Introduction
⌅In addition to the serious effect of deforestation on biodiversity loss, deforested ecosystems lose their ability to provide multiple ecological services, such as erosion control, wind protection, carbon sequestration, and air purification (Imai et al., 2012Imai N, Seino T, Aiba SI, Takyu M, Titin J, Kitayama K, 2012. Effects of selective logging on tree species diversity and composition of Bornean tropical rain forests at different spatial scales. Plant Ecol 213: 1413-1424. https://doi.org/10.1007/s11258-012-0100-y; Adekunle et al., 2013Adekunle VAJ, Olagoke AO, Ogundare LF, 2013. Logging impacts in tropical lowland humid forest on tree species diversity and environmental conservation. Appl Ecol Environ Res 11: 491-511. https://doi.org/10.15666/aeer/1103_491511). Deforestation is the conversion of forests to non-forest land uses (e.g., agriculture or housing); in this process, forest cover is gradually lost until only 10 % or less of the original cover is conserved (Chakravarty et al., 2012Chakravarty S, Ghosh SK, Suresh CP, Dey AN, Shukla G, 2012. Deforestation: causes, effects and control strategies. Glob Persp Sust For Manag 1: 1-26. https://doi.org/10.5772/33342). The ecological consequences of deforestation on plants through selective logging, firewood extraction, grazing, and land clearing for permanent agriculture involve changes in succession patterns and species composition. Deforestation favors pioneer species due to their ability to colonize recently disturbed sites, while late successional species are more sensitive to disturbance, which considerably reduces their populations, even to the point of local extinction (Figueroa-Rangel et al., 2008Figueroa-Rangel BL, Willis KJ, Olvera-Vargas M, 2008. 4200 years of pine-dominated upland forest dynamics in west-central México: human or natural legacy? Ecol 89: 1893-1907. https://doi.org/10.1890/07-0830.1; Chakravarty et al., 2012Chakravarty S, Ghosh SK, Suresh CP, Dey AN, Shukla G, 2012. Deforestation: causes, effects and control strategies. Glob Persp Sust For Manag 1: 1-26. https://doi.org/10.5772/33342; Pessoa et al., 2017Pessoa MS, Hambuckers A, Benchimol M, Rocha-Santos L, Bomfim JA, Faria D, et al., 2017. Deforestation drives functional diversity and fruit quality changes in a tropical tree assemblage. Perspec Plant Ecol Evol Syst 28: 78-86. https://doi.org/10.1016/j.ppees.2017.09.001).
In conifer forests, many species of Pinus are considered invasive and colonizers in areas with severe deforestation or disturbance because they can germinate and establish successfully in open areas (Richardson & Bond, 1991Richardson DM, Bond WJ, 1991. Determinants of plant distribution: evidence from pine invasions. The Amer Nat 137(5): 639-668. https://doi.org/10.1086/285186; Galindo-Jaimes et al., 2002Galindo-Jaimes L, González-Espinosa M, Quintana-Ascencio P, García-Barrios L, 2002. Tree composition and structure in disturbed stands with varying dominance by Pinus spp. in the highlands of Chiapas, Mexico. Plant Ecol 162: 259-272. https://doi.org/10.1023/A:1020309004233). Previous investigations have reported an increase in pine abundance in conifer forests due to land clearance for logging and agriculture; as a result of deforestation, drier environments have a positive effect on pine dominance (Figueroa-Rangel et al., 2008Figueroa-Rangel BL, Willis KJ, Olvera-Vargas M, 2008. 4200 years of pine-dominated upland forest dynamics in west-central México: human or natural legacy? Ecol 89: 1893-1907. https://doi.org/10.1890/07-0830.1). Furthermore, pine-dominated forests are characterized by drier and hotter conditions along with more compact, less acidic, and less fertile soils, which reduce microbial biomass, deteriorate the microhabitat, and restrict the establishment of other tree species (Galindo-Jaimes et al., 2002Galindo-Jaimes L, González-Espinosa M, Quintana-Ascencio P, García-Barrios L, 2002. Tree composition and structure in disturbed stands with varying dominance by Pinus spp. in the highlands of Chiapas, Mexico. Plant Ecol 162: 259-272. https://doi.org/10.1023/A:1020309004233).
Mexican conifer forests cover 16 % of the country, with a deforestation rate of 0.8 % per year, being the second-most-affected vegetation type from original cover loss (27 %), only after tropical forests (42 %) (Guerra-De la Cruz & Galicia, 2017Guerra-De la Cruz V, Galicia L, 2017. Tropical and highland temperate forest plantations in Mexico: Pathways for climate change mitigation and ecosystem services delivery. For 8: 489. https://doi.org/10.3390/f8120489). The conifer forests of central Mexico are particularly important in terms of biodiversity due to the environmental conditions conferred by the convergence of the Nearctic and Neotropical zones (Morrone, 2005Morrone JJ, 2005. Hacia una síntesis biogeográfica de México. Rev Mex Biod 76: 207-252. https://doi.org/10.22201/ib.20078706e.2005.002.303). However, it is one of the areas of Mexico with the greatest conversion of forest cover to different land uses (Reyes-González & Rhodes, 2015Reyes-González J, Rhodes A, 2015. Conservación de la biodiversidad en el Eje Neovolcánico: Colaboración interinstitucional en un territorio biodiverso y proveedor de servicios ambientales. Red de Gestión Territorial para el Desarrollo Rural Sustentable. México. Boletín 2: 7-8.; FAO & PNUMA, 2020FAO & PNUMA, 2020. El estado de los bosques del mundo 2020. Los bosques, la biodiversidad y las personas. Organización de las Naciones Unidas para la Alimentación y la Agricultura, Programa de las Naciones Unidas para el Medio Ambiente, Roma.).
