619

Western redcedar response to precommercial thinning and fertilization through 25 years posttreatment

Warren D. Devine and Constance A. Harrington

Abstract: There is little infonnation available on the long-tenn effects of managing western redcedar (Thuja plicata Donn ex D. Don). IJ;l a 15- to 20-year-old naturally regenerated second-growth redcedar stand on a poor site on the Olympic Peninsula of Washington, we tested crop tree (largest 250 trees·ha-1) response to precommercial thinning and fertilization in a replicated study. Fertilization treatments were N or N+P applied at study installation and year 13; precommercial thin­ning occurred at installation. Precommercial thinning without fertilization produced a sustained increase in periodic indi­vidual-tree basal area (BA) growth rate from years 3 through 25 posttreatment. However, through year 12, higher BA growth rates resulted from fertilization. During years 13 through 25, when intraspecific competition increased, the highest BA growth rate resulted from the combination of fertilization and precommercial thinning. Compared with the unthinned­unfertilized control, fertilization without thinning increased year-25 crop-tree height by 34% and BA by 137%; thinning without fertilization increased height by 11 % and BA by 91 %. Height to live-crown base was decreased by thinning but increased by fertilization, while thinning significantly increased stem taper on the lower bole. Treatment responses and fo­liar analyses indicate crop tree growth was substantially limited by nutrient availability.

Resume: II y a peu d'information disponible sur les effets a long terme de l'amenagement du thuya geant (Thuja plicata Donn ex D. Don). Dans un peuplement de thuya age de 15 a 20 ans, regenen! naturellement apres une coupe et etabli sur une station pauvre de la peninsule Olympic de l'Etat de Washington, aux Etats-Unis, nous avons mesure la reaction des ar­bres d'avenir (les 250 plus gros arbres/ha) a une eclaircie precommerciale et a une fertilisation dans un dispositif experi­mental comportant des repetitions. Les traitements de fertilisation comprenaient des applications de N ou de N et P lorsque l' etude a debute et 13 ans plus tard. L' eclaircie precommerciale a ete faite au moment ou l' etude a debute. L'ec1aircie precommerciale sans fertilisation a produit une augmentation soutenue du taux d'accroissement periodique de la surface terri ere (ST) des arbres individuels a partir de la troisieme annee apres Ie traitement jusqu'a la 2SC annee. Ce­pendant, jusqu'a la 12e annee, la fertilisation a produit un taux de croissance de la ST plus eleve. Au cours des anm!es 13 a 25, pcriode pendant laquelle la competition intraspecifique a augmente, la fertilisation et l'ec1aircie pn!commerciale combinees ont produit Ie taux d'accroissement de la ST Ie plus eleve. Comparativement au remoin non eclairci et non fer­tilise, la fertilisation sans eclaircie a augmente la hauteur de 34 % et la ST de 137 % alors que l'ec1aircie sans fertilisation a augmente la hauteur de 11 % et la ST de 91 % sur les arbres d'avenir apres 25 ans. La hauteur jusqu'a la base de la cime vivante a ete reduite par l'eciaircie mais augmentee par la fertilisation, alors que l'eciaircie a significativement aug­mente Ie defilement de Ia partie inferieure du tronc. Les reactions aux traitements et des analyses foliaires indiquent que la croissance des arbres d'avenir a ete substantiellement limitee par la disponibiJite des nutriments.

[Traduit par la Redaction]

Introduction Compared with the other major conifer species of the

Pacific Northwest, there has been little research on the man­agement of western redcedar (Thuja pUcata Donn ex D. Don) (Curtis et al. 2007). Although second-growth redcedar timber is relatively valuable and as decay resistant as old-growth wood (Freitag and Morrell 2001), concerns regarding estab-

Received 26 September 2008. Accepted 5 December 2008. Published on the NRC Research Press Web site at cjfr.nrc.ca on 7 March 2009.

W.D. Devine1 and C.A. Harrington. Pacific Northwest Research Station, USDA Forest Service, Olympia Forestry Sciences Laboratory, 3625 93rd Avenue SW, Olympia, WA 98512, USA.

lCorresponding author (e-mail: wdevine@fs.fed.us).

lishment, growth rate, and stem form historically limited management of redcedar except on sites where other species were poorly suited (Nystrom et al. 1984; Oliver et al. 1988). In recent years, planting and management of redcedar have increased, particularly in coastal British Columbia (Gonzalez 2004). Beyond the Pacific Northwest, the species has been planted for timber production throughout Europe and in New Zealand, Africa, and elsewhere (Minore 1983; Hermann 1987). In its native range redcedar occasionally occurs in pure stands, but it is most often found in mixed-species stands across a wide range of sites, including sites with low nutrient availability or poor soil drainage (Minore 1990).

The wide nutritional amplitude of redcedar al10ws it to persist on nutrient-poor soils (Weetman et al. 1988), but sev­eral studies have shown growth responses to fertilization, particularly on low-productivity sites. Most of these studies reported significant short-term growth increases of saplings

Can. J. For. Res. 39: 619-628 (2009) doi: 10.1139/X08-199 Published by NRC Research Press

620

or small trees in response to application of inorganic N or N+P fertilizers (Blevins et al. 2006; Negrave et al. 2007) or to various forms of organic amendments (McDonald et al. 1994; Prescott and Blevins 2005). Early results from the study reported here (Harrington and Wierman 1990), and from another study (Weetman et al. 1989), isolated signifi­cant growth responses to P fertilization when P was applied in addition to N. Management recommendations for redce­dar production on poor-quality coastal sites, where the spe­cies is prevalent, include an early N+P application followed by periodic applications of N or N+P (Blevins and Prescott 2002). Control of salal (Gaultheria shallon Pursh), which competes strongly for nutrients; is often necessary in young stands (Messier 1993; Prescott 1996). Few data exist on the long-term effects of fertilization on redcedar growth. Early responses to N+P were no longer significant for redcedar in a mixed-species stand after 13 years (Bennett et al. 20(3), although, in a redcedar plantation, growth increases associ­ated with fertilization were present 15 years posttreatment (Blevins et al. 2006). The growth rate of redcedar was posi­tively correlated with foliar nutrient concentration across a range of sites in the coastal Pacific Northwest (Radwan and Harrington 1986), and the species' foliar nutrient levels also have been related to soil parent material (Kranabetter et al. 2003).

There is limited information on the response of redcedar to thinning, whether in naturally established stands or in plantations. Five year results from the study reported here in­dicated that precommercial thinning in a 15- to 20-year-old natural stand significantly increased diameter growth of re­sidual trees (Harrington and Wierman 1990). Based on data from naturally e tablished redcedar stands. Nystrom et al. (1984) suggested that trees could be selected for precommer­cial thinning by age 10 or when trees reached approximately 4.5 m in height. However delaying thinni ng until after lower limbs have died could reduce stem defects such as fluting and large knots, which are exacerbated when live lower limbs are exposed to sunlight (Oliver et at 1988). Thinning also may influence taper of redcedar, as taper is reduced by maintaining closer spacings (DeBell and Gartner 1997). Spacing studies have shown negative correlations between redcedar planting density and variables including height growth, diameter growth, and individual-tree stem volume growth (Reukema and Smith 1987; Negrave et al. 2(07).

The objective of this research was to evaluate the 25 year effects of precommercial thinning and fertilization with N and N+P on growth of young western redcedar. While the study was conducted in a redcedar-dominated second-growth stand that established naturally after clear-cutting, the long­term treatment responses reported here also have implica­tions for management of redcedar in mixed-species stands and in plantations. This analysis focuses on the growth of crop trees (largest 250 trees·ha-1) selected at the time of treatment. Three and five year results from this study were published previously (Harrington and Wierman 1985, 1990).

Methods

Study site The study is located in the northwestern part of the Olym­

pic Peninsula, approximately 10 km from the Pacific Coast,

Can. J. For. Res. Vol. 39, 2009

Clallam County, Washington, USA (4S008/N, 124°38/W; elevation 100 m), in the Picea sitchensis zone (Franklin and Dyrness 1988). Under a mild maritime climate, January and August temperatures average 5 and 15 °C, respectively (measured in nearby Quillayute, Washington; Western Re­gional Climate Center 2008). Long-term mean annual pre­cipitation is 2600 mm; on average, 13% of this precipitation occurs from 1 May through 31 August (Western Regional Climate Center 20(8). The study plots are located in a 12 ha area where slope ranges from 0% to ] 0%. Soils are of the Kydaka series (Typic Humaguepts), formed in glacial lacustrine sediments (averaging 8-20 cm thick) over glacial outwash (Soil Survey Staff 2008). Soils are poorly drained, with a dense layer of glacial till of very low permeability beginning at a depth of SO to 160 em. In the Canadian System of Soil Classification, this soil series is approxi­mately equivalent to a Humic Gleysol with mor humus (Soil Classification Working Group 1998). The 50 year site index for redcedar in the study area is approximately 16-18 m (Kurucz 1978).

