Modeling of spatial distribution for scorpions of medical importance in the São Paulo State, Brazil

Research (Published online: 07-07-2015)

2.  Modeling of spatial distribution for scorpions of medical importance in the São Paulo State, Brazil - José Brites-Neto and Keila Maria Roncato Duarte

Veterinary World, 8(7): 823-830

   doi: 10.14202/vetworld.2015.823-830

---

Abstract

---

Aim: In this work, we aimed to develop maps of modeling geographic distribution correlating to environmental suitability for the two species of scorpions of medical importance at São Paulo State and to develop spatial configuration parameters for epidemiological surveillance of these species of venomous animals.

Materials and Methods: In this study, 54 georeferenced points for Tityus serrulatus and 86 points for Tityus bahiensis and eight environmental indicators, were used to generate species distribution models in Maxent (maximum entropy modeling of species geographic distributions) version 3.3.3k using 70% of data for training (n=38 to T. serrulatus and n=60 to T. bahiensis) and 30% to test the models (n=16 for T. serrulatus and n=26 for T. bahiensis). The logistic threshold used to cut models in converting the continuous probability model into a binary model was the “maximum test sensitivity plus specificity,” provided by Maxent, with results of 0.4143 toT. serrulatus and of 0.3401 to T. bahiensis. The models were evaluated by the area under the curve (AUC), using the omission error and the binomial probability. With the data generated by Maxent, distribution maps were produced using the “ESRI® ArcGIS 10.2.2 for Desktop” software.

Results: The models had high predictive success (AUC=0.7698±0.0533, omission error=0.2467 and p<0.001 for T. serrulatus and AUC=0.8205±0.0390, omission error=0.1917 and p<0.001 for T. bahiensis) and the resultant maps showed a high environmental suitability in the north, central, and southeast of the state, confirming the increasing spread of these species. The environmental variables that mostly contributed to the scorpions species distribution model were rain precipitation (28.9%) and tree cover (28.2%) for the T. serrulatus and temperature (45.8%) and thermal amplitude (12.6%) for the T. bahiensis.

Conclusion: The distribution model of these species of medical importance scorpions in São Paulo State revealed a higher environmental suitability of these species in the regions north, central, and southeast of the state, warning to emergencies actions for prevention and surveillance from scorpion stings in several counties. There is also a need to best conservation strategies related to neighboring territories, with the implementation of new environmental protected areas and measures of spread control of these species in urban areas of several counties.

Keywords: binomial probability, environmental variable, georeferencing, maxent, Tityus bahiensis, Tityus serrulatus.