Population dynamics in a 6-year old coppice culture of poplar. I. Clonal differences in stool mortality, shoot dynamics and shoot diameter distribution in relation to biomass production

Biomass and Bioenergy 24 (2003) 81 – 95

Population dynamics in a 6-year old coppice culture of poplar. I. Clonal di$erences in stool mortality, shoot dynamics and shoot diameter distribution in relation to biomass production I. Laureysens∗ , W. Deraedt, T. Indeherberge, R. Ceulemans Department of Biology, University of Antwerp (UIA), Universiteitsplein 1, B-2610 Wilrijk, Belgium Received 29 April 2002; received in revised form 1 August 2002; accepted 9 August 2002

Abstract Poplar trees have the capacity to regrow a number of shoots after being coppiced. In April 1996, a high density 1eld trial with 17 di$erent poplar (Populus) clones was established in Boom (Belgium) on a former waste disposal site. At the end of the establishment year (December 1996), all plants were cut back to a height of 5 cm to create a coppice culture. Four years after the 1rst coppicing in January 2001, the stand was cut back again. During 6 years (1996 –2001), shoot diameters and number of stools and shoots were assessed every year for all clones. Before the second coppicing, biomass production of all clones was estimated. Signi1cant clonal di$erences were found in stool mortality, number of shoots per stool and biomass production. After 6 years (December 2001), stool mortality averaged 7– 65%. After the 1rst coppicing (1997), the average number of shoots ranged between three and seven shoots per stool; after the second coppicing, the average number of shoots ranged between 8 and 19 shoots per stool. During the 4 years following the 1rst coppicing, shoot density decreased exponentially, leaving mostly one or two dominant shoots per stool by the end of 2000. The other shoots had no signi1cant in9uence on stool dry mass, since most of the surviving shoots were suppressed and small and made little contribution to total dry mass. The diameter of the dominant shoot(s) was the most important determinant of stool dry mass. Mean annual biomass production ranged from 2 to 11 Mg ha−1 . ? 2002 Elsevier Science Ltd. All rights reserved. Keywords: Populus; Short-rotation coppice; Mortality; Coppicing ability; Sprouting vigour; Competition; Self-thinning; Biomass

1. Introduction Since the 1980s, Europe has shown a renewed interest in the use and technical realisation of renewable energy as a means to reduce the CO2 -emission levels from fossil fuels into the atmosphere. The European Commission proposed a major role for biomass, in ∗

Corresponding author. Tel.: + 32-3-820-22-89; fax: +32-3820-22-71. E-mail address: [email protected] (I. Laureysens).

particular for energy crops, since they o$er a feedstock for energy and industry, and additionally o$er an alternative use for the land taken out of agricultural production [1]. When fast-growing tree crops are grown in carefully tended plantations for rotations shorter than 15 years, they are referred to as short-rotation forests (SRF) [2,3]. SRF have several additional advantages. They have a positive impact on biodiversity, nutrient capture and carbon circulation in the soil–plant atmosphere system, especially on former agricultural land. Furthermore,

0961-9534/03/$ - see front matter ? 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 1 - 9 5 3 4 ( 0 2 ) 0 0 1 0 5 - 8

82

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

they protect the soil from water and wind erosion [4–6]. However, to become economically feasible, biomass yields of energy crops need to be increased. High biomass production can be achieved by optimising plant genotype to site and cultural management. Frequently applied agricultural management practices are site preparation, irrigation, fertilisation, high planting density and short-rotation coppicing. A planting density of about 10 000 –20 000 cuttings per hectare has been used to ensure an optimal utilisation of space and maximise potential yield [7]. Coppicing refers to the cutting of a tree at the base of the trunk, utilising the ability of the trees to regenerate from the cut stump, resulting in the emergence of new shoots from the stump and/or roots [8]. Coppicing is frequently applied at the end of the establishment year to promote sprouting of many shoots per cutting (which is supposed to increase 1nal biomass production) [7,9]. The early growth rate of coppice shoots is faster than that of seedlings or cuttings, because they bene1t from the existing root system and the rapid development of a high leaf area index. A high number of shoots per unit area gives a rapid leaf area development, leading to fast crown closure and eGcient utilisation of space. In addition, coppice shoots have an early onset of growth and have a continuous development of new nodes until late into the fall [8,9]. Several authors have shown that Populus coppices well [10–12], although others such as HervHe and Ceulemans [13] have shown that it does not always perform well under a short-rotation coppice regime. The latter authors found that after 2 or 3 years of growth, most non-coppiced poplars produced significantly more stem volume than those which were coppiced. Poplar has a vigorous regrowth of shoots after coppice, but competition for light results in the mortality of the smallest shoots in a stand, resulting in loss of biomass [13]. Moreover, in comparison with willow, poplar clones have not been selected for their coppicing ability. In Europe, they are mainly grown for production of veneer [14,15]. Most of the poplar clones examined in this study were bred and selected at the Institute of Forestry and Game Management (Geraardsbergen, Belgium), producing straight and tall trees with little branching resulting in high quality wood.

Few of these clones have been studied for their performance in a coppice culture. In addition, few studies have quanti1ed the sprouting capacity of poplar and its mortality rate of stools and shoots. Therefore, the objectives of this paper were: (1) to examine clonal and species di$erences in stool mortality and resprouting capacity of 17 poplar clones in short-rotation coppice culture; (2) to examine clonal di$erences in the growth and number of shoots; (3) to examine the relationship of sprouting vigour and shoot size to total above ground biomass production. 2. Materials and methods 2.1. Experimental set-up and plant material In April 1996, a short-rotation coppice plantation was established on an industrial zone of Boom near Antwerp, Belgium (51◦ 05 N; 04◦ 22 E). The plantation is situated on an old household waste disposal site, which was covered with a 2 m thick layer of sand, clay and mixed rubble. The soil is characterised by a high bulk density (heavy clay–loam) and high Ca-levels/high pH. Bulk density ranges between 1.221 and 1:621 g cm−3 ; pH ranges between 7.3 and 8.1. The soil upper horizons contain between 0.8% and 1.8% organic matter. The nutrient and mineral reserves are extremely high in comparison with forest soils, but moderate in comparison with agricultural soils [16]. The site is situated at 5 m above sea level and has a temperate climate, with a mean temperature of 10:6◦ C. The 6 years of the experiment (1996 –2001) were rather wet, with a mean annual precipitation of 870 mm. Prior to planting, the area was levelled and cleaned of large stones, plastic, metal and other debris. A rotor tiller was used in the early spring of 1996 for 1nal pre-planting soil preparation. Seventeen poplar clones (Populus), belonging to di$erent species and interspeci1c hybrids were studied, i.e. P. trichocarpa T. & G. × P. balsamifera L. (T×B) clone Balsam Spire; P. trichocarpa × P. deltoides Marshall (T × D) clones BeauprHe, Boelare, Hazendans, Hoogvorst, Raspalje and Unal; P. trichocarpa (T) clones Columbia River, Fritzi

