Can we expect to protect threatened species in protected areas? A case study of the genus Pinus in Mexico

Podemos proteger especies en riesgo en áreas protegidas? Un estudio de caso del genero Pinus en México

Jesús Aguirre Gutiérrez* and Joost F. Duivenvoorden

Institute for Biodiversity and Ecosystems Dynamics (IBED), Universiteit van Amsterdam, Science Park 904, 1098 HX, Amsterdam, The Netherlands. *Correspondent: jeaggu@gmail.com

Abstract

The distribution of 56 Pinus species in Mexico was modelled with MAXENT. The pine species were classified as threatened according to IUCN criteria. Our aim was to ascertain whether or not threatened pine species were adequately represented in protected areas. Almost 70% of the species had less than 10% of their modelled distribution area protected. None of the pine species reached their representation targets. Threatened pine species were less widely distributed, occurred at lower maximum elevations, and were less well represented in protected areas than other pine species. The results suggest that the present system of protected areas in Mexico fails to protect pine species adequately. Conservation targets should be especially directed to species with narrow distributions which occur at low altitudes, such as Pinus. attenuata, P. cembroides subsP. cembroides var. lagunae, P. radiata var. binata, P. rzedowskii, and P. muricata.

Key words: MAXENT, species distribution model, IUCN, conservation.

Resumen

La distribución de 56 especies del género Pinus en México fue modelada por medio de MAXENT. Nuestro objetivo principal fue investigar si las especies de pino clasificadas por IUCN como amenazadas tienen una representación adecuada en las áreas protegidas de México. Se encontró que casi el 70% de las especies tienen menos del 10% de su distribución modelada protegida. Ninguna de las especies alcanzó el nivel de representación propuesto como adecuado. Se observó que las especies de pino clasificadas como amenazadas tienen un distribución más estrecha, ocurren a menores elevaciones máximas y se encuentran menos representadas en las áreas protegidas en comparación con las otras especies de pino modeladas. Los resultados sugieren que la red actual de áreas protegidas en México no protege adecuadamente el genero Pinus. Proponemos que los esfuerzos de conservación deben estar dirigidos especialmente a especies con distribución reducida y que se encuentran principalmente distribuidas a bajas altitudes, como por ejemplo Pinus attenuata, P. cembroides subsP. cembroides var. lagunae, P. radiata var. binata, P. rzedowskii and P. muricata.

Palabras clave: MAXENT, modelos de distribución de especies, IUCN, conservación.

Introduction

The same factors that increase the probability of a species being endangered also increase its risk of not being present in protected areas. The origin for this unwanted situation is twofold. First, narrowly distributed species, i.e. species with small range sizes, low population densities, or occurring in widely spaced small patches, are more vulnerable than widespread species (Gaston et al., 1997; Rabinowitz, 1981; Purvis et al., 2000). Unfortunately, narrowly distributed species are not likely to occur in protected areas (Rodrigues and Gaston, 2001). Thus, species that are potentially endangered because of their population structure are prone to be misrepresented in areas designed to protect them. Secondly, many species are severely threatened by high degrees of human disturbance. Such species are likely to be underrepresented in protected areas, as these natural reserves tend to be located in areas where human influence is minimized (Cantú et al., 2004). The aim of this study is to test these 2 propositions by means of a distribution analysis of species from the genus Pinus in Mexico.

Mexico is recognized as a mega diverse nation with a high level of endemism (Rzedowski, 1991). In Mexico, 56 pine taxa have been recorded (Farjon, 2001), a great proportion of which are endemic to the country or have quite narrow distributions (Perry, 1991; Farjon and Styles, 1997). The International Union for Conservation of Nature and Natural Resources (IUCN, 2009) listed 15 Pinus species as threatened according to 3 categories: Endangered (EN), Vulnerable (VU) or Near–Threatened (NT) (see Table 1). The Red List Criteria used by IUCN (2001) for listing species as threatened were population size, geographic range, and species probabilities of extinction. The extent of species distributions in protected areas (i.e. the degree of representation) was not part of these criteria.

