Macroecological patterns of fruit infestation rates by the invasive fly Drosophila suzukii in the wild reservoir host plant Sambucus nigra

The invasive pest Drosophila suzukii is a fruit fly native to Asia that infests a wide variety of fruits. Wild plant species are major reservoirs for D. suzukii populations but their infestation rates vary greatly among geographical areas. We aimed at disentangling the relative roles of macroclimatic, landscape and local factors in the rate of D. suzukii infestation of elderberry fruits (Sambucus nigra), a major wild host plant in Europe. We collected elderberry fruits across 215 sites distributed in 13 regions from North to South of France. We counted the number of emerging D. suzukii adults and tested for the effect of macroclimatic, landscape, local biotic and abiotic variables, as well as plant traits using linear mixed models. Site latitude and mean daily maximum temperature since the beginning of the year had respectively the strongest positive and negative effects on mean infestation rates across regions. Infestation rates also increased with forest cover in a 100 m radius around sampling sites. This study shows the importance of taking into account multiple scales and factors, from the fruit characteristics (i.e. maturity) to landscape composition and macroclimatic variables, when studying the interaction between D. suzukii and its host plants.


INTRODUCTION
The recent anthropogenic acceleration of intercontinental movements of species combined with climate change has caused an increase in the number of invasive alien species worldwide, including insect pests (Seebens et al., 2017;Skendži c et al., 2021). One of the most widespread and economically impacting invasive alien species nowadays is the fruit fly Drosophila suzukii (Matsumura, 1931), an insect pest native to Asia. Since 2008, it has spread rapidly across Europe (Cini et al., 2014) and North America , causing important damage to agricultural production with considerable economic impact Yeh et al., 2020). Indeed, this polyphagous pest infests a wide variety of cultivated , ornamental and wild fruits (Kenis et al., 2016;Poyet et al., 2015). Research on sustainable methods of biocontrol is still in progress (Ulmer et al., 2020(Ulmer et al., , 2021Wang, 2020). Given the rapidity of its expansion, disentangling the respective roles of local (fruit resource quantity and quality), landscape (land uses) and global (macroclimatic factors) conditions in determining the ecological niche of D. suzukii and driving its population dynamics is crucial to improve management strategies and better predict its future spread.
The interactions between D. suzukii and crop plants are well documented due to its important damages on fruit production. Although non-crop fleshy-fruited plant species represent major reservoirs for populations of D. suzukii (Delbac et al., 2020b;Kenis et al., 2016;Poyet et al., 2015), their role in the invasion and spread of the fly has been much less studied. Deciphering the role of non-crop plants is also crucial for understanding population dynamics across seasons.
Wild plants are often categorized as either host or non-host (Bolton et al., 2021;Yan et al., 2017), but this dichotomy overlooks the potential variability in infestation rates among wild plants in the field, both among species and among population within the same species (e.g., due to variation in climatic conditions among geographic locations or due to variations in local factors, Kenis et al., 2016). Yet, very few studies have looked at the impact of environmental factors on the infestation rates of wild and cultivated fruits by this fly (Delbac et al., 2020b;Tonina et al., 2016). More importantly, most studies exploring environmental drivers of D. suzukii abundance were conducted using traps in the field and were not based on direct measures of fruit infestation rates (dos Santos et al., 2017;Stockton et al., 2019b). Therefore, these studies do not provide accurate estimates of the contribution of a particular host reservoir to fly population dynamics.
Many factors may influence the success of fruit infestation by D.
suzukii. At macroecological scales, latitudinal and longitudinal gradients have been reported to affect the distribution of Drosophila species (dos Santos et al., 2017). Latitudinal gradients could affect D.
suzukii either directly through abiotic factors (e.g., temperature or humidity) or indirectly through biotic interactions (e.g., through the indirect effects of temperature or humidity on the host plants of D. suzukii). Latitudinal gradients are often correlated with macroclimatic variables, including temperature and rainfall, which are important drivers of the ecological niche of D. suzukii based on species distribution modelling (SDM) at large geographical scales (dos Santos et al., 2017;Langille et al., 2017;Ørsted & Ørsted, 2019).
Macroecological studies show larger populations of D. suzukii in places with mild winters and humid environments (Ørsted & Ørsted, 2019).
Drosophila suzukii abundance generally increases with latitude (to a certain degree) in the northern hemisphere, as observed in North America (Langille et al., 2017) and Europe (Ørsted & Ørsted, 2019).
Experimental studies are concordant with these macroecological observations, which suggests that latitudinal effects are partly due to direct effects of abiotic variables. Indeed, experimental studies show an optimal growth of D. suzukii populations around 22 C with a relative tolerance to cold but frost susceptibility (Jakobs et al., 2015;Stephens et al., 2015;Tochen et al., 2014) and a preference for high relative humidity (Tochen et al., 2016). Furthermore, because of the effect of temperature, increasing altitude may locally have the same effect as increasing latitude, and it has been shown that D. suzukii can move to higher altitude particularly during summer Tait et al., 2018). Latitudinal gradients can also impact the interactions between phytophagous insects and their host plants (Woods et al., 2012). For example, rates of herbivory decrease with latitude in other plant-insect systems (Lehndal & Ågren, 2015;Więski & Pennings, 2014). For D. suzukii, macroclimatic factors could indirectly affect infestation rates through their impact on the number and quality of fruits produced by host plants (including content in toxic compounds, number of fruits or fruit size ;Senica et al., 2017;Menzel, 2021).
