Establishment of Quercus ilex L. subsp. ballota [Desf.] Samp. using different weed control strategies in southern Spain

Ecological Engineering 25 (2005) 332–342

Establishment ofQuercus ilex L. subsp.ballota [Desf.] Samp.using different weed control strategies in southern Spain

Rafael Ma Navarro Cerrilloa,∗, Benito Fragueiroa, Carlos Ceacerosb,Antonio del Campoc, Rafael de Pradod

a Departamento de Ingenierıa Forestal, Facultad de Agronomıa y Montes, Universidad de Cordoba,Avda Menendez Pidal s/n, 14080 Cordoba, Spain

b EGMASA, Av. Johan Guttemberg s/n, Sevilla, Spainc Departamento Ingenierıa Hidraulica y Medio Ambiente, Universidad Politecnica de Valencia,

Camı de Vera s/n, 46022 Valencia, Spaind Departamento de Agronomıa, Facultad de Agronomıa y Montes, Universidad de Cordoba,

Avda Menendez Pidal s/n, 14080 Cordoba, Spain

Received 17 January 2005; received in revised form 6 May 2005; accepted 7 June 2005

Abstract

Formerly cropped sites on Mediterranean areas that have been planted with hardwoods are typically associated with low treesurvival and growth, with an enduring influence of agricultural weeds in the herbaceous stratum. This study evaluates the survival

landith treeoclimateof survivalhat treeture. Theitive effectve earlygrams.

ri--.

and growth of planted Holm oak (Quercus ilex L. ballota [Desf.] Samp.) in response to three weed control treatments on a cropsite in southern Spain. The weed control treatments (cultivation, herbicide, and mulch) were applied in combination wshelters. Survival, relative growth rate, biomass, and root architecture were monitored over a 1-year period. Shelter micrwas measured 1 year after establishment. Trees in the ground management treatments consistently had greater levelsthan the control. Although these effects were not significantly different between weed control treatments, we found tshelters had a significant effect on growth and biomass. There were also significant treatment effects on root architecresponse of trees to weed control treatments and tree shelters are consistent with other studies, which suggest a posrelated to the interplay of microclimate change and resource availability. Our results suggest that weed control may improsurvival in forest plantations and, in combination with tree shelters, merit consideration in Mediterranean afforestation pro© 2005 Elsevier B.V. All rights reserved.

Keywords: Afforestation; Herbaceous weed control; Oak plantations;Quercus ilex; Holm oak

∗ Corresponding author. Tel.: +34 957 218657;fax: +34 957 218563.

E-mail address: [email protected] (R.M. Navarro Cerrillo).

1. Introduction

Since the early 1990s, afforestation of former agcultural fields has been the primary silvicultural activity on the low quality croplands of southern Spain

0925-8574/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.ecoleng.2005.06.002

R.M. Navarro Cerrillo et al. / Ecological Engineering 25 (2005) 332–342 333

Survival and tree growth rates associated with theseprojects are usually inferior to those found on for-est sites (Navarro Cerrillo et al., 2001a). Commonlycited causes include poor matching of species andsite, changes in soil properties, presence of aggressiveweed communities, and animal predation (Haywoodet al., 1997; Truax et al., 2000; South et al., 2001).While all these factors are important determinants ofstand success, efforts to increase afforestation successmust focus on those management techniques that maybe improved cost-effectively, specifically tree speciesselection and management of competing vegetation.Holm oak (Quercus ilex L. ballota [Desf.] Samp.) isa slow-growth Mediterranean species and is consid-ered desirable for new plantations, but this species hasshown low survival and growth rates when planted(Navarro Cerrillo et al., 2001c). Herbaceous weeds areknown to compete with newly planted seedlings forwater, nutrients and light (Truax et al., 2000; Doboiset al., 2000; South et al., 2001). Silvicultural practices,which are often adopted for managing ground cover,include the use of cultivation, herbicides and mulch,used alone or in combination with tree shelters (Doboiset al., 2000). When compared with plantings withoutherbaceous weed control, treated forest plantations arecommonly characterized by increased plant survivaland enhanced growth (Cogliastro et al., 1990; Devineet al., 2000; Truax et al., 2000; Ezell and Nelson,2001). Research has not yet been sufficient to deter-mine all of the causes underlying the variable growthr ,2

cialp ,1 ena tion( 00;G il-iG t,1o hasb clud-i nta-t n eta ndso d iti fac-

tors, with beneficial effects in certain cases (Gupta,1991; Adams, 1997; Haywood, 1999), but not in oth-ers (Davis, 1988a, 1988b; Haywood and Youngquist,1991; Houle and Babeux, 1994).

Tree shelters are widely used to protect treeseedlings against animal browsing and as a comple-ment to weed control. Research to date has shown thattree shelters strongly affect the microclimate surround-ing the plant. The variables affected are temperature,light intensity, or radiation, relative humidity, vapourpressure deficit, CO2 concentration and wind (Potter,1991; Kjelgren et al., 1997; Dupraz, 1997a; Duprazand Berger, 1999). Survival appears to be enhancedby tree shelters (Potter, 1991; Navarro Cerrillo et al.,2001b; Bellot et al., 2002). Stem diameter growth doesnot, however, appear to increase, and in fact, reducedgrowth of stems has been observed (Potter, 1991;Kjelgren et al., 1994; Burger et al., 1996; Mayheadand Boothman, 1997; Dupraz, 1997b).

