1. IntroductionThe natural distribution of black cherry (Prunus. serotina Ehrh.) extends from the USA to Guatemala [1] [2] . The species is subdivided into five botanical subspecies: P. serotina ssp. eximia (Small) Little, P. serotina ssp. hirsuta (Elliot) [1] , P. serotina ssp. capuli (Cav.) McVaugh, P. serotina ssp. serotina (Ehrh.) McVaugh, and P. serotina ssp. virens (Woot & Stand) McVaugh. P. serotina ssp. virens includes two botanical varieties: P. serotina ssp. virens var. virens McVaugh and P. serotina ssp. virens var. rufula McVaugh. These taxa were considered by McVaugh (1951) [1] as relatively separated by geography and having particular morphological characteristics. This classification is based on the hypothesis that the species have their origins in the northeast of the USA, from whence they migrated since the end of the Mesozoic era until the Eocene period due to the effect of glaciations.
According to Popenoe and Pachano (1922) [3] , the distribution area of P. serotina ranges from California to Florida in the USA, and in Mexico, throughout the western, eastern, and neovolcanic mountain chains: its distribution also includes the area of western Guatemala. The species has adapted to environments in Colombia, Ecuador, Bolivia, North India, and South Brazil [4] . Black cherry adapts well to temperate-cold climates (Cw0, Cw1, Cw2, and Cs) and subtropical climates type A and (A) C. Fresnedo et al. (2011) [5] and Avendaño-Gómez et al. (2015) [4] indicate that black cherries are distributed from Sonora to Chiapas, in Mexico, mostly in drylands at altitudes from 1000 to 3900 m in the Valley of Mexico, but not in proximity to the coast.
According to Startfinger (2010) [6] , Carmenen et al. (2016) [7] , and Aerst et al. (2017) [8] , Prunus serotina was introduced into Europe in the 17th century as an ornamental species. In the 19th century, the interest in black cherry introduction was focused on forestry, but the endeavor was unsuccessful. In the 20th century, the utility of the species for the forestation of disturbed or degraded areas was employed to improve soils due to the low C/N ratio of its leaves [9] . Thus, black cherry was introduced into Belgium, France, Poland, Germany, Denmark, Norway, Estonia, Lithuania, and Russia [7] [9] [10] [11] [12] . Currently, it has an important presence in Belgium, Poland, Germany, and Denmark, where it has been reported to be an invasive species out of control [9] [11] [12] . For this reason, the recent modeling of the potential distribution of black cherry in Europe has received special attention by geneticists such as Pairon et al. (2006) [13] , Pairon et al. (2010) [14] , and Guzmán et al. (2018) [15] ; from an ecological point of view, such modeling has been addressed by Carmen et al. (2016) [7] . Halarewicz et al. (2017) [16] and Aerst et al. (2017) [8] sought to predict its invasive ability and analyzed its biological control. The present study aimed to describe, on the one hand, the taxonomic richness and taxonomic diversity of P. serotina, and on the other hand, it intended to identify the potential distribution of this species based on data from verified collection sites in its natural distribution area; the data were analyzed using a prospective climate change model for the American continent, where the species is native, and for Europe, where it is invasive.
3. Results and DiscussionThe pattern of P. serotina collection sites in North America is linked to the mountain chains in the region (Figure 1). P. serotina ssp. serotina is the most widely distributed, and its presence is mainly associated with humid regions in the Sierra Madre Oriental, west-central and central Mexico, as well as the central region of the USA, from Missouri to the Appalachian Mountains, as reported by McVaugh (1951) [1] , Fresnedo (2011) [5] , and Beck (2014) [22] . In Mexico, P. serotina ssp. capuli is distributed in both the north and the south, particularly in the western central regions, where is it is often sympatric with P. serotina ssp. serotina and less frequently allopatric with P. serotina ssp. virens var. virens [4] [5] . By contrast, P. serotina ssp. eximia is endemic to central Texas, and very particularly in the Edwards Plateau, as reported by McVaugh (1951) [1] . P. serotina ssp. hirsuta is mainly distributed in the states of Georgia, Alabama, and Florida. P. serotina ssp. virens var. rufula is distributed from southeast Texas to Arizona, the adjacent region of New Mexico, and the southern part of Rocky Mountains, as noted by McVaugh (1951) [1] . P. serotina ssp. virens var. virens is distributed in Jalisco, Guanajuato, and especially in the Northwest, in the states of Durango, Sonora, and Chihuahua, as noted by Rzendowski and Calderón de Rzendowski (2005) [2] .
3.1. Taxonomic Richness and Taxonomic DiversityThere are three areas In North America where considerable taxonomic richness of P. serotina Ehrh. Figure 2(a) can be observed; the first area is located near the Mexico-US border, including the Southern Rocky Mountains of New Mexico, Arizona, and the north of Sonora. P. serotina ssp. serotina, P. serotina ssp. virens var. rufula, P. serotina ssp. virens var. virens, and sometimes P. serotina ssp. capuli are present in this area. The second area presenting considerable taxonomic richness is located in the Sierra Madre Occidental in Durango, Aguascalientes, and Jalisco. Subspecies P. serotina ssp. capuli, P. serotina ssp. virens var. virens and P. serotina ssp. virens var. rufula are present in this second area. The third and most important area of taxonomic richness is located between Nuevo Leon and Tamaulipas, in Mexico. This area comprises the northern end of the Sierra Madre Oriental (known as Sierra del Burro), where subspecies P. serotina ssp. capuli, P. serotina ssp. serotina, P. serotina ssp. virens var. rufula, and P. serotina ssp. virens var. virens are located. Interestingly, the sympatric distribution of wild subspecies P. serotina ssp. serotina and the supposedly domesticated subspecies P. serotina ssp. capuli raises questions about their being truly different subspecies, as stated by McVaugh (1951) [1] and Rzendowski and Calderón de Rzendowski (2005) [2] .
The taxonomic diversity of P. serotina is presented in Figure 2(b) as a map. This pattern is also similar to the taxonomic richness pattern, with four clearly defined areas. The area of highest taxonomic diversity is located in northern
Mexico, in the so-called Sierra del Burro (Sierra Madre Oriental, Nuevo Leon). A second area is located in the Southern Rocky Mountains in the USA, in Arizona and New Mexico, as well as in part of the Sierra Madre Occidental, in Sonora and Chihuahua (Mexico). A third area, of lower taxonomic diversity, is located in the south of the USA, in the states of Texas and Missouri. A fourth area is associated with the neovolcanic mountain chain in Mexico crossing the states of Michoacán, Queretaro, and Guanajuato.
The results of taxonomic richness and taxonomic diversity presented in the maps in Figure 1, showing important areas located in Mexico (western and eastern mountain chains), do not reflect McVaugh’s (1952) [23] description of the differences among taxa as a process occurring at the end of the Mesozoic Era and during the Eocene period, when glaciations affected flora patterns in the Nearctic. This author describes the subspeciation path from the northeast part of the USA to southern areas in the USA and Mexico. Although Rohrer (2014) [24] has a new proposal of botanical classification describing four botanical varieties: P. serotina var. alabamensis = P. serotina ssp. hirsuta, P. serotina var. capuli = P. serotina ssp. capuli, P. serotina var. rufula = P. serotina ssp. virens, and P. serotina var. serotina = P. serotina ssp. serotina + P. serotina ssp. eximia, this author does not describe differences among P. serotina ssp. subspecies populations in its extensive area of distribution, and its possible relationship with P. serotina ssp. capuli in central Mexico is not mentioned [1] . We are not in a position to suggest a paleoethnobotanical pattern for the history for P. serotine, but the high taxonomical richness and taxonomical diversity in both Mexican mountain ranges suggest a review of McVaugh’s (1952) [23] descriptions. Interestingly, Guzmán et al. (2018) [15] studied 18 populations of the four subspecies and found that ssp. virens had the highest gene diversity. The comparison of genetic diversity across the four subspecies, P. serotina ssp. capuli, P. serotina ssp. virens, P. serotina ssp. serotina, and P. serotina ssp. eximia, showed that the genetic differentiation (Gst) was even lower (16%) than for the 18 populations, but the genetic differentiation between eximia and the other three subspecies (15%) was higher than the differences among these three (6%).
3.2. Current Potential Distribution in America and EuropeCurrent potential distribution of P. serotina is presented in Figure 3 as a map of North America and Europe. Two important potential areas can be distinguished on the map of North America: the first area begins in the center of Mexico in the neovolcanic mountain chain, and extends to the western mountains of Michoacán, Guanajuato, Jalisco, Nayarit, Sinaloa, Durango, and Sonora; the area continues in southern Arizona and New Mexico. Subspecies P. serotina ssp. capuli, P. serotina ssp. serotina, and P. serotina ssp. virens are predominant in this first area. An interesting point is that the dry and hot area between both mountain ranges in northern Mexico seems to be a natural barrier for black cherry; similarly, an area in the USA where the species is absent is a wetland territory. A major second area of potential distribution for the species is centered in Kansas; the area connects with a zone of medium potential toward Texas and also with a high potential area toward Missouri and the east of the USA. Subspecies P. serotina ssp. virens var. virens, P. serotina ssp. virens var. rufula, P. serotina ssp. serotina, P. serotina ssp. hirsuta, and the subspecies P. serotina ssp. eximia inhabit
this area. It is interesting to observe that, as the model suggested, P. serotina ssp. eximia is present in the Edwards Plateau in Texas. Comanches had settlements in this area, and Hamel et al. (1973) [25] describes how black cherry was used for therapeutic and forestry purposes by the tribe.
The potential distribution in Europe predicted by the model is shown in Figure 3. P. serotina is relatively widespread in the continent. Its presence is limited by temperature in majority of countries and, particularly in France, precipitation is also a limiting factor. Aerts et al. (2017) [8] indicate that the species tolerates a wide range of wetland and dryland conditions in Europe. Apparently, it has a special relationship with sandy and acid soils. Remarkably, even though the species is subdivided into five subspecies and two botanical varieties distributed from the USA to Guatemala, European studies have only taken in account the natural conditions of subspecies P. serotina ssp. serotina, from the central and northeast areas of the USA, to estimate potential invasion in Europe ( [7] [8] [12] [13] [26] but Guzmán et al. (2018) [27] , studying morphological variability indicated that P. serotina ssp. eximia, P. serotina ssp. hirsuta and P. serotina ssp. serotina are distributed in more humid and cold environments, while P. serotina ssp. virens prefers drier and warmer environments. Subspecies capuli exhibited the greatest environmental heterogeneity. Some individuals of P. serotina in North America are calculated to be 250 years old, whereas specimens in Europe live only 30 years on average. Apparently, as indicated by Deckers et al. (2005) [9] , black cherry dispersion is favored when the forest is disturbed in this continent; moreover Carmenen et al. (2016) [7] proposed an area of potential distribution or niche of invasion by P. serotina in Europe larger than in our model of Figure 3(c). These authors suggest that the species is capable of invading virtually all Western Europe. We suppose that the difference in the projections is associated with the sources of information on the presence of the species used by Carmenen et al. (2016) [7] , who based their prediction model on the little (1971) [28] database, where P. serotina is frequently considered as a synonym of P. virginiana and P. alabamensis. This fact led the author to include areas where P. virginiana is present and P. serotina is not (north of the USA). In the Europe map, our model matches descriptions by Closset-Kopp et al. (2007) [29] , Startfinger (2010) [6] , and Halarewicz et al. (2017) [16] , but whether individuals of this species present in Europe are issued from a single taxonomic entity, or from different taxonomic entities introduced at different times remains unsettled, as noted by Pairon et al. (2010) [14] . In our opinion, it is hard to believe that all individuals of P. serotina currently present in Europe belong to a single taxonomic entity.
Several authors have studied the invasive ability of black cherry in European forests. Deckers et al. (2005) [9] confirmed the effects of landscape structure on the occurrence of P. serotina in Belgium and suggested dispersal by bird action. Additionally, a decrease in aggregation throughout the plant’s life cycle (seedlings and density of young trees) has led to conclusions associated with an allelopathic effect. Godefroid et al. (2005) [11] studied the ecological factors affecting the abundance of P. serotina in the forests of Belgium and observed a negative correlation between species richness and the abundance of P. serotina in forests. According to Closset-Kopp et al. (2007) [29] , high intensities of light are required from the growth stage of juvenile individuals to maturity and seed production, but Vanhellemont et al. (2010) [30] noted that most of the European studies on P. serotina were conducted in areas where invasion was intentional, concluding that black cherry could not be considered an invasive aggressor; instead, the authors suggested that perturbations in the forest canopy could accelerate the spread of the species and turn it invasive.
3.3. Potential Distribution of the Species under the NOAA-CCM3 ModelFigure 3(b) shows the Prunus serotina Ehrh. potential distribution map for North America under a climatic change scenario and Figure 3(d) shows an equivalent projection for Europe. This model was based on the NOAA-CCM3 climatic change model, which includes an estimated variation in climate parameters resulting from a gradual increase in atmospheric CO2 concentration until 2100 (650 ppm, and temperatures between 1.1˚C - 2.6˚C) [20] . The extension of the potential invasion area of the species is greater than the current area based on information from 1860 to 2005. This can be appreciated as an important extension of the western-central region in the Mexican states of Guanajuato, Jalisco, Michoacan, Queretaro, and the State of Mexico, as well as the northern Mexico region in the states of Durango, Sonora, and Chihuahua. This is a clear sign of the enlargement of arid areas. As noted Guzmán, that increases in temperature will imply altitudinal and latitudinal displacements from currently suitable areas and particularly P. serotina ssp. virens has the highest potential for expanding its area of distribution because this subspecies tolerates drought conditions better than the rest. Predicting the possible expansion of P. serotina ssp. capuli’s potential distribution due to climate change is difficult, since as Rzendowski and Calderon (2005) [2] and Avendaño et al. (2015) [4] have noted, this subspecies is still undergoing a domestication process.
