Application of post-emergence herbicides in the regeneration of pedunculate oak (Quercus robur L.) forests

Weeds are an important limiting factor in the development of pedunculate oak seedlings. Weed control is very important on regeneration areas, particularly in the initial stages of young growth development, when the adverse effects of weeds on oak are greatest. Effective weed control should be based on the combination of several different measures. However, because of manpower shortage, high labour costs and large areas, forest managers frequently decide to use herbicides to control weeds. On oak regeneration areas, a major problem is often broadleaved weeds that grow vigorously. Post-emergence herbicides nicosulfuron, bentazon, imazamox and tribenuron-methyl were monitored in broadleaf weed suppression in pedunculate oak stands during the regeneration process. Field studies were performed at two sites during three growing seasons, i.e. herbicide selectivity was examined on 1-, 2- and 3-year-old seedlings. The study results show that the number and fresh and dry weights of weeds were significantly reduced by tested herbicides in treated sample plots when compared with control plots. The herbicides tribenuron-methyl and bentazon had phytotoxic effects on oak seedlings. In contrast, nicosulfuron and imazamox did not show phytotoxic activity against the seedlings and can be applied in weed control in regenerated pedunculate oak forests.

Introduction

The process of oak forest natural regeneration is neither easy nor simple and requires the knowledge and control of numerous factors which can have an adverse effect on the course of regeneration (Lorimer, 1989; Dubois et al., 2000). The regeneration success could be unsatisfactory despite the most professional natural regeneration procedure. The reasons could be poor acorn yield, uneven seeding, unfavourable site conditions, low-quality and empty seeds, acorn damaged by insects or rodents or the abundance of weeds (Stojanovic and Krstic, 2000). The presence of weeds and the abundance of shrub species per unit area are the main limiting factors in pedunculate oak regeneration and survival. Weeds compete with seedlings for sunlight, water and nutrients necessary for seedling growth and development (Tobisch, 2007). Weed management is very important, particularly after the emergence of seedlings, which are unable to compete with the more vigorous and fast-growing weeds (Willoughby et al., 2007). Prior to acorn sowing, it is very important to prepare the ground, to ensure the most favourable conditions for the development of oak seedlings and saplings. Grindings left on the forest floor after stump removal should be partly or completely piled up, and then taken to forest roads and burned. Weed vegetation and shoots from smaller stumps (which are not removed) are killed with total herbicides, i.e. glyphosate-based preparations. On the prepared areas, acorn seeding is performed in the autumn or in the spring (Bobinac, 1994). Successful weed control depends on the combination of several different measures, i.e. preventive, chemical and mechanical treatments (Dey et al., 2008; Vasic et al., 2009), because otherwise weed control will be ineffective.

However, due to manpower shortage, high labour costs and large areas, forest managers frequently use herbicides to protect oak seedlings and saplings from weed competition. For many forest managers, herbicide application is a popular option and more cost-effective than other available alternative measures (McDonald and Fiddler, 1993; Wagner et al., 2004; Willoughby et al., 2007; 2009; Bolte and Lof, 2010). Not all weeds have the same ability to spread owing to differing regeneration strategies. From the suppression standpoint, perennial weeds are far greater problems. Mechanical control of perennial weeds is difficult, because it usually stimulates more intensive growth and spreading, as every cut rhizome can produce a new plant (Konstantinovic, 2008; Vasic et al., 2009). On regeneration areas in Serbia, broadleaf weeds such as Solanum nigrum L., Stenactis annua (L.) Nees., Erigeron canadensis L., Phytolacca americana L., etc. are much greater problems because they quickly form large clumps because of their characteristic vigorous development, which means that they can overtop and shade the competing oaks and threaten their survival.