A small number of studies have included the com-bined effect of weed control and tree shelters on treeestablishment (Dobois et al., 2000). Thus, the objec-tive of the present study was to gauge the effect ofweed control treatments in combination with tree shel-ters on early tree survival, growth, biomass, length androot attributes in a Mediterranean Holm oak (Q. ilex L.ballota [Desf.] Samp.) plantation.

2. Material and methods

2

(4 ousfl odi-c 001g andi agri-c outd 002.T ana drys oned ns.T .b

esponse among these studies (Ezell and Nelson001).

Mulching has been recognized as a benefiractice in agronomic and forestry systems (Adams997; Haywood, 1999). This enhancement has bettributed to the reduction of vegetative competiHaywood and Youngquist, 1991; Haywood, 20reen et al., 2003) and to an increase in the availab

ty of key soil resources such as nitrogen (Truax andagnon, 1993) and water (Haywood and Youngquis991; Wien et al., 1993). More recently, the utilityf mulching (e.g. plastic or wood-fiber sheeting)een assessed in various forestry applications, in

ng the establishment of hardwood and conifer plaions (Adams, 1997; Haywood, 1999, 2000; Greel., 2003). The success of such applications depen improvements in survival and early growth, an

s assumed that mulching should enhance both

.1. Experimental site

The experimental plot was located in CordobaAndalucia, southern Spain) (coordinates 37◦51′N and◦48′E, 92 m altitude), in an area with homogeneat loam soil (Xerofluvents). The site has been perially cropped and was left fallow during the 2000–2rowing seasons, supporting a mixture of native

ntroduced herbaceous species associated withultural bottomlands. The experiment was carrieduring the period of February 2002 to December 2he area has a dry Mediterranean climate withverage annual rainfall of 670 mm, with hot andummers and warm winters. No watering was during the trial in order to reproduce field conditiohe selected species for the study wasQ. ilex L. subspallota [Desf.] Samp.

334 R.M. Navarro Cerrillo et al. / Ecological Engineering 25 (2005) 332–342

2.2. Site preparation, treatment establishment andexperimental design

In February 2002, 9-month-old seedlings wereplanted at the site. The seedlings were obtained froma private nursery, and were supplied in 400 cm3

containers using peat-vermiculite (3:1 volume) assubstrate. The average height and basal diameterof seedlings, measured just after planting, were12.26± 0.63 cm (mean values and standard error,N = 80) and 3.49± 0.2 mm (mean values and standarderror,N = 80), respectively, with no significant differ-ences between groups of treatments.

The experiment was arranged as a multi-factorialdesign with three levels of weed control (i.e. cultiva-tion, herbicide and mulch), two levels of tree shelter(tree shelter, no tree shelter) and a control, in a ran-domized complete block with 4 blocks and 20 treereplications per factorial combination, giving a totalof 80 plants per treatment. The planting area was sub-soiled, using a ripper with a single tine, to a depth ofmore than 60 cm, and soil clods were broken up byusing a spring harrow and culta-mulcher to providea more level surface for planting. The planting wasdone by hand in a rectangular plot (10 m× 30 m) fol-lowing a systematic spatial pattern distribution with adensity of 3300 plants ha−1 (2 m× 1.5 m). The treat-ments imposed were:

1. Control: This involved maintenance of the weedent.

2 la-es.ithtnted-e onub-

3 ith atorthe

4 -ts by

Immediately after planting, ventilated 60 cm tall treeshelters with numerous holes scattered along the upperpart of the tree shelter (Forespot®) were placed over thesheltered trees and buried 50 mm into the soil to elim-inate air movement through the shelter. Shelters werethen secured to a 60 cm metal rod anchored 200 mminto the soil.

Treatments were of operational intensity, which costin a range of landowners expenses in afforestation prac-tices (cultivation =D 30 ha−1; herbicide =D 55 ha−1;tree shelter =D 0.8 per plant; mulch =D 1.1 per plant).All treatments were maintained until completion of theexperiment on December 2002.

2.3. Plant analysis

Heights and diameter above the root collar (at 1 cmabove soil) were measured and recorded at the timeof planting and during the growing season using elec-tronic calipers. Survival was assessed at the end ofeach growing season. The relative growth rates (RGR)in height and diameter were calculated at the sametime (Pearcy et al., 1989), and the survival of allseedlings was measured in summer and autumn (Julyand November 2002).

In November 2002, four seedlings in each of theseven treatments were harvested. Soil was removedfrom the roots with water. The total length, averagediameter and length of each diameter class (every2 mm) of roots was measured digitally using theW da,1

ndb ball,a wasa

2

h ion( ingS or-p cells withm mea-s ber2 d as

cover that was present at the start of the experim. Herbicide: The ground was maintained in a re

tively bare condition through the use of herbicidHerbicide application consisted of spraying woxifluorfen 24% (Goal 2XL®, BASF) applied a2 kg active ingredient ha−1. Oxifluorfen has beeshown to improve survival and growth of planHolm oak (Jimenez and Saavedra, 1999). The treatment was applied with a backpack sprayer onc3 May 2002 to control early weed growth. No ssequent herbicide applications were made.

. Cultivation was carried out using two passes wsmall rotary cultivator drawn by a 40 hp farm tracto manage weeds by shallow (5 cm depth) duringpeak of growing season in May 2002.