Prior to clear-cutting in 1961, the study area was domi­nated by redcedar ranging from 0.5 to 4.0 m in diameter. After clear-cutting, a stand dominated by redcedar origi­nated naturally from seed and, to a lesser extent, from ad­vance regeneration. At the time of study establishment in 1980, redcedar accounted for over 95% of stand .basal area (BA). Other tree species present in 1980 included western hemlock (Tsuga heterophylla (Raf.) Sarg.) and occasional Pacific yew (Taxus hrevifoLia Nutt.), Pacific silver fir (A hies amahilis (Dougl.) Forbes), Sitka spruce (Picea sitchensis (Bong.) Carriere), red alder (Alnus ruhra Bong.), and cas­cara (Rhamnus purshiana DC.). The most prevalent shrub species were salal, red huckleberry (Vaccinium parvifolium Sm.), and oval-leaf blueberry (Vaccinium ovalifolium Sm.). In 1980 there was an average of 5900 trees·ha-1 in the study area. The redcedar varied widely in size and age, but the largest 250 trees·ha- 1 averaged 5.4 ± 0.6 m in height and were predominantly in the 15 to 20 year age range.

Study design and treatments The study followed a randomized block design, with four

blocks arranged according to soil drainage, which varied with topography. Twenty-eight study plots were perma­nently established in September 1980; each had a 30 m x 50 m treatment area and an interior 20 m x 40 m measure­ment plot. Treatments consisted of combinations of precom­mercial thinning and fertilization regimes (Table 1). These treatments were selected in coordination with the site' s in­dustrial forest own'er to reflect managemenl options that were likely to be implemented operationally on that forest type. The T-O, UT-O, and UT-2NP treatments occurred on four plots each (one per block); both T-2N and both T-2NP treatments occurred on three plots each, as the treatment plots in block 1 did not receive the second of two fertilizer applications. These four plots in block 1 that received only the first fertilizer application are identified as the T -IN treatment and were excluded from statistical analyses. Mean values for this treatment are included in our results with no estimates of variance.

The precommerclal thinning treatment, applied in October 1980, favored retention of the largest trees with good form

Pu blished by NRC Research Press

Devine and Harrington 621

Table 1. Thinning and fertilization treatments.

Treatment Year-O fertilization, Year -13 ferti I ization symbol Thinning elemental rate (kg.ha-1) elemental rate (kg.ha-1)

T-O Thinned T-2N Thinned 300 N 300 N T-2N Thinned 300 urea-N 300 N T-2NP Thinned 300 N, 100 P, 129 Ca 300 N, 100 P T-2NP Thinned 300 N, 100 P, 100 K. 41 S, 129 Ca 300 N, 100 P UT-O None UT-2NP None 300 N, 100 P, 100 K, 41 S, 129 Ca 300 N, 100 P T-1Na Thinned Various

Note: Treatments given the same treatment symbol did not differ in preliminary analysis and were combined for the purpose of treatment comparisons. Unless otherwise specified, N was applied as am­monium nitrate. and K and S were applied as potassium sulfate. At year-O fertilization (March 1981), Ca and P were applied as monodicalcium phosphate, and in year 13 (June 1993), P was applied as triple­superphosphate.

"This group was composed of one replication each of the four thinned and fertilized treatments (T -2N and T-2NP), but these plots did not receive the year-13 fertilization. Because the four plots all occurred on one block, this treatment group was excluded from statistical analyses.

and produced a residual tree spacing of approximately 3.5 m x 3.5 m (807 trees·ha-1). Fertilizers listed in Table I were applied to plots by hand from 26 March through 3 April 1981 (i.e., prior to year-1 growing season) and on 21-29 June 1993 (study year 13); treatment areas received two per­pendicular passes to ensure uniformity of application. In the 1981 fertilization, monodicalcium phosphate was selected as the P source because it would not increase soil acidity; this fertilizer was not available for the 1993 fertilization, and tri­ple superphosphate was applied instead.

Data collection In March 1981, 20 redcedar crop trees were selected and

tagged in each plot (equivalent to 250 trees·ha-1) by select­ing trees that were largest in diameter and free from major defects. These crop trees are the subject of this study and, unless otherwise specified, are the only trees referred to herein. Trees were measured for diameter at breast height (DBH; measured to nearest 0.1 cm at a height of 1.3 m) and total height (nearest 0.1 m) in March 1981 (i.e., year 0) and subsequently remeasured after the growing season of study years 1 (i.e., 1981), 2, 3, 5, 10, 12, 16, 20, and 25. Height was not measured in year 12. Height to live-crown base (HLC), defined as the lowest point on the bole where live limbs occurred in three quadrants, was measured in years 16, 20, and 25. In year 25 crown radius was measured in two directions, 1800 apart, at the widest point of the crown. Because only three of the 560 trees died during the 25 year study, we did not analyze mortality.

Following the year-17 and year-25 growing seasons, stem taper was assessed by measuring stem diameter of randomly selected trees (n = 77 at year 17; n = 97 at year 25) at heights of 0.3, 1.3, 2.3, and 3.3 m using a ladder. These measurements were made on three blocks in a subset of treatments: UT-O, T-O, UT-2NP, and T-2NP. At the time of the year-25 taper measurements, diameters of two branches per tree were measured, at the proximal end, on the same trees measured for taper. These branches were selected by first determining the azimuth of greatest crown diameter and then bisecting the stem vertically using an azimuth off­set 90° from that of greatest crown diameter. On each of the

bisected halves, the largest-diameter branch was selected within 2.3 m of stem base. Branch status (Jiving or dead) was recorded.

Samples of year-10 (i.e., 1990) foliar growth were col­lected 3-4 April 1991 from 10 dominant or codominant trees per plot. These samples were collected from the upper por­tion of the crown that was exposed to sunlight. Samples were composited by plot and analyzed for total N, P, K, S, Ca, and Mg. Nitrogen analysis followed the Kjeldahl method (Bremner 1965), and other elements were determined by in­ductively coupled plasma spectroscopy (Jones 1977).

Following year 25 (May 2006), mineral soil in the 0 to 20 cm depth interval was sampled using a steel soil probe at eight systematical1y arranged locations per plot. At each location, five subsamples were col1ected, and then al1 sam­ples were composited by plot. Samples were analyzed for pH (1: 1 soil:water, v/v), total C and N (Fisons NA 1500 Elemental Analyzer, Fisons Instruments, Manchester, UK), inorganic N (NIL. and N03; Sims et al. 1995), P (Bray-1 ex­traction), S04-S (CaP04 extraction), K, and Ca (both via in­ductively coupled plasma).

Data analysis Periodic annual increment (PAl; i.e., the average annual

growth rate between measurements) of total tree height and individual-tree BA (i.e., cross-sectional area of the stem at breast height) were analyzed by repeated-measures analysis of covariance (ANCOV A) models using the Mixed proce­dure (Zar 1999; SAS Institute Inc. 2005). The covariate in each model was tree height or BA prior to the year-1 grow­ing season (i.e., year 0). Year-25 height and BA also were analyzed with ANCOVA, using year-O values as covariates. Year-25 height/diameter ratio (HD), year-IO foliar nutrient concentration, and year-25 soil chemical properties were an­alyzed using analysis of variance (ANOV A) models. In all models, block was a random effect and treatment was a fixed effect with seven levels. In the repeated-measures model of height PAl, time (i.e., measurement period) had eight levels, while in the repeated-measures model of BA PAl, time had nine levels. Post-ANOVA mean separations were made using preplanned contrasts, which are shown in

Published by NRC Research Press

622 Can. J. For. Res. Vol. 39, 2009

Table 2. P values indicating contrast significance for individual-tree basal area periodic annual increment of redcedar.

Study year(s)

Contrast 1 2 3 4-5 6-10 11-12 13-16 17-20 21-25 Not fertilized: unthinned vs. thinned (T-O vs. UT-O) 0.82 0.46 0.02 0.06 0.04 0.03 0.04 0.02 0.04 Fertilized: unthinned vs. thinned (T-2NP vs. UT-2NP) 0.81 0.64 0.07 0.06 0.14 0.07 <0.01 <0.01 <0.01 Unthinned: not fert. vs. fert. (UT-O vs. UT-2NP) 0.33 <0.01 <0.01 <0.01 <0.01 0.29 <0.01 <0.01 <0.01 Thinned: not fcrt. vs. N (T-O vs. T-2N) 0.56 <0.01 <0.01 <0.01 <0.01 0.79 <0.01 <0.01 <0.01 Thinned: N vs. N+P (T-2N vs. T-2NP) 0.36 0.18 0.10 0.30 0.65 0.36 <0.01 <0.01 <0.01

Note: Significant P values are shown in bold.

