Printer Friendly

Modelling favourability for invasive species encroachment to identify areas of native species vulnerability.

1. Introduction

The effect of alien invasive species on native ecosystems is of high conservation concern because of the possible displacement of native species and the subsequent homogenization of global biodiversity [1]. Many authors have suggested that interspecific competition is one of the most important processes that determine the final outcome of biological invasions [2]. Therefore, the assessment of interactions between native and invasive species is crucial for predicting and preventing biological invasions [2, 3]. Thus, the analysis of the settlement of one exotic species in the distribution area of another similar native species could be helpful for understanding the synecological processes underlying biological invasions.

The Mediterranean pond turtle (Mauremys leprosa, Schweiger, 1812; Geoemydidae) is the most abundant Chelonia (Testudines) in the Iberian Peninsula. However, it is an endangered species whose populations have considerably declined during the last decades [4]. The distribution range includes the Iberian Peninsula and northwest Africa [5], and it can be found in different anthropogenic microhabitats and freshwater ecosystems [6, 7] such as rice fields or irrigation ditches [8]. The largest population of the species occurs in the Iberian Peninsula where it is classified as "vulnerable" in the Spanish red book [8].

The red-eared slider (Trachemys scripta, Schoepff, 1792; Emydidae) is an invasive species that is currently breeding in the Iberian Peninsula [9,10]. Overall, the red-eared slider and the Mediterranean pond turtle occupy similar microhabitats and according to different studies the former is competing with and displacing the latter [11, 12]. Nevertheless, the nature of the competitive interactions between the native and invasive species is unclear [12,13].

In this context, analysing the biogeographical interactions between the red-eared slider and the native Mediterranean pond turtle would increase our understanding of the distributional relationships between a native and a similar foreign species. The use of species distribution modelling is an appropriate approach for this objective, because if the competitive interactions between the introduced and the native species show any environmental trends, these should be reflected in ecogeographical models [14].

Our general aim is to provide a modelling methodology to identify areas of high conservation concern regarding a native species due to the encroachment of an invasive alien species. As a case study, we modelled the favourability for encroachment of the invasive red-eared slider into the area occupied by the native Mediterranean pond turtle to identify areas of vulnerability of the latter species. We proceeded in three steps: (1) we generated an encroachment favourability model for the red-eared slider within the range of the Mediterranean pond turtle to determine its potential to invade the area of the Mediterranean pond turtle; (2) we generated an ecogeographical favourability model for the native Mediterranean pond turtle; (3) we overlapped these two favourability models and calculated the mean favourability for each species in each overlap interval; and (4) we used these values to map the areas of conflict between the two species and the zones of more conservation concern regarding the Mediterranean pond turtle.

2. Methods

2.1. Study Area. This paper focuses on Malaga province, in the south of the Iberian Peninsula (Figure 1). The native Mediterranean pond turtle is widely distributed in the area occupying different types of freshwater ecosystems [4, 15]. Malaga covers an area of about 7,300 [km.sup.2] which puts it on the limit between the landscape and regional scales at which climate, topography, and land use are thought to control species distributions [16]. Malaga province has a population of more than 1,600,000 people [17]. The climate is affected by steep gradients due to the mountainous topography, with altitudes ranging from sea level to 2,000 meters a.s.l. The general climate is Mediterranean with hot dry summers and short mild winters. A subtropical Mediterranean climate prevails in the eastern coastal areas, whereas an oceanic Mediterranean climate is more evident in the west. Annual average temperatures range between 12.5[degrees]C and 19[degrees]C.

2.2. Species Distribution. Distribution data for both species (Figure 2, maps a1 and b1) came from an intensive survey in the study area (records published in Romero et al. [6, 7] and from Pleguezuelos et al. [15]). The two species studied occupied both lotic and lentic environments. The Mediterranean pond turtle occupied 72% of the study area, whereas the red-eared slider was found in 26% of the study area.

2.3. Variables. We used an initial pool of ecogeographic variables to identify the macroenvironmental factors that affect the turtles' distribution. We divided the variables into five explanatory factors: spatial situation, human activity, topography, climate, and land use (Table 1). We preferred to use a large biogeographical resolution scale rather than a local scale, as our aim was to describe the macroenvironmental processes around the sampling points that drive the distribution of both species rather than the ecological processes of each water reservoir that govern local occupancy For this reason, we used a resolution of 100 [km.sup.2] (104 cells in the study area). In this context, we expected that the distribution models based in ecogeographical factors will affect the potential distribution of both studied pond turtles. Specifically, we worked with 70 presence records of the Mediterranean pond turtle and 24 presence records of redeared slider, 19 of which were taken from the Atlas and five were new records.