La Malinche National Park has been a Protected Natural Area since 1938 and has 206.07 km2 of conifer forest in central Mexico (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.). Although the La Malinche National Park management plan indicates restricted land-use changes and controlled extraction of forest resources (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.), these guidelines have not been observed and deforestation has increased since 1985; to date, about 54 % of its area includes cropland, induced pastures, and human settlements (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.). La Malinche National Park is an area of great ecological importance for the conservation of conifer forests in central Mexico because it maintains the diversity and endemism unique to the Trans-Mexican Volcanic Belt. However, tree diversity has only been partially described, and the effects of constant deforestation and conversion of forest use have not yet been explored (SEMARNAT & CONANP, 2013SEMARNAT & CONANP, 2013. Programa de Manejo Parque Nacional La Montaña Malinche o Matlalcuéyatl. Secretaría de Medio Ambiente y Recursos Naturales, Comisión Nacional de Áreas Naturales Protegidas. Gobierno de México.).
This study aimed to describe the diversity of trees in the La Malinche National Park, identify its association with deforestation and forest use, explore whether pine dominance increases with forest cover loss and forest use, and assess whether this pine dominance determines the abundance of other tree species. We expected to find a negative relationship of deforestation and forest use with tree diversity, a positive relationship with pine abundance, and a lower abundance of non-pine trees in areas with higher dominance of Pinus species.
Material and methods
⌅Study area
⌅La Malinche National Park is located in central Mexico, in the states of Tlaxcala and Puebla, between coordinates 19°08’-19°20’ N and 98°08’-97°55’ W (Fig. 1a,b); it comprises a total of 461.12 km2 across an altitudinal range between 2,200 and 4,461 m asl, which includes conifer forest, broadleaf forest (e.g. Quercus spp. and Buddleja cordata) and alpine grassland (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.; SEMARNAT & CONANP, 2013SEMARNAT & CONANP, 2013. Programa de Manejo Parque Nacional La Montaña Malinche o Matlalcuéyatl. Secretaría de Medio Ambiente y Recursos Naturales, Comisión Nacional de Áreas Naturales Protegidas. Gobierno de México.).
The climate in La Malinche National Park changes with altitude, but the one that prevails is temperate subhumid, with a mean annual temperature of 14 ºC to 16 ºC and summer rains, which are more intense from June to September (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.). Vegetation also varies with altitude, slope, and orography. The dominant vegetation types are pine forest, fir forest, and pine-oak forest, all of which can be classified as conifer forests (Rzedowski, 2006Rzedowski J, 2006. Vegetación de México. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO). 1st Ed dig, México. 504 pp.).
The altitudinal belt sampled represents a bioclimatic transition zone that ranges between 2,800 m and 3,200 m asl (~158 km2, 34 % of La Malinche National Park; SEMARNAT & CONANP, 2013SEMARNAT & CONANP, 2013. Programa de Manejo Parque Nacional La Montaña Malinche o Matlalcuéyatl. Secretaría de Medio Ambiente y Recursos Naturales, Comisión Nacional de Áreas Naturales Protegidas. Gobierno de México.); therefore, this area harbors the highest species diversity. Above this altitude, Abies religiosa and Pinus hartwegii dominate with monospecific forests at different elevations; below this range, deforestation is more prevalent (Rojas-García & Villers-Ruiz, 2008Rojas-García F, Villers-Ruíz L, 2008. Estimación de la biomasa forestal del Parque Nacional Malinche: Tlaxcala-Puebla. Rev Mex Cien For 33: 59-86.).
The land-use changes in La Malinche National Park that have caused the greatest forest loss are agriculture and grazing (Valdez et al., 2016Valdez PME, González GG, Morales IR, Bolaño SRY, 2016. Reserva de carbono en biomasa forestal y suelos minerales en el Parque Nacional Malinche (México). Cuadernos de Geografía: Rev Col Geog 25: 207-215. https://doi.org/10.15446/rcdg.v25n1.40382). From 1995 to 2020, La Malinche National Park has been subjected to sustained deforestation, with an annual loss of continuous forest of 0.41 km2 and an increase of 0.38 km2 of discontinuous forest and 0.10 km2 of urban areas (Table S1 [suppl](Tables S1-S4) accompanies the paper on Forest System´s website).
Sampling
⌅We spatially located three forest stands on each of four slopes in La Malinche National Park (North, East, Southeast, and West) (Fig. 1c). We randomly established five 1,000 m2 circular plots (17.8 m radius) in each forest stand (Fig. 1d), with a minimum distance of 50 m between plots. In each plot, we measured the diameter at breast height (DBH) at 1.3 m above the ground for all trees ≥1.5 m in height. We taxonomically identified each tree by comparing specimen samples with herbarium material deposited at the TLXM Herbarium of the Universidad Autónoma de Tlaxcala (Thiers, 2024Thiers B, 2024. Index Herbariorum. NYBG Steere Herbarium. https://sweetgum.nybg.org/science/ih/ [10 Jan 2024].).
Deforestation and forest use
⌅We obtained the percentage of forest cover (%FC) as an approximation of deforestation measured with a convex spherical densitometer (Model A) at four points within each plot (i.e., North, South, East, and West) at a distance of 11 m from the plot center. We estimated the percentage of forest cover in each stand as the average value of the respective plots.
