Distributional Database

We gathered presence records of 54 Quercus species reported for the state of Oaxaca. Thirty-one species were red oaks (section Lobatae), and 23 species were white oaks (section Quercus). Our species list (Table 1) somewhat differed from the list compiled by Valencia and Nixon (2004) and is discussed in detail in Appendix A. The presence records were obtained from herbarium specimens (MEXU, FCME, OAX, SERO, MO), Global Biodiversity Information Facility ( www.gbif.org), and data from monographic and floristic studies reviewed by Torres et al. (2011). The information was condensed into a single database of 2,143 non-duplicated georeferenced records. Distributional maps of each of the 54 oak species were obtained using ArcGIS ver. 9.2 (Environmental Systems Research Institute, 2006).

Table 1.

List of Oak Species Recorded for Oaxaca by Section. The Third and Fourth Columns Indicate if These Species Are Endemic to Mexico or to the State of Oaxaca.

Study Area

Our study covers the entire state of Oaxaca, located in the southern portion of Mexico between 15° 39′ and 18° 42′ N and 93° 52′ and 98° 32′ W, with an area of 95,364 km² (Espejo, López, Martínez, & Pulido, 2007; Velázquez et al., 2003). The largest number of oak records are located in five regions (biogeographic subprovinces) proposed by Ortíz et al. (2004): (a) MVO, which comprises an area of 21,263 km² and consists of mountain range systems with tectonic blocks and intermountain depressions that converge to the south in a NNW-SSE orographic axis. This subprovince is characterized by humid temperate climate (12℃–18℃ annual average, 500–1,200 mm annual rainfall), but the western portion receives more precipitation than its eastern sector; (b) SMOax, with 17,520 km² in area, with a relief amplitude of 2,500 m, conspicuous altitudinal fringes, various and asymmetric orographic axes, and short and steep slopes. It is characterized by semiwarm temperate and humid climates and is the mountain sector with higher rainfall at the statewide level, associated with Gulf trade winds (annual rainfall ranges from 2,500 mm to 4,000 mm); (c) MVC covers 6,663 km²; the terrain is also complex and the relief amplitude is high and contrasting; valleys and floodplains occur below 1,400 m. It is characterized by humid warm and semiwarm semiarid climates. It is the mountain region with higher temperature (annual mean temperature is 22℃–26℃) and lower rainfall (500–800 mm); (d) Sierra de los Chimalapas (CHIM) covers 5,816 km²; 20% of its land surface is mountainous and occurs above 1,000 m; 50% are high hills; and 30 % are low hills. This subprovince has semiwarm wet to subhumid climates (1,500–2,000 mm annual rainfall) because the Pacific winds; and (e) SMS, with 12,350 km² in area, characterized by a highly variable relief, with relatively low western mountains (<2,000 m) and higher systems at the Miahuatlán Sierra. The climate is warm and temperate subhumid (1,000–1,500 mm annual rainfall; Ortiz et al., 2004).

Species Richness, Rareness, and Irreplaceability Patterns

The use of equal-area grids has been considered as an important tool for studying biogeographic patterns in biological diversity (McAllister, Schueler, Roberts, & Hawkins, 1994). As the patterns of richness and endemism (as measured with both the rareness and irreplaceability indices) are scale dependent, we initially used three different grid sizes: (a) 205 grids of 10 × 10 min (17.6 × 18.4 km) of latitude and longitude, (b) 524 grids of 5 × 5 min (8.3 × 9.2 km), and (c) 1,316 grids of 2.5 × 2.5 min (4.15 × 4.6 km). However, our preliminary results indicated that, even though the results using the three different grid sizes were largely congruent, the 5 × 5 min was optimal because it does not exaggerate the importance of sites as when a larger grid is used, and minimizes the number of cells lacking data, that increases when the smaller grid is used. Therefore, the results presented are based on the 5 × 5 min grid, but the maps for the other two grid sizes are presented in Appendix B.

Species richness was measured counting the total number of species within each grid cell (Linder, 2001). To measure rareness, we used the weighted endemism index proposed by Crisp, Laffan, Linder, and Monro (2001), calculated by counting all species in each cell and weighting each species by the inverse of its range. Thus, a single-cell endemic would have the maximum weight of 1, a species occurring in two cells a weight of 0.5, and a species occurring in 100 cells a weight of 0.01 (Crisp et al., 2001). To obtain a rareness score for a cell, these weights were summarized for all species occurring in the cell (Crisp et al., 2001). This index tends to decrease the importance of species with broad distribution.

