Study area

The RBSAT is located in the north-western middle portion of the Great Folded Sierra or physiographic province of the Sierra Madre Oriental, in Ciudad Valles and Tamuín municipalities in the state of San Luis Potosí. It covers an area of 21 464 ha that extends between the coordinates 22°05 'and 22°24' N and 98°52 'and 99°01' W (^{Semarnap, 1994}) (Figure 1). Its area of influence is made up of nine localities: Laguna del Mante, Los Sabinos Número Dos, Gustavo Garmedia (La Unión), León García, El Sabino del Obispo, El Aguaje, Las Palmas, Estación Tamuín and Fracción la Tima (^{Durán, 2018}).

Simbología = Symbology; Área de estudio = Study area.

Figure 1 Sierra del Abra Tanchipa Biosphere Reserve and its influence zone. 

The predominant climate in the region corresponds to the warm subhumid with summer rainfall Aw1(x ') with an average annual rainfall of 965 mm and an average annual temperature of 24.5 °C (^{Inegi, 2002}; ^{García, 2004}). The Reserve is located in hydrological region 26 (RH26) Pánuco River, corresponding to the Tamuín River basin, although surface currents are scarce, it has underground rivers connected through various basements propitiated by fracturing and infiltration of water into calcareous rocks (^{Inegi, 2002}). The predominant soils correspond to the Phaeozems and Leptosols (^{Inegi, 1983}).

The dominant vegetation corresponds to the low deciduous and medium sub-deciduous forests. The first is characterized by having individuals in the arboreal stratum of 4 to 15 m that lose their leaves in the dry season, and the second by having trees up to 25 to 30 m with a dense foliage (^{Rzedowski, 2006}; ^{Miranda and Hernánez, 2016}).

Present day distribution

Consultation with taxonomists and vegetation experts was carried out, as well as a review and verification of Quercus oleoides records for the state of San Luis Potosí, in the Isidro Palacios (SLPM) Herbarium of the Instituto de Investigación de Zonas Desérticas (IIZD) de la Universidad Autónoma de San Luis Potosí (Desert Areas Research Institute (IIZD) of the Autonomous University of San Luis Potosí). The records of the species in the RBSAT were corroborated and the collection sites were georeferenced.

With these data and with the support of local guides and knowledgeable experts of the regional flora, seven field trips were made in the Reserve and its area of influence to verify the presence of the species and obtain complementary information of the sites where it grows. Additionally, 15 semi-structured interviews were applied to some inhabitants of the Laguna del Mante, Gustavo Garmendia and Los Sabinos ejidos, to know the main uses of the species, former distribution areas, its importance for the locals and actions that could be implemented for its conservation.

The points identified during the field work routes, as well as the data from the collection records were represented in the Arcgis Geographic Information System 10.3., to perform a spatial analysis and determine the main environmental characteristics of the sites where the species are found. Therefore, the following factors were considered: altitude, slope, slope orientation, soil type, geology and land uses. These layers of information with a resolution of 30 m in pixel size were obtained from the digital service of Inegi.

Potential distribution

To generate the model, seven geographic records of the species obtained in the herbarium were used, as well as data from 56 georeferenced sites in the field. The main environmental variables considered in the analysis were: i) annual average temperature; ii) annual precipitation (1970-2010) obtained from the Worldclim database (^{Fick and Hijmans, 2017}); iii) geology; iv) edaphology (^{Inegi, 2006}); v) altitude and vi) slope, generated from the digital terrain elevation model (CEM 3.0) (^{Inegi, 2013}). To estimate the contributions of each variable, as well as their respective parameters, the maximum entropy algorithm of the Maxent 3.4.1 program was used.

During the execution of Maxent, 35 randomized training points of maximum sensitivity were determined. The evaluation and contribution of each environmental variable to the prediction was based on the estimation of the percentages of relative contribution of Jackknife that define the most important variables for the model (^{Phillips et al., 2006}). Likewise, the evaluation test of receiver performance characteristic curve (ROC) and the area under the curve (AUC) were used, which establish the discrimination capacity of the model and the values of aptitude and performance (^{Hanley and McNeil, 1982}). The values of the model are interpreted as insufficient in a range of 0.50-0.70; poor of 0.71-0.80: good of 0.81-0.90 and excellent of 0.91-1.0 (^{Araujo and Guisan, 2006}).

