Experimental Examination of Vegetative Propagation Methods of Nothofagus antarctica (G. Forst.) Oerst. for Restoration of Fire-Damaged Forest in Torres del Paine National Park, Chile

Abstract

Nothofagus antarctica (Antarctic beech) is one of the main woody plants in the temperate deciduous forests and anti-boreal forests of the southern hemisphere. Since colonization of the Andean-Patagonian region by European settlers, however, stands of this species have been severely affected by fires caused by human activities, considerably reducing their area. To restore these forests to their area occupied before the fires, it is necessary to use artificial regeneration, relying on production of transplants in forest nurseries. Due to the low capacity for seed propagation, we focus on possibilities of producing seedlings by vegetative propagation. In a trial, we collected cuttings during three sets of dates, and attempted to root them using three combinations of substrate and ten combinations of stimulators. Using the most favorable combination of collection period, substrate and stimulator tested resulted in rooting of 23% of the cuttings, which exceeds the documented germination rates for this species.

1. Introduction

Broad-leaved forests that include Nothofagus antarctica (Antarctic beech) are found in the zone of temperate deciduous forests and anti-boreal forests of the southern hemisphere [1]. The species represents one of the main forest trees widely distributed in the Andean-Patagonian region [2]. Nevertheless, since colonization, stands of this species have been severely affected by fires caused by humans, considerably reducing their area. Bizma et al. [3] mention a loss of natural forests in the Aysén River (Patagonia Chilena) basin, reaching 23% in the period from 1900 to 1998. Another adverse influence was the conversion of natural forests into pastures during colonization, resulting in their current fragmentation. Up to 70% of the area of these forests is managed by silvopastoral methods [4] due to their high forage value [5].

Remainders of native Antarctic beech forests having high conservation values [6] are protected primarily in national parks, including Torres del Paine National Park in Chilean Patagonia, at 51° S latitude. Despite that paleo-environmental research indicates that fires have been a frequent natural phenomenon in the park [7], Antarctic beech forests are not naturally adapted to fire impact [8]. However, 48 severe forest fires exceeding 10 hectares in the park during the last 40 years were caused by tourists. Disastrous for the National Park Torres del Paine were the three most severe forest fires, in 1985, 2005 and 2011, which altogether affected almost 46 thousand hectares of land, of which 5730 ha were various types of Nothofagus forests [9].

Although occasionally propagation by seed occurs in native forests of N. antarctica, its main mode of natural propagation is vegetative [10]. As such, its level of seed-based reproduction contrasts with that of other South American species of Nothofagus. Natural propagation of N. antarctica via seed is apparently limited due to multiple factors. Mast years occur on average every three years [11], but this period has been lengthening in recent years (personal communication, P. Salinas). Moreover, the viability of seeds is very low, ranging from 11% to 17% [12]. Germinating capacity of live N. antarctica seeds reaches only 1.7–6.0%, although after stratification, it can reach up to 20.6% [13,14]. High temperatures from forest fires reduce the germinating capacity of seeds to zero [15]. More generally, the seeds of N. antarctica do not form persistent soil seed banks [16]. Additionally, they disperse to only a distance of 2–3 tree heights [16].

For those seeds that do germinate, seedling survival frequencies on various types of sites’ ranges are quite low, or even zero [17]. Thus, although the number of seedlings was shown to range from 0 to 197/m² depending on the shade provided by the mother stand, 70–100% seedlings were found to die within the first year [14]. Moreover, Echevarría et al. [18] showed that seedling browsing associated with high numbers of livestock suppressed their growth (and resulting forest regeneration) in spite of the fact that N. antarctica is less damaged by browsing than other species in the genus [19].

The low efficiency of propagation by seeds is compensated for by the capacity of vegetative propagation by means of root suckers and rooting of procumbent branches. However, although fire induces rejuvenation of coppice shoots from the stem base [20], the number of individuals and forest stand area regenerated in this way are apparently insufficient to fully replace the trees lost.

Given that natural regeneration is not sufficient, artificial forest regeneration is needed to restore the lost areas of N. antarctica forests, not only in the Torres del Paine National Park but also in other regions of this species’ occurrence. The production of trees from seeds in nurseries is restricted by the above-mentioned factors, including the seeds’ short-term, low viability and germinating capacity, along with seed collection, realistically being limited to mast years. However, vegetative propagation of trees for transplantation could satisfy the need for the regeneration of forest devastated by fires or grazing if it can be accomplished effectively and efficiently.