. Mulch: A 1 m2 wide strip of 40�m black polyethylene sheeting was installed by burying the sheehand to a depth of 3 cm.

in Rhizo V3.10b (Regent Instruments Inc., Cana996).

The weight of the different fractions of above aelow ground biomass (leaves, shoots, trunk, rootnd roots outside of the ball) of the 28 seedlingslso measured after drying at 80◦C for 48 h.

.4. Microclimate measurements

Air temperature (◦C, accuracy 0.1◦C), relativeumidity (RH%, accuracy 5%) and PAR radiatW m−2) were recorded inside two tree shelters ustow Away XTI, Stow Away RH (Onset Computer Coration, Cape Cod, MA) and SP1110 silica photoensors, respectively, which were then comparedeasurements made on the control treatment. The

urement period was from February 2002 to Decem002 with an automatic data-logger and recorde


10-min averages. The sensors were placed immedi-ately above the seedling, at 20 cm height above thesoil, close to the seedling leaves, and held in place bya steel rod with a standardized meteorological cover.The sensors were placed close to each other to avoidpossible differences in the microclimate resulting frommicrotopography, aspect and the presence of naturalvegetation. Vapour pressure deficit (VPD), as an indi-cator of water stress, was calculated from simultaneousvalues for relative humidity and temperature.

2.5. Statistical analysis

Survival, root architecture, biomass, and relativegrowth rates data were examined to ascertain that thevariables were normally distributed and the varianceswere homogeneous. Prior to analysis, mean survivaldata were transformed by taking the arcsin of the squareroot of percent survival. Those variables were analysedwith a randomized block design analysis of variance(ANOVA), in which comparisons of blocks and treat-ments means were made using Tukey’s test (Quinn andKeough, 2002). The height and diameter measurementswere analyzed with repeated-measures ANOVA apply-ing a Tukey’s test (Quinn and Keough, 2002). The effectof treatments on measured variables was tested for sig-nificance at the 0.05 level. After proving the data, inany case we did detect a significant block effect. Datawere stored and processed by using Microsoft Excel

2000, and the descriptive statistical analysis of the datawas done with SPSS v. 8.0.

3. Results

3.1. Seedling performance

The ground management treatments had impor-tant effects on survival. Ten months after planting,seedlings in the weed control treatments had survivalrates >80%, compared with survival rates of 35% inthe control treatment (Fig. 1). These differences werestatically significant (arcsin transformation,F = 11.12;d.f. = 6;P < 0.001) with the control (survival 35%) forall sampling times. On average, herbaceous weed con-trol increased survival by about 50% as compared withuntreated plants. However, in the comparison amongstherbaceous weed treatments, differences in survivalwere not significant and ranged between 78.7% (culti-vation + treeshelter) to 93.7% (herbicide + treeshelter).

Significant differences existed between treatments(F = 63.26; d.f. = 6;P < 0.001) regarding final height(Fig. 2). Mean height of plants was 55 and 40% greaterin the sheltered treatments relative to the untreatedcontrol in the first and second measurements, respec-tively. Maximum values in height were reached forherbicide + tree shelter and mulch + tree shelter. Thesevalues were very near to the values for cultivation + tree

atide

Fig. 1. Survival rate (%) of seedlings belonging to the different trecultivation; CT, cultivation + tree shelters; H, herbicide; HT, herbic

ments in March 2003 (mean values and standard error,N = 80) (control; C,+ tree shelters; M, mulch; MT, mulch + tree shelters).


Fig. 2. Total height (cm) of seedlings according to treatment in December 2002 (mean values and standard error, number of replicates inTable 1). Different letters (a–d) indicate significant differences between treatments atP ≤ 0.05 (ANOVA, Tukey’s test) (control; C, cultivation;CT, cultivation + tree shelters; H, herbicide; HT, herbicide + tree shelters; M, mulch; MT, mulch + tree shelters).

shelter. In contrast, final diameter responses differedfor the use of tree shelter by the end of the first grow-ing season (Fig. 3). The lowest values were obtainedin cultivation + tree shelter, herbicide + tree shelter andmulch + tree shelter, and analysis of variance revealedthat the sheltered treatment was significantly differ-ent from the unsheltered treatment (F = 49.92; d.f. = 6;

P < 0.001). The ground management treatments oftenhad no significant effect on basal diameter.

These results are corroborated by the relative growthrate calculated for the study period (10 months).Table 1shows the RGR for height and basal diameter. Theeffects of treatments on these variables were weaker,but still statistically significant (F = 7.41; d.f. = 6;

Fig. 3. Basal diameter (mm) of seedlings according to treatment in December 2002 (mean values and standard error, number of replicates inTable 1). Different letters (a–d) indicate significant differences between treatments atP ≤ 0.05 (ANOVA, Tukey’s test) (control; C, cultivation;CT, cultivation + tree shelters; H, herbicide; HT, herbicide + tree shelters; M, mulch and MT, mulch + tree shelters).


