Ammar Khalil Mohamed Ahmed·Luping Jiang·Fang Wang·Song Chen·Xueyan Zhou·Xiaona Pei·Xiyang Zhao·Guanzheng Qu

Abstract To determine appropriate quantities of water and fertilizer required for early growth of hybrid poplar cuttings,we recorded the growth traits of four clones grown under four factors(irrigation and nitrogen,phosphorus,and potassium fertilizers), each with four levels, using an orthogonal experimental design.A logistic model was used to estimate growth in height.The growth curves for tree height were sigmoid,and the model R2 values were greater than 0.9,which indicated that the fit was highly significant.ANOVA results for tree height and basal diameter indicated that all sources of variance showed significant differences(p ＜0.001).The average tree height and basal diameter for all the four clones under the different treatments ranged from 155.39 to 235.04 cm,and from 13.71 to 17.42 mm, respectively. A highly positive correlation between the extreme k value and tree height was observed,suggesting that the k value was an accurate estimation of tree height.For model parameters,the earliest average time point for the onset of the rapid growth period of poplar clones was 131 d,and the highest average increment in tree height during the rapid growth period was 138.78 cm.The highest average tree height for all clones under each factor was 219,210.51,200,and 201 cm when treated with either 1200 mL of water applied every third day,3 g of nitrogen,0 g of phosphorus,or 0 g of potassium,respectively.The most suitable treatment for the early growth of hybrid poplar cuttings,as suggested by the developed logistic model,was 1200 mL of water applied every third day and three applications of 1 g nitrogen(in the form of CH4N2O).

Keywords Fertilizer management·Growth trait·Logistic model·Poplar·Rapid growth·Variation

Introduction

Poplars (Populus spp.) are an important short-rotation forest crop(Wang et al.2015),which are recommended for bioenergy plantations because of their high biomass production(Zabek and Prescott 2007)and good coppicing ability(Toillon et al.2013).Poplar plantations are subject to a variety of silvicultural treatments including irrigation,pruning, intercultivation and fertilization (Singh and Sharma 2007).Nutrient availability and soil water content are two of the main factors that limit tree growth,especially that of saplings(Dong et al.2011).Simultaneous water and fertilizer application increases the volume growth of poplars(Wang et al.2015).Water deficits affect plant growth through a number of physiological processes,including reduced water channel activity and tissue water potential(Dong et al.2011).Nitrogen(N),phosphorus(P)and potassium(K)are essential nutrients for plant development(Pitre et al.2007).Nitrogen deficiency affects plant morphology and phenology (Pitre et al. 2007),whereas phosphorus deficiency limits growth and influences the growth rate of poplar plantations(Elferjani et al.2013). In general, fertilizer application to poplar is a common practice,and considerable variation in responses to fertilizer has been observed among poplar species(Zabek and Prescott 2007).The determination of optimal irrigation and fertilizer applications is important for poplar growth at the seedling stage(Lasa et al.2016).

The ability to model plant growth under different environmental conditions is important for the development of management strategies that facilitate the optimal utilization of fertilizer,water,land and other resources(Hart et al. 2015). The modeling of plant growth allows hypotheses to be tested,virtual experiments concerning plant growth processes to be conducted,and the outcome of a given model to be directly examined(Fourcaud et al.2008).Currently,models of functional growth traits are particularly important to integrate biological processes and environmental conditions and to provide a basis for more advanced research in plant sciences(Fourcaud et al.2008).Numerous models,including polynomial or linear models,such as the Richards,Bertalanffy,exponential,Lundqvist-Korf,and logistic models,have been proposed to estimate plant growth traits(e.g.,diameter and height)(Martins et al.2014).The logistic model provides a simplified representation of the biophysical processes that control plant growth and can be readily applied over large regions and to multiple species with few necessary modifications of the model parameters(Avanza et al.2008).Owing to this ease of application,the logistic model may be the most suitable model to describe growth stages(Avanza et al.2008).In the present research,water and fertilizer were applied to four poplar clones.The objective of the study was to compare the growth of the clones and to determine the water quantity and fertilizer concentration most suitable for the optimal growth of the clones.The results obtained would contribute to the production of strong,fast-maturing trees in northeast China.

