The changing biodiversity of Alabama Drosophila: important impacts of seasonal variation, urbanization, and invasive species

Abstract Global warming and anthropogenic disturbances significantly influence the biosphere, tremendously increasing species extinction rates. In Central Alabama, we analyzed Drosophilidae species composition change nearly 100 years after the previous survey. We found ten Drosophilid species that were not reported during the last major biodiversity studies, two of which are invasive pests. In addition, we analyzed the influence of seasonal environmental variables characteristic of the subtropical climate zone on Drosophila abundance and biodiversity. We found a significant correlation between temperature and abundance of total Drosophila as well as for six of the seven most represented species individually, with a maximum abundance at intermediate temperatures (18–26°C). In addition, temperature was positively correlated with biodiversity of Drosophila. Precipitation produced a significant effect on the abundance of five species of Drosophila, with different optima for each species, but did not affect overall biodiversity. We concluded that in the subtropical climate zone of Central Alabama, seasonal temperature and precipitation changes produce a significant effect on Drosophila abundance and biodiversity, while local land use also impacts fly abundance, contributing to an apparent shift in species composition over the last century. We expect global climate change and other anthropogenic factors to further impact Drosophila species composition in the subtropical climate zone into the future.


| INTRODUCTION
In the face of global warming, we can see the influence of changes in the world's climate on the biosphere. Currently, the global species extinction rate is 100-10,000 times higher than a predicted natural rate, which suggests that a new mass extinction is underway (Ceballos et al., 2015;Pimm, Russell, Gittleman, & Brooks, 1995). A decrease in global biodiversity may cause serious consequences for the biosphere by influencing ecosystem function and reducing overall productivity of biological communities (Cardinale et al., 2006;Gamfeldt, Hillebrand, & Jonsson, 2008;Hector et al., 1999;Tilman, Wedin, & Knops, 1996).
Decreases in ecosystem productivity can lead to serious implications for fishery, forestry, and other types of industries that depend on harvesting natural resources. In addition, many wild species of plants and animals potentially harbor undiscovered beneficial natural products; thus, their extinction could have negative effect on native human populations (Cooper, 2004;Newmark, 2002). Native species can also prevent or slow invasive pests' habitat range expansion, via interspecific competition (Commar, da Conceição Galego, Ceron, & Aparecida Carareto, 2012;Hooper et al., 2008). Consequently, biodiversity surveys may reveal early colonization of pest species and native species extinction, which makes them a useful tool for natural resource management and prevention of natural habitat degradation.
Drosophila has the potential to become a standard model organism for investigating the influence of climate and habitat change on the biological community, as most species have rather limited realized niches and are dependent on a particular group of host organisms for feeding and reproduction (Parsons, 1991;Poppe, Valente, & Schmitz, 2013). In addition, Drosophila are sensitive to shifts in essential climate variables (Parsons, 1991;Poppe, Valente, & Schmitz, 2013;Srinath & Shivanna, 2014). The influence of seasonal changes on Drosophila abundance and biodiversity has been surveyed mostly in tropical and temperate climatic zones (Dobzhansky & Pavan, 1950;Levitan, 1954;da Mata et al., 2015;Patterson, 1943;Shorrocks, 1975). The abundance of Drosophila species is influenced primarily by precipitation in tropical regions (Dobzhansky & Pavan, 1950;Guruprasad et al., 2010;Srinath & Shivanna, 2014;Torres & Madi-Ravazzi, 2006), while in temperate climate regions, seasonal changes in abundance and biodiversity of Drosophila are better explained by temperature variation (da Mata et al., 2015;Patterson, 1943;Poppe, Valente, & Schmitz, 2013).