To protect biodiversity, the IUCN proposed that 10–12% of a countries' territory should be dedicated to the establishment of protected areas (IUCN, in Rodrigues et al., 2003). However, this proposal did not take into account the distribution of biodiversity, nor did it consider specific requirements to protect species adequately, such as the size of protected areas (Rodrigues et al., 2003). Rodrigues et al. (2004a) stipulated that the total cover of protected areas in countries does not give adequate information regarding the true level of protection. In Mexico, the conservation and protection of biodiversity has not always been the primary reason to establish protected areas (CONANP–CONABIO, 2007). Despite the establishment of 166 federal protected areas which cover almost 12% of its territory (CONANP, 2009a), recent studies have reported an inadequate levels of representation for several species (Cantú et al., 2004; Riemann and Ezcurra, 2005).

In this study, the distribution of species from the genus Pinus was modelled by means of the Maximum entropy modeling software "MAXENT", (Phillips et al., 2006) to quantify the degree of representation of Pinus species in the protected areas of Mexico. We analyzed whether narrow distributions and/or occurrences at relative low elevations (high human pressure) affected the species' representation levels. We expected to find that pine taxa which were classified as threatened on the basis of IUCN criteria (Table 1) were poorly represented in the Mexican protected areas.

Materials and methods

Species distribution data. The study included the 56 pine taxa listed for Mexico in the World Checklist and Bibliography of Conifers (Farjon, 2001) (Table 1). A total of 5336 location records for these taxa were extracted from the BRAHMS database (BRAHMS database, University of Oxford, http://dps.plants.ox.ac.uk/bol/documentation). Subsequently, the records were geo–referenced to Lambert Conformal Conic and projected on a map of Mexico. The records that appeared as duplicates or were erroneously geo–referenced (1469), were deleted, leading to a total of 3 867 points used in the analysis.

Layer preparation. In total, 8 environmental layers were used in the species distribution modeling: a digital elevation model (DEM) (INEGI, 2009), slope, precipitation, temperature, soil types, humidity, soil humidity regimes and land vegetation cover (CONABIO, 2009). The DEM was downloaded at a grain size of 30 x 30 m, in fragments with a maximum size of 2 degrees square. These fragments were pasted together to obtain a single DEM layer covering the entire country. The slope layer was constructed from the DEM, using the slope function from the spatial analyst tool of the ArcInfo software version 9.3 (ESRI, 2009), at a resolution of 300 x 300 m. The layers of average annual precipitation, humidity, and soil humidity regimes were obtained in 9 classes, at a scale of 1:4 000 000. The layer of average annual temperature contained 6 categories and was also at a scale of 1:4 000 000. The land use and vegetation layer, modified by CONABIO (1999), had a scale of 1:1 000 000 and included 27 categories. Finally, the soil types layer contained 32 different categories and was available at a scale of 1:4 000 000. All layers were changed into a raster format with a resolution of 300 x 300 m at the same extent. A total of 166 federal protected areas (CONANP, 2009b) were included in a separate layer. The dataset was analyzed with the software ArcInfo version 9.3 (ESRI, 2009).

For each pine taxon a MAXENT model was generated using a cumulative output (Fig. 1). This output showed the so–called relative suitability for each grid cell, applying values from 0 to 100, with higher values meaning higher suitability.

The models generated by Maxent were evaluated by their area under the ROC curve (AUC) generated for each model, and by the test proposed by Pearson et al. (2007) in the case of species with small sample sizes. Models with AUC values from 0.75 and above were considered useful, as recommended by Elith (2000. In Phillips and Dudik, 2008). The AUC is a threshold–independent method that measures the overall model performance. AUC values range from 0 to 1, with values of 0.5 or below indicating a model that is no better than random and with a value of 1 when the model has perfect discrimination (Engler et al., 2004). The Maximum Training Sensitivity + Specificity (MTS+S) threshold criterion was used to create presence/absence maps for each pine taxon as suggested by Jimenez–Valverde and Lobo (2007). The presence/absence maps were converted to a polygon format yielding the total extent of occurrence in km² for each pine taxon. Representation levels of pine taxa were defined based on the overlap of the modelled distribution area with the layer of protected areas. Three representation classes (gap, partial gap or covered) were defined following Rodrigues et al. (2004b). Taxa for which the modelled distribution was fully outside any protected area were called gap taxa. Partial gap taxa showed a representation below a so–called target value. This value depended on the total extent of the modelled distribution of each taxon. For taxa with a total distribution of 1 000 km² or less the target representation value was defined as 100%. For taxa with a total distribution of 250 000 km² or more the target representation value was defined as 10%. For taxa with total distributions between 1 000 and 250 000 km², the target was interpolated as proposed by Rodrigues et al. (2004b). Covered taxa showed a representation equal or above the target value. All GIS analyses were carried out with ArcInfo version 9.3 (ESRI, 2009).