Table 1Average relative growth rates of height (RGRH) and basal diameter(RGR∅) according to type of treatment (mean values and standarderror)

Treatments N RGRH(cm cm−1

year−1)

RGR∅(mm mm−1

year−1)

Control 24 0.57± 0.072bc 0.94± 0.08aCultivation 60 0.49± 0.65c 0.72± 0.40bCultivation + tree

shelters60 0.69± 0.738abc 0.53± 0.52b

Herbicide 60 0.64± 0.73abc 0.95± 0.41aHerbicide + tree

shelters60 0.95± 0.58a 0.75± 0.52ab

Mulch 60 0.62± 0.75abc 0.92± 0.59aMulch + tree shelters 60 0.83± 0.60ab 0.68± 0.58b

Different letters (a–c) indicate significant differences between treat-ments atP ≤ 0.05 (ANOVA, Bonferroni test;N: number of repli-cates).

P < 0.001;F = 11.08; d.f. = 6;P < 0.001) for the posi-tive effect of tree shelters in height growth rate andnegative in base diameter growth rate in majority ofsampling dates.

3.2. Root system

There were significant differences between the con-trol and the treatment (Table 2) in either total length(F = 2.39; d.f. = 6;P = 0.007) or in the average lengthof fine roots (0–4 mm) (F = 2.50; d.f. = 6;P = 0.005).On average, root architecture was weakly affected byherbaceous weed control as compared to untreatedplants. However, in the comparisons amongst shelteredplants within the herbicide and mulch treatments, wefound significant differences in total length and fineroots (0–4 mm), as well as an increase of length of fineroots for the control.

3.3. Biomass

The above-ground and root biomass were similaramong the control and weed management treatment,but plants growing with tree shelters had lower biomassvalues (Table 3). For above-ground biomass, signifi-cant differences existed among treatments (F = 2.35;d.f. = 6;P < 0.001), with significant differences for themulch treatment and the cultivation + tree shelteredtreatment (Table 3). The underground biomass washigher in the control and herbicide than the cultivationand mulch treatments. Plants growing with tree shel-ters had lower root biomass values (F = 48.78; d.f. = 6;P < 0.001) (Table 3).

3.4. Microclimatic characteristics

The annual pattern in air temperature showed differ-ences between inside and outside tree shelters (Table 4).The air temperature inside the tree shelters was higherthan that outside throughout the year, with the meantemperature being 0.82◦C higher than in the control.Ventilation of shelters reduced the difference betweenthe air temperature inside and outside by a maxi-mum of 1.76◦C during the summer. Temperatures werenever high enough to affect the plants, with an abso-lute maximum of 44.57◦C inside the tree shelters inAugust. During the spring, air temperature inside thetree shelter rose more rapidly, while outside it rosemore gradually and was higher than inside by the latea eren no fort hent thati

Table 2Average root morphological characteristics according to type of treatm

Treatments Total length(m)

Average diameter(cm)

Control 4.29± 0.61ab 0.11± 0.04aCultivation 4.65± 0.36ab 0.10± 0.01aCultivation + tree shelters 4.48± 0.97ab 0.11± 0.06aHerbicide 5.29± 0.27a 0.09± 0.005aHerbicide + tree shelters 3.80± 0.37ab 0.09± 0.004aMulch 4.16± 0.72ab 0.11± 0.005aMulch + tree shelters 3.39± 0.52b 0.09± 0.001a

Different letters (a and b) indicate significant differences between trea

fternoon. Differences in minimum temperatures wot so marked (0.20◦C) with a maximum variatiof 0.35◦C, which we assume is a positive effect

he plant. This pattern was similar in autumn, whe temperature outside the shelter was similar tonside.

ent (mean values and standard error,N = 4)

Length of roots(m) 0 <∅ < 2 mm



4.03± 0.60ab 0.21± 0.02a 0.01± 0.005a4.47± 0.33ab 0.13± 0.02ab 0.02± 0.001a4.34± 0.96ab 0.09± 0.002ab 0.02± 0.009a5.04± 0.20a 0.19± 0.02ab 0.03± 0.006a3.61± 0.34ab 0.14± 0.05ab 0.02± 0.009a3.94± 0.72ab 0.15± 0.03ab 0.04± 0.001a3.28± 0.47b 0.07± 0.004b 0.02± 0.009a

tments atP ≤ 0.05 (randomized block design ANOVA; Tukey’s test).


Table 3Average above-ground, root biomass, and root-to-shoot ratio according to type of treatment (mean values and standard error,N = 4)

Treatments Above-ground biomass (g) Root biomass (g) Root-to-shoot ratio

Control 4.22± 0.011a 3.71± 0.06a 0.87± 0.018aCultivation 4.07± 0.007a 2.98± 0.13b 0.73± 0.015bCultivation + tree shelters 3.87± 0.033b 2.71± 0.04bc 0.70± 0.016aHerbicide 4.12± 0.008a 3.56± 0.04a 0.86± 0.019bHerbicide + tree shelters 4.07± 0.007a 2.87± 0.05abc 0.70± 0.010bMulch 3.43± 0.046c 2.60± 0.03c 0.75± 0.011bMulch + tree shelters 3.37± 0.028c 2.55± 0.05c 0.75± 0.012b

Different letters (a–c) indicate significant differences between treatments atP ≤ 0.05 (ANOVA, Tukey’s test).

Table 4Monthly average temperature, maximum and minimum temperature, and vapour pressure deficit between the inside tree shelter and the outside

Month (2002) Average temperature(◦C)

Average maximumtemperature (◦C)

Average minimumtemperature (◦C)

Average pressuredeficit (kPa)

Outside Inside Outside Inside Outside Inside Outside Inside

March 14.49 15.51 20.51 21.89 9.38 9.49 0.36 0.44April 15.86 16.90 22.22 23.59 10.21 10.35 0.46 0.55May 19.41 20.46 26.82 28.29 11.93 12.12 0.92 1.05June 25.09 26.26 32.82 34.40 17.21 17.41 1.36 1.53July 28.05 29.28 36.89 38.65 18.61 18.96 1.84 2.07August 26.76 27.94 34.98 36.54 18.88 19.20 1.55 1.75September 22.91 24.03 29.51 31.05 17.29 17.56 0.81 0.95October 18.73 19.90 24.85 26.42 14.10 14.29 0.43 0.53November 13.87 14.93 18.89 20.20 9.20 9.25 0.28 0.35December 12.30 13.32 16.24 17.44 8.94 9.08 0.14 0.20

Data recorded between 01-03-2002 and 31-12-2002. Measurements performed at 0.20 m height.

The average values of relative humidity remainedquite similar inside the tree shelters, although relativehumidity (%) was slightly lower inside the tree shel-ters than outside (data not included). However, the airinside the tree shelters had a higher vapour pressuredeficit due to the higher temperature (Table 4) with themean 0.13 kPa higher than control. The average differ-ence in vapour pressure deficit was also similar in eachseasonal analysis. In summer, values of vapour pressuredeficit were in the range of 0.19–3.07 kPa (tree shelter)and 0.1–2.74 kPa (outside). In the spring and autumn,the vapour pressure values converged (Table 4).

Levels of total radiation varied along the seasons,with an average of 9.57 MJ m−2 day−1 (outside)and 6.71 MJ m−2 day−1 (inside). Maximum valueswere obtained in summer with 12 MJ m−2 day−1 out-side (maximum daily radiation 1329.22 W m−2) and6.8 MJ m−2 day−1 inside the tree shelters (maximumdaily radiation 931.93 W m−2). This pattern was sim-ilar in the autumn and spring outside the tree shelter,with minimum values of daily radiation of 427 W m−2

outside and 151 W m−2 inside the tree shelters. In gen-eral, a strong reduction in radiation levels inside thetreeshelters was observed (Fig. 4).

4. Discussion

The application of weed control treatments causedlarge variations in the survival and growth ofQ.ilex seedlings. Survival rates significantly differedamong the various treatments and were similar to thatreported previously (Navarro Cerrillo et al., 2001c).After the summer, with a relatively wet summer period(110 mm of precipitation) the survival rate in the con-trol treatment was 35.6% percent. The control treat-ment retained high densities of herbaceous plants onthe plots, which caused sustained losses in survival.In contrast, seedlings in the weed control treatmentsusually had good survival. This would suggest an ini-tial need to control weeds to allow for greater resourceavailability for the seedlings (Truax et al., 2000; Ezell


Fig. 4. Annual pattern of the daily total radiation (MJ m−2 day−1) and maximum radiation (W m−2) between the inside of treeshelter and theoutside. Data recorded between 01-03-2002 and 31-12-2002. Measurements performed at 0.20 m height.

and Nelson, 2001). The generally low levels of sur-vival in the control treatment could have been relatedto diminished levels of belowground resources (Woodand Yeiser, 2000).

A positive survival response to cultivation is oftenassociated with the control of a well-established andaggressive herbaceous layer in place prior to tillage.Repeated cultivation during the first several grow-ing seasons is more often associated with a positiveresponse in tree survival than the single applicationused here (Cogliastro et al., 1990; Devine et al., 2000).However, the costs associated with this and other sil-vicultural practices typically limit practitioners to onecompetition control treatment at or near the time ofestablishment. The use of herbicides had similar effectson survival in terms of reduced ground plant cover,and this is consistent with previous literature indicatingthat herbicide additions have mainly direct effects onearly plantations through modifying vegetation com-petition (Truax et al., 2000; Ezell and Nelson, 2001).Some studies have also noted increased tree survivalusing mulch, particularly under conditions of persis-

tent low soil moisture (e.g. regions with “summer dry”climates and/or thin soils (Haywood, 2000; Tarara,2000)). In this study, a detectable mulching effect wasalso observed with regard to survival, which may sug-gest that the intensity of weed control of poly mulchmay be sufficient forQ. ilex establishment on thiscondition. It must be emphasized that this study hasbeen conducted in a former cropped land. Thus, theweed control practices management recommendationsderived for this study cannot be extrapolated to othersituations (e.g. afforestation of mountain areas).

Further insights into the effects of weed controlmanagement practices on the early establishment offorest plantations can be gained by considering growthresponses. In the present study, this was done for above-ground growth, biomass and root architecture. Theeffect of weed control treatments was often lowest inthose variables related to growth. Different weed con-trol treatments did not clearly favour different growthpatterns and root architecture. This suggests that alter-ation of belowground resources related to weed controlis likely to have no great consequences on plant growth.


However, these effects had consistent consequencesrelated to the interaction between the use of tree shel-ters and different types of weed control (cultivation,mulch and herbicide). The absence of a survival effectof tree shelters found in this study supports previousfindings that survival can be independent of tree shel-ter (Potter, 1991; Burger et al., 1996; Bellot et al., 2002)and tends to be linked most closely to weed control. Asshown in previous studies, in this trial the plants withintree shelters grew more in height than non-protectedplants, although there is a reduction of stem growth insheltered trees (Potter, 1991; Kjelgren and Rupp, 1997;Dupraz, 1997a; Bergez and Dupraz, 2000). Shorteningthe period of vulnerability is a useful silvicultural strat-egy in croplands where pastoral uses may be important.In terms of above-ground growth, growth rates weremost consistently linked to the use of tree sheltersrather than to the levels of weed control. However,unprotected plants developed the greatest base diame-ter after the first year of planting (Kjelgren and Rupp,1997; Bergez and Dupraz, 2000). Also, there was sig-nificantly lower accumulative root biomass in shelters,which contrasts with previous results.Ponder (1995)and Bellot et al. (2002)reported a reduction in thetotal biomass of different tree and shrubs species, andthis was consistent with the inhibitory effect detectedin other research (Burger et al., 1997; Svihra et al.,1996). However, differences between the architectureof the roots in different weed control practices werenot apparent. The use of ventilated tree shelters hasb s,1 geza rr didn wtho ot tob

y ber icro-c ;K aa ;K 99A cli-m l.,1 herv lter.H side

the ventilated shelters was only slightly higher thanunder open conditions (∼1.6◦C), and absolute max-imum temperatures were always below 45◦C, simi-lar to those reported in other studies with ventilatedtree shelters (Dupraz, 1997a,b; Dupraz and Berger,1999; Navarro Cerrillo et al., 2001a). Radiation levelsinside the tree shelters (∼60% of outside) in Mediter-ranean areas do not appear to limit photosyntheticprocesses (Bellot et al., 2002). In the hottest period(July–August), environmental conditions inside theshelters may influence physiological processes, partic-ularly photosynthesis that could imply a certain loss ofgrowth (Werner et al., 1999; Bellot et al., 2002). How-ever, the increased growth during early spring due togreater water availability in sheltered plants may com-pensate for the reduction of growth in summer whenphysiological processes may be limited in Mediter-ranean vegetation (Bellot et al., 2002).

The removal of dominant agricultural weedsdirectly through the use of different weed control treat-ments in combination with tree shelters may facilitatethe early establishment of a wider array of both treesand shrub forest species, and this has important impli-cations for afforestation programs. Land owners whowish to increaseQ. ilex survival and growth shouldconsider either herbicide treatment or cultivation. Forall weed control treatments, herbicides offer the mostpromise in terms of cost-effectiveness but results aresensitive to the economic and agricultural conditionsof landowners. For trained farmers with herbicide usese her-b ricesl rst pli-c lm itialg tingt entsi forl on,w fD th.

cedi dya ons,s , andf petew ate

een described as a key issue (Mayhead and Jenkin992; Dupraz, 1997b; Swistock et al., 1999; Bernd Dupraz, 2000; Bellot et al., 2002). However, ouesults show that the ventilation of the tree sheltersot improve either the diameter or the biomass grof protected trees. Therefore, ventilation appears ne clearly favourable on tree growth.

Generally speaking, the treeshelter effect maelated to the relationship between changes in mlimate conditions inside the tree shelter (Potter, 1991jelgren, 1994; Navarro Cerrillo et al., 2001)nd plant physiological conditions (Dupraz, 1997a,bjelgren and Rupp, 1997; Dupraz and Berger, 19).s in previous studies conducted in Mediterraneanates (Costelo et al., 1996; Navarro Cerrillo et a998, 2001a) we found higher temperatures and higapour pressure deficit (VPD) inside the tree sheowever, the maximum temperature recorded in

xperience, our results indicate good feasibility foricide treatments combined with tree shelters at p

ess thanD 300 ha−1 the first year. The following yeahe weed control cost would be just the herbicide apation with a cost ofD 55 ha−1. An increase of survivaore than 50% and shortening the period of inrowth appear to be an important threshold affec

he economic attractiveness of using these treatmn croplands in Mediterranean areas. An alternativeow trained farmers could be the use of cultivatiith an initial cost ofD 270 ha−1 and following years o30 ha−1, but lower increased of survival and growThe results obtained on this study should be pla

n the context of the climatic conditions of the sturea and croplands. Under harsh climatic conditiuch as those of semi-arid Mediterranean areasorestry areas, existing vegetation may not comith introduced plants, and may, indeed, facilit


their establishment, growth and physiological status(Maestre et al., 2001; Castro et al., 2002).

5. Conclusion

Our study provides clear evidence that differenttreatments of weed control in forest plantations differ interms of their effects on the survival and growth, as wellas on biomass and root architecture ofQ. ilex. Previousexperiments have compared different kinds of soil man-agement but have not included the tree shelters factor.We have shown that a combination of weed control andventilated shelter significantly improved the survivaland growth ofQ. ilex trees when growing in a for-mer cropped land. Manipulation of the levels of weedcompetition is likely to be more important in influenc-ing the survival than are other components of forestsplantations, such as the direct effects of tree shelter, atleast in the system that we considered. The trial resultsshow that the cultivation and herbicide treatments withtree shelters are the most favourable for the plantingof Q. ilex in Mediterranean environments. The mulchtreatment appears to be the second most appropriatefor plantingQ. ilex seedling in farmer-cropped areas inMediterranean environments.

Acknowledgements

theA deA ain-i ba.W fromV er-s r forh

R

A oakr. 21,

B 2.