The procedure for obtaining the variables related to spatial distribution, topography, climate, and human activity has been described in Barbosa et al. [18], Castro et al. [19], and Munoz et al. [20]. Land-use variables were extracted from Corine Land Cover [21]. We reclassified the categories into those shown in Table 1. The new land-cover classes, initially in polygon shape-file format, were processed using ESRI ArcMap 9.2 software. Polygons were converted to raster of 1000 x 1000 m using a spatial-analyst tool. Finally, the surface area of each category was calculated for each 100-[km.sup.2] square in Malaga province.

2.4. The Favourability Models. We analysed the environmental factors that influence the cooccurrence of the two species, first modelling each species separately. We used univariate logistic regression to obtain the statistical potential occurrence of each species as a function of each ecogeographic variable [22-24] in order to determine the predictor variables that were significantly related to the distribution of each species. To control the increase in type I errors due to multiple tests [25], we only accepted those variables that were significant under a false discovery rate of q < 0.05, using the Benjamini and Hochberg procedure [25]. We then performed forward-backward stepwise logistic regression on the subset of significant predictor variables to obtain a multivariate logistic model. These models were named the Mauremys-Ecogeographical model and the Trachemys-Ecogeographical model.

We then performed a similar procedure to assess the presence or absence of the red-eared slider in the range of the Mediterranean pond turtle, with the aim of determining the factors that drive the encroachment of the red-eared slider into the range of the Mediterranean pond turtle. This model was named the Trachemys-Encroachment model.

We used the favourability function (F) to identify the areas that are favourable to the species [23], regardless of the presence/absence ratio. Favourability was easily calculated from the probability obtained from the logistic regressions according to the expression

F = [P/(1 - P)]/[(n1/n0) + (P/[1 - P])], (1)

where P is the probability of a species being present, n1 is the number of observed presences in the geographical region, and n0 is the number of observed absences. We applied this favourability function to the three models described above.

In addition, we obtained a synecological favourability model for the red-eared slider by performing a logistic regression of its presence/absence using the values of F-Mauremys-Ecogeographical model as a predictor variable and applying the favourability function to obtain the F-Trachemys-Synecological model.

2.5. Model Assessment. The goodness-of-fit of the models was assessed using the Hosmer and Lemeshow test (the test statistic also follows a chi-square distribution; low P values would indicate lack of fit of the model [26]). The discrimination capacity of the models was evaluated with the area under the curve (AUC) of the receiving operating Characteristic (ROC). To evaluate the classification capacity of the models, we obtained a set of measures of classification based on the 0.5-favourability threshold (as favourability is independent of prevalence, 0.5 is the favourability value at which both sensitivity and specificity are equal): the correct classification rate (CCR), sensitivity, specificity, and Cohen's Kappa [27].

2.6. Interspecific Relationships. We used Pearson's chi-squared test to determine whether the distribution of the two species overlapped more than would be expected at random. Based on the values of the F-Mauremys-Ecogeographical model and the F-Trachemys-Encroachment model, we calculated favourability overlap between the two species [14] in the Mauremys range, indicating the degree to which the local favourability is similar for the two species. To obtain this, we calculated the fuzzy intersection between the fuzzy sets of areas favourable to each species (minimum favourability value for the two species at a given location), which can be used to identify the fuzzy set of areas simultaneously favourable for the two species; we also calculated the fuzzy union (maximum favourability value), which can identify the fuzzy set of areas favourable to either species [28]. We obtained the favourability overlap as the intersection divided by the union [29, 30]. This overlap ranges from 0, indicating no overlap, to 1, indicating the highest favourability overlap. We calculated the mean favourability values in each overlap interval for each species. Finally, we used these values to map the areas of conflict between the two species and the zones of more conservation concern regarding the Mediterranean pond turtle.

3. Results

We obtained a significant favourability model of the distribution of the Mediterranean pond turtle in Maalaga province using all the available data, the F-Mauremys-Ecogeographical model, whose logit function was

Y = +2.572 * HumI - 1.871 * Lo + 0.040 * SE - 13.495. (2)

The variables that explained the distribution of the Mediterranean pond turtle were humidity index (HumI), longitude (Lo), and southward exposure degree (SE).

The model's goodness-of-fit statistic showed no significant differences between the observed and expected values (Hosmer and Lemeshow: chi-square = 5.513, df = 8, and P = 0.702).

