Study area

The La Trinidad community is part of the Santiago Xiacuí municipality in the Ixtlán de Juárez District, in the Northern Sierra region of the state of Oaxaca. Geographically, it is located at the coordinates 17Ί609.43" N and 96°250>0.77" W. The climate, according to the classification proposed by Köppen and modified by ^{García (1988)}, is C (w₂)(w)big, described as subhumid temperate -the most humid of all temperate climates-, with abundant rains in the summer; it comprises 30 to 59 days of winter rains, with a 150-200 mm precipitation. The total annual precipitation is 1 115.6 mm; the period with the most intense rains is June through October (^{UZACHI, 2003}).

According to the ^{FAO/UNESCO classification (1990)}, the soil types found in the region include Lithosols, Rendzines and Cambisols. In the conifer forest area, the soils are acid and very thin (^{UZACHI, 2003}). The altitude ranges between 2 000 and 3 000 m (^{UZACHI, 2003}). The area comprises the following types of vegetation: pine forest, humid and subhumid pine-oak forest, and mountainous, mesophytic forest (^{Rzedowski, 1978}; ^{UZACHI, 2003}).

From the information obtained from the Programa de Manejo Forestal (Forest Management Program) (^{UZACHI, 2003}), forest land tenure is communal. The total forest surface is 805 ha, of which 643.5 ha are under management, and the rest are protected or conservation areas, or else, areas reserved for other productive activities. Yearly cut volumes range between 2 000 and 3 000 m³, with 40-year periods. The Método de Desarrollo Silvícola (MDS) (Silvicultural Development Method) is utilized with regenerative strip clear-cutting, combined with the Método Mexicano de Ordenación de Bosques Irregulares (MMOBI), (Mexican Method for Ordering Irregular Forests) and with the group-selection regeneration method. The main commercial species are Pinus patula Schiede ex Schltdl. et Cham., Pinus ayacahuite Ehrenb. ex Schltdl. and Pinus pseudostrobus Lindl.

Field work

Two adjoining areas with similar climate, exposure and slope, separated by a creek (≈30 m), were located. Both were intervened in the year 2009 by means of a regenerative clear-cutting treatment in which all the commercial and non-commercial sized trees -pines, oaks and other broadleaves- were extracted. The clear-cutting areas include a rectangle space no larger than 1 ha, surrounded by conifer and broadleaf temperate forests.

The residues were ground and piled perpendicularly to the slope; however, only in one of the sites were the residues burnt, as a complementary treatment in order to prepare the site after exploitation. In this type of practices, all the surface (topsoil, dead leaves and undergrowth) and fine aerial materials (branches, thin stems) are burnt; these burn rapidly, so that only thick materials remain on the ground (^{UZACHI, 2003}).

One year after the intervention, in 2010, P. patula and P. pseudostrobus were planted in the two described sites, with an average density of 1 600 individuals per hectare. The utilized plants originated from collected seeds within the same forest in the study area, and were produced in the communal nursery of La Trinidad. Seedling emergence and seedling survival were not monitored after the planting; only clearings were applied in order to control the weeds and shrubs; hence the relevance of assessing these life history characteristics in this study.

Seed bank

Four 100 m² circular plots were located at random. The seed bank in these was first assessed. For this purpose, four 4 m² subplots were established; thus, the total sampling surface for each condition was 16 m². The seeds in found the topsoil and in the first 5 cm of depth of the subplots were counted. The samples were sifted, and the seeds were separated and classified as full or void (^{Daskalakou and Thanos, 1996}; ^{Tíscar, 2007}), using a gravity separator based on the seed weight, and those seeds containing an embryo were regarded as the heaviest and therefore as the ones with the highest germination potential.

Emergence and survival

Subsequently, the entire natural repopulation of pine species present from the start of the study (in December 2011), as well as that which has emerged since 2012 was identified within each 100 m² plot. However, only the data of the seedlings of this last year were registered, whereas the emergence and survival time were recorded on a monthly basis. The causes of mortality, when this occurred, were classified as drought (for cotyledons or the dry structure of the individual), physical damage (for seedlings that died due to damage during the clearing performed in July 2012), and unknown cause (for individuals not found in the sampling sites).

Statistical analyses

The data were analyzed using a generalized linear model (GLM) based on the Poisson distribution with the link log function (^{Kerr et al, 2008}; ^{Otto et al, 2010}). A significance level of α<0.05 was considered; this made it possible to detect differences between the areas with (CQ) and without (SQ) residue burning. The proposed method is the most appropriate for count variables (^{Balzarini et al., 2008}; ^{Kerr et al., 2008}).