Materials and methods

Experiment materials

The plant material consisted of four poplar clones[Populus simonii×P.nigra‘Baicheng-1'(XH),Populus pseudosimonii×P. nigra ‘Baicheng-1' (XQH), Populus simonii×P. nigra ‘Bailin-3' (BL3), and Populus×Xiaozhanica‘Baicheng-5'(BC5)],which were selected as improved cultivars in 2002 in Jilin Province,northeast China.Two-hundred plants of each clone were propagated from cuttings in a greenhouse of Northeast Forestry University [45°78′N, 126°61′E; 127 m a.s.l.(Wang et al.2016a,b)]on 3 April 2016.Each cutting was planted in a 25.5-cm deep pot with top and base diameters of 22.6 and 17 cm,respectively.The growing medium was comprised of a mixture of soil,vermiculite,perlite,and silver sand(2:1:1:1,w/w,7.2 kg pot-1).When the plants were approximately 10 cm in height,112 plants of each clone with the same height were selected.An orthogonal experimental design was used,incorporating four factors and four levels of each factor(16 treatments in total;‘‘Table 1''),with seven replications of each treatment for each clone.The plants were treated with different volumes of water(300 mL,0.042 L kg-1;600 mL,0.083 L kg-1;900 mL, 0.125 L kg-1; and 1200 mL, 0.167 L kg-1)every third day and different concentrations of N fertilizer(CH4N2O;0,1,2,or 3 g),P fertilizer(Ca(H2PO4)2·H2O;0,1.43,2.86,or 4.29 g)or K fertilizer(K2SO4;0,0.133,0.266,or 0.399 g)were applied with water at three times(Fertilizer was applied on 4 June 2016 and every 15 days thereafter) during the experimental period. The water quantities and fertilizer concentrations were selected based on the recommendations of Zhao(2010).

Measurement of growth traits

Tree height(H)and basal diameter(BD)were measured 18 days after planting in the greenhouse(21 April 2016,the 110th day of the year)and every 7 days thereafter.The final BD measurements were recorded on 10 September 2016(the 252nd day of the year).Given that the trees were capping before 19 July(the 200th day in a year),tree height was investigated until that day.A ruler and Vernier calipers were used to measure H and BD,respectively.

Table 1 Water and fertilizers applied on poplar clones in the different treatments

Statistical analyses

All statistical analyses were conducted using SPSS(version 13.0;SPSS Inc.,Chicago,IL,USA)and OriginPro(version 8.5; OriginLab Corporation, Northampton, MA, USA)software packages.The significance of fixed effects was tested by analysis of variance(ANOVA)F tests.Variation among clones at different time points was determined by ANOVA following Hansen and Roulund(1997),using the formula:

where yijis the performance of an individual of clone i in treatment j,μ is the overall mean,αiis the clone effect,βjis the treatment effect,αβijis the random effect of clone i in treatment j,and εijis the random error.

A logistic function was used to fit the annual plant growth curve using the following logistic equation from Yang(2006):

where y is the cumulative growth in tree height;t is the growth period;a and b are parameters that were calculated with the least squares method(Zhou et al.2004);and k is the extreme value of annual growth in tree height calculated with Eq.3(Zhang and Gai 1994;Zhou et al.2004):

where y1,y2and y3are the corresponding tree heights at three time points(t1,t2and t3)with 2t2=t1+t3.

The second derivative was calculated for Eq.2,when d2y/dt2=0 and t0=ab-1,where t0(the rapid growth time point)is the date of maximal growth in tree height.

Calculation of the third derivative for Eq.2 yielded the formula:

Thus,t1and t2were calculated as:

where t1and t2were the start and end dates,respectively,of the rapid height growth stage.

Tree growth was divided into three phases by t1and t2.The first phase(0-t1)was the first slow growth phase,and during this period,the growth rate gradually increased.The time point at which the average annual growth rate of the tree was attained represented the beginning of the second phase(t1-t2),during which the growth rate was higher than the average annual growth rate of the tree.When the growth rate declined to less than the average annual growth rate,plant development entered the third phase(t2to capping),in which the growth rate gradually decreased until ultimately capping was attained(Liu et al.2014).

The period of rapid height increment and specific parameters are important for tree development when the growth pattern conforms to a sigmoid(‘S'-shaped)curve.Except for k,the time points t0,t1and t2,RR(the period of rapid growth increment), GR (total growth increment during the rapid growth period),GD(daily growth increment during the rapid growth period)and RRA(the ratio between GR and total height increment)were calculated according to the growth trait during the whole growth process(Liu et al.2014).

The phenotype correlation rA(xy)of traits x and y was calculated following Pliura et al.(2007):

where σ2a（xy）is the phenotypic covariance for traits x and y,σ2a（y）is the phenotypic variance for trait y,and σ2a（x）is the phenotypic variance for trait x.

Table 2 ANOVA analyses of tree height(H)and basal diameter(BD)under different treatments

Results

ANOVA

ANOVA results for the H and BD of the four poplar clones under the different treatments are presented in Table 2.All variance sources showed significant differences(p ＜0.001).

Average final H and BD

The average final H and BD of each clone under the different treatments are shown in Tables 3 and 4.Treatment 3(300 mL water,2 g N,2.86 g P,and 0.266 g K)showed the lowest average H(155.39 cm)of all the treatments,with an average H of 154.73, 146.90, 160.74, and 159.21 cm for the XH, XQH, BL, and BC5 clones,respectively.Treatment 14 showed the highest average H of the clones,which exceeded the lowest average H by 51.26%.Treatment 10 showed the highest average BD(17.42 mm), which exceeded the lowest average BD(13.71 mm in Treatment 1)by 27.06%.Treatments 15(17.31 mm)and 14(16.87 mm)also showed high BDs,which were higher than the lowest average BD by 26.25%and 23.04%,respectively.

Construction of growth model for average H of all clones under the different treatments

The average H growth curves for each clone in each treatment are shown in Fig.1.All curves were sigmoid in shape.The average H values of clones XH,XQH,BL3,and BC5 were higher in treatments 10,16,14,and 14,among the 16 treatments. All parameters used for modeling average H in the different treatments are listed in Table 5.The highest average H and k values(in Treatment 14)were 235.05 cm and 240.64 cm, respectively. These values exceeded the lowest average H and k values(in Treatment 3)by 51.25%and 52.17%,respectively.The maximum difference between average H and k values was 8.68 and the minimum difference was-1.01,which were observed in Treatments 10 and 1,respectively.The coefficient of determination(R2)values of H under the different treatments ranged from 0.985 to 0.996,which indicated that the matching effect was significant.

Relationship between measured and predicted tree height

The relationship between the measured values of H and the predicted(k)values is represented in Fig.2.The linear regression model for k and H was k=1.056H-6.289.The R2values were higher than 0.98,which indicates asignificant and strong relationship between the k and H values.

Table 3 Average tree height(H)for the four poplar clones under different water and fertilizer treatments

Table 4 Average basal diameter(BD)for the four poplar clones under different treatments

Clone growth parameters under the different treatments

Growth parameters of each clone under the different treatments are summarized in Table 6.The latest time points for t0,t1and t2were the 159th,140th,and 178th day(Treatment 10 in clone 1),respectively,which were later than the earliest time points for t0(145 days,Treatment 1 in clones 3 and 4),t1(130 days,Treatment 1 in clone 2)and t2(158 days,Treatment 1 in clone 3)by 14,10,and 20 days,respectively.The longest period for RR among the clones in the different treatments was 40 days(Treatment 12 in clone 4),which was longer than the shortest RR by 14 days(Treatment 1 in clone 3).The largest value of GR was 157.19 cm(Treatment 10 in clone 1),which was higher than the smallest GR(Treatment 3 in clone 2)by 70.33 cm.The highest GD value of 4.40 cm was observed under Treatment 11 in clone 4,which was higher than the lowest GD value by 1.88 cm(Treatment 1 in clone 2).The highest value of RRA was 61.12%(Treatment 12 in clone 4),which was higher than the lowest RRA value(56.45%,Treatment 1 in clone 2)by 4.67%.Considering all clones collectively in each treatment present in Table 7,Treatment 10 showed the latest average time point for t0,t1and t2(155,137,and 173 days,respectively).These time points were later than the average earliest time points for t0(147 days,Treatments 1 and 3),t1(131 days,Treatments 1,3,and 4)and t2(162 days,Treatment 1)by 8,6,and 11 days, respectively. The longest average RR period(37 days,Treatment 14)for all clones was greater than the shortest average RR(31 days,Treatment 1)by 6 days.The largest average GR(Treatment 14)was higher than the smallest(Treatment 3)by 47.64 cm.The highest average GD(3.85 cm,Treatment 15)was higher than the lowest average GD(2.84 cm,Treatment 4)by 1.1 cm.The highest average RRA was 60.17%(Treatment 8),which was higher than the lowest average RRA(57.39%,Treatment 1)by 2.78%.

Correlation between measured and predicted parameters

The correlation coefficients among tree height(H)and the model parameters ranged from -0.097 (GD×RR) to 0.998 (GR×H) as shown in Table 8. All correlation coefficients were significant and positive except for the correlations GD×RRA(0.234)and GD×RR(-0.097).Correlation coefficients between H and each growth parameter ranged from 0.448 (H×RR) to 0.998(H×GR), and all correlations were significant and positive.

Fig.1 The average height of the four poplar clones under 16 treatments(a,b,c,and d,average heights of the clones XH,XQH,BL3 and BC5,respectively,under different treatments)

Optimal water and fertilizer treatments for all the clones

The average heights for all the clones under each level of water or fertilizer treatment are shown in Fig.3.For the water treatment, the highest average H (219 cm) was observed under irrigation with 1200 mL water,which was greater than the lowest average H (164.5 cm, under 300 mL water)by 54.5 cm(Fig.3a).The highest average H under each fertilizer treatment was 210.51 cm(N:3 g),200 cm(P:0 g)and 201 cm(K:0 g),which were greater than the lowest average H(185.97 cm,N:0 g;194.66 cm,P:4.30 g;and 195.2 cm,K:1.2 g)by 24.54,5.64,and 6.2 cm,respectively.

Discussion

Model efficiency

ANOVA is an important analytical method for estimating the extent of variability and plays an important role in tree breeding(Zhao et al.2014).In the present study,highly significant differences among the poplar clones under the different water and fertilizer treatments(p ＜0.01)were observed,which indicated that the measured growth traits(H and BD)were highly responsive to the water and fertilizer treatments.The results agree with Van et al.(2008),who observed significant irrigation×fertilizer interaction effects on both H and BD of hybrid poplars,which indicated that evaluation and selection of optimal irrigation and fertilizer treatments are important.

Table 5 Growth model parameters for all clones under different treatments

Fig.2 Correlations between the measured value and predicted(k)values of linear regression model in tree heights of different clones under different treatment

Mathematical models have been applied to describe and evaluate plant growth stages under field conditions and silvicultural management for many years(Sharma et al.2017).Although many models have been proposed to simulate plant growth,a logistic model is regarded as the most suitable one to describe a slow-fast-slow growth pattern because it reflects the highest significance of the model parameters in relation to the environment,growthperiod,and prediction of height growth(Avanza et al.2008).The present study provided insight into the use of height to estimate plant growth stage.All models were fitted with height under different water and fertilizer treatments and high R2values(＞0.9)were indicative of strong correlations between all model parameters.The results agreed with Zhao and Zhang(2013),who applied a similar model to describe Populus tomentosa tree growth and reported that the simulation coefficients of the logistic model were higher than 0.9 for seedling growth traits.Thus,logistic regression models are suggested to be the most suitable models for estimation of growth processes.