In general, vegetative propagation of broadleaves is little used in forest nursery practice because of its high demands and relatively low success [21,22,23,24]. However, based on the vegetative propagation abilities of N. antarctica, this species is well-suited to this approach. Our study therefore focuses on identifying the optimal procedures for regenerating N. antarctica by means of vegetative propagation of cuttings, which has not been sufficiently explored.

2. Materials and Methods

2.1. Experimental Design

Our propagation activities were conducted in the forest nursery at Puerto Natales (51°42′27″ S, 72°27′26″ W, 30 m a.s.l.), using woody shoots of Nothofagus antarctica collected from plots in Torres del Paine National Park (50°56′16″ S, 72°47′25″ W, 220 m a.s.l.) affected by fire in 2005. We collected one-year-old woody shoots of 50–100 cm in length, depending on the size and condition of the individual parent trees. The cuttings were taken from the lower parts of the plant. Collecting was carried out in the morning using garden scissors, and the cut shoots were then bound into bundles and placed in closed dark plastic bags to safely transport them to the nursery. Upon arriving at the nursey, the shoots were sliced into individual woody cuttings whose length depended on the location of the last bud and averaged about 10 cm.

The shoots were collected in three periods during the months of September and October 2008 with the onset of spring. Dates of collection were 18 September (experimental period 1), 28 September (experimental period 2) and 3 October (experimental period 3). The material collected differed phenologically: during the first collection period, of the woody shoots, the buds were still dormant, during the second collection there was already a white stripe apparent in the upper parts of the buds, and during the third collection, the buds in the apical parts of the plant had already partly broken. Immediately after each date of shoots’ collection, the trials were established using the combination of substrates and hormones described below.

For propagation, we used Styroblock polystyrene planting containers (Beaver Plastics, Ltd., Acheson, AB, Canada; volume 200 mL), three combinations of substrates and ten combinations of root stimulators. Based on our long-term experiences, the individual combinations of substrate contained two components (common commercial perlite with size of grains 2–6 mm and black peat) in the following ratios: 9:1 (experimental variant A), 5:5 (experimental variant B) and 3:7 (experimental variant C). The stimulators comprised the following ratios of hormones: IAA (indole acetic acid), IBA (indole butyric acid), NA (nicotinic acid) and NAA (naphthalene acetic acid). Combinations of stimulators were as follows: 4000 ppm IBA + 500 ppm NA (experimental variant 1), 4000 ppm IAA + 500 ppm NA (experimental variant 2), 9000 ppm IBA + 500 ppm NA (experimental variant 3), 6000 ppm IBA + 500 ppm NA (experimental variant 4), 6000 ppm IAA + 500 ppm NA (experimental variant 5), 1500 ppm IBA + 1500 ppm NAA + 500 ppm NA (experimental variant 6), 3000 ppm NAA (experimental variant 7), 9000 ppm IAA (experimental variant 8), 6000 ppm NAA (experimental variant 9) and the control, which had no root stimulator (experimental variant 10). All planting containers used were placed on growing benches and covered with transparent foil, and irrigation was applied every morning and evening.

Altogether, we tested 90 combinations of substrate, stimulator and date of cutting collection, and tested 84 cuttings (planting container capacity) for each of them. The variants were registered as three-digit codes, in which the first digit was substrate combination, the second digit was stimulator combination and the third digit was the date of cutting collection (phenological stage).

The first check on the survival of cuttings was carried out from 3 to 23 December 2008 in the order of the periods in which the cuttings were collected. Survivors were defined as live cuttings with already developed roots or with a lightly colored callus section, which is a sign of future successful rooting. Each survivor was then transplanted into a Patrik-Mini-type planting container (block with 50 holes with a volume of 100 mL)—all into the same substrate, commonly used in nurseries for the production of containerized planting stock (60% sawdust, 35% earth, 5% urea). The containers were labeled to indicate the initial combination of experimental variants. The second check was carried out from 3 to 23 March 2009, when the transplants were sorted into “rooted” and “dead”. Rooted transplants were considered to represent the final product usable for afforestation.