First-year seedling survival was analyzed by the SAS LIFETEST procedure survival function (SAS version 9.0). This procedure estimates the probability of survival and carries out a comparison between survival curves in order to detect whether or not there are significant statistics between them, utilizing statistical tests such as Wilcoxons (^{SAS, 2002}). The equation is the following:

Where:

The curves were compared using Wilcoxons test (α < 0.05) (^{Castillo, 2013}). "

Seed bank

A larger amount of seeds was obtained from the seed bank without residue burning (SQ) than from the bank with residue burning (CQ), with significant differences (z = 3.35, p<0.01) (Figure 1). The estimated average number (±DE) of seeds per hectare was 6 875 ± 5 153.88; the average number of full and of void seeds obtained was 5 625 ± 3 750 and 1 250 ± 1 443.37, respectively (Figure 1). In contrast, the average in the treatment without burning was 60 625 ±43 892.62 seeds per hectare; which the mean of full and void seeds corresponds to 28 125 ± 15 728.82 and 32 500 ± 35 997.68, respectively (Figure 1). These results agree with those of studies performed by ^{Korb et al. (2004)} and Creech (2009), according to whom residue burning eliminates or negatively affects the seed bank due to the high temperatures that it generates, of above 270 °C (^{PNUMA, 2005}).

Figure 1 Number of seeds of Pinus spp. species per hectare in areas treated with regenerative clear-cutting with residue burning (CQ) and without residue burning (SQ) in the temperate forest of La Trinidad, Ixtlán, Oaxaca. 

It is important to consider that pines do not form permanent seed banks (^{Tomback et al., 2001}; ^{Carrillo et al, 2009}); besides, the lapse between intervention and burning at the time of the sampling was two years (2010 and 2011). During this period, the rain of seeds from the surrounding trees provided the study areas with a seed deposit -a more important source than the soil seed reservoir (^{Dalling, 2002}; ^{Castillo, 2013}). This may be the reason why CQ plots had a higher percentage of full seeds (81.82 %) than of void seeds (18.18 %), unlike SQ plots, where there are no significant differences between the number of full seeds (53.60 %) and that of void seeds (43.39 %).

The renewal of the seed bank after the burning of the residues and the subsequent basket placement increases the number of seeds in the SQ area; however, their low viability (^{Tomback et al., 2001}; ^{Pausas et al., 2004}; ^{Carrillo et al., 2009}; ^{Marañón et al., 2013}) results in a larger number of void seeds. In both areas, the seed rain is a major contribution to the repopulation of pines, as is shown by various researches (^{Izhaki et al., 2000}; ^{Dalling, 2002}; ^{Larson and Franklin, 2005}; ^{Otto et al, 2010}; ^{Marañón et al, 2013}). Notwithstanding, the average number of seeds from both sites is lower than that determined by ^{Castillo (2013}) for P. patula in the state of Hidalgo, and higher than that observed by ^{Carrillo et al. (2009}) in the state of Puebla; therefore, it would be important to make a projected evaluation of the factors that have an impact on the annual production of seeds of the pine species in the studied area.

Seedling emergence

Significant differences in seedling emergence (z= 10.535, p= 0.0472) and survival rates (z= 10.54, p= 0.0472) were observed between SQ and CQ conditions. The density of emerged seedlings per hectare for the SQ condition was 1 450 ± 640.31, with an average survival of 800 ± 489.89 seedlings ha^-1 at the end of the assessed period.

Conversely, the density of the seedlings emerged per hectare in the CQ condition was 1 000 ± 547.72, with an average survival of 625 ± 450 seedlings ha^-1 (Table 1).This figure is lower than that recorded in pine forests under forest management and in areas where there have been fires. ^{Rebottaro and Cabrelli (2011)} document in Argentina 27 000 seedlings ha^-1 in a P. elliottii Engelm. plantation exploited in strips, or in the P. patula plantation (1 875 ha^-1) studied by ^{Castillo (2013)}, in a site with exploitation of parent trees in the state of Hidalgo, Mexico. It is suggested that the low seed density in CQ is associated to the reduced number of seeds observed in the seed bank. In this respect, ^{Pausas et al. (2003)} consider that a density of less than 4 500 seedlings ha^-1 is low and results from fires affecting the seed bank.

Table 1 Density (±DE) of seedling emergence and survival per hectare in intervened areas with regenerative clear-cutting accompanied by residue burning in La Trinidad, Xiacuí, Oaxaca. 

*CQ = With residue burning; SQ = Without residue burning.

