Description of the study area

The research was carried out in the community of Xaltepuxtla, belonging to the municipality of Tlaola, Puebla. It has an extension of 40 hectares that are part of the River Basin of the Necaxa River. This zone is located between the extreme coordinates 97° 58’ 5.303” west longitude and 20° 11’ 23.06” north latitude, 97° 57’ 30.836” west longitude and 20° 10’ 57.124” north latitude, at an altitude of 1 280 m. The climate is semi-warm, with little annual oscillation of the average monthly temperatures. The average annual rainfall is 1 992.4 mm and the average annual temperature is 17.5 °C. (SMN, 2016). The predominant soil type is mollic andosol and slight slopes until very pronounced ranging from 2% to more than 45%. The conditions under study are ornamental system (SO) with 37.3 ha, silvopastoral system (SS) with 0.43 ha and remnants of mesophilic mountain forest (RBMM) with 2.27 ha.

Sampling design

The polygonal and surfaces of each condition was delimited by digitizing the study area with the COGO tool of Arc map 10.1; from azimuth and distance from the boundaries of the property, rectified in the field with transects with a Garmin global geo-positioner (GPS) (^{Ruiz, 2016}). The sampling design used was systematic sampling with predetermined distances and in arrangements in a manner of grid (^{Ralph et al., 1996}), obtaining 22 units of choice or sites with distances between them of 141.4 m. Vegetation was sampled using quadrant methods with central point for tree vegetation (^{Cottam and Curtis, 1956}) and embedded boxes for shrub and herbaceous vegetation (^{Oosting, 1956}).

Variables evaluated

In each of the sites, cartesian coordinates were recorded with a GPS, as well as habitat variables such as exposure, slope and thickness of mulch and dasometric variables such as total height (H), clean shaft height (HFL) and diameter at 1:30 m. of the ground (DN). In the case of shrubs, the crown diameter (DC), number of branches for those species with foliage use, height (H) and number of individuals were measured for each individual, for the herbaceous, only the percentage of land covered was measured by them.

In each site, specimens of the different species found were collected for subsequent identification in the herbarium of the Division of Forest Sciences of the Autonomous University Chapingo (UACH).

Analysis of data

Frequency of observation. To know the percentage frequency of species wealth by condition and by site, the observation frequency index (FO) (^{Salas and Orduña, 1993}) was used, with the total number of woody individuals registered per species among the total number of individuals per 100.

Wealth of species. The first-order non-parametric Jacknife estimator was used, using presence-absence or abundance data by species (^{Colwell, 2004}), which represents one of the most accurate and least biased indexes (^{Palmer, 1990}), using the computational program ‘EstimateS 6b1a’.

Relative abundance. To estimate the relative abundance of the species, present in the different sites and conditions, the abundance index defined by ^{Carrillo and Cuarón (2000)} was calculated; ^{Aranda (2000)}, modified for this study.

Diversity of species Based on a review of the characteristics and assumptions of various species diversity indices (alpha diversity), the non-parametric index known as the Shannon-Wiener diversity index was chosen, using the ‘EstimateS’ version 6b1a (^{Colwell, 2000}). To determine significant differences in the wealth, relative abundance and diversity of species recorded by condition and by site, a non-parametric test of Kruskal-Wallis (^{Zar, 1999}) was performed, using the statistical software JMP IN version 4.

To determine if all species were considered in wealth, relative abundance and diversity of species recorded by condition and site, a ‘Chi square of Pearson’ analysis was carried out, (^{Molinero, 2003b}): with a significance of α= 0.05, said analysis was used to assess whether the differences between the observed and expected frequencies are attributed to chance, under the hypothesis of independence. This analysis was carried out using JMP IN software version 4.

Poisson regression analysis (ARP)

An analysis was carried out in order to establish possible associations between the dasometric and habitat variables on the abundance of the woody species with the highest ecological value recorded in the three conditions evaluated. First, the effect between habitat and dasometric variables and the abundances of different woody species with greater ecological value was analyzed, using a Poisson regression model (^{González-Oreja, 2003}).