Areas with aptitude for conservation

A multicriteria spatial evaluation (CME) was carried out, considering continuous and categorical variables, for the selection of elements that allowed to measure and evaluate. This implied the multiplication of the value of each standardized criterion at a common scale and a range of weighting, in order to have multiple alternatives based on the previous principles (^{Malczewski, 1999}; ^{Eastman, 2012}).

The variables chosen for the analysis were grouped into three components: environmental, anthropogenic pressure and subsidies and social lag (Figure 2) and were based on the criteria by ^{Moffett and Sarkar (2006)}, ^{Fabbro (2014)} and ^{Hernández (2014)}. These were: connectivity, slope, elevation, distance to the Reserve, land use, susceptibility to fires, degree of degradation, distance to agricultural areas (^{Inegi, 2017}), distance to roads, distance to localities (^{Inegi, 2011}), subsidies (payment for environmental services, PROCAMPO and PROGAN) and social lag index. These variables were spatially represented in the Arcgis 10.3 program, in which the UTM geographic projection was assigned. Later they were transformed into a raster format and represented in criteria maps. These layers were exported to the Idrisi Selva program, where the units of measure of all the criteria were standardized on the same numerical scale.

Componentes para la conservación = Components for conservation; Condiciones para la conservación = Conditions for conservation; Componente ambiental = Environmental component; Conectividad = Connectivity; Pendiente= Slope; Elevación = Elevation; Distancia a la Reserva = Distance to the Reserve; Uso de suelo y vegetación = Land use and vegetation; Rasgos antrópicos = Anthropic traits; Componente de presión antrópica = Anthropic pressure component; Susceptibilidad a incendios = Susceptibility to fires; Distancia a zonas agropecuarias = Distance to agricultural areas; Distancia a vías de acceso = Distance to roads; Distancia a localidades = Distance to localities; Grado de degradación = Degree of degradation; Efectos para la conservación = Effects for conservation; Componente subsidios y rezago social = Component of subsidies and social lag; Pago por servicios ambientales = Payment for environmental services; Índice de rezago = Social lag index.

Figure 2 Conceptual scheme used for the assessment of priority areas for conservation. 

In the multicriteria evaluation module, the components were integrated in a weighted linear summation according to the following formula:

Where:

S = Cell fit

W _i = Asigned weight to the i factor

X _i = i factor value

The estimation of the weights and the weighting of each criterion involved the consult of experts, technical personnel of the RBSAT Management and inhabitants. Fifteen opinion matrixes and added value were applied for each criterion. The data obtained was analyzed and verified by Analytic Hierarchy Process (AHP), which consists of evaluating direct values or pairwise comparisons and inconsistencies (^{Saaty, 2008}). The values represented in a numerical scale (1 to 9) according to their importance, were incorporated into a comparison matrix (^{Eastman, 2012}). The weights were integrated and calculated in the Decision Support Weight module, to average the results of all the criteria at a consistency level lower than 0.1 (^{Clark Labs, 2015}).

The standardization consisted of transforming each criterion into comparable units in the Fuzzy module at a byte scale (0-255), where 0 represents the minimum value of the alternative and 255 the maximum value for conservation. The values of the comparison matrix (AHP) derived from each questionnaire were agreed to calculate the weighted values of relative importance. Starting from it, a cartographical composition of non -optimun values <44 and very optimum >195 were derived for conservation. Later, the information was standardized in four levels of importance (unfit, low, medium and high).

Distribution and use of Quercus oleoides

In the RBSAT and its area of influence, 56 sites with Quercus oleoides were identified, mainly in the ejidos of Laguna del Mante, Gustavo Garmendia and Los Sabinos Dos (Figure 3). In these spaces, shallow soils (Rendzic leptosols) predominate, altitudes below 350 m, flat terrains or with gentle slopes located in the lower parts of the sierra. The holm oaks of the RBSAT mostly develop on rocky strata of limestone and limestone-shales.