For that reason, the purpose of this study was to examine the application of post-emergence herbicides nicosulfuron, bentazon, imazamox and tribenuron-methyl in the suppression of broadleaf weeds in the regeneration of pedunculate oak forests. Nicosulfuron and tribenuron-methyl are a group of post-emergent herbicides that contain sulphonylurea herbicide and act by inhibiting the enzyme acetolactate synthase (O’Sullivan et al., 2000; Skrzypczak et al., 2011; Filimon et al., 2011). Like other sulfonylurea herbicides, nicosulfuron and tribenuron-methyl are active at low rates, have low potential for ground water contamination and have low mammalian toxicity, characteristics that are desirable for herbicides used in an integrated weed management system (O’Sullivan and Bouw, 1998). Bentazon is a post-emergence herbicide used in early spring to early summer in many plantations (Huber and Otto, 1994) for the control of numerous broadleaf weeds. Bentazon is a benzothiazolinone contact herbicide acting as a photosynthetic electron transfer inhibitor (Macedo et al., 2008). Imazamox is member of the imidazolinone family of herbicides. It inhibits acetolactate synthase (ALS) and is the key enzyme in the biosynthesis of branched-chain amino acids: valine, leucine, and isoleucine (Stidham and Singh, 1991; Tan et al., 2005). It can be applied for early post-emergence weed control in lucerne and various grain legumes. Imidazolinone herbicides show high biological activity in plants and soils at low concentrations (Ahrens, 1994).

Materials and methods

Herbicide efficacy and selectivity in the regeneration of pedunculate oak forests were investigated during the period 2007–2009. Experimental plots were established at two sites within the Forest Estate Sremska Mitrovica. One experiment was established in the Forest Administration area of Morovic (44° 55′ 39.85″ N, 19° 6′ 30.39″ E), and a second trial was conducted in the Forest Administration area of Visnjicevo (44° 56′ 43.74″ N, 19° 15′ 48.97″ E). Based on the morphological structure and properties, the study soils are classified as loamy soils. Soil pH was 5.3 at Visnjicevo and 5.7 at Morovic. CaCO₃ concentrations were 0.8 per cent at Visnjicevo and 2.1 per cent at Morovic. The supply of humus in the soils was good: humus content accounted for 6.2 per cent at Visnjicevo and 6.2 per cent at Morovic.

Prior to acorn sowing, sample plot preparation included brush removal and herbaceous weed control by overall spraying of the herbicide glyphosate at 2.88 kg a.i. ha⁻¹. Branchwood was piled and removed because it can disturb the movement of seeding machinery. The prepared plots were seeded using a seed sower at 500 kg of acorn per hectare. At Visnjicevo, acorns were sown in the autumn of 2006, immediately after acorn collection. At the Morovic site, acorn seeding was performed early in the spring of 2007. During the winter, acorns were kept at a temperature of 4°C. On the seeded areas, the seedlings were 1 year old by the end of the 2007 growing season, 2 years old by the end of 2008 and 3 years old by the end of 2009. Sample plots were not irrigated or fertilized either before or after sowing. The experiments were established using a randomized block design with four blocks, the size of each experimental plot was 30 m² (3 × 10 m), so each treatment covered an area of 120 m². Each spring, the treatments were applied always on the same seedlings. Baseline data on the experiments and the dates of herbicide application are presented in Table 2. Herbicides nicosulfuron (Motivell; ISK Biosciences Europe, Belgium), bentazone (Basagran; BASF AG, Ludwigshafen, Germany), imazamox (Pulsar-40; BASF Agro, Wadenswil, Switzerland) and tribenuron-methyl (Granstar 75-WG; DuPont International Operations, Switzerland) were applied with a CP-3 sprayer, each year after weed emergence. Water consumption was 300 L ha⁻¹ (3 L per 100 m²). The main data on the study herbicides are presented in Table 1.The study included two types of control, which were not treated with herbicides, of which Control I consisted of mechanical control only, and the objective of Control II was to examine the effect of weeds on seedling growth without the application of any chemical or mechanical protection measures. Mechanical weeding consisted of weed suppression by one-hand hoeing during the growing seasons.

The analysis of plant number per unit area was repeated two times, 15 and 30 days after herbicide treatment. The number of weeds was determined in each elementary plot sized 2 × 1 m². The measurements of weed fresh and dry weight were performed after the last assessment of herbicide efficacy. Dry weight was measured after the fresh weed mass was dried to oven dry weight in the dryer. The study data were statistically processed and presented in the tables. Baseline data on the experiments are presented in Table 2. To assess the effect of the study herbicides on pedunculate oak growth and development, at the end of the growing season during the three study years, seedling height and root collar diameter growth were measured on 20 randomly selected pedunculate oak seedlings in both treated plots and control plots. Data were log-transformed and analysed using ANOVA at P <0.05 significance level, and the differences among means were tested using Duncan's test.

Results

On both sites, there were a total of 34 plant species classified into 17 families (Table 3). The greatest number of species belonged to the family Asteraceae (26.5 per cent) and Poaceae (11.8 per cent). Families Polygonaceae and Rosaceaesu account for 8.8 per cent, families Lamiaceae and Plantaginaceae 5.9 per cent and other families 2.9 per cent. Table 3 presents weed species identified at the study sites. Both sites were characterized by abundant presence of broadleaf weeds; 24 species were identified at the Visnjicevo site and the most abundant were Ranunculus ficaria L., Ambrosia artemisiifolia L., Sonchus arvensis L., S. nigrum L. and Symphytum officinale L., S. annua (L.) Nees., E. canadensis L. Glechoma hederacea L. and Rubus caesius L. Grassy weeds were of the species Echinochloa crus-galli (L.) Beau., Setaria viridis (L.) P.B. and Festuca sylvatica Poll. (Vill.). At the Morovic site, 22 species were identified and the most abundant species were Polygonum lapathifolium L., S. nigrum L., S. arvensis L., Galium aparinae, Chenopodium polyspermum L., S. annua (L.) Nees., S. arvensis L., A. artemisiifolia L., P. lapathifolium L., Symphitum officinale L. and S. arvensis L. The grass species E. crus-galli (L.) Beau., S. viridis (L.) P.B. and Sorghum halepense (L.) Pers. were also present.

After herbicide treatment, the number of weeds per unit area was significantly reduced at both the study sites during the study period when compared with the untreated control (Tables 4 and 5). The greatest reduction in the number of weeds at both study sites was attained using the herbicide tribenuron-methyl, followed by bentazon. However, tribenuron-methyl had phytotoxic effects on 1-, 2- and 3-year-old pedunculate oak seedlings, which was manifested in the form of chlorosis and necrosis and degradation of juvenile foliage. Also, oak formed multiple stems and developed many lateral shoots. Bentazon caused leaf necrosis in oak seedlings of all ages. By the end of the growing season, leaf necrosis disappeared, but it led to unpleasant visual effects during the growing season. The herbicides nicosulfuron and imazamox were effective in the reduction of weed density when compared with the control and achieved almost the same effects. Nicosulfuron and imazamox did not have a phytotoxic effect on pedunculate oak. The number of weeds was lower on mechanically controlled plots, when compared with control plots without the application of any chemical or mechanical protection measures. Mechanical measures such as uprooting, digging and mowing were effective in weed control, but their effect was not long-lasting because weed vegetation reproduced relatively quickly.

When compared with no-herbicide treatments at the study sites, all the study herbicides significantly reduced the fresh and dry weight of weeds (Tables 6 and 7). At the site Visnjicevo, in 2007, the most efficacious herbicide in the reduction of weed fresh and dry weight was nicosulfuron. Fresh weight was reduced by 85.9 per cent and dry weight by 84.0 per cent. In the other treatments, the reduction in fresh weight ranged from 81.7 to 72.0 per cent and the reduction in dry weight from 82.5 to 72.4 per cent. In 2008 and 2009, the most efficacious reduction in weed fresh and dry weight was attained using the herbicide tribenuron-methyl (Table 6). Compared with other study herbicides in 2008 and 2009, nicosulfuron showed significantly lower efficacy in fresh and dry weight reduction, but compared with no-herbicide treatments, its efficacy was high and was >75.0 per cent. At the site Morovic in 2007, the highest efficacy in the reduction of weed fresh and dry weight was reached using imazamox. The lowest efficacy was shown using the herbicide bentazon. In 2008 and 2009, the highest efficacy in the reduction of weed fresh and dry weight was attained by tribenuron-methyl (Table 7). When compared with control plots, weed fresh and dry weight was also significantly reduced using other study herbicides. Weed weight was also reduced by the application of mechanical measures, but they were less efficacious than the herbicides.

In 2007, the values obtained by measuring the height of 1-year-old seedlings in treated plots did not show statistically significant differences (Table 8). Although during 2007 the lowest values at both sites were measured on control (24.0 cm at Visnjicevo and 11.4 cm at Morovic) and in the treatment with tribenuron-methyl (24.3 cm at Visnjicevo and 11.8 cm at Morovic), the values were not statistically significant. The measurements of 2- and 3-year-old seedlings show statistically significantly lower values both in the treatment with tribenuron-methyl and in control. In control plots without the application of any chemical or mechanical protection measures, vigorous weeds inhibited normal seedling development, which had an adverse effect on height growth. The negative effect of tribenuron-methyl phytotoxic activity resulted in considerably lower seedling heights. Although bentazon caused leaf necrosis of oak seedlings, it did not result in inhibitory effect on oak seedlings growth. The effect of nicosulfuron and imazamox on oak seedling growth was not negative. Mechanical weed control made it possible for the seedlings to have undisturbed development and to attain the appropriate heights. At the site Visnjicevo, the measured weights of 1-, 2- and 3-year-old seedlings were higher than the weights measured at the site Morovic.

The analysis of data obtained by the measurement of root collar diameter growth of 1-year-old seedlings at the study sites shows that the values were not statistically significant (Table 9). The differences were statistically significant in the measurements of 2- and 3-year-old pedunculate oak seedlings. The measured seedling diameters in the variant with tribenuron-methyl (4.7 mm at Visnjicevo, and 2.7 mm at Morovic) were significantly larger when compared with other investigated variants with herbicides. Tribenuron-methyl, in addition to causing the degradation of top foliage, also led to the formation of multiple stems. Pedunculate oaks formed multiple stems and developed many lateral shoots, which made the measured seedling diameters larger than those in control plots and other variants with herbicides. The lowest values during the study period at both study sites were measured in control. The presence of weeds inhibited the normal seedling development, which resulted in seedlings with lower diameter growth. Although bentazon caused leaf necrosis, this phenomenon did not have an adverse effect on root collar diameter growth. On the plots where weeds were mechanically suppressed, seedling development was normal. During the study years, root collar diameters of seedlings measured at Morovic were lower than that at Visnjicevo.

Discussion

Many activities in forestry, such as clear cutting, thinning, reforestation and regeneration, can bring about the changes in site conditions. This leads to disturbed edaphic and hydrological conditions, and light regime. In the modified environments, numerous weed species find favourable conditions for their development (Vasic et al., 2012). The changes develop very quickly, particularly during the first years after forest cutting. During forest regeneration, sunlight intensity is increased (Pattison et al., 1998; Levine and Feller, 2004), as well as the quantity of soil nutrients resulting from the accelerated decomposition of forest litter (Huebner and Tobin, 2006). In such modified conditions, regenerating areas are invaded by herbaceous vegetation species, mainly nitrophilous, ruderal and invasive plant species, which can be detrimental to the development of pedunculate oak seedlings. In the study sample plots, abundant broadleaf weed species compete with oak seedlings for moisture, nutrients and sunlight. Nilsson and Örlander (2003) report that, after felling and during regeneration, ground vegetation becomes the main consumer of nutrients that are necessary for seedling development. The presence of grass weed species can be hazardous to seedling survival on regenerating areas (McCarthy et al., 2011). However, grass weeds can readily be suppressed with herbicides, such as fluazifop-P-butyl, cycloxydim and propaquizafop, which are selective in oaks (Clay et al., 2006; Willoughby et al., 2006; Novak et al., 2007). There have been many studies (Porterfield et al., 1993; Willoughby et al., 2006; Jinks et al., 2006) of the efficacy and selectivity of pre-emergence herbicides to broadleaf species. However, it is only the annual weeds that can be controlled by pre-emergence herbicides. Their successful activity requires good soil preparation, and the maximal efficacy of herbicide treatment depends on sufficient rainfall. In contrast to agricultural soil, forest soil is not sufficiently prepared, the forest floor is covered with stump grindings and dead foliage, and the irrigation of regenerated areas is rare. In such conditions, pre-emergence herbicides cannot attain high efficacy. For that reason, this study was focused on the herbicides registered for post-emergence application and intended for the control of a great number of broadleaf weeds.

The study results show that the applied herbicides decreased significantly the number of weeds and their fresh and dry weights in treated plots when compared with untreated areas. Seedling survival and growth is improved by herbicide application (Collet et al., 1998; Coll et al., 2003). The herbicides nicosulfuron and imazamox were not phytotoxic to 1-, 2- and 3-year-old oak seedlings. Tribenuron-methyl was the most effective herbicide in weed reduction in sample plots during all study years, but it was phytotoxic to oak seedlings. Dixon et al. (2006) investigated the selectivity of tribenuron-methyl to pedunculate oak and other forest species and found that tribenuron-methyl applied at the rates of 0.015 and 0.045 g ha⁻¹ caused the symptoms of phytotoxic activity on seedling leaves in the form of chlorosis and shoot dying. As a result of tribenuron-methyl phytotoxic activity, the measured seedling heights were smaller and root collar diameters were significantly larger, because the plants formed multiple stems and analysis showed these differences were significant in 2- and 3-year-old seedlings.