The measure of irreplaceability was based on the corrected weighted endemism index (Crisp et al., 2001). For this, we constructed a distribution matrix of all oak species at the whole Mexican and Central American level, based on the herbaria, literature, and online (Global Biodiversity Information Facility) records. In this matrix, we considered the geographical units identified by Torres et al. (2011), and we determined the number of units in which each species is recorded (Table 2). The irreplaceability index was calculated by assigning a value from 1 to 10 to each species, depending on the number of geographical units in which it is present. Species restricted to a single unit were weighted with a 10, whereas species present in 10 or more areas were weighted with a 1. In this way, species with local endemism had a high weight factor, while species with a wide geographical distribution had a low weight factor. The weight factor calculated for each species was multiplied by the inverse of their ranges obtained in the previous step. Therefore, the irreplaceability of an area increases when it contains species not found in other geographical units, and it decreases when it contains species found in other geographical units. The weighted irreplaceability values were summed across species to determine the areas with high overall irreplaceability importance.

Table 2.

Matrix Used to Correct the Rareness Index Based on the Presence or Absence of Species in 15 Geographical Units: (SMOcc) Sierra Madre Occidental, (SMOr) Sierra Madre Oriental, (TMVB) Trans-Mexican Volcanic Belt, (MEC) Meseta Central, (SMS) Sierra Madre del Sur, (SMOax) Sierra Madre de Oaxaca, (SMCG) Sierra Madre de Chiapas-Guatemala, and (COMA) Montañas de Comayagua.

The patterns of species richness and endemism (rareness and irreplaceability) were examined by mapping the three indices described earlier and detecting in maps the concentration of cells with the highest values.

Species Turnover

Beta diversity, or species turnover, is a measure of differentiation in species composition among distinct areas or along gradients (Koleff, Gaston, & Lennon, 2003; Rodríguez, Oyama, MacGregor, & González, 2015). This measure increases when there are few shared species between two areas. We determined the species turnover patterns based on the proposals of many authors (Baselga, 2010; Koleff et al., 2003; Lennon, Koleff, Greenwood, & Gaston, 2001), using Sørensen index (β_SOR) and Jaccard index (β_JAC). β_SOR is a widely-used measure to reflect the proportion of species shared between two communities (Baselga, 2010), but it is related to species richness (Koleff et al., 2003) while the β_JAC actually describes the gain or loss of species without the bias of richness (Koleff et al., 2003; Lennon et al., 2001). β_SOR and β_JAC indices were computed with the betapart package for the R software (Baselga & Orme, 2012). Both turnover measures were calculated for each pair of cells. The final turnover index for each cell was the average of all turnover values of the cell with the other cells.

Elevational Patterns

To analyze elevational patterns, we built a matrix of species absence or presence using intervals of 100 m of altitude across the full elevational distribution for the genus (200–3,000 masl). The number of species occurring in each elevational interval was calculated. To evaluate rareness patterns, for each species we counted the number of elevational intervals in which it is present and weighted each species by the inverse of this number. To obtain a rareness score for each elevational interval, these weights were summarized for all species. In the case of irreplaceability, in this second matrix, the weighted inverse of rareness of each species was weighted by the correction factor for each species, and the sum of corrected values of all species for each interval was used to determine its irreplaceability index. The turnover between elevational intervals was computed through β_SOR and β_JAC indices computed with the betapart package for R (Baselga & Orme, 2012).

Habitat Loss

The amount of habitat loss for oaks was computed by means of cartographic overlaying of two comparable databases (same geographic scale, same vegetation typology) for the most recent decade. For this, we worked with maps in digital format (scale 1:250,000) from the National Forestry Inventory for 2000 (Mas et al., 2002; Secretaría de Medio Ambiente y Recursos Naturales, 2001) and INEGI cartographic data series IV for 2010. Vulnerability was determined separately for the two sections within the genus based on habitat loss percentage or coverage change (primary to secondary; Myers et al., 2000). First, we selected vegetation polygons with presence of oaks (pine forests, pine-oak forests, oak-pine forests, oak forests, mountain cloud forests, high evergreen forests, medium evergreen tropical forests, medium deciduous forests, deciduous forests, and medium deciduous forests) in both 2000 and 2010 inventories and then calculated their areas. Next, we calculated the percentage of habitat loss for the polygons between both interval years. In the 2010 polygons, we also calculated the percentage coverage of secondary and primary vegetation (Myers et al., 2000).