Zona de estudio = Study zone; Área núcleo = Core área; Cuerpo de agua = Body of wáter; Localidad urbana = Urban locality; Curvas de nivel = Level curves; Vías de comunicación = Communication roads; Brecha = Dirt road; Terracería = Dirt; Carretera estatal = State highway; Carretera federal = Federal highway; Vía férrea = Railroad track; Sitios con presencia de Quercus oleoides = Sites with the presence of Quercus oleoides.

Figure 3 Presence of Quercus oleoides Schltdl. & Cham. in the RBSAT and its area of influence. 

In most of the studied sites few adult individuals (3 to 5) are regularly found. The oldest live in the Los Sabinos ejido with average heights of 35 m and DAP of 331 cm, while in Gustavo Garmendia some sites gather specimens up to 25 m in height and DAP of 220 cm. The youngest and the lowest are present in the ejidos of Laguna del Mante and Gustavo Garmendia where, in addition, sprouts are numerous. An environmental gradient related to the physiognomic characteristics of the individuals was identified, that is, those with the greatest size are grouped in the southern portion of the RBSAT, which has higher humidity and precipitation (Figure 3).

According to the information gathered in the interviews, in addition to the large areas covered by the oaks that were removed to give way to agricultural fields and pastures cultivated between 1960 and 1970, this species was widely used by the inhabitants of the area, to provide themselves with charcoal, and wood for the construction of houses, canvases and furniture making. Forest fires are another reason that villagers recognize as causes of their disappearance, which is argued by some authors such as ^{Zavala (2000)}, ^{Peña and Bonfil (2003)} and ^{Encina et al. (2007)} who state that fire is a factor that favors the presence of certain oak species.

Once these forests are distroyed, the presence of fires or periodic burning and browsing of cattle in the undergrowth, make their recovery even harder to achieve (^{Flores, 2007}), by means of growth or regeneration from the stumps, as the species’ rhythm is rather slow.

This could explain why the oaks of interest only persist as dispersed relicts in the less accessible and more restricted areas of the area of influence of the RBSAT. Most of the human communities are unaware of the effects associated with deforestation, but recognize that it is a native species, which must be protected.

Currently, oak wood is used in construction of houses, furniture in general, as firewood, in addition to the production of charcoal and domestic utensils, which demonstrates the importance of this species in the tropical regions of the country (^{Pérez et al., 2000}; ^{Pennington and Sarukhán, 2005}; ^{Rzedowski, 2006}; ^{Pérez and Dávalos, 2008}; ^{Montoya, 2009}).

Potential distribution

Among the different algorithms for predicting ecological niches, Maxent is one of the most widely used and its main applications allow the identification of protection and conservation zones of threatened species, the potential distribution of flora and fauna, the effects of climate change and the distribution potential for pests and infectious diseases (^{Guisan and Zimmermann, 2000}; ^{Dudík et al., 2004}; ^{Elith et al., 2006}; ^{Phillips et al., 2006}; ^{Phillips and Dudík, 2008}; ^{Austin, 2007}; ^{Mateo et al., 2011}; ^{Cruz et al., 2014}). Therefore, it was chosen to determine the potential distribution of Quercus oleoides.

The Jackknife test, which determines the most important variables for the model (^{Phillips et al., 2006}), showed that the probability of presence of the species of interest increases in areas with shallow soils, as well as in areas of smooth or flat slope and when the altitude is less than 600 m. Of the total of the georeferenced sites in the field, 76 % are located at altitudes between 300 and 325 m. Likewise, 89 % of the field records indicate that the species develops on flat or almost flat lands, on slopes no greater than 14° where the soils are usually deeper due to the accumulation of materials. In both cases, they coincide with the points made by ^{Rzedowski (2006)} and ^{Pennington and Sakukhán (2005)}.