Knowing that deforestation is directly related to forest use (Vieilledent et al., 2018Vieilledent G, Grinand C, Rakotomalala FA, Ranaivosoa R, Rakotoarijaona JR, Allnutt TF, et al., 2018. Combining global tree cover loss data with historical national forest cover maps to look at six decades of deforestation and forest fragmentation in Madagascar. Biol Con 222: 189-197. https://doi.org/10.1016/j.biocon.2018.04.008) but that the latter (forest use) is a continuous and historical process (Table S1 [suppl](Tables S1-S4) accompanies the paper on Forest System´s website) for which we have no record, we used three indicators of forest use to obtain current values and assess how they affect species diversity and dominance. These indicators are: logging, ocoteo (cuttings of pine wood fragments that are impregned with resin and used by local human populations to light fires), and firewood extraction. To this end, we used the relative density and basal area of damaged individuals (juveniles and adults), compared to undamaged trees, for each indicator (Sapkota et al., 2010Sapkota IP, Tigabu M, Odén PC, 2010. Changes in tree species diversity and dominance across a disturbance gradient in Nepalese Sal (Shorea robusta Gaertn. f.) forests. J For Res 21(1): 25-32. https://doi.org/10.1007/s11676-010-0004-4). We measured firewood extraction as the number of tree stumps <20 cm DBH, while logging considered tree stumps >20 cm DBH, and ocoteo included the number of individuals of Pinus spp. with cut marks in the bark.
To estimate the extent of each type of forest extraction, we first calculated the relative density and basal area of damaged individuals per plot with the following formulas:
where RDdind is the relative density of damaged individuals, ΣRDdind is the sum of the relative density of damaged individuals, TD is the total density; Gdind is the relative basal area of damaged individuals, ΣGdind is the sum of the relative basal area of damaged individuals, and TG is the total basal area. Subsequently, we estimated relative values by dividing the values of the plot of interest by the lowest value (density and basal area) for the entire La Malinche National Park. Then, we calculated the overall value for each forest stand as the arithmetic mean of the respective plots, representing the sum of density and basal area (Sapkota et al., 2010Sapkota IP, Tigabu M, Odén PC, 2010. Changes in tree species diversity and dominance across a disturbance gradient in Nepalese Sal (Shorea robusta Gaertn. f.) forests. J For Res 21(1): 25-32. https://doi.org/10.1007/s11676-010-0004-4).
Tree composition
⌅We estimated diversity for each forest stand by obtaining rarefaction curves and Hill numbers (Hill, 1973)Hill MO, 1973. Diversity and evenness: a unifying notation and its consequences. Ecol 54: 427-32. https://doi.org/10.2307/1934352 of order q, or the effective number of species, defined as , where qD is the effective species richness, and q indicates the relative abundance of species i. When q = 0, species abundances are not considered, so 0D is species richness. If q = 1, species are weighed proportionally to their frequencies (common or typical species), so 1D is equivalent to the exponential of Shannon’s entropy (Shannon’s diversity index). Finally, for q = 2, diversity (2D) represents the dominant species and is equivalent to the inverse of Simpson’s diversity index (Chao et al., 2014Chao A, Gotelli NJ, Hsieh TC, Sander EL, Ma KH, Colwell RK, et al., 2014. Rarefaction and extrapolation with Hill numbers: a framework for sampling and estimation in species diversity studies. Ecol Monogr 84: 45-67. https://doi.org/10.1890/13-0133.1). We constructed interpolation or extrapolation curves relative to the sample size at each sampling site with a 95 % confidence level and 1,000 bootstrap replicates. Additionally, we estimated the completeness of the sampling (coverage) for species.
We identified dominant species through rank-abundance curves for each species. Furthermore, we determined the dissimilarity between forest stands according to the order q using Jaccard’s index (J) for q = 0 with the formula: , where is the average species richness of the plots in the forest stand, S is the species richness of all forest stands pooled, and N is the number of forest stands (Chiu et al., 2014Chiu CH, Jost L, Chao A, 2014. Phylogenetic beta diversity, similarity, and differentiation measures based on Hill numbers. Ecol Monogr 84: 21-44. https://doi.org/10.1890/12-0960.1). For q = 1, the Horn index (RO ) was used: , where Hmax is the value when the forest stands (X and Y) have no species in common; Hobs are the species observed in the forest stands (X + Y); and Hmin is the value of the forest stands that have the same species in equal proportion. For q = 2, the Morisita-Horn index (MH) was used: , where HGS is gamma (γ) and alpha (α) diversity with the Gini-Simpson index, and N is the number of forest stands (Chiu et al., 2014Chiu CH, Jost L, Chao A, 2014. Phylogenetic beta diversity, similarity, and differentiation measures based on Hill numbers. Ecol Monogr 84: 21-44. https://doi.org/10.1890/12-0960.1).