Prioritizing Conservation Areas

Conservation priority areas were defined at the statewide level. First, results obtained with grid cells of 5′ × 5′ (to minimize biases and overstatement) in the different analyses (for the whole genus and by section) were transformed to a ratio scale along the following lines: (a) Richness, rareness, irreplaceability, and turnover criterion. Values from 0 to 5 were assigned to each cell, where 0 corresponds to cells with low indices (richness, rareness, irreplaceability, or turnover) and 5 corresponds to cells with highest indices in each case. (b) Vulnerability criterion. It was incorporated using a vegetation cover loss index. First, vegetation polygons of the 2000 year with confirmed oaks presences were selected and intersected with the year 2010. The resulting polygons were in turn intersected with the cell grid, and vegetation percentage still classified as primary for 2010 was calculated for each cell. We assigned values from 0 to 5, where 0 corresponds to cells with less primary coverage and 5 to cells with more primary coverage. To do this, we evaluated patterns of richness, rareness, turnover, and habitat loss for the genus and its two sections. (c) Systematic conservation planning. According to the guidelines proposed by Margules and Pressey (2000) and Sarkar et al. (2006), non-protected areas in a network of protected natural areas (PNAs) are priority to conservation, so in this section we used polygons of protected areas enacted at the federal, state, municipal, and community levels. Initially, we gave a protection value of 1 to each cell. Then, we subtracted or added from this value on the basis of their degree of protection. We considered the areas with official decree as having the maximum degree of protection. Therefore, we quantified the area of each cell that has federal protection and, for example, if the percentage of federal protection coverage in a cell is 10%, we subtracted 0.1 from the baseline protection value for the cell. For areas with state or local protection, the coverage percentage was transformed into decimal added to the initial value of each cell. A similar procedure was used for communal protected areas. After the subtraction of costs, the final value of maximum conservation of each cell was transformed into a scale of 0 to 5, where 0 represents those cells whose total surface has federal protection, and 5 represents those cells with no areas under federal or community protection. (d) Areas prioritization. Finally, an average of all the values of all criteria (richness, rareness, turnover, irreplaceability, vulnerability, and effective conservation) was calculated according to each criterion listed earlier. The average value is a direct measure of the prioritization of areas, which was made at both the genus and the section levels. All indices in this study were computed in R software (R Development Core Team, 2011), and the results were visualized in ArcGIS ver. 9.2 (Environmental Systems Research Institute, 2006).

Richness Patterns

The SMOax is the subprovince with the highest richness of oak species (38 species). This subprovince includes the Ixtlán district (with 19 species), the Teotitlán district (with 14 species, mainly in the municipalities of San Francisco Huehuetlán and Santa María Teopoxco), and the Cuicatlán, Centro, and Villa Alta districts (with 12 species each one). The MVO subprovince has 29 species (the Juxtlahuaca and Tlaxiaco districts have 12 species and the Sola de Vega district 11 species), SMS has 25 species (the Miahuatlán and Yautepec districts have 12 species), and MVC has 20 species. Eleven species are recorded in CHIM (Figure 2; Table 3).

Figure 2.

Patterns of richness and rareness for the Lobatae section, the Quercus section, and the whole genus in Oaxaca using a 5 × 5 min grid. The cells with greater richness or rareness are indicated in darker colors. Black lines delimit the physiographic subprovinces, and gray lines delimit the districts of Oaxaca.

Table 3.

Areas in Oaxaca Representing the Most Important Centers of Oak Species Richness, Rareness, or Irreplaceability at the Genus and Section Levels.

For white oaks, the districts with higher richness are Teposcolula, Coixtlahuaca, and Nochixtlán in MVO with seven species, Ixtlán in SMOax also with seven species, and the Juxtlahuaca and Mixtepec districts in MVC with six species (Figure 2). For red oaks, the districts with higher richness are Ixtlán (13 species), Teotitlán (12 species), and Cuicatlán (11 species) in the SMOax; Sola de Vega (nine species) and Juxtlahuaca—Mixtepec (seven species) in the MVO; Juchitán (eight species) in CHIM, and Miahuatlán and Yautepec (seven species) in the SMS (Figure 2; Table 3).