The variables that explain the greater variability to select the model were then determined (^{Rotenberry and Wiens, 1980}). Thus, the abundance of the woody species (y) was modeled before the dasometric and habitat variables (x), making an adjustment of the variable ‘y’ to the variables ‘x’, using a generalized linear model (^{McCullagh and Nelder, 1989}) that will explain the relationship between a response variable ‘y’ and a set of explanatory variables ‘x’ through a linear relationship.

In this analysis, those in which p< 0.05 (^{McCullagh and Nelder, 1989}; ^{González-Oreja, 2003}) were considered statistically significant coefficients. In this way, woody species were obtained that had an effect on independent variables (dasometric and habitat). This analysis was performed using the GLM procedure of the statistical software R-version 2.6.1. (^{Dalgaard, 2006}).

Species wealth

19 herbaceous species, 9 shrubs, 9 trees and one tree were recorded; all of them distributed in 15 families and 18 genders (Table 1).

20 species were found so that, with the sampling effort carried out, 19% of the species theoretically present in the sites considered are known (Figure 2).

Figure 2 Rarefaction of the diversity of woody species in the 22 study sites. 

In the wealth of species by condition an estimated 15 species is indicated, so up to now, with the sampling effort carried out, 100% of the theoretically present species in this condition are known, for the silvopastoral system 2 species were registered, which It shows a sampling effort of 10% and for the RBMM, 9 species were observed, indicating a sampling effort of 29% (Figure 3).

Figure 3 Rarefaction of the diversity of woody species in the three study conditions.  

The present work uses rarefaction curves for the first time with the Jackniffe 1 estimator for silvopastoral systems, so that beta-level registrations are obtained in this type of forest. The trends in recorded wealth coincide with what was reported by ^{Chiarucci et al. (2003)} who point out that data from large samples analyzed using Jacknife 1 present more precise values of wealth.

Relative abundance index (IAR) by sites

The results of the IAR for the frequency data recorded by species in all sites by condition, suggest that the species of greater relative abundance were: Chamaecyparis. lawsoniana (IAR= 0.96), Rhododendron simssi (IAR= 0.46), Chamaecyparis. thyoides (0.25), the values of the rest of the species show lower values (Figure 4a). By condition Chamaecyparis lawsoniana presented greater abundance (IAR= 0.96) in the ornamental system, (IAR= 1.86) for silvopastoral and (IAR= 1.09) for RBMM, the value of the rest of the species exhibited a differential trend in each condition (Figure 4b).

Figure 4a and b Relative abundance of the species registered in the study sites and conditions. 

Species diversity

The results of Shannon-Wiener for the data of the different conditions under study suggest that the one with the greatest diversity was: the RBMM (H’= 2.92), followed by the silvopastoral system (H’= 2.61) and finally the ornamental system (H’= 2) (Figure 5).

Figure 5 Diversity of woody species in one of the conditions. 

Statistical analysis

The results of Kruskal Wallis for data on the wealth, abundance and diversity of woody species per site in the conditions under study suggest that there are no significant differences, therefore, the wealth, abundance and diversity of species found in the 22 sites are they are subject to the same levels of disturbance (Table 2).

X² square test for sites

The results of X² square for data on the wealth, abundance and diversity of woody species per site, under the conditions under study, suggest that not all the species that potentially exist in the study area were recorded (Table 3).

X² square for conditions

The results of X² square for the conditions evaluated suggest that not all species that potentially exist in the study area with R square of 1 and a Pearson value of 0.19 for each of the conditions were recorded.

Species of higher ecological value

The selection of the species with the highest ecological value was made with the indices of wealth, abundance and diversity, through the analysis and interpretation of these indices, the wealth allowed us to know the species existing in each study condition, the relative abundance the number of individuals by species and diversity the condition of each plant community (Table 4).

Four dasometric and habitat variables presented a degree of association with the abundance of the registered woody species, with an AIC= 174.94 (Table 5).