In addition, Drosophila has the potential to be a model organism for studying the influence of anthropogenic disturbance, particularly urbanization, on species biodiversity and abundance. Endemic Drosophila species are often very sensitive to natural habitat deterioration, can form very specific host associations, and are not able to adapt to human environmental modifications (Avondet, Blair, Berg, & Ebbert, 2003;Dobzhansky & Pavan, 1950;Ferreira & Tidon, 2005;van Klinken & Walter, 2001;Parsons, 1991;Shorrocks, 1975). As a result, a decrease in abundance of endemic species of Drosophila could indicate that a habitat is recently or currently being disturbed by human influence. The opposite is true for cosmopolitan species of Drosophila that are often abundant in urban environments (Avondet, Blair, Berg, & Ebbert, 2003). During a biodiversity survey in an urban area in Brazil, it was shown that exotic species of Drosophila contributed to over 90% of the total abundance in this type of environment (Ferreira & Tidon, 2005).
Surveys of wild populations of cosmopolitan Drosophila are especially important as there are two known pest Drosophilids: Drosophila suzukii and Zaprionus indianus. Both of these pest species are polyphagous and are very efficient in colonizing new types of environments (Burrack, Smith, Pfeiffer, Koeher, & Laforest, 2012;Commar et al., 2012). Drosophila suzukii originated in Asia and has recently quickly expanded its range, being first reported in the continental United States in 2008 and in Europe in 2009 (Berry, Anthony, Newfield, Ornsby, & Armstrong, 2012). In new habitats, D. suzukii broadened its host range and became a significant pest of berries and soft flesh fruits such as plums, peaches, and nectarines (Berry et al., 2012;Burrack et al., 2012). Zaprionus indianus is believed to have originated in Africa where it was not considered a significant pest (Commar et al., 2012;Joshi, Biddinger, Demchak, & Deppen, 2014). However, in 1999 was reported in South America, where it became a serious pest of figs (Commar et al., 2012;Vilela, 1999). In 2005, Z. indianus was reported in North America, Florida, where it continued to expand its habitat and host range (Commar et al., 2012;van der Linde et al., 2006).
To the best of our knowledge, there is very little accessible information on the influence of seasonal climate variables on abundance of Drosophila in subtropical regions, specifically Alabama. The average annual temperature for Central Alabama is 18.3°C and varies monthly from 1.6 to 33.7°C, with average annual precipitations of 1395 mm, ranging from 90 to 142 mm per month (NOAA 2014(NOAA -2015. The last major Drosophila biodiversity survey reported the presence of 26 Drosophilidae species in Alabama (Sturtevant, 1916(Sturtevant, , 1918(Sturtevant, , 1921.
However, the influence of climatic variables on abundance and biodiversity was not evaluated. Due to the relatively high seasonal variation in temperature and precipitation in Alabama, we hypothesized that both of these climatic variables will significantly influence Drosophila abundance and biodiversity. In addition, we hypothesized that abundance of cosmopolitan Drosophila species would be higher in urban settings than in industrial or minimally disturbed rural environments.

| Sample collection
We collected samples from 23 sites in and around Tuscaloosa, AL ( Fig. 1 and Table S1). Collection sites were chosen based on land-use type. We sampled seven sites that are used for industrial production or storage of industrial products, eight urban parks, three nonurban parks (a biological station, an arboretum, and a state park), and five sites that did not fall into any category (a highway rest station, an apartment complex, an archeological park, a roadside, and a farm, Table S1). Latitude and longitude of each collection site were recorded with a Garmin GPS navigator (Table S1). Samples were collected from banana and mushroom traps left overnight and collected in a time range from 7:00 a.m. to 11:00 a.m. the following morning (18-24 hr collection period total) using an alcohol aspirator (Markow & O'Grady, 2005). Samples were stored in 70% ethanol at −20°C. We made 16 collection trips from July 2014 to May 2015. Collection trips were performed once per month except for the time period from August 2014 to November 2014, during which collections were performed twice per month. For the period from June 2014 to December 2014, we sampled from five to seven randomly selected sites from our 23 collection sites, then starting in January 2015, we chose six collection sites to focus on and visited them monthly.