Regarding the F-Trachemys-Ecogeographical model, no variable was significant under a false discovery rate of q < 0.05. For this reason, we could not obtain a model that performed significantly better than a randomly generated model. The presences of the red-eared slider tended to overlap with the M. leprosa presences, and the distributions of the two species had an overlap value higher than that expected at random (Pearson [chi square] = 9.3962, P value = 0.002174). We found a significant favourability model when analysing the distribution of the red-eared slider over the Mediterranean pond turtle range, the F-Trachemys-Encroachment model (Hosmer and Lemeshow: chi-square = 8.961, df = 8, and P = 0.346), with the following logit function:

Y = -0.370 * HJan + 1.881 * LK + 29.943. (3)

Thus, the variables that explained the distribution of the red-eared slider were mean relative air humidity in January (HJan) and reservoirs and lakes (LK).

Favourability maps resulting from both models are shown in Figure 2. According to the models, the most favourable conditions for the Mediterranean pond turtle in Malaga were located in the southwest of the province, whereas the most favourable conditions for the red-eared slider were in central to eastern Malaga.

The logit functions obtained for the F-Trachemys-Synecology model presented the form

Y = +13.416 * F-Mauremys-Ecogeographical - 11.452 * [F-Mawremys-Ecogeographical.sup.2] - 4.431. (4)

Discrimination and classification assessment of the three models are shown in Table 2. Regarding sensitivity, specificity, and CCR, the F-Mauremys-Ecogeographical model and the F-Trachemys-Encroachment model correctly classified the presences in grids of high favourability and correctly classified the absences in grids with low favourability. The Mauremys model was the best model for classifying presences and absences (highest Kappa, sensitivity, specificity, and CCR) and discrimination capacity (AUC) was always higher than 0.8, that is, excellent, according to Hosmer and Lemeshow [26]. The F-Trachemys-Synecological model had the highest sensitivity values and the lowest kappa, specificity, CCR, and AUC values.

We obtained the intersection (or minimum according to fuzzy logic [30]) between the Mediterranean pond turtle and the red-eared slider models; the areas of greatest interaction are shown in Figure 2(c).

The relationship between the F-Trachemys-Synecology model and the F-Mauremys-Ecogeographical model is represented in Figure 3. Favourability values for the Mediterranean pond turtle and red-eared slider increased with a positive correlation until reaching a favourability value of 0.6 for both species. Subsequently, although the favourability values for the Mediterranean pond turtle continued to increase, the favourability values for the red-eared slider decreased. In addition, the relationship between the F-Mauremys-Ecogeographical model and F-Trachemys-Encroachment model according to the overlap values is represented in Figure 4. In grid cells where overlap values were low (less than 0.4), higher mean favourability values were obtained for the Mediterranean pond turtle than for the red-eared slider. In grid cells with range overlap values between 0.4 and 0.8, the mean favourability values for both species were intermediate and similar. Finally, when overlap values were more than 0.8, the favourability values were very high for both species. The types of relationships between overlap and favourability values for both species are shown in Figure 5, highlighting the areas of conservation concern for the Mediterranean pond turtle due to the encroachment of the red-eared slider.

4. Discussion

4.1. Distribution Models and the Importance of the Variables Entered for Both Species. The lack of a significant ecogeographical model for the red-eared slider suggests that, not being a native species, its distribution is not yet in equilibrium with the environment which is a common attribute for poor dispersers such as reptiles [31]. However, the significant F-Trachemys-Encroachment model obtained in our study area proved that the distribution of the red-eared slider is not random, but closely related to the distribution of the Mediterranean pond turtle. Encroachment models of this kind have not been widely used, but we consider that they may become useful tools in studies of native and invasive species with overlapping distributions.

The results of this study revealed differences in environmental requirements between the native and introduced species. Polo-Cavia et al. [12] had already detected physiological differences between both species, with heating and cooling rates being lower for the red-eared slider. In this study, the significant ecogeographical variables for the Mediterranean pond turtle were only associated with natural explanatory factors (climatic, topographical, and geographical). However, the significant variables for the red-eared slider encroachment model also suggested the influence of human activities, implying that the encroachment of the red-eared slider into the range of the Mediterranean pond turtle could be due to the release of pet turtles at water points [6,7,32,33], with natural factors being of secondary importance. Real et al. [33] also found, at national scale, that the distribution pattern of this invasive pond turtle was mediated by human activities. Given that, as Jeschke and Strayer [34] proposed, to be considered invasive, a species should undergo a human-induced introduction in a host area and later should establish and spread, the combination of human and natural factors in the F-Trachemys-Encroachment model suggest that the human factor points to introduction while the natural factor points to establishment and spread.

