Natural regeneration of Pinus spp. around seed production areas and orchards

This study aims at determining the establishment potential of the main species of the Pinus genus planted in Brazil and broaden the knowledge on the conditions that facilitate the invasion. The density of regenerants neighboring pine stands planted with 35 populations of Pinus spp. (P. caribaea (var. bahamensis and var. caribaea), P. elliottii, P. kesiya, P. merkussi, P. oocarpa, P. patula, P. pseudstrobus, P. strobus, P. taeda, and P. tecunomanii) was evaluated in three locations in the state of São Paulo. The vegetal covers neighboring the pine stands were evaluated regarding native vegetation occupation, open area, and eucalypts plantation. The census of regenerants was conducted up to 100 m away from the pine stand edges; samples were collected over an area divided into four 25m-wide transects. In each transect, the regenerants were quantied and classied into three size classes: sprouts ( ≤ 0.3 m), seedlings ( ≥ 0.3 m ≤ 1.0 m), and trees (> 1.0 m). The results indicate that Pinus elliottii is the species with the greatest invasion potential in the studied areas since it adapts to different environments. Areas without established vegetation favored regeneration, indicating the effect of the environment on regeneration occurrence. The number of regenerants increased near the seed source but decreased signicantly between 50 and 100 m away, despite the directional effect of the prevailing winds and the long-distance dispersal capacity of the genus.


Introduction
Most of the economically important forestry and agronomic crops are grown outside their region of origin and considered exotic species. It becomes, therefore, necessary to know the regeneration potential of the species in different environments since a few species of commercial importance have become invasive, causing problems in different ecosystems (Richardson and Rejmánek 2011; Blackburn et al. 2014).
In Brazil, the exotic forest species belonging to the Eucalyptus and Pinus genera stand out due to the great commercial value and for being important sources of raw material for the industrial forest sector. These species were introduced several times in the country while many introduced populations were maintained as Seeds Production Area (SPA) and Seed Sedlings Orchards (SSO) in experimental stations to serve as seed sources for breeding programs and even commercial planting. Thus, the fact that several species may be found in the same environmental conditions allows studying the regeneration of several species in the same place, as performed for eucalypts (Miolaro et al. 2017).
The species of the Pinus genus were introduced and commercially planted especially in the South and Southeast regions of Brazil (Shimizu et al. 2018) due to good adaptability to environmental conditions and great availability of seeds in SPA and orchards. Paradoxically, these same adaptability characteristics allow the species/genus to become invasive if they are "locally" adapted (Andrade-Restrepo et al. 2019).
There is great ecological diversity among the 105 species of the Pinus genus (Richardson and Bond 1991). Of this total, few species were considered as potentially invasive. The genus is registered as an invader in several parts of the world, such as New Zealand, Australia, South Africa, Argentina, and Chile (Simberloff et al. 2009). The genus has been registered as a potential invader of open areas, whether degraded or naturally occupied by herbaceous-shrubby vegetation (Zanchetta and Diniz 2006;Simberloff et al. 2009). In Brazil, the described invasions are mainly caused by the two most widespread species of the genus, P. taeda and P. elliottii (Ziller and Galvão 2002;Bechara et al. 2013;Brewer et al. 2018;Ramos et al. 2019). An e cient control strategy has been eliminating adult pine trees that function as seed sources and not allowing the next generation to reach maturity ledding to changes in the population structure of pines in protected area (Dechoum et al. 2019). However, other species must be studied in different environments to verify the potential for regeneration and implementation.
This work aims to determine the density of Pinus spp. in three different sites in the State of São Paulo, around homogeneous plantations in the reproductive phase, to identify the factors involved in the regeneration process.

Study sites
The Experimental Station for Forest Sciences (Estação Experimental de Ciências Florestais, EECF) is The Anhembi Experimental Station for Forest Sciences (Estação Experimental de Ciências Florestais, EECF) is located in the municipality of Anhembi, SP (22º latitude and 48º longitude; 500m altitude). Soil is predominantly Latosol and Quartzarenic Neosol. The climate is Cwa (Aw), with humid and hot summer, and cold and dry winter, as well as 23°C average annual temperature (Alvares et al. 2013). The average annual rainfall is 1,100 mm but a 20 mm water de cit is observed in the dry period, between May and August (Ferez et al. 2015). A season characteristic is grazing for grass control, that is, the area is released for cattle after the forest plantations reach arboreal size.