The herbicide bentazon caused leaf necrosis, which disappeared by the end of growth period, but it led to an unpleasant visual effect during the growing season. Lawrie and Clay (1994) studied the effect of bentazon on oak seedlings and found that it caused leaf necrosis. If applied to soybean, bentazon can also cause temporary phytotoxicity. Soybean is initially damaged, but it rapidly adjusts to tolerate bentazon, because of induction of genes involved in glucoside conjugation of acyl hydroxybentazon, which is not toxic to plants (Sterling and Balke, 1989). Willoughby et al. (2004) reported that it is possible that herbicide application can cause damage to planted broadleaves, but in some cases, such losses are significantly lower than the losses caused by the absence of herbicide treatments.

Du7ring the study period, the number of weeds and fresh and dry weights of weeds in mechanically controlled plots were lower than those in control plots without the application of any herbicide or mechanical protection measures. This method of weed control is rather expensive, because it requires the engagement of a substantial number of workers, and it is insufficiently effective in the case of weeds with high regeneration power, when mowing and cutting should be repeated several times (Vasic et al., 2009). There are also cases when, due to intense weed competition, mechanical measures lead to accidental damage to seedlings hidden in dense weed canopy (Thiffault and Roy, 2011). Their research shows that the applied mechanical measures enable normal seedling development and the adequate height and diameter growth. Gourley et al. (1990) also claim that mechanical weed control makes it possible for planted seedlings to develop normally and attain sufficient heights.

In control plots without herbicide treatment or mechanical protection measures, abundant weeds overtopped and shaded the seedlings and inhibited their normal development and height and diameter growth. In addition to competition for moisture and nutrients, weeds also compete for sunlight. Most studies on tree species light requirement show that seedling growth decreases with the decrease in the available sunlight (Balandier et al., 2006; Modry et al., 2004; Dobrowolska, 2008).

The measurement of seedling heights and root collar diameters shows that there are differences between the vegetation in study locations. In addition to acorn quality, sizes of the acorns, sowing conditions and climatology of the year of sowing (Rubic, 1997; Navarro et al., 2006), the time of sowing is an important factor that can affect acorn germination and emergence. Acorn sowing in the autumn has its advantages over sowing in the spring (Dey et al., 2008). Autumn sowing makes acorn germination and emergence possible as soon as the conditions (temperature and moisture) are favourable. Acorns sown in the autumn are less attractive to rodents than spring-planted acorns (Dey et al., 2008; Birkedal et al., 2009). Soil characteristics can also affect pedunculate oak development. Soil analysis shows that the sites do not differ by physicochemical characteristics, except for CaCO₃ concentrations. According to Valdecantos et al., (2006), germination rate and seedling growth are affected by high CaCO₃ concentration in the soil, because Phosphorous and Potassium assimilation is disturbed. It is possible that high CaCO₃ concentration in the Morovic soil (CaCO₃ concentrations were 0.8 per cent at Visnjicevo and 2.1 per cent at Morovic) affected on the seedling growth, but this hypothesis is by all means still to be verified.

The herbicides nicosulfuron and imazamox can be applied as post-emergence herbicides in pedunculate oak forest regeneration. Based on the study results, macro-experiments were conducted in the area of Forest Estate Sremska Mitrovica over the period 2009–2011. The area of 10 ha was treated with herbicides nicosulfuron and imazamox, and the attained selectivity and efficacy in weed control confirmed the results of our research. The application of herbicides reduced the weediness, particularly in the early stages of oak seedling development when the unfavourable effect of weeds is the most harmful.

The herbicides tested in this research are not on the current FSC list of ‘Highly Hazardous' pesticides (FSC Pesticide Policy, 2007), in view of the fact that the forests of the Forest Estate Sremska Mitrovica are certified and hold a FSC certificate. Therefore, the application of nicosulfuron, bentazone, imazamox and tribenuron-methyl is allowed for use in FSC certified forests. Also, the study herbicides are allowed for use in the EU (http://ec.europa.eu/sanco_pesticides/public/index.cfm).

Conflict of interest statement

None declared.

Funding

This paper was realized as a part of the project ‘Studying climate change and its influence on the environment: impacts, adaptation and mitigation’ (43007) financed by the Ministry of Education and Science of the Republic of Serbia.

Acknowledgement

We also acknowledge the comments of two anonymous reviewers, their comments really helped to improve the paper.

References