Differences in area protected between taxa classified as threatened and not threatened were tested by means of a Mann–Whitney test (Zar, 1984). Linear regression was used to examine how the area protected depended on distributional extent.

Results

All MAXENT models performed well as assessed by the AUC on the test data, which was above 0.76 for all species with the exception of 7 pine taxa. These had small sample sizes (6 or less points) and, therefore, the test proposed by Pearson et al. (2007) was applied (Table 1).

The total extent of the modelled species distributions was highly related to the total extent inside protected areas (Fig. 2). All species were classified as partial gap taxa. The representation levels did not go beyond 70% (Table 1). Thirty nine taxa showed a representation below 10%, 14 taxa had a representation between 10% and 20%, and only 3 taxa were represented at levels of 20% or above.

All threatened taxa had at least 90 percent of their extent of modelled occurrence below 2800 m. Pinus species which prefer lower altitudes were sparse in protected areas (Fig. 3). The maximum elevation of the modelled distribution of threatened pine species was considerably lower than that of the non–threatened pine species (Mann–Whitney test U= 173.5, p<0.05).

Discussion

Overall, our results suggest that the present system of protected areas in Mexico fails to adequately protect pine species. Almost 70% of the analyzed species had less than 10% of their modelled distribution area protected. None of the taxa reached their representation targets (following Rodrigues et al., 2004b), implying that their modelled distribution had insufficient overlap with protected areas. In view of the economic importance of pines for Mexico, this outcome is troublesome, even more so if we take into account that the deforestation rate for the country is extremely high (Semarnat, 2009; INEGI, 2007).

Threatened pine species were less widely distributed than other pine species. Presumably this is because species that occur in scattered and sparsely distributed populations are prone to extinction due to anthropogenic or stochastic effects (Gaston, 2003). Also, pine species distributed at higher altitudes were better represented in protected areas than the ones distributed at lower elevations. This corroborates conclusions by Cantú et al. (2004) that high elevation areas were better represented in the Mexican protected areas. Further, it agrees with observations by CONABIO et al. (2007) that pine–oak forests in Mexico are poorly represented in protected areas and that this vegetation type should receive a conservation priority also because of its high biological value. Our results indicated that conservation targets should be directed particularly toward species with narrow distributions, which occur at low altitudes, such as Pinus attenuata, Pinus cembroides ssp. lagunae, Pinus muricata, Pinus radiata var. binata, and Pinus rzedowskii. Specifically, the populations of Vaca Pinta and Chiqueritos in Michoacán for P. rzedowskii, Vereda and Picacho in Baja California Sur for P. cembroides ssp. lagunae and Colorado, Baja California, for P. muricata have been proposed by Delgado et al. (2008) as important regions for conservation following phylogenetic, genetic, and demographic information. Arguably, some of the important areas for establishing new protected areas with a focus on the conservation of the genus Pinus are the Sierra Madre Occidental and Sierra Madre del Sur, where many of the Mexican pine taxa exist and where protected areas are scarce.

Threatened pine species occurred at lower maximum elevations that non–threatened species, and were less well represented in protected areas. This substantiates the fact that at low elevations in Mexico, where human population density is highest, and where land cover changes and fragmentation impacts are most severe, more protection is needed (Fjeldså and Rahbek, 1998; Balmford et al., 2001; Menon et al., 2001; Cantú et al., 2004).