The red-eared slider was introduced via pet shops and its release is usually facilitated by their owners due to the nuisance caused when the turtles grow up [33]. Pet species would be difficult to manage if environmental education is lacking. Most people are unaware of the consequences of exotic species introductions and of their own crucial role in this issue. Environmental education is needed in this respect, as an important step in the management of the red-eared slider and other pet species intentionally released.

4.2. Interspecific Relationships and Areas Vulnerable to Invasion. Our results indicate that the areas of low favourability for the Mediterranean pond turtle were also areas of low favourability for the red-eared slider (Figure 3). These could be areas whose environmental characteristics make them inappropriate for the community of pond turtles in general. On the other hand, the areas of high favourability for the Mediterranean pond turtle were areas of low favourability for the red-eared slider. In areas where this occurs the presence of the native species could be hindering the settlement of the exotic species. Thus, the exotic species is not expected to generate a significant conservation risk for the native species in these areas. This confirms the finding of Martinez-Silvestre et al. [35], who analysed the interactions between both species, showing that the Mediterranean pond turtle is able to outcompete the red-eared slider and maintain its presence in the habitat, at least in some circumstances.

Richerson and Lum [36] proposed the favourableness hypothesis, suggesting that when environmental conditions approach the lethal limits for a species, organisms must increase the variety of adaptations and energy and material resources devoted to coping with these extreme conditions, to the detriment of those resources devoted to coadaptive accommodation with other species. Conversely, when all environmental conditions approach the optima for the species, organisms can devote a higher amount of energy, matter, and genome to coadaptive adjustments to other species. Consequently, several species may more easily coexist when the environmental conditions are favourable to all of them. Bearing the favourableness hypothesis in mind and making use of the favourability overlap proposed by Acevedo et al. [14], three situations result from the interaction between the two species (see Figure 4). The first situation shows areas where favourability was greater for the Mediterranean pond turtle than for the red-eared slider, corresponding to the areas oflow favourability overlap (represented in gray in Figure 5). We expect that the native species will not experience major problems due to the invasive species in these areas. In the remaining areas, the species compete on a similar basis regarding environmental favourability for the two species. However, two different scenarios arise from the favourability overlap shown in Figure 4 (interaction situations 2 and 3). Interaction situation 2 occurs in the areas of intermediate overlap (shown in black in Figure 5), where favourability for both species was intermediate and similar. In these areas, the red-eared slider may outcompete and displace the Mediterranean pond turtle, because the latter is not in optimal environmental conditions. These areas could correspond to areas where the exotic species is regularly released and where favourability for the Mediterranean pond turtle is not very high. Efforts to study, prevent, and mitigate the possible displacement of the native species should be concentrated in these areas, which composed 37.5% of the study area and 513% of the area occupied by the Mediterranean pond turtle. Interaction situation 3 affects areas of maximum overlap (just two grids shown in dark grey in Figure 5), where both species have very high favourability and it is more likely that they could coexist according to Richerson and Lum's favourableness hypothesis [36].

Some authors have discussed potential competition between both species [4,12,13], but the nature of the competition between them remains poorly understood. Although there is no direct evidence of the red-eared slider and the Mediterranean pond turtle competing for food or natural resources [4, 35], the diet of the red-eared slider strongly overlaps with that of the Mediterranean pond turtle [13]. Polo-Cavia et al. [11, 12, 37] have discussed other types of competition that affect ethological behaviour, showing that, whereas the Mediterranean pond turtle tended to avoid water containing semiochemicals released by the red-eared slider, the red-eared slider was not affected by substances released by the Mediterranean pond turtle. This effect could explain to some extent the displacement of the Mediterranean pond turtle by the red-eared slider. However, none of these competition mechanisms explains the capacity of the Mediterranean pond turtle to hinder, in its most favourable areas, encroachment by invasive pond turtles. The main cause of low favourability for the red-eared slider in most areas favourable to the Mediterranean pond turtle could be the lack of reservoirs. In any case, our methods led to the detection of areas of low concern related to the potential immediate encroachment by the invasive species.

4.3. Implications for the Conservation of the Mediterranean Pond Turtle. The approach used in this study can be used to identify areas highly favourable to the red-eared slider, whether or not this species is already present. These areas are of concern for the conservation of the Mediterranean pond turtle, except in the few locations where favourability for the latter species is also high. Thus, this study adds to our knowledge regarding the biogeographical relationship between these native and exotic species and may help to develop strategies to manage the alien species.