2.2. Data Analysis

Vegetative propagation success was evaluated in terms of the proportion of rooted cuttings of the total number of cuttings. The influence of possible explanatory factors (substrate type, stimulator and period of shoot collection) was analyzed using generalized linear models (quasibinomial family and logit link function) [25,26], with the factors’ significance assessed using the Likelihood-ratio test (LR-test) [26].

In addition to this analysis, we also used generalized linear models to evaluate the influence of each studied factor on the share of cuttings that were still alive at the end of the experimental period (the share of “survivors”) and their influence on the ratio of rooted cuttings to the number of surviving cuttings. We performed these additional analyses because of differences between the first and the other two dates of the experiment in the survival and therefore rooting of cuttings.

Significance of differences between the reference and other factor levels was assessed by one-dimensional Wald tests ([26], p 11). Significance of the factors as such was not inferred based on these Wald tests to circumvent the increase of the probability of a type-I error (due to the performance of multiple Wald tests). LR-tests were preferred in this regard.

These analyses were preceded by a simple analysis using box plots (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8 and Figure 9), elaborated as a breakdown of specific factor values. In this way, we were able to obtain at least a general idea already before the test itself, which could eventually be used in the selection of reference levels of each factor.

Initially, models for each explained variable were constructed incorporating all factors, including a possible stimulator–substrate interaction. However, interactions involving establishment date were not explored because there was no specific explanation for them and any eventually detected interactions could not be of practical use in future applications. Factors and interactions that were found to be non-significant at α = 0.05 by the LR-test were excluded from the model, with model simplification continuing until it included only significant factors.

3. Results

3.1. Proportion of Rooted Cuttings from the Total Number of Cuttings

A box plot showing the relationship between the substrate and the outcome (Figure 1) reveals that the highest share of rooted cuttings was achieved for Substrate A. No difference is apparent between Substrates B and C. As for the factor of the stimulator, the best results were achieved for Stimulators 1 and 3 (Figure 2). In the box plot regarding the cutting collection date, we can see a notably lower share of rooted cuttings for Period 1, while for the remaining two periods, the shares of rooted cuttings are rather similar (Figure 3).

Values and tests of parameters for the resulting model of the share of rooted cuttings are presented in Figure A1. As it shows, the best result was achieved for the combination of Substrate A, Stimulator 3 and Period 2, with a model share of 23% of rooted cuttings from all cuttings.

Stimulators denoted as numbers 5, 6 and 10 produced significantly different outcomes than the Stimulator 1 (p = 0.009891, 0.013485 and 0.036291, respectively), while the others did not. The difference between Substrate A and both Substrates B and C is statistically significant (p = 0.001630 and 0.000721, respectively), indicating that the use of Substrate A provides demonstrably higher shares of rooted cuttings. Period 1 yielded a share of rooted cuttings statistically significantly lower (p = 1.98 × 10⁻⁸ and 3.65 × 10⁻⁷, respectively) than the other two dates.

Results of the LR-test of factor significance are presented in Figure A2. As it shows, all considered factors were found to be statistically significant.

3.2. Proportion of Surviving Cuttings from the Total Number of Cuttings

The graphical analysis (Figure 4, Figure 5 and Figure 6) suggests that the highest proportion of surviving cuttings was grown in Substrate A, the lowest in Substrate C, and an intermediate share in Substrate B. In comparing the proportions in relation to the different stimulators, the situation is relatively less clear because the shares of surviving cuttings are rather similar among them. Comparison of the shares of surviving cuttings from different collecting dates reveals a considerably lower share associated with Period 1.

The values of the explanatory model are presented in Figure A3. The highest model share of survived cuttings (62.49%) corresponds to Substrate A, Stimulator 8 and Period 2. The Wald test did not reveal any significant differences between the reference stimulator and the other stimulators. The reference Substrate A and Period 1 statistically significantly differ from the other variations of each of these factors (p = 0.000878 for Substrate B and 4.31 × 10⁻⁸ for Substrate C; p < 2 × 10⁻¹⁶ for Period 2 and p < 2 × 10⁻¹⁶ for Period 3).

The general influence of factors was evaluated by the LR-test shown in Figure A4. As it shows, all considered factors were found to be statistically significant. However, according to the test results, survival relates much more to both the substrate and the date of the collection of cuttings.