Effect of deforestation or forest use on species diversity
⌅We assessed the relationship of total tree species diversity (0D, 1D, 2D) with deforestation (forest cover percentage; %FC) and forest use indicators (firewood extraction, logging, ocoteo) as explanatory variables using generalized linear mixed-effects models and the Poisson family for richness, and Gaussian family for abundant and dominant species, and whose random variable was the stand (i.e, 0D, 1D or 2D ~ (Logging, ocoteo, Firewood, or %FC)*Slope+(1|stand)). With the same models using the plot as random variable and the Poisson distribution, we also studied the effect of deforestation and forest use on pine abundance (only Pinus spp.), which included the five species recorded at La Malinche National Park (Pinus ayacahuite, P. hartwegii, P. leiophylla, Pinus pseudostrobus, and P. teocote) (i.e., Pinus spp. ~ (logging, ocoteo, firewood, or %FC)*Slope+(1|stand)). We tested the statistical significance of each model by contrast with the null model (dependent variable ~ 1+(1|stand)) (Gualchi, 2019Gualchi S, 2019. Introducción a los modelos de efectos mixtos. https://gesel.github.io/materiales/4modelos_mixtos.pdf [10 Jan 2024]). For those response variables with more than one model, we used the corrected Akaike Information Criterion (AICc) to identify the model with the best fit to the data (Hurvich & Tsai, 1991Hurvich CM, Tsai CL, 1991. Bias of the corrected AIC criterion for underfitted regression and time series models. Biome 78(3): 499-509. https://doi.org/10.1093/biomet/78.3.499). Finally, we assessed the contribution of each variable of the statistically significant and optimal models (according to the AICc) with a Permuted Multivariate Analysis of Variance (PERMANOVA).
In addition, to evaluate whether pine dominance (Pinus spp.) determines the abundance of other tree species, we performed generalized linear models using the Poisson family (i.e., other non-pine trees ~ Pinus spp.). As P. montezumae was the dominant conifer in almost all forest stands, we explored whether its abundance alone governs the abundances of other tree species (i.e., other non-pine trees ~ P. montezumae). We determined the statistical significance of all generalized linear models using a likelihood-ratio test. We performed all the analyses with the R v. 4.1.1 program, using the following libraries: AICmodavg, Biodiversity, car, corrplot, lme4, lattice, MASS, MuMIn, permute, Vegan, vegdist, and iNEXT (R Core Team, 2023R Core Team, 2023. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.r-project.org/ [10 Dec 2023].).
Results
⌅Deforestation and forest use
⌅Forest cover values ranged between 35.8 % and 74.4 % (Table 1); Southeast-3 and East-1 were the forest stands with the lowest and highest percent cover values. In general, there were differences in intensity of use according to slope exposure; on average, there was higher ocoteo (50.2) than firewood extraction (11.7) and logging (11.2); there was no firewood extraction in West-3 and Southeast-2, while no logging was detected in East-2, West-2, and West-3, and we did not record ocoteo (0.0) in East-2 and West 2.
| Forest stand | Firewood extraction | Logging | Ocoteo | %FC | 0D | 1D | 2D | %SC |
|---|---|---|---|---|---|---|---|---|
| East-1 | 10.6 | 10.0 | 17.4 | 74.4 | 5 | 2.6 | 2 | 100 |
| East-2 | 16.1 | 0.0 | 0.0 | 70.0 | 11 | 4 | 3.2 | 100 |
| East-3 | 2.0 | 18.4 | 51.7 | 52.2 | 4 | 1.5 | 1.3 | 98 |
| West-1 | 11.7 | 25.0 | 9.1 | 72.9 | 5 | 2.3 | 1.8 | 99 |
| West-2 | 11.9 | 0.0 | 0.0 | 54.9 | 6 | 1.9 | 1.5 | 100 |
| West-3 | 0.0 | 0.0 | 127.6 | 49.6 | 8 | 2 | 1.5 | 98 |
| North-1 | 45.6 | 11.7 | 0.5 | 68.0 | 11 | 1.7 | 1.3 | 100 |
| North-2 | 7.4 | 20.2 | 32.7 | 59.9 | 2 | 1.1 | 1.0 | 100 |
| North-3 | 5.0 | 16.1 | 179.2 | 70.9 | 3 | 1.3 | 1.1 | 98 |
| Southeast-1 | 1.1 | 10.8 | 23.5 | 55.2 | 6 | 2.5 | 2.2 | 99 |
| Southeast-2 | 0.0 | 20.6 | 99.0 | 63.9 | 7 | 3 | 2.4 | 98 |
| Southeast-3 | 29.2 | 1.1 | 61.6 | 35.8 | 6 | 2.5 | 1.7 | 100 |
| Average | 11.7 | 11.2 | 50.2 | 60.6 | 6.2 | 2.2 | 1.8 | 99.2 |
%FC: percentage of forest cover. 0D, observed species richness. 1D, observed diversity of common or typical species. 2D, observed diversity of the dominant species. %SC, relative sampling coverage, in percentage.
Tree composition
⌅We recorded a total of 16 tree species: Alnus jorullensis, Abies religiosa, Arbutus xalapensis, Buddleja cordata, Comarostaphylis discolor, Cupressus benthamii, Pinus ayacahuite, Pinus hartwegii, Pinus leiophylla, Pinus montezumae, Pinus pseudostrobus, Pinus teocote, Prunus serotina, Quercus laurina, Quercus crassipes, and Quercus dysophylla; of these, P. montezumae was the dominant species (Abundance = 2219 individuals in total), and Q. dysophylla and P. ayacahuite were the least abundant (1 individual of each species) (Table S2 [suppl](Tables S1-S4) accompanies the paper on Forest System´s website).
Tree species richness ranged from 0D = 2 in North-2 to 0D = 11 in East-2 and North-1. Common species ranged from 1D = 1.1 in North-2 to 1D = 4.0 in East-2. The dominant species were more diverse in East-2 (2D = 3.2) and less diverse in North-2 (2D = 1.0) (Table 1). Unlike the diversity of common and dominant species, species richness did not reach the asymptote in all forest stands. The closest curves corresponded to East-1, East-2, North-1, and West-2 (Fig. 2), but the %SC was never below 98 % (Table 1).
