = 64.13, d.f. = 3, p < 0.001) and the ratio was significantly different between all sampling dates. The green/woody biomass ratio was 0.41 ± 0.03 in January 2019, increasing to 0.83 ± 0.03 in May 2019, and declining to 0.67 ± 0.03 in September 2019 and to 0.524 ± 0.03 in January 2020.

Associations between spacing, precipitation and growth parameters with ‘KX4-Hawaii’ growth and yield

The growth parameters measured were significantly correlated with biomass yield (Table 6). There were moderate positive correlations between plant height, canopy width, and light interception with green biomass yield per plant. Correlations between these variables and the woody biomass yield per plant were strong and positive, apart from that with canopy light interception which was moderate in strength (rho = 0.499, d.f. = 2, p < 0.001). All correlations between the recorded growth parameters and the green:woody biomass ratio were negative and the strength was weak for canopy light interception (rho = −0.346, d.f. = 42, p = 0.022), moderate for canopy width (rho = −0.455, d.f. = 43, p = 0.002), and strong for plant height (rho = −0.645, d.f. = 43, p < 0.001).

When the data for each spacing treatment were pooled, all recorded growth and biomass yield variables were significantly correlated with the cumulative precipitation received during each cutting interval (Table 7). These correlations were all positive, apart from the strong negative correlation with the green:woody biomass ratio (rho = −0.686, d.f. = 43, p < 0,001). Notably, the correlation with plant height (rho = 0.757, d.f. = 43, p < 0.001) was strong as compared to the moderate correlations with canopy width (rho = 0.581, d.f. = 42, p < 0.001) and light interception (rho = 0.434, d.f. = 42, p = 0.003). The correlation with green biomass yields was weak (rho = 0.375, d.f. 43, p = 0.011) compared to the strong correlation with woody biomass yields (rho = 0.641, d.f. = 43, p < 0.001). There were no significant correlations between any recorded variable and the length of the interval between data collection dates apart from the moderate negative correlation with the green:woody biomass ratio (rho = −0.516, N = 43, p < 0.001). When correlation analyses were conducted using data from each spacing treatment individually, the canopy light interception of the 75 cm spacing treatment was not significantly correlated with precipitation (rho = 0.415, d.f. = 13, p = 0.124), and the green biomass yield per plant of the 50, 75 and 100 cm treatments were not significantly correlated with precipitation (rho = 0.300, d.f. = 13, p = 0.277; rho = 0.411, d.f. = 13, p = 0.128; and rho = 0.419, d.f. = 13, p = 0.120 respectively) (Table 7).

Multiple linear regression modelling of spacing, spacing², and the duration of each cutting interval, and the amount of precipitation received was not significant when fitted to the green biomass yield per plant (adjusted R² = 0.100, d.f. = 4,40, p = 0.083) (Table 8). However, the regression model fitted to the woody biomass yield per plant was significant (adjusted R² = 0.100, d.f. = 4,40, p = 0.083) and spacing (p = 0.004), spacing² (p = 0.005), and precipitation (p < 0.001) were significant predictors of woody biomass yield. This model explained 47.9% of the variation in woody biomass yields. However, the duration of the interval between cuts (p = 0.508) was not a significant predictor of woody biomass yield, likely because the intervals only ranged from 113 to 123 days; a maximum difference of 10 days. There was a quadratic response of woody biomass yield per plant to plant spacing (spacing, B = 8.96e−2; spacing², B = −5.87e−4), indicating that yield increased initially with spacing before declining. Woody biomass yields also increased with increasing precipitation (B = 2.23e−3).

Propagation of Leucaena ‘KX4-Hawaii’

Air layering is a low-tech method of propagation with a higher than 90% success rate for ‘KX4-Hawaii’ propagation under ideal conditions (Idol, Youkhana & Santiago, 2019), so it was used to produce ‘KX4-Hawaii’ planting material in this study. However, the survival rates (76.25% and 84.6%) achieved were lower than the 90% success rate reported by Idol, Youkhana & Santiago (2019). It should be considered that the rooted air layers directly planted in the field, though irrigated initially, were propagated under rain-fed conditions and this may explain the lower success rates. Any differences between these findings and those of Idol, Youkhana & Santiago (2019) in Hawaii might be due to the use of larger diameter branches for air layering in this study. Idol, Youkhana & Santiago (2019) also mentioned that adequate soil moisture was needed for the best success rate in producing rooted air layers, and the stock trees used in this work were not irrigated. Furthermore, Hawaii is a volcanic island with volcanic soils, while Barbados’ soils are mainly of limestone origin (Ahmad, 2011). Generally, there is agronomic interest in modern seedless Leucaena cultivars, but breeders should consider more efficient multiplication methods for large scale distribution and planting.

Effects of plant spacing on ‘KX4-Hawaii’ growth and yield

Overall, the 75 cm spacing resulted in the greatest ‘KX4-Hawaii’ growth and biomass production on aper-plant basis. Research has shown that Leucaena does not compensate for wider plant spacings by producing more lateral stems (Guevarr, 1976). Therefore, when space is no longer a limiting factor, further stem growth that would normally drive a bigger canopy and greater biomass does not occur. Leucaena biomass yield per plant (effects of experimental treatments at the individual plant level) and per unit area or length of hedgerow (yields under a production setting) are both helpful metrics to measure plant productivity in response to experimental treatments. Our results showed that there were no significant effects of plant spacing on the biomass yield per plant, but there were significant effects of plant spacing on the biomass yield per meter of hedgerow. Previous research has shown that Leucaena biomass yield per plant did not significantly increase with intra-row plant spacings above 25 cm (Chotchutima et al., 2013; Tuncay & Rüstü, 1989), and 50 cm was the smallest plant spacing used in this current study. As plant spacing did not significantly affect biomass yield per plant, a higher biomass yield per length of hedgerow with smaller plant spacings was due to greater plant densities.