Table 3. P values indicating contrast si&1J1ificimce for height growth periodic annual increment of redcedar.

Study year(s)

Contrast 1 2 3 4-5 6-10 11-16 17-20 21-25 Not fertilized: unthinned vs. thinned (T-O vs. UT-O) 0.42 0.47 0.16 0.20 0.49 0.46 0.37 0.82 Fertilized: unthinned vs. thinned (T-2NP vs. UT-2NP) 0.57 0.06 0.77 0.67 0.98 0.72 0.66 0.45 Unthinncd: not fCl1. vs. fert. (UT-O vs. UT-2NP) 0.08 <0.01 <0.01 <0.01 0.25 <0.01 <0.01 0.65 Thinned: not fert. vs. N (T-O vs. T-2N) 0.73 <0.01 <0.01 0.10 0.36 0.07 <0.01 0.83 Thinned: N vs. N+P (T-2N vs. T-2NP) 0.03 0.21 0.96 0.45 0.82 0.16 0.83 0.30

Note: Significant P values are shown in bold.

Table 4. P values indicating contrast significance for individual-tree basal area (BA), height, and height/diameter ratio (RD) of redcedar at study year 25.

Variable

Contrast BA Height HD Not fertilized: unthinned vs. thinned (T-O vs. UT-O) Fertilized: unthinned vs. thinned (T-2NP vs. UT-2NP) Unthinned: not fert. vs. fert. (UT-O vs. UT-2NP) Thinned: not fert. vs. N (T-O vs. T-2N)

0.07 <0.01 <0.01 <0.01

0.08 0.83

<0.01 <0.01

<0.01 <0.01 <0.01

0.09 0.20 Thinned: N vs. N+P (T-2N vs. T-2NP)

Note: Significant P values are shown in bold.

Tables 2, 3, and 4. These contrasts test both T-2N treatments and both T-2NP treatments simultaneously, as preliminary analysis showed no differences between the two T-2N treat­ments or the two T-2NP treatments. Where ANCOVA mod­els indicated significant treatment x time interactions, these were interpreted by examining contrast significance in com­bination with plotted treatment means. Because stem taper and branch diameter data were collected on a subset of treat­ments, these models consisted of a two-way ANOV A, with thinning (presence vs. absence) and fertilization (none vs. N+P) as fixed effects. Analysis of variance also was used to compare 25 year BA growth of crop trees among the four year-O BA quartiles; these BA quartile assignments were made for the crop trees within each plot. Multiple compari­sons among crop tree quartiles were made using the Bonferroni test. For each ANOVA model assumptions in­cluding normality of residuals, treatment homo cedasticity, and lack of covariate x treatment imeraclion were mel. A minimum confidence level of 95% was used in all analyses.

Results

Individual-tree basal area Basal area PAl was signit1cantly affected by treatment

and time (F = 90.9, P < 0.01; F = 70.1, P < 0.01, respec­tively); additionally, there was a significant treatment x

0.04 0.11

time interaction (F = 4.9, P < 0.01). Year-25 BA was signif­icantly influenced by treatment (F = 14.8, P < 0.01).

Without fertilization, BA PAl was significantly greater with thinning than without in year 3 and from . year 6 through year 25 (Table 2, Fig. 1). Year-25 BA in unfertil­ized treatments did not differ significantly with thinning (Table 4). Under N+P fertilization, BA PAl was greater with thinning than without from years 13 through 25. At year 25, BA was significantly greater in the T -2NP treat­ment (810 ± 32 cm2 (mean ± SE» than in the UT-2NP treat­ment (632 ± 38 cm2).

In the absence of thinning, 2NP fertilization was associ­ated with an increase in BA PAl in years 2-10 and 13-25 compared with no fertilization. Year-25 BA in un thinned treatments also was greater with 2NP fertilization compared with no fertilization (632 ± 38 vs. 267 ± 38 cm2). Among thinned treatments, BA PAl was greater under N fertilization than without fertilization in years 2 through 10 and 13 through 25. Similarly, year-25 BA among thinned treatments was greater under N fertilization than where unfertilized (740 ± 32 vs. 510 ± 38 cm2). Also among thinned treat­ments, BA PAl was greater under N+P fertilization than under N-only fertilization during years 13 through 25. At year 25, BA was greater in the former treatment than in the latter treatment (810 ± 32 vs. 740 ± 32 cm2). BA PAl in the T-1N treatment was similar to that in the T-2N treatment

Published by NRC Research Press

Devine and Harrington

Fig. 1. Basal area (BA) and BA periodic annual increment (PAl), with standard error bars, for redcedar under various combinations of fertilization and thinning (Table 1). Treatment contrasts appear in Tables 2 and 4. T and F indicate timing of thinning and fertilization, respectively.

800 "1'- ToO ___ T-2N

-+- HNP -6- uroO

600 -0- UT-2NP - T-1N C'II

E () -<C 400 III

200

0

50 -l' :;;40 ~

C'I

E30 ~

f 20 <C III

10

0 0 5 10 15 20 25

Study year

following the , first fertilization but followed the same trend as that in the T -0 treatment after the second fertilization.

In each of the fertilized treatments, 25 year BA growth was greatest for trees that were in the fourth crop tree BA quartile at the time study initiation (Fig. 2). Basal area growth among the first through third crop tree quartiles did not differ. In the unfertilized treatments, there were no sig­nificant differences in 25 year BA growth among crop tree BA quartiles.

Height Height P AI was significantly influenced by treatment and

time (F = 33.3, P < 0.01; F = 38.7, P < 0.01, respectively), and there was a significant treatment x time interaction (F = 1.6, P = 0.04). Year-25 height also was significantly af­fected by treatment (F = 18.4, P < 0.01).

Whether with or without fertilization, height PAl (years 1 through 25) and total height at year 25 were not signifi­cantly affected by thinning (Table 3, Fig. 3). Without thin­ning, height ' PAl was significantly greater under N+P fertilization than under no fertilization during years 2 through 5 and during years 11 through 20. Also, year-25 height was greater under N+P fertilization (15.9 ± 0.5 m) than in the unfertilized treatment (11.9 ± 0.5 m). With thin-

623

ning, height P AI was greater in the N fertilization treatments than in the unfertilized treatments, in years 2-3 and 17-20. Year-25 height was greater in the thinned N fertilization treatments (15.2 ± 0.4 m) than in the thinned unfertilized treatment (13.2 ± 0.5 m). Within the thinned treatments, there was no difference in year-25 height between the N fer­tilization treatment and the N+P fertilization treatment; the only height P AI difference between these treatments oc­curred in year 1. Height PAl in the T-IN treatment followed a trend similar to that in the T-2N treatment through year 10 but declined more rapidly than in the other fertilized treat­ments after year 10.

Height/diameter ratio There was a significant treatment effect on year:.. 25 lID

(F = 9.4, P < 0.01). Under N+P fertilization, year-25 HD was significantly lower with thinning than without (56.9 ± 2.2 vs. 67.8 ± 2.2) (Table 4, Fig. 4). Without fertilization, year-25 HD also was significantly lower with thinning (50.7 ± 2.0 vs. 61.8 ± 2.2). In the absence of thinning, N+P fertilization was associated with significantly lower lID, compared with the unfertilized treatment (61.8 ± 2.2 vs. 67.8 ± 2.2). Among thinned treatments, year-25 HD did not differ significantly by fertilization treatment.

Stem taper Year-I7 stem taper, calculated as the ratio of stem diame­

ter at heights of 0.3, 1.3, or 2.3 m relative to stem diameter at 3.3 m, was greater with thinning at stem heights of 0.3 and 1.3 m (Table 5). Also in year 17, stem taper at a height of 2.3 m diminished with fertilization. At year 25, stem taper was significantly greater with thinning at stem heights of 2.3 and 1.3 m.

Crown Mean crown radius at year 25 was not significantly af­

fected by thinning (F = 0.7, P = 0.41), but crown radius was significantly greater (F = 16.6, P < 0.01) in the N+P treat­ment (3.4 ± 0.1 m) than in the unfertilized treatment (2.9 ± 0.1 m). Height to live-crown base was affected by a signifi­cant treatment x year interaction (F = 2.0, P = 0.04): treat­ment affected HLC in year 25 but not in years 16 or 20 (Fig. 5). In year 25, lll.C in the UT-2NP treatment was sig­nificantly greater than that in the T-2NP treatment (F = 17.9, P < 0.01) and the UT-O treatment (F = 20.7, P < 0.01). Year-25 lll.C also was significantly greater in the T-2N treatment than in the T-O treatment (F = 5.0, P = 0.03).