Regarding dissimilarity for the order q = 0, Jaccard’s similarity index (SJ ) showed a higher dissimilarity between East-2 and North-3 (SJ = 0.923), while East-1 and East-3 were less differentiated (SJ = 0.200) (Table S3a [suppl](Tables S1-S4) accompanies the paper on Forest System´s website). When the common (q = 1) and dominant species (q = 2) were weighted, the similarity index showed high similarity between North-2 and North-3 (SH = 0.001, SMH = 0) and greater dissimilarity between East-1 and Southeast-1 (SH = 0.741, SMH = 0.740) (Table S3b,c [suppl](Tables S1-S4) accompanies the paper on Forest System´s website).
Effect of deforestation and forest use on species diversity
⌅Of the 12 models constructed to assess the influence of deforestation and forest use on total tree diversity, none was statistically significant (Table 2a). In contrast, pine dominance (Pinus spp.) was affected by the interaction of slope with forest use and with percentage of forest cover; the latter was chosen as the optimal model according to AICc (Pinus spp. ~ FC*Slope+(1|stand); Table 2b), however the PERMANOVA did not show significant differences between groups (stands), with a marginal contribution of forest cover (r2 = 0.01; Table S4 [suppl](Tables S1-S4) accompanies the paper on Forest System´s website). Nonetheless, the data showed a trend of higher pine dominance in forest stands with lower forest cover (Fig. 3).
| Model | r2m | r2c | AICc | p | |
|---|---|---|---|---|---|
| (a) Diversity ~ deforestation/forest use | |||||
| 0D | ~ Logging*Slope+(1|stand) | 0.58 | 0.58 | 19.08 | 0.07ns |
| ~ Ocoteo*Slope+(1|stand) | 0.62 | 0.62 | 18.80 | 0.07ns | |
| ~ Firewood*Slope+(1|stand) | 0.48 | 0.48 | 24.25 | 0.07ns | |
| ~ %FC*Slope+(1|stand) | 0.01 | 0.01 | 19.84 | 0.75ns | |
| 1D | ~ Logging*Slope+(1|stand) | 0.88 | 0.88 | 20.23 | 0.44ns |
| ~ Ocoteo*Slope+(1|stand) | 0.94 | 0.94 | 20.08 | 0.42ns | |
| ~ Firewood*Slope+(1|stand) | 0.82 | 0.82 | 20.33 | 0.45ns | |
| ~ %FC*Slope+(1|stand) | 0.01 | 0.01 | 14.92 | 0.65ns | |
| 2D | ~ Logging*Slope+(1|stand) | 0.92 | 0.92 | 20.11 | 0.75ns |
| ~ Ocoteo*Slope+(1|stand) | 0.98 | 0.98 | 20.03 | 0.75ns | |
| ~ Firewood*Slope+(1|stand) | 0.76 | 0.76 | 20.41 | 0.79ns | |
| ~ %FC*Slope+(1|stand) | 0.48 | 0.48 | 21.17 | 0.87ns | |
| (b) Pinus spp. ~ deforestation/forest use | |||||
| Pinus spp. | ~ Logging*Slope+(1|plot) | 0.34 | 0.92 | 363.98 | 0.003** |
| ~ Ocoteo*Slope+(1|plot) | 0.37 | 0.91 | 369.01 | 0.026* | |
| ~ Firewood*Slope+(1|plot) | 0.50 | 0.91 | 344.23 | 0.000*** | |
| ~ %FC*Slope+(1|plot) | 0.49 | 0.92 | 335.14 | 0.000*** | |
0D: observed species richness. 1D: typical or common species diversity. 2D: dominant species diversity. %FC: percent forest cover. r2m: determination coefficient of the model without the effect of the random variable. r2c: determination coefficient of model including the effect of the random variable. AICc: corrected Akaike Information Criterion (Hurvich & Tsai, 1989Hurvich CM, Tsai CL, 1991. Bias of the corrected AIC criterion for underfitted regression and time series models. Biome 78(3): 499-509. https://doi.org/10.1093/biomet/78.3.499). p: p-value relative to the null model (dependent variable ~ 1+(1|stand or plot)); *: p<0.05; **: p<0.01; ***: p<0.001; ns, not significant.
Finally, P. montezumae alone and all pine species combined (Pinus spp.) affected the abundance of B. cordata (Table 3b; Fig. 4a); furthermore, when we assessed only the influence of P. montezumae on the abundance of the rest of the species, we also detected an effect on P. hartwegii (Table 3b; Figure 4b). In the remaining four species (A. jorullensis, A. religiosa, A. xalapensis, P. teocote), we detected a significant relationship and optimal models according to AICc, but the explanation of the data by the model was very low (r2 = 0.2-0.10) (Table 3a,b); therefore, we deemed these relationships inconclusive.