Although the nature of competition between these species remains unclear, the Mediterranean pond turtle is at high conservation risk from the red-eared slider. This work not only pinpoints areas where the current presence of the alien species could negatively impact the native species, but also areas where their presence could have a negative impact in the future. The red-eared slider could displace the Mediterranean pond turtle in these areas. Thus, management efforts should be focussed on the areas of highest concern.

The methodology developed in this study could be useful to researchers focused on assessing the potential impact of biological invasions on native species in a global context.

http://dx.doi.org/10.1155/2014/519710

Conflict of Interests

The authors declare that they have no conflict of interests regarding the publication of this paper.

Acknowledgments

David Romero was supported by a Grant from the Ministerio de Educacion: AP2007-03633. This study was supported by the project CGL2009-11316 (Ministerio de Ciencia e Innovacion, Spain, and FEDER). The authors thankJuan J. Castillo for his help with field sampling and his comments on a previous version of the paper. The authors would also like to thank Andrew Paterson for style corrections. The authors also thank S. Coxon for his help in the English revision of the paper.

References

[1] M. Lopez-Darias, J. M. Lobo, and P. Gouat, "Predicting potential distributions of invasive species: the exotic Barbary ground squirrel in the Canarian archipelago and the west Mediterranean region," Biological Invasions, vol. 10, no. 7, pp. 1027-1040, 2008.

[2] J. D. Parker, D. E. Burkepile, and M. E. Hayt, "Opposing effects of native and exotic herbivores on plant invasions," Science, vol. 311, no. 5766, pp. 1459-1461, 2006.

[3] J. C. Boez, J. Olivero, C. Peteiro, F. Ferri-Yonez, C. Garcia-Soto, and R. Real, "Macro-environmental modelling of the current distribution of Undaria pinnatifida (Laminariales, Ochrophyta) in northern Iberia," Biological Invasions, vol. 12, no. 7, pp. 2131-2139, 2010.

[4] J. M. Pleguezuelos, "Estatus y Conservacion de las tortugas marinas en Espana," in Atlas y Libro Rojo de los Anfibios y Reptiles de Espana, J. M. Pleguezuelos, R. Marquez, and M. Lizana, Eds., Direccioon General de Conservacioon de la Naturaleza, Madrid, Spain, 2002.

[5] P. Segurado and D. Figueiredo, "Coexistence of two freshwater turtle species along a Mediterranean stream: the role of spatial and temporal heterogeneity," Acta Oecologica, vol. 32, no. 2, pp. 134-144, 2007.

[6] D. Romero, P. Ferri, J. C. Boez, and R. Real, "Indicios de reproduccion de Trachemys scripta elegans en lagunas artificiales de Malaga," Boletin de la Asociacion Herpetologica Espanola, vol. 21, pp. 100-101, 2010.

[7] D. Romero, J. C. Boez, F. Ferri, J. Bellido, J. J. Castillo, and R. Real, "Nuevas citas de Galapago leproso (Mauremys leprosa (Schweiger, 1812)) y de Galapago Americano (Trachemys scripta) en la provincia de Moalaga," Boleton de la Asociacioon Herpetologica Espanola, vol. 22, pp. 104-107, 2011.

[8] E. Da Silva, "Mauremys leprosa (Schweiger, 1812)," in Atlas y Libro Rojo de los Anfibios y Reptiles de Espana, J. M. Pleguezuelos, R. Morquez, and M. Lizana, Eds., Direccion General de Conservacioon de la Naturaleza, Madrid, Spain, 2002.

[9] X. Capalleras and M. A. Carretero, "Evidencia de reproduccion con exito en libertad de Trachemys scripta en la Peninsula Iberica," Boletin de la Asociacion Herpetologica Espanola, vol. 11, pp. 34-35, 2000.

[10] J. M. G. de Berrazuela, T. Marron, M. Perianes, A. J. Gordillo, and J. del Moral, "Poblaciones asilvestradas en Cantabria de Trachemys scripta elegans y su potencial reproductor," Boletin de la Asociacion Herpetologica Espanola, vol. 18, pp. 34-37, 2007

[11] N. Polo-Cavia, P. Lopez, and J. Martin, "Interspecific differences in chemosensory responses of freshwater turtles: consequences for competition between native and invasive species," Biological Invasions, vol. 11, no. 2, pp. 431-440, 2009.