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

Pauley and Trichobel; P. deltoides × P. nigra L. (D×N) clones Gaver, Gibecq and Primo; P. deltoides × P. trichocarpa (D×T) clones IBW1, IBW2 and IBW3; and native P. nigra (N) clone Wolterson. In April 1996, all clones were planted as 25 cm long dormant, unrooted hardwood cuttings, after being soaked in water for 24 hours. Cuttings were planted manually to a depth of 20 cm, leaving one or two buds above the soil surface. They were planted in a double-row design with alternating inter-row distances of 0.75 and 1:5 m, and a spacing of 0:9 m between cuttings within the rows, accommodating an overall planting density of 10 000 cuttings per hectare. A randomised block design was used with 17 clones × 3 replicate plots (except for clone Hoogvorst with six replicates, and for clones IBW1 and Raspalje with two replicates) according to a protocol prescribed by the British Forestry Commission [17]. However, for years 1997 and 1998, one replicate plot of both clones Balsam Spire and Hoogvorst was omitted from the shoot density results because of missing data. Individual plot size was 9 m × 11:5 m, containing 10 rows of 10 trees each. In the centre of each plot, 36 trees (6 × 6) were used as assessment trees to avoid border e$ects [18]. 2.2. Management regime April and May 1996 were characterised by a severe drought, so the plantation was irrigated shortly after planting to promote better establishment. In December 1996, all trees were cut back to a height of 5 cm above soil level to create a coppice culture, i.e. a stool composed of a stump with its shoot(s). The cuttings that did not survive the 1rst growing year were replaced in the spring of 1997 with new 25 cm long hardwood cuttings (40 cm for the clones with a mortality rate higher than 10%). These replanted trees were not included in the biomass production data and shoot density assessments. In January 2001 the stand was cut back again to a height of 5 cm above soil level. Mechanical weeding was done frequently during the 1996 and 1997 growing seasons. In February 1997, a 5 cm thick layer of mulch was applied to reduce weed growth and enhance the acidity of the soil. On three occasions, in June 1996, June 1997

83

and April 2001, weeds in between the trees were treated with a mixture of glyphosate (at 3:2 kg ha−1 ) and oxadiazon (at 9:0 kg ha−1 ), as the mechanical weed control was not e$ective. These herbicides were applied using a spraying device with a hood-covered nozzle to minimise impact on the trees. During the 1998, 1999 and 2001 growing seasons, the understory vegetation was mechanically cut to ground level with a trimmer. No fertilisation or irrigation was applied after the establishment of the experiment. 2.3. Measurements and harvesting Stool mortality was determined annually by counting the number of dead stools among the 36 assessment trees at the end of each growing season. Each year, the overall shoot density was determined by counting the number of surviving shoots in the assessment plots. Shoot diameter (d) of the living shoots was measured every year at the end of the growing season with a digital calliper (Mitutoyo, type CD-15DC, UK) at 22 cm above soil level [19]. In December 2000, i.e. at the end of the fourth growing season after coppicing, shoot diameter and dry mass production were determined for all 17 clones. Shoot diameter at 22 cm above soil level was measured for all living and dead shoots with a digital calliper. For stems thicker than 3 cm, diameter was measured in two perpendicular directions, and the mean value was used in further calculations [19]. Five shoots per clone were harvested. The shoots were selected using the technique of the quantils of the total, so that the sampled shoots represented the total basal area and its variation for each clone over the three replicates [20,21]. The sampled shoots were cut at about 5 cm above soil level, and separated into stems and branches. To estimate the dry mass of the stem, a subsample was taken from the lower half of the stem [22]. Dry mass (DM) of stems and branches was determined after drying in a forced air oven at 105◦ C until constant mass was reached. Allometric power equations (DM = d , with  and  as regression coeGcients) were calculated per clone to estimate dry mass of all living shoots as a function of stem diameter at 22 cm. The resulting plot estimates were extrapolated to produce dry mass per hectare per year.

84

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

2.4. Statistical analysis An analysis of variance (ANOVA) was used to test the signi1cance of di$erences in stool mortality and shoot density between clones (17) and parentages (6). The analysis was performed with the SAS statistical software package (SA System 6.12, SAS Institute Inc., Cary, NC) using the mixed procedure [23]. A randomised complete block design was applied, with clone or parentage as a 1xed factor, and replicate as a random factor. Satterthwaite’s procedure was used to obtain the denominator degrees of freedom. Least squares means were pairwise compared for clones/parentages, and were considered signi1cant when the P-value of the ANOVA t-test was ¡ 0:05. Data were tested for normality by the Shapiro–Wilk statistic (proc univariate in SAS); shoot density data were log transformed to obtain a normal distribution. Spearman’s rank correlation tests were performed with StatMost 2.50 (DataMost Corporation, Salt Lake City, USA), with correlations considered signi1cant at the P ¡ 0:05 level. 3. Results 3.1. Stool mortality From the original 10 000 cuttings per hectare planted in 1996, only 7850 survived the establishment year. Mean cutting mortality ranged from 0% for clone Fritzi Pauley (T) to 57% for clone Gaver (D × N) (Fig. 1). Clones Fritzi Pauley and Trichobel (T), Hoogvorst (T × D), Balsam Spire (T × B) and Wolterson (N) had the lowest cutting mortality, i.e. less than 12% in all replicates. Cutting establishment was bad for clones Boelare (T × D), Gaver and Gibecq (D × N), with a mean cutting mortality of more than 45% (Fig. 1). The stool mortality of the three latter clones was signi1cantly higher than 11 of the 14 other clones. Besides signi1cant clonal di$erences, considerable replicate di$erences were observed for some clones. Clones Boelare and Gaver both had one replicate with a cutting mortality of 64%, i.e. 23 dead cuttings of the originally 36 planted cuttings. The two replicates of clone Raspalje (T × D) had two and 17 dead cuttings, respectively; clone IBW1 (D × T) had two replicates with seven