Rareness Patterns

We found five main rareness regions at the level of the whole Quercus genus, that in order of importance are two districts in the SMOax (Ixtlán and Teotitlán), one in the MVO (in the border between the Coixtlahuaca and Nochixtlán districts), one in the SMS (Yautepec district), and one in CHIM (Juchitán district; Figure 2; Table 3). For the section Quercus, we identified five areas with importance for rareness: one in the north of the Ixtlán district, and four areas located within the Coixtlahuaca, Nochixtlán, and Teposcolula districts (Figure 2). For section Lobatae, the regions with high rareness values are located in order of importance in the following districts: Teotitlán, in the limits between Ixtlán and Cuicatlán, and in the south of Ixtlán in the SMOax; Yautepec, in the limits between Juquila-Miahuatlán in the SMS; and Juchitán, in the CHIM (Figure 2). The highest rareness index represented those areas where there are species with a restricted distribution within Oaxaca but without considering if those species occur in other biogeographic regions outside of Oaxaca.

Irreplaceability Patterns

The irreplaceability measure obtained considered if species occur, besides Oaxaca, in other regions. In this way, we recognized the areas with greater value in terms of irreplaceability at the level of the whole Quercus genus, that in order of importance are located in Teotitlán and Ixtlán (mainly in the northern region) in the SMOax, Yautepec and borders between Juquila-Miahuatlán in the SMS and the CHIM (see Figure 3; Table 3). For white oaks, the areas in order of importance are in the following districts: northern Ixtlán in the SMOax; Coixtlahuaca, Teposcolula, and Nochixtlán in the MVO; Villa Alta in the SMOax; and Miahuatlán in the SMS (Figure 3). For red oaks, the districts with higher importance are Teotitlán in the SMOax; Yautepec in the SMS, the limits between Cuicatlán and Ixtlán in the SMOax; Juchitán in CHIM; and the border between Juquila and Miahuatlán in the SMS (Figure 3).

Figure 3.

Patterns of irreplaceability for the Lobatae section, the Quercus section, and the whole genus in Oaxaca, using a 5 × 5 min grid. Darker colors indicate cells with greater indices of irreplaceability or turnover. In the case of turnover, the arrows indicate the areas with highest turnover among cells. Black lines delimit the physiographic subprovinces, and gray lines delimit the districts of Oaxaca.

Turnover Patterns

Turnover patterns for the genus coincide in general with the limits among the physiographic subprovinces (Figure 3). Also, high turnover values were observed between the Teotitlán district and the rest of the SMOax subprovince. A similar case was observed for the Nochixtlán district when compared with the Teposcolula and Coixtlahuaca districts (Figure 3). The areas of higher turnover for the white oaks are almost the same as those observed at the whole genus level, but the highest turnover values were found at the limits between the districts of Coixtlahuaca, Teposcolula, and Nochixtlán (Figure 3). However, in the case of the red oaks, the area with the higher turnover was localized at the Isthmus of Tehuantepec. There are also important turnover areas between the districts of Tepescolula-Tlaxiaco in the MVO, and Juquila and Putla in the SMS (Figure 3). Also, a high turnover value was observed between the SMOax and the MVO, which was not detected at the genus level or for the white oaks.

Altitudinal Patterns

The highest species richness at the whole genus level was found between 1,800 and 2,200 masl. For white oaks, the highest richness occurs between 2,000 and 2,200 masl, while red oaks showed a bimodal pattern with peaks of richness at 1,700 to 2,000 and 2,200 to 2,400 masl. In this last altitudinal interval, a decrease in the richness of white oaks was observed (Figure 4).

Figure 4.

Elevational patterns at section levels and genus. The graph shows the richness, rareness, irreplaceability, and turnover indices for each elevation range.

We also observed that oaks with restricted altitudinal distributions occur mostly between 2,000 and 2,500 masl. From 1,300 to 2,300 masl, there is an increase in the rareness index. In the red oaks, altitudes around 1,700 masl are also important in terms of rareness. The white oak species that have a very restricted geographic distribution in Mexico are found at 2,200 masl, while in the case of the red oaks, such species are found between 2,200 and 2,400 masl, with the 1,000 and 1,700 masl altitudes also being important. The greater species turnover was observed at 1,100, 1,600, 2,300, 2,600, and 2,800 masl (see Figure 4 for details).