89–

Bergez, J.E., Dupraz, C., 2000. Effect of ventilation on growth ofPrunus avium seedlings in treeshelters. Agric. For. Meteorol.104, 199–214.

Burger, D., Foister, G., Gross, R., 1997. Short and long-term effectsof treeshelters on the root and stem growth of ornamental trees.J. Arboriculture 23 (2), 49–56.

Burger, D.W., Forister, G.W., Kiehl, P.A., 1996. Height, calipergrowth, and biomass response of 10 shade tree species to treeshel-ters. J. Arboriculture 22, 161–166.

Castro, J., Zamora, R., Hodar, J.A., Gomez, J.M., 2002. Use of shrubsas nurse plants: a new technique for reforestation in Mediter-ranean mountains. Restor. Ecol. 10, 297–305.

Cogliastro, A., Gagnon, D., Coderre, D., Bheıreur, P., 1990. Responseof seven hardwood tree species to herbicide, rototilling, andlegume cover at two southern Quebec plantation sites. Can. J.For. Res. 20, 1172–1182.

Costelo, L., Peters, A., Giusti, G., 1996. An evaluation of tree sheltereffects on plant survival and growth in a Mediterranean climate.J. Arboriculture 22, 1–9.

Davis, R.J., 1988a. Sheet mulching as an aid to broadleaved treeestablishment I. The effectiveness of various synthetic sheetscompared. Forestry 61, 89–105.

Davis, R.J., 1988b. Sheet mulching as an aid to broadleaved treeestablishment II. Comparison of various sizes of block polythenemulch and herbicide treated spot. Forestry 61, 107–123.

Devine, W.D., Houston, A.E., Tyler, D.D., 2000. Growth of threehardwood species through 18 years on a former agricultural bot-tomland. Southern J. Appl. For. 24, 159–165.

Dobois, M.R., Chappelka, A.H., Robbins, E., Somers, G., Baker,K., 2000. Tree shelters and weed control: effects on protection,survival and growth of cherrybark oak seedlings planted on acutover site. New For. 20, 105–118.

Dupraz, C., 1997a. Les protections de plantsa effect de serre.Premiere partie: ce quıen present ler arbres. Rev. For. Fr. 49,417–432.

D .

D pho-ed

E afterron in.

G ethy-ed-173,

G on. 44,

H ty tongs.

H n therease

This work was made possible with the help ofgricultural, Fish and Food Department of Juntandalucia as well as Agricultural Research and Tr

ng Centre (CIFA, Alameda del Obispo) in Cordoe appreciated comments on the draft manuscripticky Cramer and Richard Hobbs (Murdoch Univity, Western Australia) and an anonymous revieweelpful comments on an earlier version.

eferences

dams, J.C., 1997. Mulching improves early growth of fourspecies in plantation establishment. Southern J. Appl. Fo44–46.

ellot, J., Ortiz de Urbina, J.M., Bonet, A., Sanchez, J.R., 200The effect of tree shelters on the growth ofQuercus coc-cifera L. seedlings in a semiarid environment. Forestry 75,106.

upraz, C., 1997b. Les protections de plantsa effect de serreDeuxiene partie: amelioration de leur efficacite par aeration opti-misee et luminosite accrue. Rev. For. Fr. 49, 519–530.

upraz, C., Berger, J.E., 1999. Carbon dioxide limitations of thetosynthesis ofPrunus avium L. seedlings inside an unventilattreeshelters. For. Ecol. Manage. 119, 89–97.

zell, A.W., Nelson, L., 2001. Weed control and crop tolerancepre-emergent and post-emergent applications of sulfometuoak (Quercus spp.) plantations. Weed Technol. 15, 585–589

reen, D.S., Kruger, E.L., Stanosz, G.R., 2003. Effects of polylene mulch in a short-rotation poplar plantation vary with wecontrol strategies site quality and clone. For. Ecol. Manage.251–260.

upta, G.N., 1991. Effects of mulching and fertilizer applicationinitial development of some tree species. For. Ecol. Manage211–221.

aywood, J.D., 1999. Durability of selected mulches, their abilicontrol weeds, and influence growth of loblolly pine seedliNew For. 18, 263–277.

aywood, J.D., 2000. Mulch and hexazinona herbicide shortetime long-life pine seedlings are in the grass stage and incheight growth. New For. 19, 279–290.


Haywood, J.D., Youngquist, J.A., 1991. Mulching effects of plantfiber and plant fiberpolyester mats combined with fertilizer onloblolly pine seedlings. Tree Plant. Notes 42, 32–35.

Haywood, J.D., Tiarks, A.E., Sword, M.A., 1997. Fertilization, weedcontrol, and pine litter influence loblolly pine stem productivityand root development. New For. 14, 233–249.

Houle, G., Babeux, P., 1994. Fertilizing and mulching influenceon the performance of four native woody species suitable forrevegetation in the subartic Quebec. Can. J. For. Res. 24, 2342–2349.

Jimenez, M., Saavedra, M., 1999. Selectividad de herbicidas ennuevas plantaciones deQuercus ilex y Pinus halepensis. In: ActasCongreso Sociedad Espanola de Malherbologıa, Logrono, pp.347–352.

Kjelgren, R., 1994. Growth and water relations of Kentucky coffeetree in protective shelters during establishment. HortScience 29,777–780.