3.3. Proportion of Rooted Cuttings from the Surviving Cuttings

The box plot (Figure 7) does not show any apparent difference in the proportion of rooted cuttings to the number of surviving cuttings in relation to the substrate. The effect of the stimulator is relatively apparent from the relevant graph (Figure 8). The box plot regarding the relationship with cutting collection date (Figure 9) suggests that a higher share of the surviving cuttings from the first period of cutting collection achieved rooting. There is an apparently higher variance of the experiment outcome in this particular period (note the higher boxplot for period one compared to the other two), suggesting that the proportion of rooted cuttings may finally be comparable among all three periods.

As it is apparent in Figure A5, and particularly in Figure A6 (LR-test of factors’ significance), period and substrate are shown as non-significant factors, and only the factor of the stimulator was found to be statistically significant. The parameterization of the resulting (simplified) model including the stimulator as the only factor is presented in Figure A7. Figure A8 shows the result of an LR-test of significance of the stimulator factor. The highest proportion of rooted cuttings from the total number of surviving cuttings was achieved with Stimulator 1 (share 0.3934). The second-best result was obtained with Stimulator 3 (0.3846). Stimulators 2, 5, 6 and 8 obtained statistically significantly lower proportions of rooted cuttings (see Wald tests in Figure A7). The lowest proportion of rooted cuttings was grown with Stimulator 5 (0.1288).

4. Discussion

There are many methods of vegetative propagation of woody plants [22,27], including by means of rooted cuttings, which is affected by multiple factors [28]. This approach has continued to develop and is used particularly in growing cultivars of fruit and ornamental woody plants [29,30,31,32,33,34]. In forestry, it is important to maintain the uniformity of genetically significant characteristics, especially in establishing plantations of fast-growing woody species [24,28,35,36]. Another possible application of this propagation method in forestry is for species with small production of seeds or with seeds having problematic germination, such as is the case for Nothofagus antarctica [12].

Hardwood broadleaves usually take root from cuttings with difficulty [22], and genus Nothofagus has been characterized in this way [21]. In this genus, N. pumilio is an example of a species known to take root very poorly [37,38,39,40]. However, in N. antarctica, we can reasonably presume a greater success in rooting of cuttings due to the species’ natural regeneration through vegetative propagation after fires [10,20,41,42].

Previous studies of artificial rooting of cuttings of N. antarctica have been performed by Acuña [21] and Salinas et al. [23]. Acuña [21] achieved success only in rooting cuttings collected from young plants arising from root suckers. Thus, for cuttings taken from annual shoots occurring in the apical parts of 3-year-old suckers, the success rate was 25%, while for cuttings from 1-year-old root suckers broken off together with the basal part, it was 43%. However, that study reported zero success in rooting cuttings from the top parts of the crowns of adult trees of generative origin. Salinas et al. [23] collected cuttings only from the top parts of selected vigorous individuals of generative origin and obtained success of 7–20%, with various stimulator concentrations.

In the present study, our success rates ranged from 0 to 23% for the 90 treatment variants. The observed variability of successful vegetative propagation (proportion of rooted cuttings from the total number of cuttings) can be explained by the treatments’ influence on two aspects of propagation—survival potential and rooting potential. The effects of the types of treatments were often quite different with respect to survival and rooting. Thus, the substrate had a demonstrable influence on survival (in terms of proportion of surviving cuttings), but not on rooting capacity (in terms of the proportion of total cuttings taking root). Similarly, a statistically significant difference in survival was observed between the first and the remaining two dates of the cutting collection, with cuttings collected in dormancy exhibiting lower survival, whereas no dependence of rooting on collection date was found. In contrast, although the effect of stimulator on survival was only marginal, rooting capacity was closely related to the stimulator used.

Quite understandably, these results would reflect the direct influences of substrate composition and cutting collection date on survival ability, whereas different stimulators prime the rooting capacity to varied extents. Kanmegne et al. [43] arrived at a similar conclusion in an experiment with the rooting of Cola anomala cuttings, in which although the stimulator IBA had no effect on cutting mortality, higher concentrations of it led to higher rooting ability.