| Sample identification
Samples were identified based on species identification keys (Markow & O'Grady, 2005;Strickberger, 1962). DNA from representative samples of each species was extracted via DNA extraction Chelex 100 protocol (Walsh, Metzger, & Higuchi, 1991), modified by usage of one Drosophila sample per DNA extraction. Extracted DNA was PCRamplified for the cytochrome oxidase 1 (CO1) gene (Reed, Nyboer, & Markow, 2007) and then sent for sequencing. Morphological species identifications were confirmed with CO1 sequence similarity to Genbank reference sequences via BLAST.

| Seasonal climatic data
Climate data were obtained from the Tuscaloosa weather station of the National Oceanic and Atmospheric Administration (NOAA).
Average temperature was calculated via averaging daily high and low temperatures. Monthly averages were calculated via averaging daily temperature and precipitation for the 2-day collection period and the preceding 28 days. In addition, we calculated the average climate variables for each month that preceded each collection month (56-28 days preceding).
F I G U R E 1 Collection sites around Tuscaloosa. Nonurban parks are marked in blue, urban parks are marked in red, sites of industrial production and industrial product storage are marked in black, and sites that do not match any of the described categories are marked in purple
Influence of climatic variables on the abundance of all Drosophila, Mycodrosophila, and Scaptomyza samples, as well as on each species individually was performed with negative binomial GLM of the following form: In our work, we refer to abundance as a number of individual specimens that were collected during a single collection trip at a given site. The influence of each climatic variable on species abundance was tested with linear and quadratic models. The final singleterm model was chosen based on the lowest Akaike information criterion (AIC) value. In addition, with linear and quadratic negative binomial GLM models, we tested the influence of the precipitation average in month previous to the collection month on the abundance of mycophagous Drosophila species: D. tripunctata and D. putrida. We also tested linear and quadratic multivariate models that combined effects of temperature and precipitation. The best multivariate model was chosen based on AIC value and single-term deletion test. The multivariate model was considered superior if it explained more deviance and has a lower AIC value than models with only one variable. Biodiversity was evaluated as species richness and the Shannon and Simpson biodiversity indices. To analyze the influence of climate variables on biodiversity, we used linear regression models of the following form: To evaluate the influence of land use on abundance and biodiversity of Drosophila, we separated 17 of our sites into three land-use categories: nonurban parks, urban parks, and places of industrial production. To evaluate sufficiency of the number of sampling sites for each category, rarefaction analyses were performed (Kindt & Coe, 2005). In addition, we developed a model that evaluated all of our sites for the presence or absence of six variables: garbage or trash cans, industrial or agricultural production, asphalt road, highway, railroad or airport, and residential or public catering buildings. The site's degree of anthropogenic disturbance (disturbance score) was the sum of factors present for each collection area and was then tested in a negative binomial GLM against Drosophila abundance per trap. Linear regression was used to test for correlations between biodiversity (as measured by species richness, Shannon and Simpson indices) and the disturbance score. In addition, using the negative binomial model, we tested the influence of latitude and longitude on the abundance and biodiversity of Drosophila.
Other Drosophilidae species that were identified but not included in statistical analyses were Scaptodrosophila latifasciaeformis, Chymomyza amoena, and Leucophenga angusta (

| Climatic variable influence on abundance
We analyzed the influence of climate variables on total abundance of Drosophila and closely related species, as well as on the abundance of the most represented species individually. It should be noted that in this manuscript, observed abundances of flies are the result of the combined effects of their actual numbers present in the environment, but also of their activity level and thus "catchability." When conditions are less favorable over the short term (e.g., too cold), many species of flies become much more difficult to trap.