[12] N. Polo-Cavia, P. Lopez, and J. Martin, "Interspecific differences in heat exchange rates may affect competition between introduced and native freshwater turtles," Biological Invasions, vol. 11, no. 8, pp. 1755-1765, 2009.

[13] N. Perez-Santigosa, M. Florencio, J. Hidalgo-Vila, and C. Diaz-Paniagua, "Does the exotic invader turtle, Trachemys scripta elegans, compete for food with coexisting native turtles?" Amphibia Reptilia, vol. 32, no. 2, pp. 167-175, 2011.

[14] P. Acevedo, A. I. Ward, R. Real, and G. C. Smith, "Assessing biogeographical relationships of ecologically related species using favourability functions: a case study on British deer," Diversity and Distributions, vol. 16, no. 4, pp. 515-528, 2010.

[15] J. M. Pleguezuelos, R. Morquez, and M. Lizana, Eds., Atlas y Libro Rojo de los Anfibios y Reptiles de Espana, Direccion General de Conservacioon de la Naturaleza, Madrid, Spain, 2004.

[16] R. G. Pearson and T. P. Dawson, "Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful?" Global Ecology and Biogeography, vol. 12, no. 5, pp. 361-371, 2003.

[17] "Instituto Nacional de Estadistica," http://wwwine.es/.

[18] A. M. Barbosa, R. Real, J. Olivero, and J. M. Vargas, "Otter (Lutra lutra) distribution modeling at two resolution scales suited to conservation planning in the Iberian Peninsula," Biological Conservation, vol. 114, no. 3, pp. 377-387, 2003.

[19] A. Castro, A. Munoz, and R. Real, "Modelling the spatial distribution of the Tengmalm's owl Aegolius funereus in its southwestern Palaeartic limit (NE Spain)," Ardeola, vol. 55, no. 1, pp. 71-85, 2008.

[20] A. R. Munoz, R. Real, A. M. Barbosa, and J. M. Vargas, "Modelling the distribution of Bonelli's eagle in Spain: implications for conservation planning," Diversity and Distributions, vol. 11, no. 6, pp. 477-486, 2005.

[21] "Corine Land Cover CLC2000, Ministerio de Fomento, Gobierno de Espana," http://www.eea.europa.eu/publications/CO-R0-landcover.

[22] J. Bustamante, "Predictive models for lesser kestrel Falco naumanni distribution, abundance and extinction in southern Spain," Biological Conservation, vol. 80, no. 2, pp. 153-160, 1997

[23] R. Real, A. M. Barbosa, and J. M. Vargas, "Obtaining environmental favourability functions from logistic regression," Environmental and Ecological Statistics, vol. 13, no. 2, pp. 237-245, 2006.

[24] P. Acevedo and R. Real, "Favourability: concept, distinctive characteristics and potential usefulness," Naturwissenschaften, vol. 99, no. 7, pp. 515-522, 2012.

[25] Y. Benjamini and Y. Hochberg, "Controlling the false discovery rate: a practical and powerful approach to multiple testing," Journal of the Royal Statistical Society B, vol. 57, no. 1, pp. 289-300, 1995.

[26] D. W. Hosmer and S. Lemeshow, Applied Logistic Regression, John Wiley & Sons, New York, NY, USA, 2nd edition, 2000.

[27] A. H. Fielding and J. F. Bell, "A review of methods for the assessment of prediction errors in conservation presence/absence models," Environmental Conservation, vol. 24, no. 1, pp. 38-49, 1997.

[28] L. A. Zadeh, "Fuzzy sets," Information and Control, vol. 8, no. 3, pp. 338-353, 1965.

[29] D. Dubois and H. Prade, Fuzzy Sets and Systems: Theory and Applications, Academic Press, New York, NY, USA, 1980.

[30] L. I. Kuncheva, "Using measures of similarity and inclusion for multiple classifier fusion by decision templates," Fuzzy Sets and Systems, vol. 122, no. 3, pp. 401-407, 2001.

[31] M. B. Araujo and R. G. Pearson, "Equilibrium of species' distributions with climate," Ecography, vol. 28, no. 5, pp. 693-695, 2005.

[32] N. Perez-Santigosa, C. Diaz-Paniagua, and J. Hidalgo-Vila, "The reproductive ecology of exotic Trachemys scripta elegans in an invaded area of southern Europe," Aquatic Conservation: Marine and Freshwater Ecosystems, vol. 18, no. 7, pp. 1302-1310, 2008.