The mean diameter of the two largest branches on the lower stem was greater (F = 9.1, P < 0.01) with thinning than without (3.5 ± 0.2 vs. 3.1 ± 0.2 cm). Branch diameter also was greater (F = 9.1, P < 0.01) for trees in the N+P treatments (3.5 ± 0.2 cm) than for unfertilized trees (3.1 ± 0.2 cm). Among Jive branches only, neither thinning (F = 3.7, P = 0.06) nor fertilization (F = 1.2, P = 0.28) atIected branch diameter, but there was a nonsignificant trend toward larger branch size with thinning (3.6 ± 0.2 vs. 3.3 ± 0.2 cm). Overall, 25 % of the branches measured were dead in the thinned treatment, while 48 % were dead in the unthinned treatment. Forty-three percent of measured branches were dead in the fertilized treatment, and 29% of measured branches were dead in the unfertilized treatment.

Published by NRC Research Press

624 Can. J. For. Res. Vol. 39, 2009

Fig. 2. Increase in individual-tree basal area (BA) after 25 years [or redcedar crop trees classified by pretreatment crop tree BA quartile. Treatments consisted of various combinations of fertilization and thinning (Table 1). Same letters denote no significant difference (P ~ 0.05) in growth among quartiJes within a treatment; ns denotes no significant differences within a treatment.

-C'I

E (.) -CI) en m CI) a-(.)

.5 <C ED a-m ~ I

II) C\I

1200

1000

800

600

400

200

0

c::=J 1 st quartile 2nd quartile

c::=J 3rd quartile _ 4th quartile

ns

T-O T-2N

Foliar and soil chemical properties

a

With the exception of Ca, all year-lO foliar nutrient con­centrations were unaffected by treatment and are presented as study means with one standard deviation. Foliar N con­centration was 7.62 ± 0.59 g·kg-1, P was 1.15 ± 0.12 g·kg- I ,

K was 3.29 ± 0.35 g·kg-1, S was 0.8,6 ± 0.05 g·kg-I, and Mg was 1.13 ± 0.09 g·kg-1• Foliar Ca concentration was greater (F = 4.6, P = 0.04) in the UT-O treatment (8.11 ± 0.25 g·kg-l )

than in the T-O treatment (7.36 ± 0.25 g·kg-1). Foliar Ca also was higher (F = 9.0, P < 0.01) in the UT-O treatment (8.11 ± 0.25 g·kg-I ) than in the UT-2NP treatment (7.07 ± 0.25 g·kg-1).

The majority of year-25 soil chemical properties were not significantly affected by treatment (P 2 0.05) and are pre­sented as overall mean concentrations with one standard de­viation. The mean value for soil pH was 4.1 ± 0.1, soil C was 63.4 ± 7.6 g.kg-I , CIN ratio was 29.3 ± 2.2, total inor­ganic N was 28.0 ± ' 7.0 mg·kg-1, Bray-l P was 2.2 ± 2.3 mg·kg-1, sulfate-S was 9.8 ± 7.2 mg·kg-1, K was 30.8 ± 4.4 mg·kg-1, and Ca was 228.1 ± 59.2 mg·kg-l . Soil total N concentration averaged 2.2 ± 0.3 g·kg-1, although N was sig­nificantly greater (F = 4.9, P = 0.04) in the T-O treatment (2.3 ± 0.1 g·kg-1) than in the UT-O treatment (2.0 ± 0.1 g·kg-1).

Discussion

Twenty-five years after initial treatment, both precommer­cial thinning and fertilization significantly increased growth of the 250 largest western redcedar per hectare in a naturally established second-growth stand. Total height of redcedar was greatest in fertilized treatments, regardless of which fer­tilizer was applied or whether fertilization was accompanied by thinning. Individual-tree BA was greatest in the thinned treatment that also was fertilized with N+P. Comparing the

b

b

ab

a

T-2NP

Treatment

a

b

b

ns b

UT-O UT-2NP

benefit of fertilization with that of precommercial thinning, fertilization without thinning increased total height by 34%, while thinning without fertilization produced a nonsignificant 11 % height increase. The combination of these treatments in­creased height by 31 % compared with the unthinned, unfer­tilized treatment. Fertilization without thinning increased BA by 137%; thinning alone increased BA by 91%. The combi­nation of these treatments increased BA by 203%.

The BA growth response to thinning differed between un­fertilized and fertilized treatments. Without fertilization, the magnitude of the thinning response remained very consistent during years 3-25. With fertilization, the thinning response was greatest in years 13-25 (Table 2). The fact that thinning benefited fertilized trees only in the second half of the study suggests that competition among these trees was greater at that time. In the same period, the smaller trees of the unfer­tilized treatments apparently had not reached the same level of intertree competition. An interaction between thinning and fertilization also occurs among Douglas-fir (Pseudot­suga menziesii (Mirb.) Franco): the individual-tree volume growth response to fertilization is greater in thinned stands than in unthinned stands because fewer trees are competing for the added nutrients (Miller and Fight 1979; Peterson and Hazard 1990).

Although thinning had no statistical1y significant effect on height growth within the fertilized or unfertilized treatments, there was a trend (P = 0.08) toward greater 25 year post­treatment height in the T -0 treatment than in the UT -0 treat­ment. Increased height growth was associated with lower stand density for unfertilized redcedar planted at 500 to 2500 trees/ha on low-fertility sites, a trend attributed to competition for soil nutrients among trees (Negrave et al. 2007). On a moderately poor site, Douglas-fir planted at comparable densities also showed a negative stand density -height relationship (Curtis and Reukema 1970). In our study,

Published by NRC Research Press

Devine and Harrington

Fig. 3. Height and height periodic annual increment (P AI), with standard en-or bars, for redcedar under various combinations of fer­tilization and thinning (Table 1). Treatment contrasts appear in Ta­bles 3 and 4. T and F indicate timing of thinning and fertilization, respectively.

16 -.-T-o ___ T·2N

-.- T·2NP 14 -/:r UT-o

-0- UT·2NP T·1N -E 12 --s::.

OJ 10 'a; :::z:::

8

6

T -..- 0.7 F I ... as CD 0.6 ~

E - 0.5

~ - 0.4 s::. OJ 'a; :::z::: 0.3

0.2

0 5 10 15 20 25

Study year

Fig. 4. Height/diameter ratio (HD), with standard elTor bars, for redcedar under various combinations of fertilization and thinning (Tab1e 1). Treatment contrasts appear in Tab1e 4. T and F indicate timing of thinning and fertilization, respectively. I

75~-----------------------------------, -.- T-o ___ T·2N

-+- T-2NP -I:r UT-o -+ -()- ~1."p _-----{

--- ==~t=t--1 70

65

c :::z::: 60

55

50 F

o 5 10 15 20 25

Study year

625

we also attribute the trend toward reduced height growth in the UT-O treatment, relative to the T-O treatment, to greater nutrient competition in the unthinned treatment where den­sity was much higher. This conclusion is supported by the fact that in our fertilized treatments, where nutrient avail­ability was presumably greater, thinning had no effect on height growth (i.e., UT-2NP vs. T-2NP). Also, during the tirst 5 years of this study, foliar nutrient content analyses in­dicated that nutrients added via fertilization were apparently depleted more rapidly in the unthinned treatment than in the thinned treatment (Harrington and Wierman 1990).

We observed no evidence of thinning shock, the tempo­rary reduction in height growth rate after thinning, which has been reported for Douglas-fir on similarly poor sites and for many other species (Miller and Reukema 1977; Harrington and Reukema 1983). The lack of thinning shock may be due to the species or to the relatively young stand age (DeBell et a1. 2002).

Foliar N concentration in year 10 was lower than most of the values reported previously for redcedar in the region (Radwan and Harrington 1986; Weetman et a1. 1988, 1989) and was slightly lower than that reported at year 5 in the same study (Harrington and Wierman 1990). The fact that foliar nutrient concentrations did not differ among fertiliza­tion treatments at year 10 was not surprising given that fo­liar nutrient concentrations among treatments were already converging at year 5 (Harrington and Wierman 1990), and periodic growth response to fertilization had also diminished substantially by year 10. According to regional guidelines, year-lO foliar N concentration was within a range indicating severe deficiency, while foliar P concentration indicated moderate deficiency (Carter 1992). On sites such as the one where this study was located, low N mineralization rates and the inherently low nutrient availability of the young gla­cially derived soil is exacerbated by root competition from vegetation such as salal, making nutrients the most growth­limiting factor for young trees including redcedar (Messier 1993; Prescott 1996).