The suitable area for black cherries in the USA could also become larger, and this could occur in two directions: the first area of expansion comprises the southern parts of the Rocky Mountains in New Mexico, Arizona, and Colorado, with small portions in Wyoming and Montana, and the second favorable area extends to the southern Appalachian Mountains in Georgia and Tennessee. We assume that P. serotina ssp. serotina, P. serotina ssp. hirsute, and P. serotina ssp. eximia will likely colonize these areas. Further to the north, the distribution of P. serotina ssp. serotina, present in the USA, mainly in the state of Missouri, is projected to enlarge toward the States of Indiana and Ohio on the east, and southwest toward Kansas and Oklahoma, despite its limited genetic variation, as stated by Beck et al. (2014) [22] . According to these authors, this subspecies is increasingly being used in forestry.
The projected growth of the species due to climate change in Europe also increased, mainly in France, Germany, and Italy, in areas surrounded by current invasions. In our study, we were also interested in pointing out climatic factors affecting the current potential distribution and those that could affect potential distribution scenarios in the future. Thus, we can see that, currently, mean monthly temperature is the limiting factor of the current invasion of P. serotina in southern England (Figure 4(c)), but according to the model, the distribution of this species will be greater in the future, limited only by precipitation seasonality (Figure 4(d)). Reinhardt et al. (2003) [31] report that the species is frequently deemed a “forest plague” in Poland, Germany, Denmark, and the Netherlands. Aerst et al. (2017) [8] noted that P. serotina is changing nitrogen, phosphorus, and carbon cycles to its own advantage, altering the photosynthetic capacity of the indigenous species; moreover, an uncontrolled invasion of European temperate forests by P. serotina would affect the long-term climate change mitigation potential of invaded forests. Evidently, the possible invasion of P. serotina will have an economical cost if measures to control it are not put in place [16] . The economic losses in Germany are calculated to be over 25 million EUR; in the Netherlands, the control of this species has cost between 150 and 1500 EUR per hectare each year, according to Startfinger (2010) [6] .
Diverse eradication strategies have been proposed to control this invasive species [29] . However, they have not become popular because of their high costs, long time for execution, and uncertain success. Finally, it seems that the best method to control it is to assume that P. serotina will continue being part of European forests, and that its impact and dominance can be reduced by controlling disturbances in the forest canopy [6] [16] [30] .
4. ConclusionPrunus serotina Ehrh. is a native species of North America that has become invasive in Europe. The present study used data from 554 confirmed collection sites. The taxonomical richness and taxonomic diversity of the plant are described in the first part of this article, and the current potential distribution and future potential distribution of the species have been estimated using 19 climatic parameters. Climate variability data revealed the regions where subspecies of P. serotina are mainly distributed are Mexico’s northwest, the northwestern part of the Central Mexican Plateau, and regions of the Mississippi River and the Great Lakes. Taxonomic richness and taxonomic diversity show similar patterns, both showing four defined areas. The first area is located in northern Mexico, in the Sierra Madre Oriental (Sierra del Burro) in Nuevo Leon; the second area is located in the southern Rocky Mountains in the USA, in Arizona, New Mexico, and part of the Sierra Madre Occidental in Sonora and Chihuahua (Mexico). A third area presenting lower taxonomic richness and taxonomic diversity is located in the south of the USA, in the States of Texas and Missouri. A fourth area is related to Neovolcanic Mountains in Mexico, in the states of Michoacán, Queretaro, and Guanajuato. The current potential distribution of P. serotina in North America shows a continuous pattern starting in the Center of Mexico and following both main Mexican mountain ranges extending to the North and tilting toward the center of the USA and the East of the country. Regions of northeast Mexico, northwestern Mexico, the Great American Basin, and the Mississippi River-Great Lakes region in the USA are shown as areas where the taxa of P. serotina are present. According to the maps obtained using our NOAA-CCM3 model, patterns will gradually expand as an effect of climate change. If the current potential species distribution in Europe includes practically all the western part of the continent, the potential effect of climate change suggests that the areas of distribution of the species will expand, especially in France, Germany, and Italy. As species, P. serotina seems to be adapted to future possible climatic conditions, the studies of the potential invasion of this species in Europe should take into account the different taxa of the species throughout America in order to achieve more taxonomically accurate conclusions about the behavior and evolution of its invasion in the Old Continent and the potential impacts of climate change.