Regarding substrates, in our study, the best survival results were achieved with Substrate A, which contained the lowest amount of peat, and of the total cuttings, both the percentage that survived and the percentage that rooted decreased with increasing proportions of peat. It is likely that the poorer results for the higher amounts of peats have to do with the organic origin of this component, as it provides better conditions for development of fungal pathogens than does the other admixture—perlite. Indeed, all the dead cuttings showed clear signs of fungal infection in their portions buried in the soil.

In contrast to the negative effect that we found for the admixture of organic material on rooting survival, other studies have found that such material can be beneficial. In particular, Acuña [21] found Nothofagus antarctica cuttings to be more successful in a pure leaf litter substrate than when mixed with perlite powder. This difference might be attributable to the difference in method of fungicide application, in that whereas we sprayed the cuttings after planting them, Acuña [21] dipped the bases of cuttings directly into the fungicide. Similarly, De Carvalho Silva et al. [44], in an experiment examining the influence of the substrate on rooting cuttings of Eplingiella fruticosa (Salzm. ex Benth.) Harley & J.F.B.Pastore, found the cuttings to survive best in a mixture that included the organic material humus (along with the commercial substrate Biomix and vermiculite), and like Acuña [21], had directly dipped the cutting bases in fungicide. Akat et al. [34] found peat to be a better substrate than perlite for rooting cuttings of two Nerium oleander L. cultivars right before; however, their methodology does not mention the use of fungicide.

In our study, the phenological stage of the cuttings showed a dominant influence on their survival, with dormant cuttings exhibiting higher mortality than those with buds already broken. Similarly, Acuña [21] achieved greater success using N. antarctica cuttings collected in spring than in winter. Monder and Pacholczak [45] also found that the phenological stage influenced rooting success for two rose cultivars, which showed a higher degree of rooting by cuttings taken before the opening of floral buds (and also 1–2 weeks after the fall of crown petals) than when the flowers were open. Cristofori et al. [30] found that Corylus avellana L. cuttings collected in June and September rooted better than those collected in July; unfortunately, they did not test cuttings collected in dormancy. Greater survival and subsequent rooting of cuttings relates to the higher content of carbohydrates, content of mineral nutrients and plant phytohormone in shoots [46].

A wide range of chemical products—both natural and synthetic—can enhance the rooting capacity of cuttings. This issue is dealt with in most studies focused on the rooting of broadleaved hardwood cuttings, with the greatest number of studies devoted to the use of various concentrations of indole-3-butyric acid (IBA), its derivatives, NAA, etc. (

Table 1. Overview of different rooting regulators used in rooting of hardwood cuttings.

Source	Rooting Regulator	Species	Concentration
Stevens and Pijut [36]	IBA potassium salt (K-IBA) and (IBA)	Juglans nigra	highest
De Carvalho Silva et al. [44]	5 concentrations of IBA	Eplingiella fruticosa	1.5 g·L−1.
Filho et al. [51]	indole-3-butyric acid (IBA)	Khaya anthotheca	2.0 g·L−1
Kanmegne et al. [43]	indole-3-butyric acid (IBA)	Cola anomala	5 a 10 g·L−1
Véras et al. [33]	indole-3-butyric acid (IBA)	Spondias mombin	3.0 g·L−1
Chater et al. [32]	indole-3-butyric acid (IBA)	Punica granatum	3.0 g·L−1
Akat et al. [34]	naphthaleneacetic acid (NAA)	Nerium olenader	0.5 g·L−1
Justamante et al. [48]	IBA + NAA	Argania spinosa	1 + 2 mg·L−1
Yusnita et al. [29]	IBA + NAA	Syzygium malaccense	1 + 1 g·L−1
Yusnita et al. [29]	NAA	Syzygium malaccense	2–4 g·L−1
Kaviani and Negahdar [49]	IBA + NAA	Buxus hyrcana	-
Cristofori et al. [30]	IBA + putrescine	Corylus avellana	1 + 1.6 g·L−1
Rugini et al. [52]	IBA + putrescine	Olea europaea	2 + 1.6 g·L−1
Gergoff Grozeff et al. [31]	sodium nitroprusside (SNP) + IBA	Prunus insititia × P. domestica	1 + 0.3 g·L−1