For the model, edaphology made the largest contribution with 47.3 %, which coincides with that indicated by ^{Márquez et al. (2005)}, ^{Pennington and Sarukhán (2005)} and ^{Rzedowski (2006)}, followed by slope (17.8 %) and altitude (12.7 %). The rest of the variables contributed less than 10 %, and although the temperature and precipitation did even less, it is considered that the warm humid climates, with an annual rainfall of 965 mm and an average temperature of 24.7 °C, influence the presence of tropical holm oaks and explain their similarity to tropical forests (^{Pennington and Sarukhán, 2005}).

The prediction capacity of the data in the distribution model showed an AUC (ROC curve) of 0.988, with a standard deviation of 0.002, interpreted as an optimal test prediction (^{Phillips et al., 2006}). As a result, Quercus oleoides could have spread over an area of 15 528 ha parallel to the sierra, from the limits of Ciudad Valles to the La Lajilla dam (Figure 4). The most favorable areas for its eventual recovery are located northwest of the RBSAT, which coincides with that reported by ^{Sabás et al. (2015)}, who recognize their presence in the area.

Zona de estudio = Study zone; Área núcleo = Core área; Cuerpo de agua = Body of wáter; Localidad urbana = Urban locality; Curvas de nivel = Level curves; Vías de comunicación = Communication roads; Brecha = Dirt road; Terracería = Dirt; Carretera estatal = State highway; Carretera federal = Federal highway; Vía férrea = Railroad track; Distribución potencial de Quercus oleoides = Potential distribution of Quercus oleoides; Valor = Value; Alto = High; Bajo = Low.

Figure 4 Potential distribution of Quercus oleoides Schltdl. & Cham. in the RBSAT and its area of influence. 

In the potential distribution by ejido, the largest area belongs to Laguna del Mante, followed by Gustavo Garmendia and Los Sabinos. According to ^{Zavala (2000)}, the anthropogenic factors are decisive for the presence of the species. Observations and records in the field allow inferring that human pressure and agricultural activities condition, to a large extent, the absence of this species in some places, despite the existence of the right environment for their development.

Areas with aptitude for conservation

The identification of areas with aptitude for conservation is based on a series of stages that includes the selection of quantifiable and variable indicators to discriminate options, whose purpose is to identify places with conditions and characteristics that should be conserved (^{Hernández, 2014}).

The procedure involves selecting criteria that can be measured and evaluated to take a decision (^{Eastman, 2012}). A cartographic composition was derived from values <44 non-optimal and> 195 very optimal for conservation outside the protection polygon of the RBSAT. Subsequently, the information was classified into four levels of importance (not suitable, low, medium and high). The unfavorable areas for conservation were discarded, as they correspond to agricultural, livestock or urban land uses (Figure 5).

Sitios con presencia de Quercus oleoides = Sites with the presence of Quercus oleoides; Zona de estudio= Study zone; Aptitud = Aptitude; Alto = High; Medio = Medium; Bajo = Low, No apto = Not suitable.

Figure 5 Areas with greater aptitude for the conservation of Quercus oleoides Schltdl. & Cham. 

The areas with values of 155-216 have high potential for conservation and are associated with ejidos supported by payment for environmental services, greater connectivity of the landscape and distance to the protected area.

The next level (115-155) indicates an average probability and is mainly associated with the connectivity and distance aspects of the protected area, which are basic for preserving priority areas (^{Moffett and Sarkar, 2006}; ^{Fabbro, 2014}; ^{Hernández, 2014}). Values below 114 are unfit or unfavorable areas for conservation, due to their proximity to areas with agricultural activities and population growth (^{Challenger and Caballero, 1998}; ^{Flores, 2007}; ^{FAO, 2016}).

The areas with greater aptitude for conservation are located to the southwest of the Reserve, in the ejidos of Laguna del Mante, Gustavo Garmendia, Los Sabinos and Las Palmas with an estimated area of 198 ha. The components that most contribute to the comparison matrix (AHP) are the environmental component by 37 %, the subsidies and the social backwardness with 34 %.