In both thinned and unthinned treatments, the positive ef­fect of the year-O N fertilization on BA PAl peaked at year 3 and remained significant through year 10 posttreatment. After the year-13 fertilization, the BA PAl response peaked during the period 3 to 6 years after treatment and remained significant through year 25. Based on redcedar site index curves, the projected age-50 height advantage associated with N+P fertilization in our unthinned treatments is 5.0 m (Kurucz 1978). The duration and magnitude of our fertiliza­tion responses are similar to or greater than responses re­ported previously for fertilized redcedar on comparable sites. On Vancouver Island, British Columbia, growth of a young redcedar plantation was increased by N but not by P over a 4 year period (Blevins et a1. 2006), while another plantation, fertilized at year 5, had a 10 year growth re­sponse to fertilization (Negrave et a1. 2007). In a young, mixed-species stand on a similar site, fertilization with N+P increased redcedar height growth during the first 3 years after treatment, but after 13 years, no effect was evident (Weetman et a1. 1989; Bennett et a1. 2003). While the fertil­ization responses in the our study were dramatic, they were of finite duration and did not produce sustained increases in periodic growth rate. This trend was especial1y apparent in

Published by NRC Research Press

626 Can. J. For. Res. Vol. 39, 2009

Table 5. Stem diameter at three heights (mean ± SE, expressed as a percentage of stem diameter at 3.3 m) and analysis of variance (ANOV A) results (P >F) for the effects of thinning and fertilization (N+P) treatments applied to redcedar in a 25 year study.

Treatment

Study Stem year height (m) UT-O UT-2NP T-O 17 2.3 111.7± 1.2 10B.6±1.2 1 14.0±1.2

1.3 130.0±2.5 130.B±2.6 136.0±2.7 0.3 167.7±3.7 161.9±3.6 177.6±3.7

25 2.3 lOB.9±1.2 lOB.3±1.1 112.0±1.2 1.3 126.5±2.9 126.6±2.7 130.9±2.B 0.3 16B.l±3.6 169.4±3.2 174.6±3.5

Note: Significant P values are shown in bold.

Fig. 5. Height to live-crown base (HLC), with standard error bars, for redcedar under various combinations of fertilization and thin­ning (Table 1).

E -

8~------------------------------____ ~

6

........ T-O ___ T-2N

.... T-2NP -A- UT-O -0- UT-2NP -- T-1N

o 4 ..I :::E:

2

O+---------------~--------------~~--J 15 20 25

Study year

the BA and height PAl responses of the T-1N treatment group (Figs. 1, 3). In contrast, sustained fertilization re­sponses to N+P have been reported for western hemlock on sites similar to ours (Bennett et aJ. 2003; Blevins et al. 2006). On those sites, nutrient deficiency had severely lim­ited productivity of hemlock, possibly because of the effect of P deficiency on tree physiology or on nutrient cyc1ing. The fact that redcedar on the same sites did not show a sus­tained response to fertilization was attributed to differences in the species' ability to compete for soil P (Blevins et al. 2006).

Two P applications resulted in a BA that was 9% greater after 25 years compared with the BA obtained with N fertil­ization alone; the benefit of P fertilization on BA PAl was greater after the year-13 application than after the initial ap­plication. The response to P fertilization was not surprising because mean foliar P concentration in year 10 was lower than nearly all of the values reported previously for redcedar stands in the coastal Pacific Northwest (Radwan and Harrington 1986; Weetman et a1. 1988). The low foliar P concentrations may have been due, in part, to the low soil pH (4.1), as soil P availability generally declines in relation to pH at pH values below 5.5 (Tisdale and Nelson ] 975).

ANOV A results

T-2NP Thinning Fenilization Thin. x fert.

111.l±1.3 0.06 6.62 0.92 139.5±2.8 <6.61 0.43 0.61 185.0±3.7 <6.()1 0.B3 0.08 109.6±1.2 6.63 0.14 0.38 133.0±2.B <6.61 0.55 0.63 172.3±3.5 0.13 0.86 0.57

Foliar P concentration for redcedar was positively conelated with pH in a regional analysis where pH ranged from 4.0 to 5.2 (Radwan and Harrington 1986). Studies conducted on similar sites have suggested that long-term improvements in site productivity may be achieved by overcoming P limita­tions (Bennett et al. 2003, 2004; Blevins et al. 2006); how­ever, in these snldies responses to P were evident in western hemlock and in herbaceous species rather than in redcedar. After 25 years, we did not observe a significant effect of N+P fertilization on tree height, compared with N-only fer­tilization, and we conclude that redcedar producti vity at our site was primarily limited by N .

In the fertilized treatments, trees that were largest at the beginning of the study had significantly greater 25 year BA growth than smaller trees (Fig. 2). In unfertilized treatments, this pattern was not significant. Thus, the increased nutrient availability apparently contributed to differentiation in BA growth among tree size classes. The magnitude of this trend was greatest in the unthinned, fertilized treatment, emphasiz­ing the importance of initial tree size at higher stand den­sities, where competition was presumably greater.

Stem taper in the lowest 3.3 m of the bole was greater in the thinned treatment than in the unthinned treatment. While stem taper is a function of many factors, including genetics, crown class, site quality, and management practices (Larson 1963), the high degree of taper that is typical for redcedar is attributed to the relative longevity of its lower branches (Oliver et a1. 1988). Unless these branches are controlled through stand density management, many persist and con­tinue to produce photosynthates that contribute to the thick­ening of the lower stem (Oliver et al. 1988). Thus, it is not surprising that our thinning treatment increased stem taper, gi ven the association among wider spacing, increased crown length (Fig. 5) (Reukema and Smith 1987; DeBell and Gartner 1997), and increased taper (Larson ] 963). Height/ diameter ratio was significantly reduced by thinning, owing to the increase in stem diameter associated with thinning. Height/diameter ratios in the thinned treatments were com­parable to those of plantation redcedar at spacings of 2.7 to 4.6 m (Reukema and Smith 1987). The fact that lID values were not excessively high, even in the unthinned treatments (i.e., <70; cf. Reukema and Smith 1987), was likely due to poor site quality and the fact that the meas­urement trees were in codominant or dominant crown posi­tions (Larson 1963).

Crown attributes 25 years posttreatment were intluenced

Published by NRC Research Press

Devine and Harrington

by thinning and fertilization treatments. Thinning was asso­ciated with decreased HLC (Fig. 5) and somewhat larger­diameter limbs within 2.3 m of the ground. These trends were likely due to greater amounts of· within- and be­tween-crown light penetration in the thinned treatments. In 35-year-old redcedar plantations, spacing (1.8 to 4.6 m) was positively related to diameter of lower branches, although there was only a 1.0 cm difference between mean branch diameter in the narrowest and widest spacings (De­Be]] and Gartner 1997). In the same study, spacing did not affect the number of limbs on the butt log. We found an association between fertillzation and increased HLC. This may have been a result of increased leaf area reSUlting from fertilization (Binkley and Reid 1984)~ greater leaf area may have increased within-crown shading, leading to mortality of lower limbs. Overall, HLC was greatest in the unthinned, fertilized treatment, where trees were tallest and crown spacing was closest. The increases in limb diameter and longevity associated with thinning in this study were small in magnitude, and by the time of harvest, they may not affect log value. Because lower limbs of redcedar per­sist at wide spacings or where trees are overtopped, the best approach for reducing these limbs in pure even-aged stands is a narrow initial spacing followed by thinning (Oliver et al. 1988).

Redcedar is usually found in mixed-species stands rather than in pure stands, and in mixed stands N+P fertilization is likely to benefit the growth of other species at least as much as that of redcedar (Bennett et al. 2003; Negrave et al. 2007). If fertilization responses of other species are superior to those of redcedar, this may have the adverse effect of pro­moting their eventual dominance over redcedar. Therefore, precommercial thinning may be an important tool for influ­encing the desired species mixture. In young mixed-species stands, where other conifers often have similar or greater height growth rates, it is important that redcedar crop trees not become overtopped. Overtopped redcedar has a reduced growth rate and does not self-prune lower limbs owing to its shade tolerance. When overtopped redcedar is released, it does not recover apical control and often develops large spreading limbs that reduce log quality (Oliver (It al. 1988).

Our treatment responses and foliar analysis indicate that growth of redcedar crop trees was substantially limited by N availability and, to a lesser extent, by P availability. Large increases in tree growth followed both fertilizer applications, with the growth response to the second application (stand age of 28-33 years) at least as great as the response to the first (stand age of 15-20 years). This suggests that additional applications also would produce significant growth responses on this nutrient-poor site. However, after the second fertilizer application, growth of fertilized crop trees also benefited from the stand density reduction associated with thinning. On sites characterized by low nutrient availability due to low mineralization rates and nutrient competition from salal (Prescott 1996), we would expect fertilization responses comparable to those reported here. Although it is uncertain what effects our treatments would have on a more productive site, a recent Vancouver Island study showed redcedar to be as responsive to N+P fertilization on a medium-fertility hemlock-amabi1is fir site as on "a low-fertility cedar­hemlock site (Negrave et al. 2007). However, the thinning

627

response and the thinning-fertilization interactions that we observed are likely to differ on more productive sites.