). Fewer studies (Table 1) have dealt with combinations of two or more growth regulators, which can sometimes be more effective than if used individually [47]. In our experiment, we used various combinations of four growth regulators (IAA, IBA, NA and NAA). As the most effective for the rooting of N. antarctica cuttings, we evaluated different concentrations of IBA (4.0, 6.0 and 9.0 g·L⁻¹) with 0.5 g·L⁻¹ NA, or separate use of NAA at 6 g·L⁻¹. From Table 1, it follows that a growth regulator most frequently used for the rooting of cuttings is IBA. Multiple studies [29,34,48,49], however, have produced results that, like ours, indicate that NAA either alone or in combination with IBA has a better influence on rooting. On the other hand, although the results of our comparisons of IBA at different concentrations showed greater success in rooting cuttings using mid- and higher-IBA concentrations, some other studies have shown lower concentrations to be more favorable [32,33,44]. This is in keeping with the general observation that successful results are often species- or even clone-dependent [50].

There is synergy between the growth regulators used (auxins) and the rooting cofactors. Nevertheless, the chemical identification of endogenous cofactors remains unclear, despite the fact that their character is that of phenolic substances of natural origin [22]. Phenolic compounds affect the concentration of indole acid (IAA), which influences the development of adventitious roots [53]. In light of this, adult tissues have a higher content of phenol than younger ones, which both makes their rhizogenic capacity greater and increases the total content of phenols at root regeneration sites in order to form an auxin-phenol conjugate; moreover, the rhizogenic capacity increases towards the base of cuttings due to decreasing inhibitors and increasing synthesis of auxin [54].

The action of phenolic compounds supports adventitious roots by at least partial protection from auxin destruction by the IAA oxidase enzyme [22]. On the other hand, it seems that the high concentration of total soluble phenols in Nothofagus pumilio (Poepp. et Endl.) Krasser inhibits the development of adventitious roots by inducing the activity of enzymes responsible for auxin oxidation [37]. Many phenolic compounds are known to participate in prime rooting if they occur in low concentrations because their boundary of toxicity is close to the optimal concentrations for rooting. The role of phenolic compounds in the development of random roots is a subject of discussion, because some authors [37] argue that the action of phenolic substances in priming roots consists of protection from auxin destruction by the AIA oxidase enzyme. However, these authors also claim that these would act as inhibitors and antagonists of auxin effects, blocking their transport and facilitating the action of the IAA oxidase enzyme.

More detailed conclusions from a further similar experiment could be obtained based on a procedure in which the individual substrate and stimulator components are expressed separately in their proportional amounts. In such an approach, the analysis of variance can be substituted by the analysis of covariance or by the logistic regression with exclusively continuous variables, whose result is a parametrized model, which allows calculation of the optimal representation of individual components of substrate and stimulator. Optimal outcomes can then be assessed in terms not only of the highest expected proportion of transplants yielded from a given number of cuttings, but for example also with regards to greatest cost efficiency (considering different input prices of substrate and stimulator components).

Although absolute control of all factors that could possibly influence the result cannot be ensured in experiments under operational conditions, the results of our experiment would be even more robust if we included additional factors. Thus, other growth stimulators, e.g., putrescine, as recommended by Rugini et al. [52] and Cristofori et al. [30], could be employed. Additionally, in greenhouse experiments with the rooting of cuttings, substrate heating and irrigation control are often used. For example, in their experiment on the rooting of Nothofagus antarctica, Salinas et al. [23] used an elevated bed along with a thermo cable installed in an underlying layer of quartz sand to achieve optimal temperature. Thus, we suggest that the results of our study, in which we compared multiple variants of different factors, should be considered in light of these other possible factors, to yield the best conditions to propagate the study species.

5. Conclusions

Crucially, in our experiment, the most successful method (using cuttings that had recently ended dormancy along with a 9:1 perlite/peat substrate and a root stimulator consisting of 9000 ppm IBA + 500 ppm NA) resulted in a rate of rooting (up to 23% of the cuttings) that exceeded even the highest documented germinating capacities for Nothofagus antarctica: 20.6% using seeds stratified for 90 days at 5 °C [55] and 18.4% using seeds stratified for 60 days at 4 °C [13]. Thus, vegetative propagation by cuttings offers promise as a highly effective approach for the production of Nothofagus antarctica planting stock. This approach would be especially advantageous in periods outside mast years given that the seeds have only short viability and the possibilities for their storage have not been studied.