[33] R. Real, A. L. Morquez, A. Estrada, A. R. Munoz, and J. M. Vargas, "Modelling chorotypes of invasive vertebrates in mainland Spain," Diversity and Distributions, vol. 14, no. 2, pp. 364-373, 2008.

[34] J. M. Jeschke and D. L. Strayer, "Invasion success of vertebrates in Europe and North America," Proceedings of the National Academy of Sciences of the United States of America, vol. 102, no. 20, pp. 7198-7202, 2005.

[35] A. Martinez-Silvestre, C. Flecha, and J. S. Massana, "Observaciones de interacciones entre Trachemys scripta elegans y Mauremys leprosa en el pantano del Foix (Barcelona)," Boletin de la Asociacion Herpetologica Espanola, vol. 23, pp. 106-109, 2012.

[36] P. J. Richerson and K. Lum, "Patterns of plant species diversity in California: relation to weather and topography," The American Naturalist, vol. 116, no. 4, pp. 504-536, 1980.

[37] N. Polo-Cavia, P. Lopez, and J. Martin, "Interspecific differences in responses to predation risk may confer competitive advantages to invasive freshwater turtle species," Ethology, vol. 114, no. 2, pp. 115-123, 2008.

David Romero, (1) Jose C. Baez, (1,2) Francisco Ferri-Yanez, (3) Jesus J. Bellido, (1,4) and Raimundo Real (1)

(1) Departamento de Biologia Animal, Facultad de Ciencias, Universidad de Malaga, 29071 Malaga, Spain

(2) Instituto Espanol de Oceanografia, IEO, 29016 Malaga, Spain

(3) Departamento de Biogeografia y Cambio Global, Museo Nacional de Ciencias Naturales, CSIC & Laboratorio Internacional en Cambio Global CSIC-PUC (LINCGlobal), Calle Jose GutierrezAbascal 2, 28006 Madrid, Spain

(4) Aula del Mar de Malaga, 29003 Malaga, Spain

Correspondence should be addressed to David Romero; davidrp@uma.es

Received 29 August 2013; Accepted 24 October 2013; Published 21 January 2014

Academic Editors: F. Garcia-Gonzalez, A. Robins, and A. Surmacki

TABLE 1: Variables used to model the distribution of Trachemys scripta
and Mauremys leprosa.

Abbreviations                   Variable Abbreviations

Spatial
situation

La                            Latitude ([degrees]N) (1)
Lo                            Longitude ([degrees]E) (1)

Topography

A                               Mean altitude (m) (2)
[D.sub.A]         Difference altitude (m) (calculated from altitude)
S                    Slope ([degrees]) (calculated from altitude)
SE                          Southward exposure degree (3)
WE                           Westward exposure degree (3)

Climate

[H.sub.Jul]      Mean relative air humidity in July at 07:00 (%) (4)
[H.sub.Ran]       Annual relative air humidity range (%) (=[absolute
                               value of HuJan - HuJul])
[T.sub.Jan]          Mean temperature in January ([degrees]C) (4)
[T.sub.Jul]           Mean temperature in July ([degrees]C) (4)
Temp                   Mean annual temperature ([degrees]C) (4)
Inso                     Mean annual insolation (h/year) (4)
SRad             Mean annual solar radiation (kWh/[m.sup.2]/day) (4)
[T.sub.Ran]             Annual temperature range ([degrees]C)
                             (=[T.sub.Jul] - [T.sub.Jan])
MP24                    Maximum precipitation in 24 h (mm) (4)
Perm                            Soil permeability (5)
HumI                              Humidity index (6)
Prec                      Mean annual precipitation (mm) (4)
AET                   Mean annual actual evapotranspiration (mm)
                                  (=min [Prec, PET]
PET               Mean annual potential evapotranspiration (mm) (4)
DHail                    Mean annual number of hail days (6)
DSno                     Mean annual number of snow days (6)
Dfro                  Mean annual number of frost days (minimum
                        temperature [greater than or equal to]
                                   0[degrees]C) (6)
[H.sub.Jan]            Mean relative air humidity in January at
                                    07:00 (%) (4)
RMP                  Relative maximum precipitation (=MP24/Prec)
ContI                           Continental index (6)
PIrr                        Pluviometric irregularity (7)
ROff                         Mean annual runoff (mm) (5)
DPre                Mean annual number of days with precipitation
                         [greater than or equal to] 0.1 mm(4)
DStrS                 Mean annual number of storm in (6) summer
Win                Mean annual number of route of winter (km/h) (6)
DOvc                   Mean annual number of overcast days (6)
DClear                   Mean annual number of clear days (6)
DFog                      Mean annual number of fog days (6)
DFogW                Mean annual number of fog days in winter (6)
DFogS                Mean annual number of fog days in summer (6)
DStor                    Mean annual number of storm days (6)