The green:woody biomass ratio responded quadratically to plant spacing and the 75 cm spacing treatment resulted in the lowest green:woody biomass ratio (it had the largest proportion of woody biomass). This was unexpected, as literature sources suggest narrower plant spacings are superior for woody biomass production (Chotchutima et al., 2013; Elfeel & Elmagboul, 2016; Prasad et al., 2011). Another factor in the lower biomass yield per plant of the 100 cm plant spacing treatment might have been because it took longer for this treatment to form a solid canopy and shade out weeds (if it ever did within 4 months). This allowed more weeds to grow inside the hedgerow, which would lead to interspecific competition that would counteract the lower intraspecific competition due to a lower plant density. Desmanthus virgatus (L.) Willd and Bothriochloa pertusa (L.) A.Camus were common weeds in the ‘KX4-Hawaii’ hedgerow in this study.

The biomass yield per meter of hedgerow responded linearly to plant spacing, with an inverse relationship between plant spacing and biomass yields. The 75 cm spacing treatment can be recommended for ‘KX4-Hawaii’ under similar conditions in the present study as it did not differ significantly from the yields per meter of hedgerow of the 50 cm spacing interval but would require less planting material due to the smaller population produced. Our findings are similar to those by Elfeel & Elmagboul (2016), who also used a 4-month pruning interval across four biomass sampling dates. Biomass yields recorded from their first biomass sampling responded quadratically to plant spacing and linearly to plant spacing on the subsequent three sampling dates, with the lowest plant spacing (40 cm) having higher biomass yields. On its initial release, the ‘KX4-Hawaii’ cultivar was noted for its wood and biofuel production (Brewbaker, 2013). However, earlier studies with Leucaena by Guevarr (1976) reported that green biomass comprised the predominant fraction when compared to woody biomass yields. This differed from later studies by Elfeel & Elmagboul (2016) who reported the production of equal proportions of Leucaena green and woody biomass. In the current study, woody biomass comprised the predominant fraction of the ‘KX4-Hawaii’ biomass produced. Therefore, the high woody biomass production here makes ‘KX4-Hawii’ a suitable cultivar under these conditions as a biofuel source.

Impacts of climate conditions during the study period

Leucaena can grow with 650 mm of annual precipitation (Orwa et al., 2009; Proverbs, 1984) but this amount is not optimal. Leucaena has been shown to respond to water availability (Maslekar, 1984; Noulèkoun et al., 2017) and thrives with over 1,200 mm of annual precipitation (Orwa et al., 2009). However, there was a drought during the study period in 2019 in Barbados (Alleyne, 2020; Smith, 2020), with the Charnocks weather station receiving only 58% (736.5 mm) of the climatological average precipitation. This was the lowest amount of precipitation received since 1942 (Alleyne, 2020). Barbados’ climatological average of 1,260 mm of annual precipitation (Layne, 2021) falls into a favorable precipitation regime for Leucaena growth. Therefore, there is potential for greater ‘KX4-Hawaii’ growth and biomass production than that seen in 2019 outside of drought conditions. Notably, the September 6th to December 31st period in 2018 (580.4 mm) was not dissimilar to the climatological average of 1991–2020 for the September to December period (605.8 mm) (Layne, 2021), and this period resulted in the highest woody and total biomass yields when the Leucaena trees were pruned in early January 2019. There were positive correlations between precipitation and all recorded ‘KX4-Hawaii’ growth and yield variables when the data for all treatments were pooled, indicating the potential for higher biomass production. However, ‘KX4-Hawaii’ was still productive under the drought conditions seen during the study and future dry conditions may not hamper the generation of useful woody biomass for agronomic purposes. Elfeel & Elmagboul (2016) also found that the rate of Leucaena biomass production was seasonal, similar to the correlation between precipitation and biomass yields in the current work. So, lower green biomass production under drought conditions was expected as reduced leaf production and increased leaf senescence in Leucaena are typical under these conditions (El-Juhany & Aref, 1999; Rosecrance, 1990). However, as woody biomass was more responsive to precipitation than green biomass, which was not significantly correlated with precipitation when the spacing treatments were analysed independently, woody biomass may make up an even larger fraction of the total biomass produced by ‘KX4-Hawaii’ under the climatological norm in Barbados as seen from the January 2019 results.

Limitations of study and future work

Several constraints were encountered during the study period that were outside the control of the authors. First, Barbados is a small island developing state with limited land availability for agroforestry, and there was limited planting material, so the study was restricted to intra-row plant spacing and the pruning regime was fixed. Second, the study had to be terminated prematurely due to national COVID-19 lockdowns. Hedgerow studies typically are run for at least 2 years and include the effect of pruning management such as that by Mullen et al. (2003). The current study lasted 16 months from the first cut to the last pruning. Nonetheless, the study provides valuable information on ‘KX4-Hawaii’, a new cultivar that has not been significantly studied, on a small island with limited land area and creates a base for future research. Future work will build on this research with experiments of longer durations investigating the effects of pruning management on the cultivar and the application of the biomass produced to the soil.