Habitat Loss

Vegetation fragments with presence of oak species in Oaxaca have lost 5.83% of their original area between 2000 and 2010 (413,123 ha). From the remaining area, 41.89% and 58.11% correspond to primary and secondary vegetation, respectively (Table 4). By section, Quercus showed a higher loss of habitat (6.47%); however, the percentage of remnants representing primary and secondary vegetation is almost equal (45.46% and 54.54%, respectively). In contrast, areas where section Lobatae occurs have been transformed into secondary vegetation at a higher percentage (62.74%).

Table 4.

Habitat Loss (in ha), Disturbance and Quercus Conservation in Oaxaca, Mexico.

Areas that have retained most of their primary vegetation cover from 2000 to 2010 for both the genus and sections levels are located in the north of the Ixtlán district (from the Macuiltianguis to the Santa María Yavesía municipalities) in the SMOax, as well as the border of the Ixtlán district with the Cuicatlán and Tuxtepec districts (Figure 5). The importance of other districts is evident, as, for example, Sola de Vega—Zimatlán, Etla, northern Tlacolula, Mixe, and western Miahuatlán. In the rest of the districts, most of the vegetation where oaks occur is secondary.

Figure 5.

Areas with primary and secondary vegetation cover and vulnerability of habitat for oaks in Oaxaca. The areas that remained with primary vegetation between the year 2000 and 2010 are considered as having a low degree of disturbance (green color) and those that were transformed from primary to secondary vegetation during the same period are considered as showing a high disturbance (red color). Black lines delimit the physiographic subprovinces, and gray lines delimit the districts of Oaxaca.

Prioritizing Conservation Areas

According to our analysis, the most important conservation areas for the genus and both sections are mainly located in the SMOax, including the municipalities of San Pedro Yolox, Macuiltianguis, San Juan Quiotepec, Ixtlán, Comaltepec, Capulálpam, Santa Catarina Ixtepeji, and Santa María Yavesía; all in the Ixtlán district, and municipalities in the Cuicatlán district, as San Felipe Usila and San Juan Tepeuxila. In the Teotitlán district, Santa María Teopoxco, San Lucas Zoquiapam, and Mazatlán are important. In the south of the SMOax, the San Andrés Yaa and Totontepec municipalities are important for conservation, in the limits between the Villa Alta and Mixe districts (Figure 6).

Figure 6.

Conservation priority areas in Oaxaca for the Lobatae section, the Quercus section, and the genus Quercus. This proposal considers a systematic conservation planning approach based on a multicriteria analysis. The protected natural areas (PNA) are shown in green and in red the indigenous conservation and community areas (ICCA). Black lines delimit the physiographic subprovinces, and gray lines delimit the districts of Oaxaca.

In the east of the SMS, there are important areas in the municipalities of San Carlos Yautepec, Santa María Ecatepec, and San Pedro Mixtepec, in the limits between the Yautepec and Miahuatlán districts; in the west, the municipalities of San Jerónimo Coatlán and San Juan Lachao in the limits between the Juquila and Miahuatlán districts. In the MVO, the area between the municipalities of Putla and Juxtlahuaca is important, as well as the Santiago Textitlán municipality in the Sola de Vega district. In CHIM, the area with greater importance for conservation is in the San Miguel Chimalapa municipality in the Juchitán district (Figure 6).

In the analysis for the Quercus section, other additional areas were identified, as: the Tepescolula and San Pedro Nopala municipalities in the Teposcolula district, San Andrés Cabecera Nueva in the Putla district, San Miguel Chicahua in the Nochixtlán district, and the Tepelmeme municipality in the Coixtlahuaca district. In the SMOax, the municipalities of Talea de Castro, Juquila-Vijanos, and Tanatze in the Villa Alta district are also of importance, as well as the Santiago Lachigurí municipality in the Tehuantepec district (Figure 6).

For section Lobatae, the area with more importance is located in SMOax in the Teotitlán district, including Santa María Teopoxco and San Lucas Zoquiapam. The rest of the priority sites are located, in order of importance, in the SMOax, SMS, CHIM, and MVO, but in this last physiographic province the only important municipality is Santiago Textitlán, in the Sola de Vega district (Figure 6).