and 17 dead cuttings, respectively. For clones Gibecq (D × N) and Fritzi Pauley (T), on the other hand, cutting survival was the same in all three replicates (single replicate data not shown). After coppicing, stool mortality increased, leaving 6635 stools per hectare at the end of the fourth year after the 1rst coppicing (2000). A lot of stools survived the 1rst year after the 1rst coppicing (1997), but not the second year (1998). During 1997, stool mortality occurred in only 22 of the 52 replicate plots, leaving about 7600 stools per hectare. Survival of almost a third of all clones (Boelare (T × D), Columbia River (T), Raspalje (T × D), IBW1 (D × T) and Wolterson (N)) was not a$ected by coppicing in either one of the replicates, i.e. none of their stools died. Only Gibecq and Primo (D × N) experienced stool mortality in all three replicates, with a mean stool mortality of 12% for both clones (single replicate data not shown). In 1998, 29 replicates experienced a stool mortality of at least 10%, leaving 6650 ha−1 . Clones Hazendans (T × D), Hoogvorst (T × D) and Wolterson (N) had a relatively low stool mortality, but all other clones had a stool mortality averaging 7–21% during 1998. Hardly any stool mortality occurred during the growing seasons of 1999 and 2000, i.e. the third and fourth year after coppicing. At the end of 2000, mean cumulative stool mortality ranged between 7% for clone Wolterson and 65% for clone Gibecq, i.e. leaving, respectively, 33 and 13 plants of the originally 36 cuttings (Fig. 1). After the second coppicing, only a few stools died, leaving about 6500 stools per hectare at the end of 2001. Clones Gaver (D × N), IBW1 (D × T) and IBW3 (D × T) lost one stool, clones Hoogvorst (T × D) and Trichobel (T) lost two stools. Highest stool mortality was observed for T × D clones Boelare and BeauprHe with respectively four and seven dead stools. The other clones did not experience any mortality of the assessment stools during this 1rst year after the second coppicing (Fig. 1). 3.2. Shoot density After planting, on average one or two shoots sprouted from the cuttings. After the 1rst coppicing, an average of 1ve shoots emerged from the stump. Clones Trichobel (T) and IBW1 (D × T) had an

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

85

Fig. 1. Stool mortality expressed as a percentage of the original 36 cuttings during the establishment year (1996), during the 1rst (1997), second (1998), third (1999) and fourth (2000) year after the 1rst coppicing, and during the 1rst year (2001) after the second coppicing. Mean values of replicates, and standard error of cumulative stool mortality (1996 –2001) are presented. BAL: Balsam Spire, BEA: BeauprHe, BOE: Boelare, HAZ: Hazendans, HOO: Hoogvorst, RAS: Raspalje, UNA: Unal, COL: Columbia River, FRI: Fritzi Pauley, TRI: Trichobel, GAV: Gaver, GIB: Gibecq, PRI: Primo, WOL: Wolterson.

average of seven shoots per stool. Clones Gaver (D × N), Gibecq (D × N), Hazendans (T × D), Raspalje (T×D) and IBW2 (D×T) had an average of three shoots per stool. All other clones ranged in between (Table 1). Besides signi1cant clonal di$erences, rather high inter-replicate variations were found. Clone Trichobel (T) had the highest inter-replicate variation, i.e. two replicates with an average number of, respectively, four and 1ve shoots per stool, and one replicate plot with an average number of 12 shoots per stool. Moreover, 12 was the highest individual plot score (single replicate data not shown). At the end of the fourth year after the 1rst coppicing (2000), shoot mortality resulted in mainly single-shoot stools for clones Hazendans (T × D), Hoogvorst (T × D), Fritzi Pauley (T) and IBW1 (D × T), leaving, respectively, 36%, 32%, 36% and 18% of the surviving shoots produced in the 1rst year after the 1rst coppicing (1997). Clones Columbia River (T) and Balsam Spire (T × B) both had an average of four shoots per stool, and clone Wolterson (N) had an average of three shoots per stool, leaving, respectively, 55%, 54% and 66% of the

surviving shoots in 1997 (data not shown). All other clones had an average of two shoots per stool at the end of 2000. After the second coppicing, an average of 13 shoots emerged from the stump, ranging between 8 for clone IBW2 (D × T) and 19 for clone Wolterson (N) (Table 1). When parentages were compared, the D×T and T × D clones had a signi1cantly lower number of shoots per stool than the other parentages (P ¡ 0:05), i.e. respectively 9 and 12 shoots per stool on average. The T clones, D × N clones, N clone Wolterson and T × B clone Balsam Spire had an average of respectively 15, 15, 19 and 18 shoots per stool. The number of shoots per stool produced after the second coppicing was not correlated with the number of shoots per stool produced after the 1rst coppicing (Table 2). In Fig. 2, the evolution in the number of shoots is shown for one clone per parentage. A decrease in shoot density from the 1rst to the fourth year after the 1rst coppicing was common for all clones, but the curve of decline di$ered considerably among clones (Fig. 3). An exponential decay curve 1t described best the evolution of shoot mortality. An exponential decay function (1) is used whenever the

86

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

Table 1 Average number of shoots per stool at the end of the 1rst year after the 1rst (1997) and the second coppicing (2001), number of shoots (2001) per unit stool basal area before the second coppicing (m−2 ), and mean and maximum shoot diameter of 17 poplar clones in a short-rotation coppice culture established in April 1996. The stand was coppiced in December 1996 and January 2001. For the number of shoots per stool and shoot density per stool, mean values with standard error (SE) of replicates are presented. Per clone, mean shoot diameter (upper quartile, Q3 ) and the diameter of the thickest shoot over all replicates is shown. T: P. trichocarpa, B: P. balsamifera, D: P. deltoides, N: P. nigra

Parentage Clone

Number of shoots per stool (1997)

Number of shoots per stool (2001)

Shoot density per stool (m−2 ) (2001)

Mean diameter (cm)

Mean (SE)

Mean (SE)

