Potential Distribution of Rattlesnakes in Mexico

Authored by Citlalli Edith Esparza-Estrada and José Jesús Sigala-Rodríguez

Rattlesnakes are an important part of the biological and cultural diversity of Mexico. It is an easily-recognizable animal, and the evolutionary story of the group points to northern Mexico as the geographical origin. Rattlesnakes are viperid snakes belonging to the subfamily Crotalinae (Ernst y Ernst 2011), where there are 11 genera and 115 species. From those, 38 species are rattlesnakes in two genera: 35 in Crotalus Linnaeus, (1758) and three in the genus Sistrurus Garman, (1883) (Campbell and Lamar 2004).

Although we have distribution maps for most species (Campbell and Lamar 2004), combining potential distribution with those maps and new records would improve the spatial distribution for most species. For this endeavor we used the ecological niche modeling technique and obtained areas for potential distribution for the genus Crotalus in Mexico.

To build the distribution models we obtained collecting records from literature, zoological collections through the Herpetological Network (HerpNet www.herpnet.org), Global Biodiversity Information Facility (GBIF www.gbif.org), the Red Mundial de Información sobre la Biodiversidad (REMIB www.conabio.gob.mx/remib/doctos/remib_esp.html) from the CONABIO (Comisión Nacional para el Conocimiento y Uso de la Biodiversidad) and unpublished records. We accumulated 15,594 presence records from 62 institutions, and after a detailed locality selection and removal of duplicate records, we ended with 8,232 point localities from the following sources: 5,562 records from HerpNET (67.57% of the total of obtained localities), 2,441 records from GBIF (29.65%), 115 records from REMIB (1.40%) and 114 new localities (1.38%) (Alvarado Díaz and Campbell 2004, Cambpell and Lamar 2008, Anderson and Greenbaum 2012, and personal records).

We georeferenced 1,718 localities (20.87% of the validated localities) that contained complete geographical information using the point method (Chapman and Wieczorek 2006), and we subjected the georeference points to a scrutinous locality validation process that consisted of five stages in order to eliminate any locality with inaccurate or potentially wrong data. After the validation processes we ended with 1,623 records. The number of validated records for species ranged from 263 for C. atrox to only six for C. stejnegeri.

With this we obtained potential distribution models using the Maxent software for 22 species: Crotalus aquilus, C. atrox, C. basiliscus, C. cerastes, C. enyo, C. intermedius, C. lepidus, C. mitchellii, C. molossus, C. oreganus, C. polystictus, C. pricei, C. pusillus, C. ravus, C. ruber, C. scutulatus, C. simus, C. stejnegeri, C. tigris, C. totonacus, C. triseriatus and C. willardi. We used the 19 environmental layers and the topography from WorldClim (Hijmans et al. 2005) to choose from the resulting model the variables that contributed the most and built new models with those ~five variables.

Maxent generated accumulative logistic models with 500 iterations per sample, response curves (omission and commission errors), individual contribution and a jackknife analysis of the environmental variables (Phillips et al. 2006).

The logistic models were converted to binary maps of potential presence/absence using two methods: maximum training sensitivity plus specificity and 10 percentile training presence, in concordance with the proposed by Liu et al. (2005). From those two, we considered the model that was more conservative to avoid the overestimation of the presence of the species.

The models obtained high values of accuracy in the validation analysis in Maxent, and there was consistency between the presence records, the environmental layers used and the distribution maps in the literature. Of the species, 68.18 percent included the temperature seasonality variable in the potential distribution model, making this one of the variables that contributed more to the model building for the rattlesnakes of Mexico.

Not surprisingly, and in concordance with the information in the literature, we observed the biggest diversity of rattlesnakes in the north-central part of Mexico, particularly in the Sierra Madre Occidental, the northwest part of the Sierra Madre Oriental, the northwest part of the Mexican Volcanic Belt and the northern portion of the Mexican Plateau. There is at least one rattlesnake species in every Mexican state and as many as 10 Crotalus species. The states with more species are Veracruz and Sonora, and the ones with less species are Chiapas, Quintana Roo, Yucatán and Tabasco. Although we are still refining the maps, the combination of our predictions with the ones from the literature add more accuracy to the distribution of the rattlesnakes in Mexico.

References
Anderson GC and E Greenbaum.2012. Phylogeography of Northern Populations of the Black-Tailed Rattlesnake (Crotalus molossus Baird and Girard, 1853), With the Revalidation of C. ornatus Hallowell, 1854. Herpetological Monographs 26(1):19-57.
Alvarado Díaz J and JA Campbell. 2004. A new montane rattlesnake (Viperidae) from Michoacán, México. Herpetologica 60(2): 281-286.
Campbell JA and OA Flores-Villela. 2008. A new long-tailed rattlesnake (Viperidae) from Guerrero, Mexico. Herpetologica 64(2): 246-257.
Campbell JA and WW Lamar. 2004. The venomous reptiles of the western hemisphere, vol. I y II. Comstock/Cornell University Press. Ithaca, NY, USA. 775 p.
Chapman AD and J Wieczorek (eds). 2006. Guide to Best Practices for Georeferencing. Copenhagen: Global Biodiversity Information Facility. California, USA. 80 p.
Ernst CH and Ernst EM. 2011. Venomous reptiles of the United States, Canada and northern Mexico. 2 vol. The Johns Hopkins University Press. Maryland, VA, USA. 387 p.
Global Biodiversity Information Facility. disponible en: http://www.gbif.org. [Data portal gbif.org accessed on Mar 22, 2012].
Herpetological Network. Available in: http://www.herpnet.org. [Data portal HerpNet accessed on mar 2, 2012].
Hijmans RJ, SE Cameron, JL Parra, PG Jones and A Jarvis. 2005 Very high resolution Interpolate climate surfaces for global land areas. International Journal of Climatology; 25: 1965-1978.
Liu C, PM Berry, TP Dawson and RG Pearson. 2005. Selecting thresholds of occurrence in the prediction of species distributions. Ecography 28: 385-393.
Phillips SJ, RP Anderson and RE Schapire. 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190: 231-259.
Red Mundial de Información sobre Biodiversidad. Available in: http://www.conabio.gob.mx/remib/doctos/remib_esp.html. [Data portal REMIB accessed on Mar 25, 2012].