| Overall fly abundance
Analyses of the influence of the climatic variables during the collection period, with a second-order negative binomial GLM (sonbGLM), showed that temperature significantly influenced total number of collected samples (p = .001) with explained deviance (ED) of 43.9% (Fig. 4A) and a maximum abundance at a monthly average temperature of 21°C. We did not find any significant correlation between amount of precipitation during collection periods and total abundance of Drosophila. Combination of variables in a multivariate model did not improve the model. Taking into account monthly averages of the climatic variables, we found a significant correlation between the total number of flies and monthly average temperature (p = .001), using sonbGLM model that explained 44.5% of the deviance (Fig. 4C).
Monthly average precipitation level significantly influenced total abundance of Drosophila (p = .032) and explained 21.1% of the deviance (Fig. 4D). In this case, the correlation is better explained by negative binomial GLM (nbGLM) then by sonbGLM. The combination of climatic variables in one multivariate model did not result in a better model.

| Monthly averages of climate variables
Considering monthly climate averages, we found that sonbGLM was the best model for explaining the bell-shape correlation between temperature and abundance of D. affinis (p = .048, ED = 26%, Fig. 5C), for D. simulans (p = 1.08e-06, ED = 66.5%), D. robusta (p = 2.91e-06, ED = 75%), and D. melanogaster (p = 3.45e-06, ED = 69.9%). In addition, we did not find any significant correlation between precipitation averages during the month preceding the month of collections, with any of the models described above.

| Influence of seasonal climatic variables on biodiversity
We found a variation in the number of species collected during different seasons. Species richness was highest during the fall (September-November) with 14 species, and lowest during the winter (December-February) with only five species. During spring (March-May) and summer (June-August), we found representatives of 13 and 11 Drosophila species, respectively (Table 1). There was a significant positive correlation between monthly average temperature and biodiversity indices including Shannon index (p = .001, Fig. 6A) and Simpson index (p = .0026, Fig. 6B). Monthly temperature also significantly influenced species richness, with a positive correlation (p = .0017, Fig. 6C).
Average temperature for each collection period showed a similar pattern of a positive correlation with the Shannon (p = .01, Fig. 6D) and Simpson (p = .03) biodiversity indices (Fig. 6E), as well as species richness (p = .007, Fig. 6F), in positive correlations. We did not find any significant correlation between amount of precipitation and Drosophila biodiversity.

| Influence of land use on abundance and biodiversity of Drosophila
Separating our sites into three categories, industrial areas, urban parks, and nonurban parks, we found a significant correlation between land use and total abundance of Drosophila species (p = .037, Fig. 7).
The analyses of most represented species individually showed a correlation between abundance of D. tripunctata (p = .008) and landuse categories. In addition, land-use type may influence abundance of D. putrida (p = .079). The lowest abundance of both of these mushroom-feeding Drosophila was recorded from industrial areas.
Disturbance score, which evaluated the presence or absence of disturbance factors such as garbage or trash cans, industrial or agricultural productions, asphalt roads, highway, railroads or airports, and residential or public catering buildings across all collection sites, only correlated with the abundance of D. putrida (p = .035), where greater disturbance produced lower abundances of the fly species. In contrast to abundance, we did not find any significant correlation between land use and biodiversity indices or species richness. In addition, we did not find any evidence to conclude that narrow range of latitude or longitude over the collection sites produced a significant effect on abundance or biodiversity of Drosophila.

| DISCUSSION
The previous Drosophila biodiversity studies in Alabama reported 26 Drosophilidae species (Sturtevant, 1916(Sturtevant, , 1918Sturtevant, 1921) compared to the 18 found in this study, a difference in species richness of only five species. This difference could be due to the sampling techniques or/and sampling area choice. In previous biodiversity surveys, Sturtevant (1916Sturtevant ( , 1918, sampling was mostly performed via catching samples in a banana bait and net sweeping of fruits and mushrooms. Sturtevant and Dobzhansky (1936) noticed that at least for D. affinis subgroup species, they did not observe a significant  In addition, Dobzhansky and Pavan (1950) (Wheeler, 1960). Of ten previously unreported species, H. duncani, S. frustfrustulifera, and D. cardini were captured only once, which makes it difficult to ascertain their permanent presence in Alabama.
Overall, it is reasonable to conclude that there has been a substantial change in Drosophilidae species diversity in Alabama over the last century.

| Invasive species
Among the newly introduced Drosophilas, the most interesting was the presence of D. suzukii and Z. indianus in our collected samples.
Both of these flies are recognized as invasive pests, and both were consistently present in our traps. We were able to find D. suzukii dur-  et al., 2014). It also was reported in Mexico (Lasa & Tadeo, 2015) and Canada (Renkema, Miller, Fraser, Légaré, & Hallett, 2013). Within the United States, Z. indianus was found on peach, raspberry, and blueberry farms (Biddinger, Joshi, & Demchak, 2012;Joshi et al., 2014). In Alabama, peach and blueberry production is not very substantial and valued at <4 million US dollars in 2014 (USDA). In addition, Z. indianus usually infests only damaged fruits and is considered a secondary pest (Joshi et al., 2014). Due to these reasons, Z. indianus would not be expected to become a major pest in Alabama. However, one possible concern could be an overlap of the host range of D. suzukii and Z. indianus in Alabama, which could allow Z. indianus to colonize fruits damaged by D. suzukii.

| Native species
Among native species, we could see some pattern between Drosophila seasonal abundance and their host dietary types. Drosophila affinis was most abundant across all months with the exception of August.
The presence of at least one food type during all the seasons, a broad range of oviposition hosts, and the relatively high cold tolerance of D. affinis could at least partially explain its highest abundance during most of the year.
Drosophila tripunctata is less discriminate in food preference and can be found on rotten fruits and slime fluxes (Carson & Stalker, 1951).
During our collections, both of these species were abundant in banana traps, suggesting that D. putrida could use rotten fruits as a food source in natural environments. Both of these species were found during nearly all collection months. Together, these generalist mycophagous species would make the second most abundant group of flies across our samples, which suggests that their diverse host range could be responsible for high abundance during all seasons.
Three relatively low abundance species were D. simulans, D. melanogaster, and D. robusta, and they all have a relatively narrow host range. Drosophila simulans and D. melanogaster primarily feed on rotten fruits (Strickberger, 1962;Sturtevant, 1916), and rotten fruit is also the preferred oviposition media for these species (Avondet, Blair, Berg, & Ebbert, 2003;Carson & Stalker, 1951). Drosophila simulans and D. melanogaster were absent from our traps from December to April, which could be correlated with fruit and berry season. Drosophila robusta feeds primarily on fruits, mushrooms, and tree saps (Carson & Stalker, 1951;Strickberger, 1962;Sturtevant, 1916), is very specific in choosing sites for oviposition, and in natural environments was reported to breed primarily on slime fluxes (Carson & Stalker, 1951;Avondet, Blair, Berg, & Ebbert, 2003). Abundance of D. robusta did not follow any obvious seasonal pattern, and the species was absent from our traps in September and October, and from January to March.
Based on our observations overall, we can conclude that the ability to use a broad host range for feeding and oviposition could play an important role in abundance of Drosophila species through all seasons in Alabama.

| Climatic impacts on abundance and biodiversity
According to our models, monthly temperature influences abundance more than the 2-day collection period temperature for total abundance of Drosophila, as well as abundance of D. tripunctata, D. melanogaster, D. robusta, and D. simulans. The effect of monthly temperature on total abundance was driven largely by the highly abundant D. affinis. The average amount of precipitation per month also influences total abundance and that of the most represented species of Drosophila, with the exception of D. affinis. The lack of a significant monthly precipitation effect on the D. affinis that we collected is especially interesting as D. affinis consisted of a substantial proportion of all collections, meaning that the other species were driving the variation in total abundance in response to monthly. We can conclude that difference in climate variables per month produces a more significant effect on the abundance of Drosophila than temperature and precipitation during collection days. In addition, we can see that most Drosophila species exhibit a quadratic response to a seasonal climate variation, suggesting the presence of an optimal climatic condition range for each species.
It was shown that precipitation had a positive correlation with mushroom's productivity, which might result in increased abundance of mycophagous flies (Krebs, Carrier, Boutin, Boonstra, & Hofer, 2008;Worthen & McGuire, 1990). Therefore, assuming that increased precipitation levels would facilitate fungi's fruiting body formation, we tested the influence of precipitation during the 56-to 28-day period preceding the collection on the abundance of mycophagous D. tripunctata and D. putrida. However, we did not find any significant correlation. The lack of correlation can probably be explained by combination of factors in mushroom and Drosophila ecology. Worthen and McGuire (1990) observed that rainfall could produce a significant effect on the following week's mushroom abundance and noticed that an individual fungi's fruit body is often short-lived. At 18°C, D. putrida and D. tripunctata egg to adult developmental time ranges from 14 to 15 days and might be shortened by warmer temperatures (Markow & O'Grady, 2005); thus, we might expect a spike in abundance to occur within 2-3 weeks of the increase in mushroom fruit bodies. Boulétreau (1978) described that almost half of female Drosophila melanogaster collected from natural populations were <24-36 hr old, while Roff (1980) suggested that in wild, adult Drosophila life span might be only few days; thus, any spike in abundance due to an increase in mushroom fruiting bodies might be expected to be of a short duration, on the order of a few days.
Based on the limited information about wild Drosophila life span and rate of mushrooms' fruit body productivity increase in the response to rainfall, it is logical that monthly precipitation during the collection month would produce more effect on mycophagous Drosophila populations than precipitation in the previous month.
However, surprisingly, the abundance of the two mycophagous species was actually negatively correlated with precipitation within the collection month (statistically significant for D. putrida). The lack of a clear association between host availability and mycophagous fly abundance at the monthly scale suggests that more granular analyses are needed in future studies to determine how and whether the fly abundance is influenced by the presence of mushrooms.
Temperature influences not only mature Drosophila activity but also its developmental time and larval survival (Crill, Huey, & Gilchrist, 1996;James, Azevedo, & Partridge, 1997). If temperature fluctuations during the month are out of a species' optimal range, then fewer mature Drosophila will develop. This could generally explain the greater influence of monthly temperature on abundance of Drosophila.
Interestingly, we did not find a significant correlation between abundance of D. suzukii or Z. indianus and any of the climate variables. In the case of Z. indianus, the major reason for this lack of correlation could be small sample size (25 flies) and our ability to find this pest fly only during fall months. However, D. suzukii was present most times of the year, which suggests that this pest species has physiological or behavioral adaptations to better resist differences in seasonal climate change than native Drosophila species.
Several previous studies reported no significant correlation between abundance of Drosophila and seasonal temperature variation (Guruprasad et al., 2010;Srinath & Shivanna, 2014;Torres & Madi-Ravazzi, 2006). The reason for the difference in our results relative to theirs could be due, in part, to the different approach in statistical analyses. Most of these studies used a linear regression model.
However, living organisms have an optimal range of climate conditions for their survival and reproduction that is not linear (Kindt & Coe, 2005), and different species of Drosophila exhibit different temperature tolerance (Goto & Kimura, 1998;Hoffmann, 2010;Kellermann et al., 2012). In addition, Poppe, Valente, & Schmitz, (2013) showed negative correlation between Drosophila abundance and maximum/ minimum temperatures, which further suggests that Drosophila are mostly abundant in a temperature range between the extreme values.
The quadratic model appears to be the most appropriate for analyzing the influence of temperature on the abundance of Drosophila species and allowed us to identify optimal condition ranges. For the most abundant species, their optimal monthly average temperature ranged from 18 to 26°C. In addition, ecological data are often over-dispersed (Kindt & Coe, 2005), and several studies indicate that a negative bino-  (Dobzhansky & Pavan, 1950;da Mata et al., 2015). In addition, several studies took into account only temperature and precipitation levels measured during collection periods (Poppe, Valente, & Schmitz, 2013;Srinath & Shivanna, 2014;Torres & Madi-Ravazzi, 2006). In this study, we indicated that changes in average monthly environmental variables can influence abundance of Drosophila in a more significant way than changes during the collection period. In addition, our survey found different species composition than found in other studies, and their inherent species-specific biology could be influenced by ecological conditions in distinct ways (Hoffmann, 2010;Kellermann et al., 2012).