Table 6 Different growth parameters for each clone under different treatments

Table 6 continued

Table 7 Average growth parameters for all the clones under different treatments

Model parameters for each poplar clone under the different treatments

Different parameters(k,t1,t0,t2,RR,GR,GD,and RRA)for growth traits were significant for the estimation of hybrid poplar tree growth under the treatments applied in the current investigation.The parameter k is essential in a growth model and is suggested to be a good predictor of growth traits relative to other parameters(Herault et al.2011).In the present study,k values were similar to measured tree heights,which are consistent with a previous study by Herault et al.(2011),who reported that k values were significantly affected by all treatments and extrinsic factors,which indicated that the model was an effective predictor of growth trajectories. We observed that approximately 20 days after the onset of rapid growth(t1),all clones in the different treatments exhibited a rapid height increment and after an average of 17 days,the growth rate stabilized in t2.This result agreed with the findings of Zhao et al.(2013),who observed that all studied poplar clones had S-shaped rapid height increment after 19 days from the onset of the rapid growth period.In the present study,the latest time point for growth rate stabilization in t2was 37 d after the start of the rapid growth period t1,and the earliest time point was 31 d after the onset of the t1phase.These results were earliest than the time points observed by Zhao et al.(2013)by 27 d(for the latest time point)and 7 d(for the earliest time point)for growth rate stabilization.This difference may result from the different plant materials,environments and treatments in the two studies.For all clones in the current study,the daily growth increment(GD)under the different treatments ranged between 2.84 and 3.85 cm.These values are higher than those reported by Zhao and Zhang(2013)and Liu et al.(2014),who observed the lowest and highest GD of 0.35 cm and 2.13 cm,respectively. We observed RRA values under the different treatments greater than 50%,which is in agreement with Liu et al.(2014).The present results indicated that the rapid growth period was extremely important for tree growth,and thus,irrigation orfertilizer could be applied during this period.All growth parameters for all clones under the different treatments were significant,which suggests that the findings may supplement theoretical considerations to determine the cultural conditions most suitable for poplar growth.

Table 8 Correlation coefficients between tree height(H)and different parameters in linear regression mode of tree height

Fig.3 The average height for all the poplar clones under different water and fertilizers treatments(a,b,c,d,average heights under different water levels,N concentrations,P concentrations and K concentrations,respectively)

Suitable irrigation and fertilizer applications for growth of poplar cuttings

A decrease in the quantity of fertilizer or irrigation applied results in a significant decline in tree growth,whereas the application of excessive amounts of water or fertilizer may lead to plant death(Isebrands et al.2007).Therefore,determining appropriate water or fertilizer application amount is important for tree growth and development(Isebrands et al.2007).In the present study,poplar growth was strongly impacted by differences in the water and fertilizer quantity.The application of 1200 mL water every third day was the most suitable level for the growth of the poplar clones.However,1200 mL was the highest level of irrigation applied,whereas Wang et al.(2016a,b)applied 1453 mL as the highest level of irrigation to investigate the effect of container size on water consumption and growth of Pinus seedlings.Thus,a higher irrigation level should also be tested in a future study.For fertilization,N showed a stronger effect on growth of poplar cuttings than P and K,among the tested concentrations. The findings are in agreement with Cooke et al.(2005),who reported that growth,growth rate and phenology of hybrid poplar are all affected by N availability more than by other fertilizers.Similarly,Guillemette and Des(2008)observed that N was the most limiting nutrient in a hybrid poplar plantation.In the present orthogonal experiment,treatment 14 was the most suitable for clones BL and BC5,while treatment 10 and 16 were the most suitable treatments for clones XH and XQH,respectively,for seedling height growth.This finding may result from the interaction between genetics and environment affecting seedling height.Overall,the most suitable treatment combination was Treatment 14(water 1200 mL,N=3 g,P=8.6 g,and K=0 g),under which the average height attained by the four clones was 235.04 cm.However,for different growth parameters,the optimal level for each treatment was water 1200 mL,N=3 g,P=0 g,and K=0 g.This combination of irrigation and fertilizer treatments was not included in the present orthogonal experiment and,thus,requires verification in a future study.

Conclusion

Water and fertilizer are important factors for plant growth and development.Therefore,the determination of appropriate irrigation and fertilizer treatments for optimal growth of poplar cuttings is important.Tree height is a vital,variable trait for the estimation of tree growth at the early stages of establishment.The present study provides insight into the effects of water and fertilizer treatments on tree height to aid in the selection of a suitable method to optimize poplar growth and development. Significant variation and positive correlations were observed among different parameters for the early growth of poplar cuttings,and R2values for simulated growth curves exceeded 0.9,which indicated a significant association between all model parameters.