| Model | r2 | p |
|---|---|---|
| (a) Other trees ~ Pinus spp. | ||
| Alnus jorullensis ~ Pinus spp. | 0.09 | 0.00*** |
| Abies religiosa ~ Pinus spp. | 0.02 | 0.00*** |
| Arbutus xalapensis ~ Pinus spp. | 0.02 | 0.00*** |
| Buddleja cordata ~ Pinus spp. | 0.27 | 0.00*** |
| Comarostaphylis discolor ~ Pinus spp. | 0.00 | 0.77ns |
| Cupressus benthamii ~ Pinus spp. | 0.02 | 0.46ns |
| Pinus pseudostrobus ~ Pinus spp. | 0.05 | 0.08ns |
| Quercus laurina ~ Pinus spp. | 0.01 | 0.19ns |
| Quercus crassipes~ Pinus spp. | 0.02 | 0.24ns |
| Quercus dysophylla ~ Pinus spp. | 0.04 | 0.50ns |
| (b) Other trees ~ Pinus montezumae | ||
| Alnus jorullensis ~ Pinus montezumae | 0.10 | 0.00*** |
| Abies religiosa ~ Pinus montezumae | 0.05 | 0.00*** |
| Arbutus xalapensis ~ Pinus montezumae | 0.02 | 0.00*** |
| Buddleja cordata ~ Pinus montezumae | 0.29 | 0.00*** |
| Comarostaphylis discolor ~ Pinus montezumae | 0.00 | 0.98ns |
| Cupressus benthamii ~ Pinus montezumae | 0.02 | 0.41ns |
| Pinus pseudostrobus ~ Pinus montezumae | 0.03 | 0.14ns |
| Quercus laurina ~ Pinus montezumae | 0.02 | 0.16ns |
| Quercus crassipes~ Pinus montezumae | 0.00 | 0.58ns |
| Quercus dysophylla ~ Pinus montezumae | 0.06 | 0.42ns |
| Pinus teocote ~ Pinus montezumae | 0.04 | 0.00*** |
| Pinus ayacahuite ~ Pinus montezumae | 0.00 | 0.93ns |
| Pinus hartwegii ~ Pinus montezumae | 0.57 | 0.00*** |
| Pinus leiophylla ~ Pinus montezumae | 0.00 | 0.78ns |
| Pinus pseudostrobus ~ Pinus montezumae | 0.00 | 0.73ns |
r2: determination coefficient of the model; p: p-value obtained using a likelihood-ratio test, *: p<0.05; **: p<0.01; ***: p<0.001; ns, not significant.
Discussion
⌅Forest use did not influence tree diversity, but did influence pine dominance, and forest stands with higher deforestation showed a higher pine dominance; therefore, our predictions were partially consistent with these findings. The strong effect of slope on pine dominance in La Malinche National Park is remarkable and agrees with previous studies in other mountain systems that have also reported the high contribution of slope on species composition (e.g., Yang et al., 2020Yang J, El-Kassaby YA, Guan W, 2020. The effect of slope aspect on vegetation attributes in a mountainous dry valley, Southwest China. Sci Rep 10: 16465. https://doi.org/10.1038/s41598-020-73496-0).
We also observed the effect of slope in the levels of dissimilarity. We observed the greater affinity in species richness between the forest stands East-1 and East-3, while we found a greater dissimilarity between East-2 and West-3. The greater species turnover between these forest stands is possibly a result of historical local disturbance processes that have modified and reduced ecological niches for various species (Monárrez-González et al., 2018Monárrez-González JC, Pérez-Verdín G, López-González C, Márquez-Linares MA, González-Elizondo MS, 2018. Efecto del manejo forestal sobre algunos servicios ecosistémicos en los bosques templados de México. Mad Bosq 24: 1-16. https://doi.org/10.21829/myb.2018.2421569), which has probably affected the distribution of species between stands. The same pattern was detected in 1D and 2D, that is, the lowest dissimilarity between North-2 and North-3, and a strong differentiation between East-1 and Southeast-1, apparently due to the loss of common and dominant species (e.g., P. teocote, Q. laurina).
The dissimilarity between pairs of communities describes how communities respond to different factors, such as anthropogenic influences (e.g., deforestation; Hubbell, 1997Hubbell SP, 1997. A unified theory of biogeography and relative species abundance and its application to tropical rain forests and coral reefs. Cor Reef 16: S9-S21. https://doi.org/10.1007/s003380050237), climate change, and environmental gradients (Myers et al., 2013Myers JA, Chase JM, Jiménez I, Jørgensen PM, Araujo-Murakami A, Paniagua-Zambrana N, et al., 2013. Beta-diversity in temperate and tropical forests reflects dissimilar mechanisms of community assembly. Ecol Lett 16: 151-157. https://doi.org/10.1111/ele.12021), conditions that are highly variable between the slopes of La Malinche National Park. Despite this, our results did not find a significant effect of deforestation and forest use on tree diversity, possibly due to the need to increase the sample size (i.e., the number of sampling plots) to improve statistical power or include other potential drivers of diversity change (e.g., selective logging, distance to human settlements, main economic activity) that could help better explain the distribution of diversity. This is the case of selective logging, which drastically reduces population density until entire populations are lost (Galindo-Jaimes et al., 2002Galindo-Jaimes L, González-Espinosa M, Quintana-Ascencio P, García-Barrios L, 2002. Tree composition and structure in disturbed stands with varying dominance by Pinus spp. in the highlands of Chiapas, Mexico. Plant Ecol 162: 259-272. https://doi.org/10.1023/A:1020309004233). Therefore, further studies should consider other variables potentially affecting tree diversity.