| Land-use impacts on abundance and biodiversity
We were able to find significant correlation between land use and abundance of Drosophila species. The highest number of samples per trap came from urban park areas and could be explained by abundance of food sources for generalist Drosophila as a result of human refuse and a relatively high amount of vegetation that could provide a shelter (Ferreira & Tidon, 2005;van Klinken & Walter, 2001). The strongest correlation between land use and individual species abundance was shown by species that tend to use mushrooms as food and breeding substrate: D. putrida and D. tripunctata. In urban areas, biodiversity of fungal communities is lower than in rural areas (Egerton-Warburton & Allen, 2000;Newbound, Mccarthy, & Lebel, 2010), which potentially can influence abundance of mycophagous Drosophila.
In contrast to abundance measures, we found no significant correlation between Drosophila biodiversity and land use. A species accumulation curve that was made via a random accumulation method suggested that if we take into account only three sites per category (which is the maximum number of sites for nonurban parks), then the most natural environment should have had the highest number of species (Fig. 8). This is consistent with the idea that in a more undisturbed environment the number of rare Drosophila species would be higher (Parsons, 1991). Unfortunately, only the urban park zones were sampled at a sufficient number of sites to saturate species detection according to rarefaction analysis, which plots the number of species as a function of the sample's number and allows evaluation of the sufficiency of sample size (Kindt & Coe, 2005, Fig. S6).

| CONCLUSION
Global climate change can cause a shift in the composition and functioning of biological communities (Cardinale et al., 2006;Parsons, 1991;Root et al., 2003), and as exotic species extend their habitat range they often replace endemic organisms (Ferreira & Tidon, 2005;Hooper et al., 2008). In this study, we analyzed biodiversity and species abundance of Drosophila in Central Alabama. Nearly half of the identified species we found were not reported during the last major biodiversity surveys (Sturtevant, 1918(Sturtevant, , 1921, which suggests a change in Alabama's Drosophila species composition overall the last 100 years. It has been shown that endemic Drosophila species are usually sensitive to climatic variable changes (Parsons, 1991), and we found a significant correlation between the most abundant endemic Drosophila species and seasonal shifts in temperature and precipitation. We also found that biodiversity overall was influenced by seasonal temperature variation. Surprisingly, analyses of invasive pest species (Z. indianus and D. suzukii) did not show any significant correlation between their abundance and seasonal climate variables, suggesting that other factors, such as human influence, drive their abundances.
Urbanization often leads to destruction of natural habitats and significant changes in biodiversity and abundance of endemic and specialist species (Ferreira & Tidon, 2005;Parsons, 1991), and in our study, we found a significant correlation between land-use type and abundance of Drosophila. In addition, the majority of collected flies were representatives of generalist species. Our results suggest a significant change in Drosophila species composition and the absence of many historically endemic species in the subtropical region of Alabama. To better survey the whole biodiversity of Drosophila in Alabama, more collections should be performed across the state. Given the growing level of urbanization, we expect that cosmopolitan species of Drosophila such as D. simulans, D. suzukii, and especially Z. indianus will become more abundant in Alabama and could establish themselves as dominant species in urban environments.
F I G U R E 8 Species accumulation curves. Made with random accumulation method via 10,000 permutations for land-use categories: Circles represent industrial areas, triangles represent nonurban parks, and crosses represent urban parks