The monthly record of seedling emergence and survival during 2012 evidenced seedling emergence from January to July under both conditions, although the highest incidence occurred in April and May in CQ and SQ, respectively (Figure 2a and ^b). A smaller emergence occurred in September under CQ and SQ conditions, as well as in September and November under SQ conditions (Figure 2b). Seedling emergence is related to adequate temperature and humidity levels for its development (^{Daskalakou and Thanos, 2004}; ^{Lee et al, 2004}; ^{Tíscar, 2007}; ^{Castillo, 2013}). In the study area, humidity does not appear to be a limiting factor due to the prevailing type of climate [C(w₂)(w) big] with a moderate to intense rainy season beginning in May and finishing in mid-October (a monthly average of 137.22 mm), whereas in January through April the mean is 17.3 mm (^{Serrano et al., 2005}). It has been suggested that the highest emergence that has occurred in April and May is related to an increase in temperature, to a change in the photoperiod (11.5 to 12.4 hours, since April), in combination with the availability of moisture (information by the weather station in Ixtlán de Juárez, Oaxaca, for the 1960s and 1995s).

Figure 2 First-year seedling emergence and survival of pine species by condition in sites with regenerative clear-cutting accompanied by residue (a) and without residue burning (b) in La Trinidad, Oaxaca . 

Seedling mortality in CQ occurred in April, June, August, November, and December. Its causes were droughts (33.33 %) and physical damage (20 %), while a higher proportion of dead seedlings was due to unknown causes (46.66 %) (Table 2). Likewise, under SQ conditions, seedling mortality occurred in March, June, July, August and November due to droughts (26.92 %), physical damage (26.92 %) and unknown causes (46.15 %) (Table 2).

Table 2 Number of seedlings (± standard deviation) by mortality cause in intervened areas with regenerative clear-cutting and residue burning in La Trinidad, Xiacuí, Oaxaca. 

Seedling survival

No significant differences in seedling survival were detected between the CQ and SQ conditions. The probability of survival at the end of the assessment period for CQ was 0.441, with 62 % of the living individuals registered at the beginning of the study. In SQ, the probability of survival at the end of the study was 0.502, with 55 % of the registered living individuals (Figure 3). In this regard, ^{Pausas et al. (2003)} suggest that there is not always a clear relationship between seedling survival and the forest fire factor.

Figure 3 First-year seedling survival of pine species in sites with regenerative clearcutting accompanied by residue burning (CQ) and without residue burning (SQ) in La Trinidad, Xiacuí, Oaxaca.  

Survival rates in this study were above those registered for other pine forest areas. ^{Lee et al. (2004)} record a survival rate of 10 % for P. densiflora Siebold & Zucc. after 160 days, and of 0 % after 180 days in forests without a canopy. ^{Tíscar (2007)} indicates a survival rate of 55.8 % after 45 days, and of 5.56 % after 100 days for P. nigra subsp. salzmannii (Dunal) Franco seedlings. However, this is similar to the survival rate documented by ^{Daskalakou and Thanos (2004)} for P. halepensis Miller (60 %), and by ^{Castillo (2013)} for P. patula (53 %). One of the main causes of seedling mortality is drought (^{Otto et al., 2010}). In the case of La Trinidad, Oax., drought also was one of the identified causes, although the largest percentage of mortality could not be ascribed to a specific cause because the seedlings were simply not under observation during the measurement period.

Local forest management recommendations

This study regards the seed bank present in the forest floor in sites with regenerative clear-cutting as a significant precursor to the repopulation of pine species. However, it is also advisable to carry out research on the dynamics of dispersion of seeds from the adjoining stands toward the intervened areas with regenerative clear-cutting, as well as on the germination capacity of the full seeds from both the canopy and the soil banks. Future studies must consider comparing the soil bank seed in unintervened pine forests in order to have a better reference on the natural dynamics of pine species in the area of interest. Besides, due to the variability of the pine species as to seed production (^{Nyland, 1996}; ^{Smith et al., 1997}) and to the fluctuating environmental conditions that are appropriate for their establishment (^{Johnson and Fryer, 1996}), it is advisable to assess the stages of the repopulation process (dissemination, germination, emergence and others) for several years in order to obtain reliable behavioral patterns. It is worth noting that this type of studies do exist for certain species of the Pinus genus (^{Koskela et al., 1995}; ^{Tomback et al., 2001}; ^{Daskalakou and Thanos, 2004}; ^{Hancock et al, 2005}; ^{Otto et al, 2010}; ^{Rebottaro and Cabrelli, 2011}).

The use of natural repopulation in forests under timber exploitation is a mainstay, if forest management toward the preservation of natural genetic variability (^{Larsen and Nielsen, 2007}) is desired. This is relevant, considering that the Northern Sierra region of Oaxaca, where the study area is located, has a high diversity of Pinus spp. (^{Torres, 2004}). Studies like the present one contribute significantly to the incorporation of natural repopulation to local forest management, with or without residue burning.