Although deforestation and forest-use variables did not affect tree diversity, they are important indicators of forest deterioration. For example, logging can lead to the homogenization of communities (McKinney & Lockwood, 1999McKinney ML, Lockwood JL, 1999. Biotic homogenization: a few winners replacing many losers in the next mass extinction. Tren Ecol Evol 14(11): 450-453. https://doi.org/10.1016/S0169-5347(99)01679-1) by ecological drift, as it induces the loss of underrepresented species and drastically decreases richness by random fluctuations (Gilbert & Levine, 2017Gilbert B, Levine JM, 2017. Ecological drift and the distribution of species diversity. Proc Roy Soc B Biol Sci 284: 20170507. https://doi.org/10.1098/rspb.2017.0507). Unlike logging or firewood extraction, ocoteo has no apparent effect on the structure of pine populations since it does not eliminate trees; nevertheless, it can significantly affect populations by weakening the vigor of trees in the long term until they die or are felled by the wind (Díaz-Carranza et al., 2022Díaz-Carranza G, Andrés Hernández AR, Guillén S, Rivas-Arancibia SP, Montoya Esquivel A, 2022. Dendrometric variables and traumatic resin ducts in pine species associated with wood-stripping: a traditional practice in the conifer forests of Central Mexico. Bot Sci 100(1): 42-55. https://doi.org/10.17129/botsci.2822). In La Malinche National Park, P. montezumae is the pine species most intensively exploited by ocoteo (Díaz-Carranza et al., 2022Díaz-Carranza G, Andrés Hernández AR, Guillén S, Rivas-Arancibia SP, Montoya Esquivel A, 2022. Dendrometric variables and traumatic resin ducts in pine species associated with wood-stripping: a traditional practice in the conifer forests of Central Mexico. Bot Sci 100(1): 42-55. https://doi.org/10.17129/botsci.2822), and, according to Acuña (1984)Acuña R, 1984. Relaciones geográficas del siglo XVI: Tlaxcala. Instituto de Investigaciones Antropológicas, UNAM, México. https://doi.org/10.22201/iia.9688374911p.1985 and González (2006)González JA, 2006. El ambiente y la agricultura en Tlaxcala durante el siglo XVI. Persplatin 3: 19-46., oak trees in La Malinche National Park have been used for carbon, firewood, and wood for construction since the conquest. Therefore, the evaluation of the effect of these types of forest use on tree diversity was a necessary aspect to study for the conservation of the remaining forests, and its effect on the abundance of pine species is relevant to report.
It has already been shown that areas cleared by deforestation favor the establishment and dominance of pioneer species, such as pine, while shade-tolerant species are lost (McKinney & Lockwood, 1999McKinney ML, Lockwood JL, 1999. Biotic homogenization: a few winners replacing many losers in the next mass extinction. Tren Ecol Evol 14(11): 450-453. https://doi.org/10.1016/S0169-5347(99)01679-1; Pessoa et al., 2017Pessoa MS, Hambuckers A, Benchimol M, Rocha-Santos L, Bomfim JA, Faria D, et al., 2017. Deforestation drives functional diversity and fruit quality changes in a tropical tree assemblage. Perspec Plant Ecol Evol Syst 28: 78-86. https://doi.org/10.1016/j.ppees.2017.09.001; Monárrez-González et al., 2018Monárrez-González JC, Pérez-Verdín G, López-González C, Márquez-Linares MA, González-Elizondo MS, 2018. Efecto del manejo forestal sobre algunos servicios ecosistémicos en los bosques templados de México. Mad Bosq 24: 1-16. https://doi.org/10.21829/myb.2018.2421569). Our results showed a pattern of higher dominance of Pinus species associated with lower percent forest cover, which coincides with the findings reported by other authors who have documented the prevalence of these species in deforested areas (Galindo-Jaimes et al., 2002Galindo-Jaimes L, González-Espinosa M, Quintana-Ascencio P, García-Barrios L, 2002. Tree composition and structure in disturbed stands with varying dominance by Pinus spp. in the highlands of Chiapas, Mexico. Plant Ecol 162: 259-272. https://doi.org/10.1023/A:1020309004233; Figueroa-Rangel et al., 2008Figueroa-Rangel BL, Willis KJ, Olvera-Vargas M, 2008. 4200 years of pine-dominated upland forest dynamics in west-central México: human or natural legacy? Ecol 89: 1893-1907. https://doi.org/10.1890/07-0830.1).
Specifically, P. montezumae was the dominant species in most forest stands. Ern (1976)Ern H, 1976. Descripción de la vegetación montañosa en los estados mexicanos de Puebla y Tlaxcala. Willd Beiheft 10: 1-128. proposed that P. montezumae, like other pine species, has historically benefited from fires and forest burning in La Malinche National Park, favoring its high dominance between 2800 m and 3200 m asl. Nevertheless, the dominance of this pine species is relatively recent; Ern (1976)Ern H, 1976. Descripción de la vegetación montañosa en los estados mexicanos de Puebla y Tlaxcala. Willd Beiheft 10: 1-128. indicated that in the past, the climax forest communities between 2300 m and 2700 m asl were A. religiosa, Cupressus lindleyi, and P. pseudostrobus. Even A. religiosa was present from 2400 m asl with large populations outside ravines of the wettest slopes - the only sites where it currently occurs (S. George-Miranda, pers. obs.; Rzedowski, 2006Rzedowski J, 2006. Vegetación de México. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO). 1st Ed dig, México. 504 pp.). In addition, Acuña (1984)Acuña R, 1984. Relaciones geográficas del siglo XVI: Tlaxcala. Instituto de Investigaciones Antropológicas, UNAM, México. https://doi.org/10.22201/iia.9688374911p.1985 reported the abundant presence of firs and cypresses in 1577, indicating a major shift in the floristic composition in La Malinche National Park that favored the dominance of Pinus species. Moreover, the scarcity we found of B. cordata, Quercus crasssipes, Q. laurina, and P. serotina may be explained because the study area is above the optimal altitudinal limit reported for these species (<2700 m asl; Ern, 1976Ern H, 1976. Descripción de la vegetación montañosa en los estados mexicanos de Puebla y Tlaxcala. Willd Beiheft 10: 1-128.). Another aspect that may account for the dominance of P. montezumae is the expansion of dry environments, which has intensified due to the historical increase of fires, deforestation, and climate change (Figueroa-Rangel et al., 2008Figueroa-Rangel BL, Willis KJ, Olvera-Vargas M, 2008. 4200 years of pine-dominated upland forest dynamics in west-central México: human or natural legacy? Ecol 89: 1893-1907. https://doi.org/10.1890/07-0830.1; Richardson & Bond, 1991Richardson DM, Bond WJ, 1991. Determinants of plant distribution: evidence from pine invasions. The Amer Nat 137(5): 639-668. https://doi.org/10.1086/285186).