Mean (SE)

Mean (Q3 )

Balsam Spire

4 (0)

18 (3)

10.3 (1.6)

2.0 (2.8)

BeauprHe Boelare Hazendans Hoogvorst Raspalje Unal

5 6 3 5 3 4

10 11 15 13 12 10

T

Columbia River Fritzi Pauley Trichobel

6 (1) 4 (1) 7 (3)

15 (2) 15 (1) 15 (5)

D×N

Gaver Gibecq Primo

3 (0) 3 (0) 5 (2)

D×T

IBW1 IBW2 IBW3

N

Wolterson

T×B

T×D

(1) (2) (1) (1) (1) (0)

(2) (2) (2) (1) (0) (1)

4.1 4.1 4.2 3.4 5.9 5.0

y = y0 + A1 e−((x−1)=k) ;

9.8 11.6 11.5 12.6 10.8 10.7

4.9 (0.5) 5.9 (0.5) 6.3 (0.4)

2.8 (4.0) 4.5 (6.7) 3.7 (6.2)

10.7 8.5 11.1

18 (3) 13 (1) 15 (5)

6.9 (1.2) 10.4 (0) 6.2 (0.5)

3.7 (6.4) 2.8 (4.0) 4.1 (6.1)

9.7 7.7 10.5

7 (3) 3 (0) 4 (0)

9 (2) 8 (1) 9 (1)

6.6 (0.2) 6.6 (1.7) 5.7 (0.7)

4.2 (5.8) 3.2 (4.9) 2.9 (4.3)

7.9 8.1 8.0

4 (1)

19 (2)

7.0 (1.8)

3.0 (4.5)

8.9

(1)

where y0 is the o$set, A1 is the amplitude, k is the decay coeGcient and x is, respectively, the 1rst, second, third and fourth year after coppicing. The lower the decay coeGcient, the greater the reduction in number of shoots per hectare. The decay coeGcient di$ered between clones, but also between replicates. For clones Balsam Spire (T × B), Gibecq (D × N), Hoogvorst (T × D) and IBW2 (D × T), the exponential decay curve 1t of one replicate is shown in Fig. 3a, the corresponding parameters y0 ; A1 , and k are shown in Table 3. Curve 1ts were signi1cant (P ¡ 0:05) for most replicate plots, but were based

3.9 3.5 6.6 5.7 3.5 4.0

8.3

(6.0) (5.8) (8.7) (8.0) (6.0) (5.7)

rate at which a property y changes is proportional to its quantity, implying a constant decay rate

(0.4) (0.6) (0.8) (0.2) (0.9) (0.8)

Maximum diameter (cm)

on only four data points. Some replicates had a rather weak exponential decay curve 1t, because of a higher decay rate during the 1rst year (e.g. clone Wolterson) or during the second year after the 1rst coppicing (e.g. clone Columbia River; Fig. 3b). The number of shoots produced after the second coppicing (2001) was much higher than the number produced after the 1rst coppicing (1997), because of the larger diameter of the stumps (Table 1). For all clones, the latter assumption was supported by a signi1cant correlation between number of shoots per stool at the end of 2001 and total stool basal area at 22 cm before the second coppicing (Table 4). As a result, well performing clones with large diameters and low stool mortality, e.g. clones Hoogvorst and Hazendans (T × D), had a high number of shoots per stool in

∗

P ¡ 0:05.

n = 52 (n = 50 for shoot density 1997 and number of shoots per stool 1997).

∗∗ P 6 0:01. ∗∗∗ P ¡ 0:001.

0.4047∗∗

0.6794∗∗∗ 0.6756∗∗∗

0.6058∗∗∗

−0:6267∗∗∗ −0:6879∗∗∗

0.5746∗∗∗

−0:6377∗∗∗ −0:6723∗∗∗

0.7413∗∗∗ 0.1586 0.4299∗∗

0.2191

−0:4608∗∗∗ −0:3957∗∗

0.2629 0.8175∗∗∗ 0.4385∗∗ −0:0128

0.4533∗∗∗ −0:2837∗ −0:3875∗∗

0.2087 0.2703 0.6220∗∗∗ 0.2902∗ 0.3609∗∗

−0:0601 −0:0737 −0:1126

0.2978∗ 0.1925 0.2057 0.3276∗ 0.2367 0.2894∗

0.5243∗∗∗ −0:2151 −0:2508

Cum. stool Shoot Shoot Shoot No. of shoots No. of shoots No. of shoots No. of shoots mortality density (ha−1 ) density (ha−1 ) density (ha−1 ) per stool per stool per stool per stool 1996 –2000 1996 1997 2000 1996 1997 2000 2001

Biomass production 1997–2000 −0:6151∗∗∗ −0:7421∗∗∗ 0.5622∗∗∗ Cutting mortality 1996 0.8940∗∗∗ −0:8946∗∗∗ −0:8129∗∗∗ Cum. stool mortality 1996 –2000 Shoot density (ha−1 ) 1996 Shoot density (ha−1 ) 1997 Shoot density (ha−1 ) 2000 No. of shoots per stool 1996 No. of shoots per stool 1997 No. of shoots per stool 2000

Cutting mortality 1996

Table 2 Spearman’s rank correlation coeGcients among biomass produced after 4 years (1997–2000), cutting mortality at the end of the establishment year (1996), cumulative stool mortality at the end of the fourth year after the 1rst coppicing (1996 –2000), number of shoots per hectare at the end of the establishment year (shoot density 1996), number of shoots at the end of the 1rst year after the 1rst coppicing (shoot density 1997), number of shoots at the end of the fourth year after the 1rst coppicing (shoot density 2000), number of shoots per stool at the end of 1996, number of shoots per stool at the end of 1997, number of shoots per stool at the end of 2000 and number of shoots per stool at the end of 2001

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95 87

88

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

Fig. 2. Shoot demography, presented as the log number of shoots per stump (a) and the log number of shoots per hectare (b), for six poplar clones in a short-rotation coppice culture. The stand was established in April 1996 and coppiced in December 1996. Mean values of replicates and standard error are presented. BAL: Balsam Spire, HOO: Hoogvorst, COL: Columbia River, GIB: Gibecq and WOL: Wolterson.

2001. But the latter clones did not have a high shoot density per stool, explaining why the number of shoots produced after the 1rst and the second coppicing were not correlated. The number of shoots per unit basal area after the second coppicing was negatively correlated with the number of shoots per stool after the 1rst coppicing (R = −0:3926). The low-performing clone Gibecq (D × N), characterised by high stool mortality, had the highest number of shoots per unit basal area, i.e. 10 shoots m−2 (Table 1). This might suggest a positive correlation between stool mortality/light availability and sprouting vigour. However, a Spearman’s rank correlation test could not con1rm this hypothesis. The low-performing clone Boelare, also characterised by high stool mortality, had a relatively low number of shoots per unit basal area, while for clone Wolterson with the lowest stool mortality the opposite was true (Table 1). The average number of shoots per stool produced in the 1rst year after the 1rst coppicing (1997), on the other hand, was negatively

Fig. 3. Exponential decay curve 1t for the decrease in shoot density during the 4 years (1997–2000) following the 1rst coppicing for six poplar clones in a short-rotation coppice system. Graph (a) presents four clones with a good curve 1t, graph (b) presents two clones with a weak curve 1t.

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95 Table 3 Parameters of an exponential decay 1t describing the decline in the number of shoots as a function of time (1997–2000) for four poplar clones in a short-rotation coppice culture, where y0 is the o$set, A1 is the amplitude, k is the decay coeGcient and R2 is the coeGcient of determination. T: P. trichocarpa; B: P. balsamifera; D: P. deltoides; N: P. nigra Clone

y0

A1

k

R2

Balsam Spire (T × B) Gibecq (D × N) Hoogvorst (T × D) IBW2 (D × T)

115 19 50 35

57 16 276 49

0.27 1.66 0.57 1.66

1 1 1 1

correlated with cutting mortality (R = −0:2837; Table 2). Signi1cant di$erences in shoot density per stool were found between clones, but also between parentages. The T×D clones had a signi1cantly lower

89

number of shoots per basal area than the other parentages; T × B clone Balsam Spire had a signi1cantly higher shoot density per stool. 3.3. Shoot diameter distribution For all replicates, a frequency distribution was calculated, showing the distribution of shoot diameters at the end of the fourth year after the 1rst coppicing (2000). In Fig. 4, the diameter frequency distribution of one clone per parentage is shown, each distribution presenting the shoot diameters of all replicates of the relevant clone. Clones Hoogvorst (T × D; Fig. 4), Hazendans (T × D), Fritzi Pauley (T) and IBW1 (D × T) had a rapid elimination of the shoots in the 4 years following the 1rst coppicing, leaving mostly single-stem stools with large diameters at the end of the fourth year after

Table 4 Spearman’s rank correlation coeGcient between stool basal area before the second coppicing (2000) and number of shoots per stool at the end of the 1rst year after the second coppicing (2001), between number of shoots per stool and their maximum diameter at the end of the fourth year after the 1rst coppicing (2000) and between stool dry mass and its number of shoots at the end of 2000. Parentage

Clone

T×B

Balsam Spire

T×D

BeauprHe Boelare Hazendans Hoogvorst Raspalje Unal

Sample size

Basal area 2000× no. of shoots per stool 2001

No. of shoots per stool 2000× max. shoot diam. 2000

Stool dry mass 2000× no. of shoots per stool 2000

70

0.4218∗∗∗

55 31 69 164 38 76

0.6344∗∗∗ 0.3970∗ 0.6958∗∗∗ 0.6557∗∗∗ 0.7343∗∗∗ 0.5199∗∗∗

0.3113∗ −0:1777 0.1217 0.0045 0.5339∗∗∗ 0.3470∗∗

0.3936∗∗ −0:1099 0.1690 0.0465 0.5715∗∗∗ 0.4213∗∗∗

0.0414

0.1227

T

Columbia River Fritzi Pauley Trichobel

61 79 80

0.3469∗∗ 0.5454∗∗∗ 0.6006∗∗∗

−0:1585 0.0929 0.0816

−0:0186 0.1150 0.1255

D×N

Gaver Gibecq Primo

32 32 49

0.5789∗∗∗ 0.6444∗∗∗ 0.6836∗∗∗

0.0413 −0:1146 0.2744∗

0.1151 −0:0250 0.3937∗∗

D×T

IBW1 IBW2 IBW3

32 46 63

0.7160∗∗∗ 0.7169∗∗∗ 0.7030∗∗∗

0.1974 0.4632∗∗ 0.4496∗∗∗

0.2735 0.5105∗∗∗ 0.5144∗∗∗

N

Wolterson

88

0.4407∗∗∗

0.0129

0.1259

∗ P ¡ 0:05.

∗∗ P 6 0:01.

∗∗∗ P ¡ 0:001.

90

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

Fig. 4. Diameter distribution of shoots for six poplar clones in a short-rotation coppice culture at the end of the fourth year after coppicing. Each distribution presents the number of shoots of all replicates in 15 intervals. Total number of shoots (n), mean shoot diameter (dmean ) and maximum shoot diameter (dmax ) are shown as the inset numbers.

coppicing. By comparison, Balsam Spire (T × B), Columbia River (T) and Wolterson (N) still had a majority of stools with more than three shoots per stump, resulting in a high shoot density. Since stools had a maximum of two large-diameter ‘dominant’ shoots (data not shown), a high proportion of shoots was found in the small diameter classes for the latter clones with a consequently low mean diameter (Fig. 4). Fig. 4 also shows the poor growth rate of clones Gibecq (D × N) and IBW2 (D × T) charac-

terised by a high representation of shoots in the small diameter classes, a low number of shoots because of high stool mortality, and a low mean and maximum shoot diameter. Mean diameters ranged from 3 cm for clones Gibecq (D × N) and IBW3 (D × T) to 7 cm for clone Hazendans (T × D). In Table 1, the upper quartile (Q3 ) is shown for the diameter distribution over all replicates of each clone, below which 34 of the shoot diameters is present. Maximum plot diameter ranged from 8 cm for clones Balsam Spire (T × B), IBW1

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

91

Fig. 5. Above-ground woody biomass production of 17 di$erent poplar clones by parentage after 4 years of growth in a coppice culture. Mean values of three replicates and standard error bars are presented. BAL: Balsam Spire, BEA: BeauprHe, BOE: Boelare, HAZ: Hazendans, HOO: Hoogvorst, RAS: Raspalje, UNA: Unal, COL: Columbia River, FRI: Fritzi Pauley, TRI: Trichobel, GAV: Gaver, GIB: Gibecq, PRI: Primo, WOL: Wolterson, T: P. trichocarpa, B: P. balsamifera, D: P. deltoides, N: P. nigra.

(D × T), IBW2 (D × T), IBW3 (D × T) and Gibecq (D×N) to 13 cm for clone Hoogvorst (T×D; Table 1). In comparison with the inter-replicate variation observed for stool mortality and shoot density, variation in mean and maximum shoot diameter was relatively low. Therefore, shoots of all replicates were presented in one frequency distribution (Fig. 4). 3.4. Biomass production Mean annual above-ground biomass yields at the end of the 1rst rotation (2000) are shown in Fig. 5. The T × D clones Hazendans and Hoogvorst performed best, with a mean biomass yield of respectively 10:8 Mg ha−1 year −1 and 10:1 Mg ha−1 year −1 . The T clones Fritzi Pauley, Columbia River and Trichobel, and N clone Wolterson also had a relatively high performance, i.e. 8.1, 7.8, 8.4 and 8:1 Mg ha−1 year −1 , respectively. Lowest performance was obtained from clones Gibecq (D×N) and IBW2 (D×T), i.e. 1.6 and 2:5 Mg ha−1 year −1 . A relatively high inter-replicate variation was found for some clones, but could not be clearly related to soil characteristics. Inter-replicate dissimilarity in soil characteristics ranged between 13% and 28%; inter-replicate variation in biomass

production ranged between 0% and 83%. Soil characterics, biomass production and its inter-replicate variation are discussed in further detail by Laureysens et al. [16]. 3.5. Determinants of biomass production For the 52 plots, an examination was made of the correlation (n = 52) of total biomass with cumulated stool mortality at the end of 2000, with shoot density in the establishment year (1996), in the 1rst (1997) and the fourth year after coppicing (2000), with mean number of shoots per stool in 1996, 1997 and 2000, and with parameters y0 ; A1 and k of the exponential decay 1t. Individual stool mass was correlated (n=number of stools of the relevant replicate plots) with the number of shoots, mean shoot diameter and maximum shoot diameter at the end of 2000. A strong negative relation between biomass and stool mortality was expected and con1rmed by a Spearman’s rank correlation test (R = −0:742). Other signi1cant correlations with biomass were found for shoot density in 1996 (R = 0:562), 1997 (R = 0:575) and 2000 (R = 0:405). However, these were probably mainly the result of the strong relation between

92

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

biomass and stool mortality, since no signi1cant correlation between biomass and mean number of shoots per stool in 1996 and 2000 was found. Number of shoots per stool in 1997, on the other hand, had a signi1cant impact on biomass production (R = 0:453) (Table 1). The parameters of the exponential decay 1t did not have a signi1cant in9uence on biomass production. Individual stool mass was correlated with mean shoot diameter (R = 0:556) and maximum shoot diameter (R = 0:978), but not with number of shoots per stool in 2000. When clones were considered separately, a signi1cant correlation between stool dry mass and mean/maximum shoot diameter was observed for all clones, but the role of shoot density in 2000 differed between clones. For a few clones, i.e. BeauprHe (T × D), IBW2 (D × T), IBW3 (D × T), Primo (D × N), Raspalje (T × D) and Unal (T × D), a signi1cant correlation was found between stool dry mass and number of shoots per stool at the end of 2000. These were the same clones that had a signi1cant correlation between number of shoots per stool in 2000 and maximum shoot diameter (Table 4). 4. Discussion and conclusion Mean annual biomass production ranged between 1.6 and 10:8 Mg ha−1 . Highest performance was achieved by T × D clones Hazendans and Hoogvorst, T clones Fritzi Pauley, Columbia River and Trichobel, and N clone Wolterson with mean annual biomass production ranging between 7.8 and 10:8 Mg ha−1 . In other studies, 4-year rotation yields between 5 and 28 Mg ha−1 year −1 have been observed, depending on clone, soil fertility and climate. High annual biomass yields of 20 Mg ha−1 or more were mostly obtained in more favourable climates and with higher soil fertility than this site [24–27]. Compared to these results from the literature, low performance was obtained by the T×D clone Boelare, D×T clones IBW1, IBW2 and IBW3, and D×N clones Gaver and Gibecq with mean annual biomass production ranging between only 1.6 and 3:6 Mg ha−1 . The clones with the lowest performance were, not surprisingly, the clones with a high cumulative stool mortality, i.e. averaging 33– 65%. These mortality percentages were rather high given that in general, stump survival of poplar

averages 92–99% [28], although Armstrong [29] also reported a mortality of up to 85% for poplar, depending on clone and site. In the current study, highest mortality occurred during the establishment year (i.e. cutting mortality) and during the second year after the 1rst coppicing. High cutting mortality of the D × T and D × N clones might be explained by the bad rooting capacity of the mother P. deltoides, given the high bulk density of the soil [30]. Stool mortality during the second year after 1rst coppicing was similar for all clones, and might be explained by a heavy rust infection and/or competition, since no extreme weather conditions were observed. In Belgium, the summer of 1998 was characterised by heavy rust infection (Melampsora larici-populina), because of a high atmospheric humidity. Rust in combination with the bark-killing fungus Discosporium populeum caused high mortality rates in several poplar plantations (M. Steenackers (2001), pers. comm.). Rust damages the photosynthetic leaf area of a shoot, resulting in an insuGcient nutrient supply for the root system. This might have caused the death of the smallest stools in the stand, because of depletion of the root system or indirectly as a result of the decline in growth rate. Verwijst [31] showed Melampsora spp to have a negative impact on the height of infected willow shoots, thereby decreasing their ability to compete for light and making them more susceptible to self-thinning. A high stool mortality, resulting in a lower canopy density with higher light availability, had no signi1cant in9uence on the number of shoots per unit stool basal area after the second coppicing. However, it cannot be ruled out that within a clone, stool mortality can have an in9uence on sprouting vigour. The number of shoots per stool was highly clone speci1c. In addition, leaf and crown dimensions were considerably di$erent between clones, causing a di$erence in light availability for similar stool mortality. The relation between sprouting vigour and stool mortality within a clone could not be examined because of a limited number of replicates. Cutting mortality, on the other, had a negative impact on the number of shoots produced after the 1rst coppicing, contradicting a study of Ashby et al. [32] on silver maple (Acer saccharinum L.). The latter authors found a higher number of shoots per stool in provenances with a relatively low density of surviving trees prior to coppicing. Studies that link sprouting vigour to stool density for poplar

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

were not found. In this study, clone and parentage determined sprouting vigour of the stumps, ranging between 3 and 7 shoots per stool after the 1rst coppicing and between 8 and 19 shoots per stool after the second coppicing. Within a clone, the number of shoots per stool produced after coppicing was positively correlated with the size of the stool before coppicing, as reported in other studies. These suggested that larger stools leave a larger surface area to sprout from and have a larger below-ground store of reserves, resulting in the production of numerous large shoots after coppicing [12,32–34]. However, the highest performing clones Hoogvorst and Hazendans (T × D) had stools with large diameters, but with rather low shoot densities per stool. The low-performing clone Gibecq (D × N), on the other hand, had the highest number of shoots per unit basal area at the 1rst year after the second coppicing. During the 1rst year after coppicing, the shoots with the highest growth rate in a stool became the dominant shoots, suppressing all other shoots in the following years, as also reported by Auclair and Bouvarel [35] for poplar and Rullier-Breval [36] for chestnut (Castanea sativa). In general, these suppressed shoots died with a constant decay rate during the 4 years following coppicing. Shoot mortality is shown to be mainly caused by competition for light [31,37]. However, the rate of decline differed signi1cantly among clones. Clones Hoogvorst (T × D), Hazendans (T × D), Fritzi Pauley (T) and IBW1 (D × T) had mainly single-stem stools in the fourth year after coppicing (2000), while the clones Columbia River (T) and Wolterson (N) had an average of four shoots per stool at the end of 2000. In this fourth year, the number of shoots per stump had no signi1cant impact on stool dry mass for most clones. Individual stool weight was mostly determined by the diameter of its one or two dominant shoots, since a stool had a maximum of two dominant shoots with large diameters at the end of 2000. Suppressed shoots were always small and had only a small contribution to total stool weight. However, the mean number of shoots per stool produced in the 1rst year after the 1rst coppicing was shown to correlate with total biomass production after 4 years. This supports the 1ndings of Tschaplinski and Blake [38], who demonstrated the role of the suppressed shoots in a stool. They showed early removal of all

93

shoots but the dominant (tallest) shoot resulted in reduced vigour and viability of the dominant shoot, and in carbohydrate depletion of the lower stem and roots. Therefore, they concluded that the suppressed shoots support the growth of the dominant shoot by supplying carbon to the lower stem and roots. The results of this study suggest, however, that the role of the suppressed shoots is clone speci1c, since the rate of decline of these shoots di$ered significantly between clones. For clones Hoogvorst and Hazendans, with single-stem stools at the end of the fourth year after coppicing, the role of these suppressed shoots seemed to be limited to the 1rst year. Other clones, e.g. Columbia River (T) and Wolterson (N), kept several surviving, but suppressed, shoots after 4 years. Further research is necessary to examine the physiological role of the suppressed shoots, and the reason for the di$erences among clones. In terms of stool mortality and biomass production a relatively high variation between replicates was found for some clones. This was also in agreement with the observations of Armstrong et al. [39] for poplar and by Tahvanainen [40] for willow. The latter author found, as in this study, some variation within similar climatic conditions, but failed to link it with soil characteristics. This implies that the performance of some clones is highly dependent on other factors beside the known soil or climatic determinants. Therefore, it might be more useful to select for clones with a wide range of growing conditions, rather than looking for ‘the right clone for the right place’. In addition, a relatively high variation in sprouting vigour between replicates was found for some clones, indicating that clone–site interactions are a$ecting the population dynamics of poplar short-rotation coppice systems. The majority of clones included in the present study were not selected for their coppicing ability, but for the production of straight, long stems with little branching. Nevertheless, some of these clones proved to be 1t under a coppice regime for at least 6 years with two coppices. Their strong apical dominance ensures a rapid shoot elimination, resulting in large, single-stem stools with high productivity and low stool mortality. However, further research is necessary to determine whether the yield and survival of these clones is kept up with multiple rotations.

94

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

Acknowledgements This study is being supported by a research contract with the Province of Antwerp. All plant materials were kindly provided by the Institute for Forestry and Game Management (Geraardsbergen) and by the Forest Research, Forestry Commission (UK). The project has been carried out in close co-operation with Eta-com B., supplying the grounds and part of the infrastructure, and with the logistic support of the city council of Boom. We gratefully acknowledge T. and P. Laureysens for help with the harvest, H. Weerts and P. Missoul for help with the 2001 diameter assessments, as well as T. Verwijst and E. Casella for useful comments and suggestions on an earlier version of the manuscript. References [1] European Commission. Energy for the future—Renewable sources of energy. White paper for a community strategy and action plan 1997, COM (97) 599 1nal (26/11/97). [2] Cannell MGR, Smith RI. Yields of minirotation closely spaced hardwoods in temperate regions: review and appraisal. Forest Science 1980;26:415–28. [3] Deraedt W, Ceulemans R. Clonal variability in biomass production and conversion eGciency of poplar during the establishment year of a short rotation coppice plantation. Biomass and Bioenergy 1998;15:391–8. [4] Perttu KL. Ecological, biological balances and conservation. Biomass and Bioenergy 1995;9:107–16. [5] Isebrands JG, Karnosky DF. Environmental bene1ts of poplar culture. In: Dickmann DI, Isebrands JG, Eckenwalder JE, Richardson J, editors. Poplar culture in North America. Ottawa, Canada: NRC Research Press, 2001. p. 207–18. [6] Gordon JC. The productive potential of woody plants. Iowa State Journal of Research 1975;49:267–74. [7] Macpherson G. Home-grown energy from short-rotation coppice. Ipswich: Farming Press Books, 1995. p. 214. [8] Blake TJ. Coppice systems for short-rotation intensive forestry: the in9uence of cultural, seasonal and plant factors. Australian Forest Research 1983;13:279–91. [9] Sennerby-Forsse L, Ferm A, Kauppi A. Coppicing ability and sustainability. In: Mitchell CP, Ford-Robertson JB, Hinckley T, Sennerby-Forsse L, editors. Ecophysiology of short rotation forest crops. Oxford: Elsevier Applied Science, 1992. p. 146–84. [10] DeBell DS, Alford LP. Sprouting characteristics and cutting practices evaluated for cottonwood. Tree Planter’s Notes 1972;23:1–3. [11] Wendel GW. Stump sprout growth and quality of several Appalachian hardwood species after clear cutting. Research paper NE-329. USDA-FS, 1975. p. 9.

[12] DeBell DS, Radwan MA, Reukema DL, Zasada JC, Harms WR, Cellarius RA. Increasing the productivity of biomass plantations of alder and cottonwood in the Paci1c Northwest. Annual Technical Report. Paci1c Northwest Research Station, Olympia, Washington, 1988. p. 135 + app. [13] HervHe C, Ceulemans R. Short-rotation coppiced vs. non-coppiced poplar: a comparative study at two di$erent 1eld sites. Biomass and Bioenergy 1996;11:139–50. [14] Stevens M, Van Acker J, Steenackers J. Populierenonderzoek—Technologisch onderzoek helpt eigen houtvoorziening te verhogen. Het Ingenieursblad 1992;6: 19–25. [15] Terrasson D, Valadon A. Le marchHe des grumes de peuplier (The market of poplar wood). Comptes H rendus de l’AcadHemie d’ Agriculture de France—Etat et perspectives de la populiculture. Pulnoy: S.P.E.I., 1995. p. 23–30. [16] Laureysens I, Bogaert J, Blust R, Ceulemans R. Biomass production of 17 poplar clones grown in a short-rotation coppice culture in relation to soil characteristics. Forest Ecology and Management, submitted for publication. [17] Armstrong A. The United Kingdom network of experiments on site/yield relationships for short rotation coppice. Forestry Commission Research Information Note 294. Forestry Commission, Edinburgh, 1997. [18] Zavitkovski J. Small plots with unplanted plot border can distort data in biomass production studies. Canadian Journal of Forest Research 1981;11:9–12. [19] Pontailler JY, Ceulemans R, Guittet J, Mau M. Linear and non-linear functions of volume index to estimate woody biomass in high density young poplar stands. Annales des Sciences ForestiVeres 1997;54:335–45. [20] CermHak J, Kucera J. Scaling up transpiration data between trees, stands and watersheds. Silva Carelica 1990;15:101–20. [21] CermHak J, Riguzzi F, Ceulemans R. Scaling up from the individual tree to the stand level in Scots pine. I. Needle distribution, overall crown and root geometry. Annales des Sciences ForestiVeres 1998;55: 63–88. [22] Telenius BF. Implications of vertical distribution and within-stand variation in moisture content for biomass estimation of some willow and hybrid poplar clones. Scandinavian Journal of Forest Research 1997;12: 336–9. [23] Littell RC, Milliken GA, Stroup WW, Wol1nger RD. SAS system for mixed models. Cary, NC, USA: SAS Institute Inc., 1996. p. 633. [24] Bungart R, HWuttl RF. Production of biomass for energy in post-mining landscapes and nutrient dynamics. Biomass and Bioenergy 2001;20:181–7. [25] Strong T, Hansen E. Hybrid poplar spacing/productivity relations in short rotation intensive culture plantations. Biomass and Bioenergy 1993;4:255–61. [26] Heilman PE, Stettler RF. Nutritional concerns in selection of black cottonwood and hybrid clones for short rotation. Canadian Journal of Forest Research 1986;16:860–3. [27] Heilman PE, Ekuan G, Fogle D. Above- and below-ground biomass and 1ne roots of 4-year-old hybrids of Populus

I. Laureysens et al. / Biomass and Bioenergy 24 (2003) 81 – 95

[28]

[29] [30] [31] [32] [33]

trichocarpa × Populus deltoides and parental species in short-rotation culture. Canadian Journal of Forest Research 1994;24:1186–92. Ceulemans R, McDonald AJS, Pereira JS. A comparison among eucalypt, poplar and willow characteristics with particular reference to a coppice, growth-modelling approach. Biomass and Bioenergy 1996;11:215–31. Armstrong A, Johns C, Tubby I. E$ects of spacing and cutting cycle on the yield of poplar grown as an energy crop. Biomass and Bioenergy 1999;17:305–14. Dickmann DI, Stuart KW. The culture of poplars in Eastern North America. East Lansing, Michigan: Michigan State University Publications, 1983. p. 168. Verwijst T. In9uence of the pathogen Melampsora epitea on intraspeci1c competition in a mixture of Salix viminalis clones. Journal of Vegetation Science 1993;4:717–22. Ashby W, Bresnan D, Kjelgren R, Roth P, Preece J, Huetteman C. Coppice growth and water relations of silver maple. Biomass and Bioenergy 1993;5:317–23. Verwijst T. Cyclic and progressive changes in short-rotation willow coppice systems. Biomass and Bioenergy 1996;11: 161–5.

95

[34] Solomon DS, Blum BM. Stump sprouting of four Northern hardwoods. US Forest Service Research Paper NE-59. US Forest Service, Washington, 1967. p. 13. [35] Auclair D, Bouvarel L. Biomass production and stool mortality in hybrid poplar coppiced twice a year. Annales des Sciences ForestiVeres 1992;49:351–7. [36] Rullier-Breval B. Croissance d’un taillis de chataignier aprVes H coupe. Etude au cours des trois premiVeres annHees. Doctoral thesis, UniversitHe de Paris-Sud, Orsay, France, 1985. p. 155 + app. [37] Weiner J, Thomas SC. Size variability and competition in plant monocultures. Oikos 1986;47:211–22. [38] Tschaplinski TJ, Blake TJ. Growth and carbohydrate status of coppice shoots of hybrid poplar following shoot pruning. Tree Physiology 1995;15:333–8. [39] Armstrong A, Broadmeadow M, Evans S, Henshall P, Lonsdale D, Matthews R, Straw N, Tubby I. Yield models for energy coppice of poplar and willow. Interim Report. Forestry Commission, Edinburgh, UK, 2000. p. 106. [40] Tahvanainen L. Diameter growth models induced by competition for four Salix clone monocultures. Biomass and Bioenergy 1996;11:167–75.