The Ecological Station and State Forest of Angatuba (Estação Ecológica (EEc) e Floresta Estadual (FE)
de Angatuba) are located in the municipalities of Angatuba and Guareí, SP (23º latitude; 48º longitude). Soil is classi ed as a Hydromorphic Red-yellow Latosol. The climate is classi ed as Cwa, with 19 °C and 1215 mm average annual temperature and rainfall, respectively.
In the Angatuba FE, four populations of P. elliottii implemented in 1964, 1972, and two in 2007 were selected and are referred to as P. elliottii I, II, III, and IV to differentiate among themselves.

Statistical analysis
The Anhembi Experimental Station was removed from the statistical analysis because no regenerants were found. The analyzed data were adjusted using the negative binomial distribution because the variance greater than the average for all evaluated characteristics and effects of the model does not allow using the Poisson distribution. Also, the occurrence and abundance of tree seedlings are better described by negative binomial models compared to the Poisson models (Zhang et al. 2012). To adjust the model, analyses were performed for each location, testing the effects of species distance and direction. Additionally, the effects of species and direction over distance were also tested. Con dence intervals (95%) were estimated to infer the differences in the levels of the effects, that is, differences between species and directions for each regenerant class size (sprouts, seedlings, and trees). The analyses were performed using the software R (R Core Team 2019).
The neighboring vegetation was not statistically analyzed due to the lack of a minimum number of repetitions in each condition. Here, it should be noted that the used data originated from a previously conducted census and not an experiment implemented exclusively for this study.

Results
No regenerants were recorded in Anhembi. However, in Angatuba station, an approximate density of 500 sprouts per hectare for population III of Pinus elliotii was observed in the 0 -25m transect in the eastern direction (L). Nevertheless, because it consisted of only a single record in the data set, the Angatuba site was also excluded from sprout analysis. On the other hand, in Itatinga, sprout density was higher closest to the pine stands (the 0 -25m transect) and decreased signi cantly moving away from the edge of the pine stand. The density of sprouts was signi cantly higher for P. elliottii compared to P. caribaea in the 0 -25 m transect while the highest number of sprouts was mostly observed in the South (S) and West (O) directions. Also, the sprout density of Pinus oocarpa was signi cantly higher (except for the hybrid P. caribaea var. hondurensis x P. caribaea var. caribaea) in the 25 -50 m transect, but not affected by direction ( Figure 2). Further, neither regeneration nor sprouts were observed in the last two (50 -75 and 75 -100 m) transects.
In Angatuba, regenerant seedlings were found only up to 50 m ( rst two transects) from the seed source, with no other occurrence in the farther transects ( Figure 3A). In Itatinga, the regenerant seedlings were mainly concentrated in the rst transect up to 25 m ( Figure 3B). In both sites, the population of regenerant seedlings decreased as the distance from the seed increased. The population II of P. elliottii regenerated the most in Angatuba whereas P. elliotti was the most regenerant species in Itatinga. Furthermore, the highest number of regenerants were recorded in the East (L) and South (S) directions in the Angatuba and Itatinga sites, respectively ( Figure 3).
The regeneration in the tree developmental stage was lower in Itatinga than Angatuba (Figure 4). In Angatuba, the regeneration in the tree phase was similar to the seedling phase, with a predominance of the population II of P. elliottii in the East direction.