Acknowledgements We thank Rayonier Inc. for study installation and Green

Crow Company for ongoing support of the study after the land was sold. Special thanks to Charles Wiennan (formerly of Rayonier) and Harry Bell (Green Crow) for financial sup­port and to the many people who participated in implement­ing· the treatments and collecting data. We thank Doug Mainwaring, Peter Gould, and Mel Scott for manuscript re­views.

References Bennett, J.N., Blevins, L.L., Barker, J.E., Blevins, D.P., and

Prescott. C.E. 2003. Increases in tree growth and nutrient supply still apparent 10 to 13 years following fertilization and vegeta­tion control of salal-dominated cedar-hemlock stands on Van­couver Island. Can. J. For. Res. 33: 1516-1524. doi: 10.1 1391 x03-069.

Bennett, J.N., Lapthorne, B.M., Blevins, L.L., and Prescott, C.E. 2004. Response of Gaultheria shallon and Epilobium angustifo­Hum to large additions of nitrogen and phosphorus fertilizer. Can. J. For. Res. 34: 502-506. doi:l0.1139/x03-219.

Binkley, D., and Reid, P. 1984. Long-term responses of stem growth and leaf area to thinning and fertilization in a Douglas­fir plantation. Can. J. For. Res. 14: 656-660. doi: 10.1 139/x84-lI8.

Blevins, L.L., and Prescott, c.E. 2002. Salal Cedar Hemlock Inte­grated Research Program update #2: silvicultural practices for regeneration of cedar-hemlock sites in Coastal British Colum­bia. Faculty of Forestry, The University of British Columbia, Vancouver, B.C.

Blevins, L.L., Prescott, C.E., and Van Niejenhuis, A. 2006. The roles of nitrogen and phosphorus in jncreasing productivity of western hemlock and western redcedar plantations on northern Vancouver Island. For. Ecol. Manage. 234: 116-122. doi:lO. 1016/j.foreco.2006.06.029.

Bremner, I.M. 1965. Total nitrogen. In Methods of soil analysis (chemical and microbiological properties). Edited by C.A. Black, D.D. Evans, L.E. Ensminger, 1.L. White, and F.E. Clark. Academic Press, New York. pp. ] 149-1178.

Carter, R 1992. Diagnosis and interpretation of forest stand nutri­ent status. In Forest fertilization: sustaining and improving nutri­tion and growth of western forests. Edited by H.N. Chappel, G.F. Weetman, and RE. Miller. Institute of Forest Resources No. 73. University of Washington, Seattle, Wash. pp. 90-97.

Curtis, R.O., and Reukema, D.L. 1970. Crown development and site estimates in a Douglas-fir plantation spacing test. For: Sci. 16: 287-301.

Curtis, RO., DeBell, D.S., Miller, RE., Newton, M., St. Clair, J .B., and Stein, W.l. 2007. Silvicultural research and the evolution of forest practices in the Douglas-fir Region. USDA For. Servo Gen. Tech. Rep. PNW -696.

DeBell, J.D., and Gartner, B.L. 1997. Stem characteristics on the lower log of 35-year-old western redcedar grown at several spa­cings. West. J. Appl. For. 12: 9-14.

DeBell, D.S., Harrington, C.A., and Shumway, J. 2002. Thinning shock and response to fertilizer less than expected in young Douglas-fir stand at Wind River Experimental Forest. USDA For. Servo Res. Pap. PNW-547.

Franklin, J.F., and Dyrness, C.T. 1988. Natural vegetation of Ore-

Published by NRC Research Press

628

gon and Washington. Oregon State University Press, Corvallis, Ore.

Freitag, C.M., and Morrell, JJ. 2001. Durability of a changing wes­tern redcedar resource. Wood Fiber Sci. 33: 69-75.

Gonzalez, J.S. 2004. Growth, properties and uses of western red ce­dar (Thllja pUcata Donn ex D. Don). 2nd ed. Forintek Canada Corp. Spec. Publ. SP-37R.

Harrington, C.A., and Reukema, D.L. 1983. Tnitial shock and long­term stand development following thinning in a Douglas-fir plantation. For. Sci. 29: 33-46.

Harrington, c.A., and Wierman, C.A. 1985. Response of a poor­site western redcedar stand to precommercial thinning and ferti­lization. USDA For. Servo Res. Pap. PNW-339.

Harrington, C.A., and Wierman, C.A. 1990. Growth and foliar nu­trient response to fertilization and precommercial thinning .in a coastal western red cedar stand. Can. 1. For. Res. 20: 764-773. doi: 10.1139/x90-101.

Hermann, R.K. 1987. North American tree species in Europe. J. For. 85: 27-32.

Jones, lB., Jr. 1977. Elemental analysis of soil extracts and plant tissue ash by plasma emission spectroscopy. Commun. Soil. Sci. Plant Anal. 8: 349-365. doi:1O.1080/00103627709366727.

Kranabetter, J.M., Banner, A., and Shaw, J. 2003. Growth and nu­trition of three conifer species across site gradients of north coastal British Columbia. Can. J. For. Res. 33: 313-324. doi:10. 1139/x02-188.

Kurucz, J.F. 1978. Preliminary, polymorphic site index curves for western redcedar - Thuja plicata Donn - in coastal British Columbia. Forest Research Note No.3. Macmillan Bloedel Ltd., Vancouver, B.c.

Larson, P.R. 1963. Stem form development of forest trees. For. Sci. Mon. 5: 1-41.

McDonald, M.A., Hawkins, BJ., Prescott, C.E., and Kimmins, J.P. 1994. Growth and foliar nutrition of western red cedar fertilized with sewage sludge, pulp sludge, fish silage, and wood ash on northern Vancouver Island. Can. J. For. Res. 24: 297-301. doi:l0.1139/x94-042.

Messier, C. 1993. Factors limiting early growth of western redce­dar, western hemlock, and Sitka spruce seedlings on ericaceous­dominated clearcut sites in coastal British Columbia. For. Ecol. Manage. 60: 181-206. doi:1O.1016/0378-1l27(93)90080-7.

Miller, RE., and Fight, RD. 1979. Fertilizing Douglas-fir forests. USDA For. Servo Gen. Tech. Rep. PNW-83.

Miller, R.E., and Reukema, D.L. 1977. Urea fertilizer increases growth of 20-year-old thinned Douglas-fir on a poor quality site. USDA For. Servo Res. Note PNW-291.

Minore, D. 1983. Western redcedar - a literature review. USDA For. Servo Gen. Tech. Rep. PNW-150.

Minore, D. 1990. Thuja pUcata Donn ex D. Don - western redce­dar. In Silvics of North America. Edited by RM. Burns and B.H. Honkala. U.S. Dep. Agric. Agric. Handb. 654. pp. 590-600. .

Negrave, RW., Prescott, C.E., and Barker, J.E. 2007. Growth and foliar nutrition of juvenile western hemlock and western redce­dar plantations on low- and medium-productivity sites on north­ern Vancouver Island: response to fertilization and planting density. Can. J. For. Res. 37: 2587-2599. doi:l0.l139/X07-089.

Can. J. For. Res. Vol. 39,2009

Nystrom, M.N., DeBell, D.S., and Oliver, C.D. 1984. Development of young growth western redcedar stands. USDA For. Servo Res. Pap. PNW-324.

Oliver, C.D., Nystrom, M.N., and DeBell, D.S. 1988. Coastal stand silvicultural potential for western red-cedar. In Western red ce­dar: does it have a future? Edited by N.J. Smith. The University of British Columbia, Faculty of Forestry Publications, Vancou­ver, B.C. pp. 39-46.

Peterson, C.E., and Hazard, J.W. 1990. Regional variation in · growth response of coastal Douglas-fir to nitrogen fertilizer in the Pacific N0l1hwcst. For. Sci. 36: 625-640.

Prescott, C.E. 1996. A field guide to regeneration of salal-domi­nated Cedar-Hemlock (CH) sites in the CWHvml. Faculty of Forestry, The University of British Columbia, Vancouver, B.C.

Prescott, C.E., and Blevins, L.L. 2005. Eleven-year growth re­sponse of young conifers to biosolids or nitrogen and phos­phorus fertilizer on northern Vancouver Island. Can. 1. For. Res. 35: 211-214. doi:10.1139/x04-146.

Radwan, M.A., and Harrington, c.A. 1986. Foliar chemical con­centrations, growth, and site productivity relations in western red cedar. Can. J. For. Res. 16: 1069-1075. doi:1O.1139/x86-185.

Reukema, D.L., and Smith, J.H.G. 1987. Development over 25 years of Douglas-fir, western hemlock, and western redcedar planted at various spacings on a very good site in British Co­lumbia. USDA For. Servo Res. Pap. PNW-381.

SAS Institute Inc. 2005. The SAS System for Windows. Version 9.1. SAS Institute Inc., Cary, N.C.

Sims, G.K., Ellsworth, T.R., and Mulvaney, R.L. 1995. Microscale determination of inorganic nitrogen in water and soil extracts. Commun. Soil. Sci. Plant Anal. 26: 303-316. doi:1O.l0801 00103629509369298.

Soil Classification Working Group. 1998. The Canadian system of soil classification. 3rd ed. Agriculture and Agri-Food Canada, Publ. 1646.

Soil Survey Staff. 2008. Official soil series descriptions. USDA­NRCS. [Online.] Available from http://soils.usda.gov/technical/ classificationlosdlindex.html [accessed 9 June 2008].

Tisdale, S.L., and Nelson, W.L. 1975. Soil fertility and fertilizers. 3rd ed. Macmillan, New York.

Weetman, G.F., Radwan, M.A., Kumi, J., and Schnorbus, E. 1988. Nutrition and fertilization of western red cedar. In Western red cedar: does it have a future? Edited by N.J. Smith. The Univer­sity of British Columbia, Faculty of Forestry Publications, Van­couver, B.c. pp. 47-59.

Weetman, G.F., Fournier, R., Barker, J., and Schnorbus-Panozzo, E. 1989. Foliar analysis and response of fertilized chlorotic wes­tern hemlock and western red cedar reproduction on salal-domi­nated cedar-hemlock cutovers on Vancouver Island. Can. J. For. Res. 19: 1512-1520. doi: 10. 11 39/x89-230.

Western Regional Climate Center. 2008. Washington climate sum­maries. [Online.] Available from http://www.wrcc.dri.edu/ summary/c1imsmwa.html [accessed 9 June 2008].

Zar, J.H. 1999. Biostatistical analysis. 4th ed. Prentice Han, Upper Saddle River, N.J.

Published by NRC Research Press

Response to Reviewers' Comments

Manuscript title: Western redcedar response to precommercial thinning and fertilization through 25 years post-treatment

Authors: Devine, Warren D., and Harrington, Constance A.

Date: 12 March 12,2009

Reviewer: Peter Gould (PNW Research Station)

Reviewer Gould suggested that we present the effects of initial tree size on 25-year growth. We did this analysis and added what now appears as Figure 2. Reviewer Gould suggested that we present an analysis of the change in the Weibull distribution over the course of the study. This was done, however it will appear in a second manuscript that is currently in progress and not in this manuscript.

Reviewer: Doug Mainwaring (Oregon State University)

Reviewer Mainwaring suggested that we not use quadratic mean diameter, but use individual tree basal area as a growth metric instead. We followed this suggested and performed the analyses again using basal area. Reviewer Mainwaring also suggested several changes to the Results and Discussion section to clarify the manuscript. These suggestions were followed.

Reviewer: Mel Scott (British Columbia Ministry of Forests)

Reviewer Scott indicated that we elaborate on the phenomenon of increased height in response to thinning, as a possible result of reduced nutrient competition. In response, we have clarified our discussion of this point in the Results and Discussion sections. Scott stated: Lines 426 to 428 state that fertilizer applications will likely benefit growth of other species at least as much as redcedar and you reference two articles including Negrave et al. The report, Negrave et al2007 deals with fertilizer using N+P+K(planting) and N + P. I had a discussion with Rod Negrave last month and he indicated that fertilizing a mix of western hemlock and redcedar, on HA phase sites, with nitrogen alone would probably result in a preferential response by redcedar. I suspect that is because of the erratic response by Hw to fertilization with N alone. In response, we clarified the discussion of this point, with attention to the additional information regarding Negrave's findings.

Below are the comments of the anonymous journal reviewers

Referee #1 Comments General comments The aim of the study was to evaluate the effects of precommercial thinning (PCT) and fertilisation oa growth, stem form, and branchiness of western redcedar in Washington, US. This

is a traditional growth and yield study reporting the properties of an empirical data set. One stand density (3.5 x 3.5-m spacing) was compared with a control (no PCT), in addition to three different fertilisation (no, N, N+P) treatments 25 years after treatment onset. The topic is interesting and important, and it fits in the scope of Can. J. For. Res. The main merit of the manuscript is that it reports results on western redcedar, for tree species with little information on its management compared to other conifer species. Furthermore, the results are based on a long-term experiment. Most of the previous papers on redcedar management report short-term effects of fertilisation and thinning. In general, the results are, however, definitively nothing new. There are tons of studies reporting the effects of stand density, PCT, thinning and fertilisation on tree growth, stem taper, and branch properties; and lots of equations and models are established for predicting their effects in other regions or for other tree species. Including only one stand density after PCT does not enable a deeper analysis on the relationship between stand density and stem growth and properties (along a continuum of different densities). Comment: The introduction is relatively short; it is straightforward and gives some factual background information. The cited references are topical and refer to original publications, but the literature review is not comprehensive. A fatal deficiency is that the forest management practices applied in the US are not described at all (neither in the introduction, nor in the material and methods). The relevance of the PCT and fertilisation treatments applied in the study in relation to those applied in forestry practice remains therefore unclear. Are the treatments based on some instructions for forestry practice? For an international reader, more background information should definitively be given. Response: We have added information to the Introduction (Lines 60-64) describing current management recommendations for the species. We have amended the Methods, Lines 126-128, to indicate that these treatments in this study were developed specifically to reflect the most likely operational management scenarios.

Comment: Only 20 largest trees per plot (equivalent to 250 trees/ha) were selected as sample trees, and the results are solely based on these trees. Thus, the results do not show the treatment effects at stand level (all smaller trees were neglected; there were 5900 trees/ha). It would also be interesting to receive information on the development of the whole stand (not only on the largest trees). Response: This study produced too much data to condense into one manuscript. Thus, in this first year-25 manuscript, we focus on the intensively measured crop trees and on variables related to their stem and crown growth. There is a second manuscript in progress that examines all of the trees on the study plots. This second manuscript analyzes stand dynamics, including per-hectare BA, diameter and height distributions, and ingrowth of cedar as well as other species during the 25-year study.

Comment: The discussion is reporting the well-known facts that spacing and fertilisation have an effect on radial increment, stem form, etc. The authors mainly repeat the main results. No innovative conclusions are brought out for discussion. The discussion could be tremendously improved by connecting the results to theoretical considerations and conceptual framework. Implications for management of redcedar stands could also be derived. Response: We have made changes to improve the scope of the Discussion section. In particular we have revised the last two paragraphs to include more management implications. While we

attempted to broaden the scope, the core of the research was still inherently an applied study, owing to the nature of the treatments - for example, the fertilization treatments did not attempt to bracket the full range of likely fertilization response, and thinning was not applied at multiple intensities to gauge the response-of the species. Rather, the treatments were -based on industry practices.

Comment: The foliar and soil nutrient concentrations were not affected by the fertilisation. However, growth reactions were found. The contradiction should be discussed. Response: We have explained this apparent contradiction now (Line 3537-340).

Comment: To conclude; the study is well designed and documented. The manuscript is logical and well written (but see the comments above & below). The results are logical and clearly presented according to the research questions posed. The conclusions are predominantly grounded in the data and do not go beyond it. In my opinion, the manuscript is valuable confirmation of the existing knowledge which merits its publication.

Specific comments Comment: Throughout the whole manuscript. Avoid value judgements like 'treatment x IMPROVED property y' or 'treatment x had a POSITIVE effect on property y'; better 'treatment x INCREASED property y'. The value judgements may result in strange statements; for example on page 11 it is said that PCT had a positive influence on branch diameter (branches were thicker, which normally is considered detrimental). Response: Phasing was changed throughout. The only cases in which terms "positive" or "negative" were retained were when they were used to describe statistical relationships (e.g., two variables were positively correlated);

Comment: Page 3, L 57-60. Instead of just listing the previous research results focusing on management of redcedar stands, their results should be described. Response: This sentence was deleted because results from all of the studies it cited were mentioned later in the Introduction anyway.

Comment: P 7, L 152. Crown radius was measured in two directions. How were the directions selected? Response: This was clarified (Line 152).

Comment: P 8, L 183-185. The PCT and fertilisation treatments were used as fixed with 7 levels. Alternatively, a factorial design could be used, i.e. PCT with two levels (no, PCT), and fertilisation with three levels (no, N, N+P). The chosen approach should be justified. Response: We did not use a factorial analysis because some of the fertilization treatments were applied only within the thinned treatment (not in the unthinned treatment). Although some fertilization treatments were combined here, originally 5 levels of fertilization were applied to thinned plots and only 2 levels were applied to unthinned plots. Instead of using the factorial treatment approach, we relied on pre-planned contrasts to test specific hypotheses that were of interest.

Comment: P 10, L 222-225. Twenty largest trees out of 5900 trees/ha were selected as sample trees (see page 7). Therefore, it should be absolutely impossible that basal area increase is reported for the PRE-treatment basal area quartiles 1-3 (smaller trees)? Response: The quartiles were assigned only within the crop trees (i.e., largest 250 trees per plot at study initiation). This has been clarified (Lines 195-198,223-226). We now refer to the four quartiles as ~~crop tree BA quartiles" instead of just "BA quartiles."

Comment: P 12-13, L 276-289. The nutrient concentrations in foliage and soil could be shown in a table; easier to read. -Response: The reason that we did not put this values into a table is that nearly all of the analyses shown for foliage and soil are different (e.g., total S for foliar vs. sulfate-S for soil). Therefore, we would need to put each into a separate table with one data column and many rows per table. We did not think that adding two tables was warranted, but if the editors feel that we should, then we will do this.

Comment: P 13, L 298. Remove 'only'. Usually thinning does not increase height increment. Thus, an increase of 11 % is not 'only'. Response: This change has been made.

Comment: P 13, L 306-307 (and elsewhere). 'Type A treatment response' is not an unambiguous term. Response: This term was removed from the manuscript. Now we simple describe it as a "sustained" response or a sustained increase in growth rate.

Comment: P 14, L 313. The Latin name of Douglas-fir is ... Response: This has been added.

Comment: P 14, L 324. Thinning increased tree height (page 13, line 298), but it did not increase height increment (sic.). Response: On line 302, we have clarified that the 11 % height difference was not statistically significant. Also see 316-318 for our explanation of this apparent discrepancy.

Comment: P 14, L 329-330. The results by Curtis and Reukema (1970) were similar to WHAT; (1) results of this study or (2) results by Negrave et al. (2007). (1) and (2) contradict with each other. Why (1) and (2) contradict? Response: We have made clarifications to this paragraph to show that the non-significant trend that we reported was similar to the trends reported earlier by Curtis and Reukema and by Negrave et al. (Lines 316-322).

Comment: P 17, L 391-395. Longer crowns in thinned stands may be related to higher stem taper. However, there are other complementary explanations (mechanical resistance against wind, etc.). The discussion here is too simplistic; shade-tolerant lower branches are not indeed an adequate explanation for stem taper. Response: This paragraph has been revised to provide a clearer description of these relationships (Lines 389-403).

Referee #2 Comments This paper is a 25 year response update on the 5 year response for the same experiment published in CJFR>in 1990; There has been an additional fertilizer application. It should be published basically as is. There have been very few fertilizer trials with western red cedar. This is the oldest. The data is well presented and the literature is up to date. The results presented confirm that the response pattern is "classic": i.e. the temporary response to N in basal area and additional height growth.

Comment: This report does not present the per hectare BA responses as did the 5 year CJFR paper. No attempt is made to forecast what the long term yield increases might be or whether cedar fertilization might be a good investment. Response: There is a second manuscript from this study that is currently in progress. This second manuscript analyzes stand dynamics and variables expanded to a per-hectare basis, including basal area. It examines all trees (i.e., all species) that were on the study plots at year 25. Because the volume of data warranted two manuscripts, we chose to divide the data into 1) tree-Iev~l responses of crop trees, and 2) stand development and per-hectare responses.

Comment: Fertilizer response is often modeled as an increase in Site Index. The 4 mete.r increase in dominant height at age 25 is very attractive. Response: We added the projected age-50 height benefit from fertilization to the discussion (Line 350-352). Because trees were 15-20 years old at the time of treatment, the 4 meter increase actually occurred at age 40-45. The projected age-50 increase is 5.0 meters.

Comment: The match of the site description to the Canadian system, given in the earlier paper was not repeated ... it should be. Response: See line 105-106 for the Canadian equivalent.

Comment: Nor is there any attempt to say how portable the response may be to other sites growing western red cedar. It is possible that that cedar may be as generally responsive as is Douglas fir. A speculation or even an opinion would be helpful. . Response: Speculation regarding these responses on other sites has been added to the last paragraph of the Discussion section.

Comment: Whether P additions cause more N mineralization as suggested by Kranabetter et al is alluded to, but it may be speculative at this stage. Operational aerial fertilization, presently underway in BC, uses both N and P additions. Response: The authors deemed the mention of this P-facilitation unnecessary and removed the sentence from the discussion. We also noted that both N and P are applied operationally (Line 62).

Referee #3 Comments General comments

The manuscript presents data on the long-term effects of fertilization and thinning on growth of second-growth western red cedar. The ability to sustain current rates of cedar harvesting is a concern, but few long-term data indicate how silviculture treatments might affect productivity and supply. Hence, this particular study is timely and a valuable contribution to the literature on cedar silviculture literature. The study is of long duration, stands are older at final measurement (40-45 years) than those reported previously, and there are interesting contrasts in the response

. of cedar to thinning and fertilization. The design seems clear and the analysis and interpretation of the data seem reasonable. The responses raise interesting questions about the ultimate effects of treatments on wood and log quality and value of second-growth cedar. The study is also useful in allowing comparison of treatment responses by cedar to its more-often studied associates, Douglas-fir and western hemlock. In general, the manuscript is well-organized and well-written. Some wording pertaining to the interpretation of effects of fertilization and thinning is confusing. The discussion section might benefit from some additional discussion of operational implications

Specific comments Comment: Line 133- As stated here, the TI-N data weren't subjected to statistical analyses and the responses in that treatment also aren't referred to in results. However, they are shown in Figures 1,3,4, and 5. This unexpected inclusion of the data isn't mentioned in the figure legends. Authors need to decide whether it should be reported or not. Either remove or deal with it consistently. Response: This treatment has been added to Methods (L. 130-133). It also has been added to the Results section (Lines 220-222 and lines 241-243). The figure captions refer the reader to Table 1, where this treatment is described along with the other treatments.

Comment: Line 156- Used ladders to measure diameters at 33m height? Response: Yes, a ladder was used on those trees (Line 157).

Comment: Line 307- Add a reference for "Type A" response or don't use terminology. Here and later (line 365), the "Type A" label is immediately followed by saying that it means "sustained". Why not just say the response was sustained over the course of the study? I don't see what the "Type A" label adds. Response: The phrase "Type A" was removed from the manuscript.

Comment: Line 310-312- This could be clarified. I'm not sure what "optimized" means. Yes, thinning effect was greater after the second fertilization (than in unfertilized plots), as indicated by P AI, but over the total 25 year period, the absolute effect of thinning appears as great or (possibly) greater in the unfertilized plots than in the fertilized, perhaps because young unfertilized trees were so much smaller than others at the time of the second fertilization. Over 25 years, I don't know if growth response to thinning was "optimized" in conjunction with fertilization; it certainly wasn't maximized. Response: This paragraph has been revised to clarify its intended meaning (Lines 306-315).

Comment: Line 351-353- Hypothesis might be correct. Unfortunately, treatments didn't include addition ofP alone or other ratios ofN and P and there were no analyses of foliar elemental concentrations and contents or N allocation in foliage after year 13 fertilization. The latter would

indicate if excess N was taken up relative to P or other elements. Are there any data on N allocation in cedar in relation to- foliar elemental concentrations? Perhaps this possibility could be briefly discussed? Response: Based on comments from other reviewers, this speculative sentence has already been removed from the manuscript.

Comment: Line 365-374- Same comment as previously with respect to "Type A" label. Is it even necessary? Why not just say "sustained" or "didn't persist as long", etc.? Response: The phrase "Type A" was removed from the manuscript.

Comment: Line 402-413 - Thinning and fertilization seemed to have similar effects on maximum diameter of lower branches, but fertilization resulted in greater mortality of lower branches than thinning did. Does the greater of live and larger-diameter lower branches in the thinned treatment reduce wood/log quality and value? Or is the importance of this also a function of rotation length? What is the anticipated rotation length for 2nd-growth cedar on these sites? Response: We modified this paragraph to briefly address the issue of lower limbs and wood quality (Lines 404-419).

Comment: Line 419-427- It's important to emphasize that understanding species-specific respQnses to thinning and fertilization is important for effective management of mixed-species stands. Fertilization appears to enhance growth of cedar more than thinning did, but the response was of shorter duration and might be of greater magnitude for associated species. Again, is effect on wood quality/log value also a consideration? Response: We have revised the paragraph to include the issue of mixed species stands and log quality (Lines 420-429).

Comment: Line 653-655,662-664,666-668- Perhaps mention that TI-N data are shown and why. Also, arrows could be added to the T and F labels and they could be moved to more appropriate and visible locations, e.g., towards the top of the lower graph in each figure. Response: Mention of T -1 N data has been added to the Methods and Results sections. Thus, it should no longer a mystery to see it in the figures. The authors believe that the reference to Table 1 in the figure captions should be sufficient to tell the reader what that treament is. Arrows were added and T and F labels were moved to more visible locations in the figures.