Aljos Farjon has kindly allowed access to Brahms database. Guido van Reenen has helped with the MAXENT modeling. CONABIO has made available the environmental layers dataset. We kindly thank Maria Pontes and Arthur Wuster for their help with the language corrections. This study was part of the MSc study in Tropical Ecology at the Universiteit van Amsterdam by the first author.

Literature cited

Balmford, A., J. L. Moore, T. Brooks, N. Burgess, L. A. Hansen, P. Williams and C. Rahbek. 2001. Conservation Conflicts Across Africa. Science 291:2616–2619.         [ Links ]

BRAHMS and BRAHMS online. 2009. Department of Plant Sciences, University of Oxford. http://dps.plants.ox.ac.uk/bol/documentation/Default.aspx; 06.IV.2009.         [ Links ]

Cantú, C., R. G. Wright, J. M. Scott and E. Strand. 2004. Assessment of current and proposed nature reserves of Mexico based on their capacity to protect geophysical features and biodiversity. Biological Conservation 115:411–417.         [ Links ]

Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO). 1999. "Uso de suelo y vegetación modificado por CONABIO". Escala 1: 1000000. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad. Ciudad de México, México.         [ Links ]

Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO). 2009. http://conabioweb.conabio.gob.mx/metacarto/metadatos.pl; 17. III. 2009.         [ Links ]

Comision Nacional de Areas Naturales Protegidas (CONANP). 2009a. http://www.conanp.gog.mx/q_anp.html. 26.II.2009.         [ Links ]

Comision Nacional de Areas Naturales Protegidas (CONANP). 2009b. http://www.conanp.gob.mx/sig/informacion/info.htm. 18.V.2009.         [ Links ]

CONANP–CONABIO. 2007. Mapa ANP para los análisis de vacíos y omisiones en conservación. Comisión Nacional de Áreas Naturales Protegidas y Comisión Nacional para el Conocimiento y Uso de la Biodiversidad, México.         [ Links ]

Comisión Nacional para el Conocimiento y Uso de la Biodiversidad (CONABIO)– Comisión Nacional de Áreas Naturales Protegidas (CONANP)– The Nature Conservancy (TNC)–PRONATURA–FCF, UANL. 2007. Análisis de vacíos y omisiones en conservación de la biodiversidad terrestre de México: espacios y especies. Programa México, Pronatura, A.C., Facultad de Ciencias Forestales and Universidad Autónoma de Nuevo León, México. 127 p.         [ Links ]

Delgado, Patricia, Luís Eguiarte, Francisco Molina–Freaner, Elena Alvarez–Buylla, and Daniel Piñero. 2008. Using phylogenetic, genetic and demographic evidence for setting conservation priorities for Mexican rare pines. Biodiversity and Conservation 17:121–137.         [ Links ]

Elith, J. 2000. Quantitative methods for modeling species habitat: comparative performance and an application to Australian plants. In Quantitative Methods for Conservation Biology, S. Ferson and M. Burgman (eds.). Springer New York. p. 30–58. In S. Phillips and M. Dudik. 2008.         [ Links ]

Engler, R., A. Guisan and L. Rechsteiner. 2004. An improved approach for predicting the distribution of rare and endangered species from occurrence and pseudo–absence data. Ecology 41:263–274.         [ Links ]

ESRI. 2009. ESRI Inc. Redlands, CA, USA.         [ Links ]

Farjon, A. and B. T. Styles. 1997. Pinus (Pinaceae). New York Botanical Garden Press, New York. 291 p.         [ Links ]

Farjon, A. 2001. World Checklist and Bibliography of Conifers. Royal Botanic Gardens, Kew, UK. 309 p.         [ Links ]

Fjeldså, J. and C. Rahbek. 1998. Continent–wide conservation priorities and diversification processes. In Conservation in a changing world, G. M. Mace, A. Balmford and J. R. Ginsberg (eds.). Cambridge University Press, Cambridge, UK. p. 139–160.         [ Links ]

Gaston, K., T. M. Blackburn and J. H. Lawton. 1997. Interspecific Abundance–Range Size Relationships: An Appraisal of Mechanisms. Journal of Animal Ecology 66:579–601.         [ Links ]