Land cover

Natural
vegetation

PAST                                 Pasture (8)
Bush                                   Bush (8)
ScleV                       Sclerophyllous vegetation (8)
NMead                            Natural meadows (8)
DForeH                    Deciduous forest and hardwood (8)
MxtFor                             Mixed forest (8)
ConiFor                           Conifer forest (8)
Mountain                          Mountain areas (8)

Crops

IHerRice            Irrigated herbaceous crops and rice fields (8)
DHerVin                 Dry herbaceous crops and vineyard (8)
Olive                                 Olive (8)
Fruit                                 Fruit (8)
Agricul                         Agricultural areas (8)
AgricuLNa           Agricultural land with natural vegetation (8)

Water surfaces

LK                             Reservoirs and lakes (8)
RV                                    Rivers (8)
OW                                Other wetlands (8)

Other use

BeachD                          Beaches and dunes (8)
Urban                               Urban zone (8)
Mining                        Mining area and dumps (8)
LinearUr                   Urban linear infrastructures (8)

Other human
activities

[D.sub.hi]             Distance to the nearest highway (km) (1)
U100                Distance to the nearest urban centre with more
                          than 100 000 inhabitants (km) (1)
U500                    Distance to the nearest urban centre
HPd                        Human population density in 2000
                        (number of inhabitants/[km.sup.2]) (9)

(1) Spanish National Geography Institute (IGN), Road map, Iberian
Peninsula, Balearios and Canary Island, National Geographic
Institute, Ministry of Development, Madrid, Spain, 1999.

(2) United States Geological Survey GTOPO30, Land Processes
Distributed Active Archive Center, EROS Data Center,
http://lta.cr.usgs.gov/GTOPO30.

(3) T. G. Farr and M. Kobrick, "Shuttle Radar Topography Mission
produces a wealth of data," EOS Transaction of the American
Geophysical Union, vol. 81, pp. 583-585, 2000.

(4) I. Font, Climate Atlas of Spain, Madrid, National Meteorology
Institute, 1983.

(5) IGME, National Hydrogeological Map (2nd ed), Explanation of
useful rainfall maps, of hydrogeological survey and of synthesis
of aquifer systems, Madrid, Spanish Mining and Geology Institute,
Ministry of Industry and Energy, 1979.

(6) I. Font, Climatology of Spain and Portugal, Ediciones
Universidad de Salmanca, 2000.

(7) J. L. Montero de Burgos and J. L. Gonzalez-Rebollar,
Bioclimatic diagrams, Madrid, the National Institute for the
Conservation of Nature, Espana, 1974.

(8) Corine Land Cover CLC2000, Ministry of Development, Government
of Spain, http://www.eea.europa.eu/publications/COR0-landcover.

(9) Oak Ridge National Laboratory, LandScan 2000 Global Population
Database, Tennesse, USA, Oak Ridge National Laboratory, 2001.

TABLE 2: Comparative assessment of the models developed to
Mauremys leprosa and Trachemys scripta according to
classification, discrimination, and parsimony criteria.

                F-Mauremys-     F-Trachemys-   F-Trachemys-
              Ecogeographical   Encroachment    Synecology

Kappa              0.447           0.221          0.167
Sensitivity        0.700           0.667          0.720
Specificity        0.794           0.625          0.519
CCR                0.731           0.635          0.567
AUC                0.806           0.852          0.658
AIC               109.943          66.345        111.963

CCR: correct classification rate; AUC: area under the ROC
(receiving operating characteristic); AIC: Akaike information
criterion.
COPYRIGHT 2014 Hindawi Limited
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research Article
Author:Romero, David; Baez, Jose C.; Ferri-Yanez, Francisco; Bellido, Jesus J.; Real, Raimundo
Publication:The Scientific World Journal
Article Type:Report
Geographic Code:4EUSP
Date:Jan 1, 2014
Words:5713
Previous Article:Parallel simulation of HGMS of weakly magnetic nanoparticles in irrotational flow of inviscid fluid.
Next Article:Phytochemical analysis, antioxidant, antistress, and nootropic activities of aqueous and methanolic seed extracts of ladies finger (Abelmoschus...
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |