Tamaulipan thornscrub after fire: an analysis of the composition of species

Forest fires are a natural element in many terrestrial ecosystems, in Mexico one of the main causes of loss of vegetation coverage is related to them. Nevertheless, no research has been carried out in the Tamaulipan thornscrub communities regarding post-fire regenerative strategies of plant species, therefore, is important to generate knowledge about the ecological role of fire in them. The aim of this study is to characterize the arboreal and shrub vegetation of a Tamaulipan thornscrub community in the Northeast of Mexico after a fire. To determine the composition, diversity and structure of the floral community consisting of trees and shrubs, 6 square-shaped sampling sites of 1,600 m2 were established 3 years after the fire. Margalef (DMg), Shannon-Weiner (H’) and Pretzsch (A) indices were calculated to determine the richness, diversity and vertical stratification of species, respectively.23 species, 21 genera and 14 families of vascular plants were registered. The most species were Fabaceae (6) and Rhamnaceae (3) and Cordia boissieri was the specie with greater ecological weight (IVI = 20.3%). The indexes of richness and true diversity (DMg = 3.16; D = 16.44) showed high values, which suggests that the post-fire area has a good regeneration and a high species diversity. With these results we can conclude that the fire is an important agent of change in successional stages of Tamaulipan thornscrub with high values of regeneration after a surface fire.


Introduction
The forest fires are a natural element in many terrestrial ecosystems and they directly influence on their structure, composition and ecological dynamics (Agee, 1993;Pyne et al., 1996;Iñiguez et al., 2008;Carrillo et al., 2012). After a forest fire has been extinguished, open spaces are generated by the elimination of vegetation cover, which enhances the germination and regrowth of pre-existing individuals or present species in the soil seed bank (Pausas, 2004;Calvo et al., 2008;Sang et al., 2008;Vidal and Reif, 2011;Massi et al., 2017).
In Mexico one of the main causes of loss of vegetation coverage is related to forest fires, mostly due to anthropogenic causes, approximately 90% of recorded fires (México, 2006;Rodríguez-Trejo, 2008). Its relevance is attributed to direct contributions to soil erosion  and to the modification of the community structure (González Tagle et al., 2016). Therefore, is important to generate knowledge about the ecological role of fire in terrestrial ecosystems, in order to contribute to development plans and programs of sustainable management, adapted to specific ecosystems .
Tamaulipan thornscrub (MET) is an abundant ecosystem historically used by local people in northeastern Mexico, in activities such as logging and forage for cattle; it has high specific richness and density of shrubs and trees (Heiseke and Foroughbakhch, 1985;Alanís Rodríguez et al., 2008;Mora Donjuán et al., 2013).
Due to the lack of knowledge about the species fire ecology and the responses of MET to wildfires, it is important to study the fire effects over the MET regeneration and how is the resulting ecosystem in terms of forest structure and species composition. We believe that the regeneration of the vegetation after a wildfire can register high ecological values and similar records of species diversity and vertical distribution. Therefore, the aim of this study is to characterize the arboreal and shrub vegetation of Tamaulipan thornscrub in the Northeast of Mexico after a fire, through: 1) ecological parameters of abundance (ARi), dominance (DRi), frequency (FRi) and the importance values index (IVI); 2) species diversity indexes such as those of Shannon and Margalef, and 3) the vertical species distribution index (A).

Study area
The present research was carried out in a Community of the Tamaulipan thornscrub (MET) in the municipality of Montemorelos, Nuevo Leon (northeastern Mexico). The study area is located between the coordinates 25° 2' north latitude and 99° 41' west latitude (Figure 1), with an altitude of 410 m.a.s.l., with arid desert-hot climate (600 and 1000 mm of annual precipitation) with the highest incidence of rain in the month of September, and an annual average temperature of 18 °C. It belongs to the system of lower hills with soft slope, where the predominant soil types are the regosols.
In the summer of 2010 occurred a surface and crown mixed fire, affecting the vegetal elements. The fire extension was about 3,600 hectares and burned the aerial parts from most the trees and shrubs, with the exception of the Yucca filifera Chabaud species, due to its height, it did not suffer damage to the leaves, as observed in Figure 2.

Vegetation analysis
In April 2013, 6 square-shaped sampling sites of 1,600 m 2 were established into a stand of Ejido "Las Huertas". At the sampling sites, the regeneration of shrubs and trees was evaluated 3 years after the fire occurred. The distribution of the sampling sites was at random scheme.
To obtain the minimum number of sampling sites, a species-area curve was elaborated (Mostacedo and Fredericksen, 2000). This was verified with a Bray-Curtis ordination model, a graphic representation of the variation in vegetal composition, known as one of the methods to evaluate beta-diversity. The model is based on an algorithm that permits to analyze the similarity of the sampling sites, calculating the percentage of similarity (0-100%) in a distance unit (Beals, 1984), in this case it was developed using the program BioDiversity Pro 2.0 (McAleece et al., 1997).
At each sampling site a census of shrubs and trees was carried out, taking into consideration the basal diameter at ten centimeters of height (d 0.10 ) and the number of sprouts presented by each individual, the crown diameter through a tape measure, measuring the occupied space in north-south and east-west directions and the total height. Moreover, a botanical identification of all evaluated species was performed in the herbarium of the Faculty of Forestry Sciences of the Autonomous University of Nuevo León.

Data analysis
The abundance of every species (A i ) was determined, according to the number of trees per hectare, their dominance (D i ), as a function of the crown area (AC), and their frequency (F i ) based on their presence in the sampling sites. The results were used to obtain a taxon-weighted value called Relative Importance Value Index (IVI), which acquires percentage values on a scale of 0 to 100 (Mostacedo and Fredericksen, 2000).
The formulas used for calculating the indices are described below.
Absolute (A i ) and relative (AR i ) abundances were evaluated using the following Formulas 1 and 2: where: A i = is the absolute abundance; AR i =is the relative abundance of species i relative to total abundance; N i = is the number of individuals of species I; S = the sampling surface (ha -1 ).
The absolute (D i ) and relative (DR i ) dominance was evaluated by the following Formulas 3 and 4: where: D i = is the absolute dominance; DR i =Is the relative dominance of species i with respect to total dominance; AC = The crown area of species I; S = the surface (ha -1 ).
The absolute (F i ) and relative (FR i ) frequencies were obtained with the following Formulas 5 and 6: where: F i = is the absolute frequency; FR i = Is the relative frequency of species i with respect to the total frequency; P i = Is the number of sites where species i is present; NS = The total number of sampling sites. The importance value index (IVI) is defined by Moreno (2001), as shown in Equation 7: To evaluate the diversity of regenerated species, the method proposed by Jost (2006), who recognizes "true diversity" using effective species numbers, is used as a modification to the Shannon entropy index. For this analysis the order q = 1 was taken into account, where the proportional abundance of each species is considered (Jost, 2006). The Equation 8 is represented as: where: 1 D is true diversity, exp is the exponential, S represents the number of species present, ln equals Natural logarithm of pi which is the proportion of species, pi=ni/N, where ni is the number of individuals of the species i and N is the total number of individuals.
To estimate species richness, the Margalef index (D Mg ) was used by equations (Magurran, 1989).
where: S is the number of species present, N is the total number of individuals and ni is the number of individuals of species i. For the characterization of the vertical structure of the species, the Pretzsch index (A) was used (Pretzsch, 1996;Del Río et al., 2003). This index uses a vertical stratification of the forest in which the height of the forest is divided into three zones (Pretzsch, 1996;Jiménez et al., 2001); Zone I: 80-100% of the maximum height of the area; Zone II: 50-80%, and zone III: 0 to 50%. To define the zones the height of the individual is taken as high as 100% and from there the others are estimated. The A index uses values between 0 and a maximum value (A max ); a value A = 0 means that the stand consists of a single species occurring in a single stratum. A max is reached when all species occur in the same proportion both in the stand and in the different strata (Pretzsch, 1996;Corral Rivas et al., 2005).
This index (A) serves to determine the structural diversity in the vertical distribution of the species and is calculated with the following Formula 10: where: S = Number of species present; Z = Number of height stratas; p ij = Percentage of species in each zone, and is estimated by the following Equation 11: where: n i,j = Number of individuals of the same species (i) in zone (j); N = Total number of individuals.
In order to compare the Pretzch index it is necessary to standardize it and this is done by the value of A max , which is calculated as shown in Equations 12 and 13: Then the value of A can be standardized according to: From the previous, it can be made comparisons between stands in which different species occur.

Floristic composition
There were found 12 orders, 14 families, 21 genera and 23 species. The families Fabaceae and Rhamnaceae registered more species with 6 and 3 respectively, followed by the Euphorbiaceae and Rutaceae families with two species each; for the remaining families only one species was recorded (see Table 1). According to the Bray-Curtis ordination model, it was determined that the evaluated areas have an intermediate floristic similarity (Figure 3). In every comparison a similarity between 45.16% and

Ecological parameters
Abundance. For the number of individuals per hectare, three species were the most abundant. Cordia boissieri A.DC. was the most abundant species with 105 individuals per hectare (N ha -1 ), followed by the Vachellia rigidula (Benth.) Seigler & Ebinger with 75 Nha -1 and Leucophyllum frutescens (Berland.) I.M. Johnst. with 71.9 N ha -1 (as shown in Table 2). These three species represents the 34.2% of the total abundance in the community.  Dominance. Cordia boissieri is the most dominant species and covers 236.5 m 2 ha -1 which represent a 39.3% of the studied area. Vachellia rigidula with 74.8 m 2 ha -1 corresponded to 12.4% of the studied area, and Havardia pallens (Benth.) Britton & Rose covered 63.8 m 2 ha -1 with a relative dominance of 10.6%. It is important to highlight that the previous three species represented 62.3% of the total community coverage, while the remaining 20 species represent 37.7%.
Importance value index (IVI). The greatest ecological value, obtained by the importance value index was 20.3% represented by C. boissieri species of IVIrel. The following species in terms of importance were V. rigidula with a 10% and H. pallens with 8.9%. The previous three species together sum up 39.2% of IVIrel. The species with lower ecological value is the Y. filifera and P. leavigata with only 0.5% and 1%, respectively of IVIrel (as shown in Table 2).

Biological diversity
According to the Margalef richness index (D Mg ), the evaluated area obtained 3.16 of D Mg . In relation to True Diversity index ( 1 D), it showed a value of 16.44.

Vertical distribution of species (A)
The analysis of the vertical distribution was made by defining three layers, high, medium and low. The upper layer is formed by Yucca filifera with 2 individuals per hectare (N ha -1 ) in this classification. The middle layer is formed by three species Eysenhardtia polystachya (Ortega) Sarg., Vachellia rigidula and Cordia boissieri with 2, 2 and 3 N ha -1 respectively (as shown in Table 3). In the low layer, 99% of the individuals, including all species  Table 3). The value of the A index was 2.84 with an Amax value of 4.23.

Discussion
In the present investigation, Fabaceae represented the highest number of species, similar to that reported by Alanís Rodríguez et al., (2008) andJiménez Pérez et al. (2012) for the MET and by Gómez et al. (2009) and Hahn and Gómez (2008) in other shrublands, who stated that this family is the one with the highest number of species in these ecosystems. The 23 species recorded indicate a similarity in richness compared to mature MET communities, as reported by Mora Donjuán et al. (2013) and Marroquín-Castillo et al. (2016), who reported 21 and 23 species, respectively for MET reference areas. Of the aforementioned species, V. rigidula is the only species with the highest abundance in MET areas with different types of disturbances such as extensive cattle grazing and killing (Alanís Rodríguez et al., 2008). The highest values reported for this parameter in the MET in different areas with livestock and agricultural history are frequently supported by the species V. farnesiana and V. rigidula (Alanís Rodríguez et al., 2008;Pequeño Ledezma et al., 2012;Jiménez Pérez et al., 2012), being the latter one occuring in the burned area and with high values of importance (see Table 2). Different from Mora Donjuán et al. (2013) for an area with no recent disturbances, where the dominant species were Vachellia amentacea DC., Diospyros texana Scheele and C. boissieri, it is important to highlight that C. boissieri with the greatest dominance was present at the site of fire. On the other hand, the less dominant species belongs to the family Liliaceae represented by Yucca filifera, with 0.1 m 2 ha -1 of coverage, followed by Prosopis aculeata and P. laevigata with 0.7 and 3.3 m 2 ha -1 respectively (see Table 2). It is important to mention that the total covered area recorded was 602.2 m 2 ha -1 , which represents an extremely low coverage, since 93.7% of the surface is found without vegetation due to the fire. The importance value index of the present study reports C. boissieri as the most important species (see Table 2), since in other MET study C. boissieri has not been reported with the highest importance value index (Alanis Rodríguez et al., 2008;Jiménez Pérez et al., 2012;Pequeño Ledezma et al., 2012;Mora Donjuán et al., 2013). The previous has great relevance, since it supposes that it is a pioneering post-fire species of the Tamaulipan thornscrub. Although there is no specific information on C. boissieri's regrowth capacity, it is documented that the genus Cordia has good resprouting capacity after fire (González Tagle, 2005;Alanís Rodríguez et al., 2011).
For the estimation of biodiversity, the Margalef index indicates a high species richness (D Mg 3.16) when compared to other regenerated areas of the MET Mora Donjuán et al., 2013) or studies in temperate forest ecosystems such as Alanís Rodríguez et al. (2010), with 1.59 and González Tagle (2005) with 1.5 as maximum values in this index. For the true diversity, which encompasses the abundance of species, it shows a value of 16.44. This means that the evaluated plant community has a diversity equal to that which would have a theoretically community of 16.44 species where all of them had the same abundance (Jost, 2006). Prezsch values indicate that the community evaluated had a high presence of individuals in the lower stratum (III) and a low presence in the rest of the heightzones. Based on the table of vertical strata and the results of Pretzsch index A, it was determined that the plant community is formed by a single cohort (III), where the individuals suffered the loss of the aerial part after the fire and are currently resprouting from the base. The remaining individuals in higher Cohorts I and II are those who survived the fire.
With our results we highlight the following: the studied community showed a specific richness of 23 species which is a high value when compared with other studies. The Cordia boissieri recorded a 20.3% of IVI being the species that showed a greater ecological weight in the study area. In addition, it is concluded that this species is pioneer associated with the fires in this area. The indexes of richness and true diversity (D Mg = 3.16; 1 D = 16.44) showed high values, this suggests that the post-fire area has a good regeneration and a high species diversity. Regarding the species vertical distribution index in the burned area prevails the stratum III (10 cm -2.23 m), this indicates that the surface fire affected the aerial part of trees and shrubs and currently there is a dominant strata represented by individuals with regrowth.