Kjelgren, R., Rupp, L.A., 1997. Establishment in treeshelters. I. Shel-ters reduce growth, water use, and hardiness, but not droughtavoidance. HortScience 32, 1281–1283.

Kjelgren, R., Cleveland, B., Fouth, M., 1994. Establishment of whiteoak seedling with three post-plant handling methods on deep-tilled minesoil during reclamation. J. Environ. Horticult. 12,100–103.

Kjelgren, R., Montague, D.T., Rupp, L.A., 1997. Establishment intreeshelters. II. Effect of shelter colour on gas exchange and har-diness. HortScience 32, 1284–1287.

Maestre, F.T., Bautista, S., Cortina, J., Bellot, J., 2001. Potential forusing facilitation by grasses to establish shrubs on a semiariddepredated steppe. Ecol. Appl. 11, 1641–1655.

Mayhead, G.J., Boothman, I.R., 1997. The effect of treeshelter heighton the early growth of sessile oak (Quercus petraea (matt.)Liebl.). Forestry 70, 151–155.

Mayhead, G.J., Jenkins, T.R., 1992. Growth of young Stikaspruce (Picea sitchensis (Bong.) Carr.) and the effect of sim-

462–

N es:

.

N o dealuc-8

Navarro Cerrillo, R.M., Oliet, J., Contreras, O., 2001b. El uso detubos protectores con cuatro especies forestales en andalucıaoccidental: supervivencia y crecimiento. In: S.E.C.F. (Ed.), IIICongreso Forestal Espanol, Junta Andalucıa, vol. 3. Granada,25–28 September, pp. 916–922.

Navarro Cerrillo, R.M., Carrasco, P., Amores, R., Palacios, G.,2001c. Seguimiento de trabajos de forestacion de tierras agrariasen Andalucıa: El caso de Huelva. In: S.E.C.F. (Ed.), III Con-greso Forestal Espanol, Junta Andalucıa, vol. 3. Granada, 25–28September, pp. 745–750.

Pearcy, R.W., Ehleringer, J., Mooney, H.A., Rundel, P. (Eds.), 1989.Plant Physiological Ecology: Field Methods and Instrumenta-tion. Chapman and Hall, London, 441 pp.

Ponder, F., 1995. Shoot and root growth on northern red oak plantedin forest openings and protected by treeshelters. N.J.A.F. 12 (1),36–42.

Potter, M.J., 1991. Tree Shelters. Forestry Commission Handbook,vol. 7. HOSM Publications, London, 89 pp.

Quinn, G.P., Keough, M.J., 2002. Experimental Design and DataAnalysis for Biologists. University Press, Cambridge, 527 pp.

Swistock, B.R., Mecum, K.A., Sharpe, W.E., 1999. Summer tem-peratures inside ventilated and unventilated brown plastic treeshelters in Pennsylvania. N. J. Appl. For. 16, 7–10.

South, D.B., Zwdinski, J.B., Kotze, H., 2001. Early growth responsesfrom weed control and planting large stockes ofPinus radiate aregreater than that obtained from mechanical soil cultivation. NewFor. 22, 199–211.

Svihra, P., Burger, D., Harris, R., 1996. Tree shleter affect on rootdevelopment of redwood trees. J. Arboriculture 22, 174–179.

Tarara, J.M., 2000. Microclimate modification with plastic mulch.HortScience 35, 169–180.

Truax, J.R., Gagnon, D., 1993. Effect of straw and black plasticmulching in the initial growth and nutrition of buttermut, whiteas hand buroak. For. Ecol. Manage. 57, 17–27.

Truax, J.R., Lambert, F., Gagnon, D., 2000. Herbicide-free planta-mesic

W Twovoidring

W lchnted

W ultantmu-13.

ulated browsing, staking and tree shelters. Forestry 65,463.

avarro Cerrillo, R.M., Martinez, M., Guzman, R., 1998. El uso dtubos invernaderos en trabajos de forestacion de tierras agrariaanalisis de algunos ensayos, vol. 55/98. In: Consejerıa de Agri-cultura y Pesca, Junta de Andalucıa, Informaciones Tecnicas, p50.

avarro Cerrillo, R.M., Oliet, J., Contreras, O., 2001a. El ustubos protectores con cuatro especies forestales en andıaoccidental: estudio microclimatico. In: S.E.C.F. (Ed.), III Congreso Forestal Espanol, Junta Andalucıa, vol. 3. Granada, 25–2September, pp. 812–815.

tions of oaks and ashes along a gradient of open to forestedenvironments. For. Ecol. Manage. 137, 155–169.

erner, C., Correira, O.A., Ryel, R.J., Beyschlag, W., 1999.different strategies of Mediterranean macchia plants to aphoto-inhibitory damage by excessive radiation levels dusummer drought. Acta Oecol. 20, 15–23.

ien, H.C., Minotti, P.L., Grubinger, V.P., 1993. Polyethylene mustimulates early root growth and nutrient uptake of transplatomatoes. J. Am. Soc. Horticult. Sci. 118, 207–211.

ood, R., Yeiser, J.L., 2000. Herbaceous weed control and respine seedling growth with new Oust, Velpar, and Escort forlations: year two results. Proc. S. Weed Sci. Soc. 54, 109–1

Documents

Establishment of Quercus ilex L. subsp. ballota [Desf.] Samp. using different weed control strategies in southern Spain