Our study detected a positive association between P. montezumae and B. cordata, possibly due to their high tolerance to degraded soils (Mendoza-Hernández et al., 2010Mendoza-Hernández PE, Orozco-Segovia A, Pisanty I, 2010. Germination, emergence, and survival of Buddleja cordata in an urban forest. Ecol Rest 28(3): 263-265. https://doi.org/10.3368/er.28.3.263). In contrast, the abundance of P. hartwegii decreased when the dominance of P. montezumae increased, this is potentially explained because fire exerts a strong adverse impact on P. hartwegii while it seems to favor P. montezumae (Endara et al., 2012Endara AAR, Maass SF, Nava BG, Valdez JI, Fredericksen TS, 2012. Effect of human disturbance on the structure and regeneration of forests in the Nevado de Toluca National Park, Mexico. J For Res 23: 39-44. https://doi.org/10.1007/s11676-012-0226-8). Nevertheless, it is worth mentioning that the altitudinal distribution of P. hartwegii (3000-4300 m asl; Manzanilla-Quiñones et al., 2019Manzanilla Quiñones U, Delgado Valerio P, Hernández Ramos J, Molina Sánchez A, García Magaña JJ, Rocha Granados MDC, 2019. Similaridad del nicho ecológico de Pinus montezumae y P. pseudostrobus (Pinaceae) en México: implicaciones para la selección de áreas productoras de semillas y de conservación. Act Bot Mex 126: e1398. https://doi.org/10.21829/abm126.2019.1398) barely covered a small part of the study area (2800-3200 m asl). Likewise, the altitudinal displacement of P. montezumae resulting from climate change probably causes the observed relationship - a hypothesis that should be tested in a future study. The case of species that showed a low explanation of their models (A. jorullensis, A. religiosa, A. xalapensis, and P. teocote) deserves to be studied in further detail with a greater sampling effort and considering additional factors that may influence diversity (e.g., environmental characteristics, nutrients) since the negative relationships of P. montezumae with A. religiosa and P. teocote suggest an incipient displacement that should be identified to preserve forest diversity in La Malinche National Park.
The low abundance of other species in the National Park, such as Q. dysophylla, highlights the importance of La Malinche National Park as a diversity reservoir of tree species and the need to establish immediate strategies to increase the abundance of underrepresented species. A suggestion to address this aspect is to plan reforestation with these species, including monitoring over at least two years to improve success, promote increased tree cover, and preserve tree diversity. An issue with this approach is that the species produced in Mexican nurseries are those characterized by rapid and easy growth, and these species are not always suitable for reforestation. Therefore, additional studies are required to improve the propagation of tree species found in low abundance for use in reforestation projects. The case of P. ayacahuite is especially interesting because, according to Ern (1976)Ern H, 1976. Descripción de la vegetación montañosa en los estados mexicanos de Puebla y Tlaxcala. Willd Beiheft 10: 1-128., this species was not part of the local forest vegetation in La Malinche National Park; therefore, the few recorded individuals might have resulted from reforestation efforts in the area.
Although the altitudinal belt where the sampling was carried out has always been covered by conifer forests, the results of our study indicate a trend of shift toward the dominance of Pinus spp. and a major effect of slope on species composition. Even though the forest use indicators do not consider the history of continuous use of forest resources in La Malinche National Park (Table S1(Tables S1-S4) accompanies the paper on Forest System´s website [suppl]), these quantified the intensity of use at the time of the study; therefore, they provide valuable information to achieve our objectives. Nevertheless, given the lack of this important information, we do not rule out the need to measure the variation of forest use over time in future research.
Although we found that not only forest use modifies tree diversity, the adverse effects of forest use and deforestation on species composition warn of the need to implement immediate measures to protect it in La Malinche National Park. We propose the following actions: (1) control human activities that promote deforestation in La Malinche National Park (e.g., logging, ocoteo, firewood extraction, and changes in land use) and (2) promote reforestation using native and poorly represented species (e.g., Q. dysophylla), along with long-term monitoring (>2 years), to increase forest cover and maintain species diversity. Otherwise, the advance of the agricultural frontier and illegal deforestation (e.g., logging) will cause the loss of species diversity, the conversion to monospecific forests, and the deterioration of ecosystem services in the region (e.g., Cayuela et al., 2006Cayuela L, Golicher DJ, Benayas JMR, González-Espinosa M, Ramírez-Marcial N, 2006. Fragmentation, disturbance and tree diversity conservation in tropical montane forests. J Appl Ecol 43: 1172-1181. https://doi.org/10.1111/j.1365-2664.2006.01217.x). The present study contributes relevant knowledge on the effect of deforestation in an extremely threatened ecosystem of great importance in Mexico, which could be useful to define forest management actions for La Malinche National Park and other Mexican conifer forests.