Discussion
From the viewpoint of statistical rigor, it was not possible to analyze the effect of the neighboring vegetation, but Pinus spp. regenerants were found in all areas (open areas, as well as eucalyptus and native vegetation-covered areas), the regeneration density in each area corroborates the invasiveness theories regarding both species and environment. In the Anhembi and Itatinga stations, the results of similar observation work regarding eucalypts regenerants indicate that the exposed soil and the absence of predation by ants are fundamental for the successful establishment of sprouts (Silva et al. 2016;Miolaro et al. 2017).
In the Anhembi station, no pine regenerants were found despite the several populations of reproductive age from different species of the genus. This fact is probably related to the evaluated species, livestock management for grass control, and the annual drought. Even though the vegetation does not differ structurally between Anhembi and Itatinga, Anhembi is the only studied place that presents water de cit during the dry season. The water de cit of 20 mm in the region occurs between May and August (Ferez et al. 2015), the same period during which the Pinus spp. seeds are scattered.
In the southeastern and southern regions of Brazil, the reproductive season of P. elliottii occurs between March and May (Zanchetta and Pinheiro 2007) while P. taeda seeds start to scatter in mid-May and continue until September, with dispersal peaking in July (Jankovski 1996).
Studies on the establishment of eucalypts sprouts under controlled and eld conditions highlight the importance of air and soil moisture for the establishment of the sprouts (Silva et al. 2013(Silva et al. , 2016, similar to that observed for the Pinus genus in different studies. Furthermore, works in the literature have shown that lack or insu cient humidity in the rst year is one of the limiting factors for establishing the P. elliottii stands since high mortality rate of the regenerants occur soon after germination, when the humidity is insu cient (Jankovski, 1996). The frequent rains in the southeastern United States after the period of seed dispersal are essential for establishing the natural regeneration of Pinus taeda (Trousdell and Wenger 1963).
Additionally, P. elliotti exhibited a higher occurrence of establishment and is one of the main species of the genus that presents invasion problems (Brewer et al. 2018) . But even within the species, the changes observed in the establishment process in the different environments lead to the conclusion that the environment plays a role in this process. Already mentioned in the literature, other species such as P. oocarpa have also been shown capable of establishment as regenerant (Zenni and Simberloff 2013;Braga et al. 2014). Although several studied species did not present regenerants (P. kesiya, P. merkussi, P. pseudstrobus, P. strobus, and P. tecunomanii), it is highlighted that the species with the highest occurrence of regenerants are those more widespread in the world and with a greater number of populations evaluated in the study.
The surrounding vegetation is one of the factors strongly in uencing the establishment, in addition to the wind. In Angatuba, where open native vegetation (low and scattered trees) is present to the north of the population, a low number of regenerating seedlings (four individuals per hectare) of Pinus elliottii II was observed, however, a high number of established seedlings (> 800 per ha) up to 50 m away was observed in the east, the predominant wind direction. (Bognola et al. 2018) evaluated the dispersion of Pinus taeda seeds and concluded that seed density varied signi cantly among the cardinal points, with the majority being captured in the direction of the predominant wind direction.
Although pine seeds may disperse over a long distance (Benkman 1995) since the seeds are generally small, present wings, and can potentially be carried long distances by the wind, the density of regenerants was higher close to the seed source (up to 50 m). Similarly, Jankovski (1996) evaluated the natural regeneration of Pinus spp. and concluded that the density of regenerants decreases with the increasing distance from the pine stand matrix. As the distance increased, more seedlings and trees occurred, but no sprouts. (Pomeroy 1949) reported that more than 85% of sprouts die because the roots fail to penetrate the soil, as well as grow and develop an e cient root system. A study on Pinus elliottii one year after emergence reported that the sprout bank stabilizes at a young stage (Bourscheid and Reis 2010), considered the seedlings in the study. However, adequate conditions are necessary to reach the tree stage.
The sprouts, seedlings and trees were classi ed and analyzed separately, allowing to differentiate the density of each developmental stage. Analyzing the developmental stages separately is important because it allows proposing management of exotic invaders that is cheaper, more e cient, and less aggressive to the environment being preserved from the invasion (Boursheid and Reis 2010).
The initial establishment stages are the most amenable to economic management, during which the attention to isolated individuals or small groups can be crucial for preventing invasion (Simberloff et al. 2009). Thus, eliminating seedling regenerants is the most appropriate since the high natural mortality of sprouts and their manageable size do not impose major operational challenges. In Argentina, the early control work practiced in experimental plots of natural elds resulted in good recovery of the structure and composition of native species communities (Cuevas and Zalba 2010).
In addition to the management di culty, with the removal of trees, there is the possibility of worsening invasion after the seedling phase since, as observed in the regenerant trees, seed dispersal may occur. In