Gaston, K. 2003. The Structure and Dynamics of Geographic Ranges. Oxford University Press, Oxford, UK. 280 p.         [ Links ]

Instituto Nacional de Estadística y Geografía (NEGI). 2007. Producción maderable – especies – valor – 1980 – 2006 – nacional. www.inegi.org.mx; 06.IV.2009.         [ Links ]

Instituto Nacional de Estadística y Geografía (INEGI). 2009: www.inegi.org.mx; 06.IV.2009.         [ Links ]

International Union for Conservation of Nature (IUCN). 2001. IUCN Red List Categories and Criteria: Version 3.1. IUCN Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. 30 p        [ Links ]

International Union for Conservation of Nature (IUCN). 2009. 2008 IUCN Red List of Threatened Species. www.iucnredlist.org. Downloaded on 26 February 2009.         [ Links ]

Jimenez–Valverde, A. and J. M. Lobo. 2007. Threshold criteria for conversion of probability of species presence to either–or presence–absence. Acta Oecologica 31:361–369.         [ Links ]

Menon, S., G. Pontius, J. Rose, M. L. Khan and K. S. Bawa. 2001. Identifying Conservation Priority Areas in the Tropics: a Land–Use Change Modelling Approach. Conservation Biology 15:501–512.         [ Links ]

Pearson, R. G, C. J. Raxworthy, M. Nakamura, and A. T. Peterson. 2007. Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. Journal of Biogeography 34:102–117.         [ Links ]

Perry, J. P. 1991. The Pines of Mexico and Central America. Timber Press, Portland, Oregon. 231 p.         [ Links ]

Phillips, S. J., R. P. Anderson and R. E. Schapire. 2006. Maximum entropy modeling of species geographic distribution. Ecological Modelling 190:231–259.         [ Links ]

Phillips, S. J. and M. Dudik. 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161–175.         [ Links ]

Purvis, A., J. L. Gittleman, G. Cowlishaw and G. M. Mace. 2000. Predicting extinction risk in declining species. Proceedings of the royal society London. B. 267:1947–1952.         [ Links ]

Rabinowitz, D. 1981. Seven forms of rarity. In The biological aspects of rare plants conservation, H. Synge (ed.). John Wiley & Sons, New York. p. 205–217.         [ Links ]

Riemann, H. and E. Ezcurra. 2005. Plant endemism and natural protected areas in the peninsula of Baja California, Mexico. Biological Conservation 122:141–150.         [ Links ]

Rodrigues, A. S. L., and K. J. Gaston. 2001. How large do reserve networks need to be?. Ecology Letters 4:602–609.         [ Links ]

Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. Da Fonseca, K. J. Gaston, M. Hoffman, J. Long, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts and Y. Xie. 2003. Global Gap Analysis: towards a representative network of protected areas. Advances in Applied Biodiversity Science 5. Conservation. International.Washington DC. 98 p.         [ Links ]

Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. Da Fonseca, K. J. Gaston, M. Hoffmann, J. S. Long, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts and X. Yan. 2004a. Effectiveness of the global protected area network in representing species diversity. Nature 428:640–643.         [ Links ]

Rodrigues, A. S. L., H. R. Akçakaya, S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, J. S. Chanson, L. D. C. Fishpool, G. A. B. Da Fonseca, K. J. Gaston, M. Hoffmann, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts and X. Yan. 2004b. Global Gap Analysis: Priority Regions for Expanding the Global Protected–Area Network. BioScience 54 12:1092–1100.         [ Links ]

Rzedowski, J. 1991. Diversidad y origenes de la flora fanerogámica de México. Acta Botánica Mexicana 14:3–21.         [ Links ]

Secretaría de Medio Ambiente y Recursos Naturales (Semarnat). 2009. Compendio de estadisticas ambientales 2009. http://appl.semarnat.gob.mx/dgeia/compendio_2009/compendio_2009/10.100.8.236_8080/ibi_apps/WFServlet982a.html; 20.VII.2009.         [ Links ]

Zar, J. H. 1984. Biostatistical analysis. Prentice–Hall, Englewood Cliffs, New Jersey. 736 p.         [ Links ]