At a finer scale, the landscape composition, i.e. the cover of natural and anthropic habitats and the length of corridors (hedgerows, rivers and roads) surrounding a site, is a key driver of the diversity and abundance of insects (Bianchi et al., 2006;Burel et al., 1998) including pests (Villa et al., 2020). Landscape is an important scale of analysis for the understanding of population dynamics and the management of pest populations. Studies that examine the relationships between D.
suzukii populations and landscape features remain scarce (Delbac et al., 2020a;Schmidt et al., 2019). D. suzukii uses a wide range of non-crop hosts scattered across the invaded landscapes (Arn o et al., 2016;Delbac et al., 2020b;Diepenbrock et al., 2016;Kenis et al., 2016;Poyet et al., 2015) and its populations alternatively move between fruit crop stands and natural habitats . Some non-crop hosts could even be responsible for maintaining D. suzukii populations in landscapes and therefore for the infestation of crop plants (Diepenbrock et al., 2016). There is increasing evidence that forest habitats serve as refuges, resource reservoirs and, finally, potential sources of D. suzukii to adjacent cultivated areas (Delbac et al., 2020b;Haro-Barchin et al., 2018;Poyet et al., 2014;Urbaneja-Bernat et al., 2020). Earlier infestation risks could occur in farms in high woodland landscapes (Pelton et al., 2016). Linear elements in landscape also appear to enhance the development and spread of D. suzukii populations. Maceda-Veiga et al. (2021) found a positive association between D. suzukii captures and spatial proximity to streams. Moreover, hedgerow networks are also suspected to foster populations of D. suzukii Siffert et al., 2021). Finally, similarly to other Drosophila species, the dispersal of D. suzukii follows commercial fruit routes (Cini et al., 2014;Lavrinienko et al., 2017) ending in urban areas.
Thus, D. suzukii abundance is likely to be influenced by road and building density.
At the local scale, various biotic and abiotic environmental factors shape the dynamics of D. suzukii populations. The population size of polyphagous pests, that infest plants from different families, increases with the diversity of habitats in the very close vicinity of a sampling plot, even if this diversity also fosters the presence of natural enemies (Chaplin-Kramer & Kremen, 2012;Schmidt et al., 2019). For D. suzukii, this positive effect of the diversity of neighbouring habitats is likely linked to a local diversity of fleshy-fruited plants which increases the resource in potential hosts (Kenis et al., 2016;Poyet et al., 2015).
Moreover, the abundance of D. suzukii populations is likely to increase with both the quantity and quality of infested fruits, which are influenced by local environmental factors such as the slope level and aspect, which determine exposure to solar radiation (Kaul et al., 2001). Indeed, local factors can influence the maturation process of the fruits, which is linked to other fruit characteristics such as the colour or the sugar content, both being important cues in D.
This study aims at disentangling the relative roles of macroclimatic, landscape and local factors that could affect infestation rates of a wild host plant by D. suzukii. We focused on the elderberry (Sambucus nigra L.) because this plant is a major host present across the invaded area of D.
suzukii in Western Europe (Kenis et al., 2016) and North America (Lee et al., 2015). This approach allows us to decipher the relative importance of each environmental factor on infestation rates across a wide geographical range by fixing host plant identity. The elderberry Sambucus nigra shows high rates of infestation in the wild (Kenis et al., 2016) as well as in laboratory conditions (Poyet et al., 2015). This deciduous shrub grows in various climatic conditions, from Mediterranean to oceanic, continental and mountain areas, although hot and dry climates are less favourable for its development (Rameau et al., 2008). This species is often present in disturbed and nitrogen-rich soils. It is found in woodland margins, floodplains, hedgerows, wastelands and other habitats characterized by eutrophic soils (Atkinson & Atkinson, 2002). The infructescence corresponds to a corymbiform cyme (hereafter named corymb) that can bear large amounts of drupe-type fruits (up to more than 500 per cyme). As they ripen, the colour of the fruits changes from green to reddish and then to black. Fruits within corymbs can be heterogeneous in their maturity, with black fruits contiguous to green fruits (Künsch & Temperli, 1978).
Given the ecology of this host plant, we hypothesized that, at a global scale, the latitudinal and climatic gradients shape the level of fruit production of S. nigra and thus the reproduction opportunities for D.
suzukii on this host. More specifically, we hypothesized that D. suzukii might show an increasing infestation gradient of S. nigra with latitude.
The climatic niche of this host plant is optimal under northern temperate climates compared to the Mediterranean climates which is too dry and characterized by heat waves (Rameau et al., 2008), thus showing a larger quantity of available corymbs in the North of France. At a finer scale, we expect infestation rates to be modulated by landscape and local environmental factors that condition the presence of D. suzukii populations and/or their accessibility to elderberry fruits. Particularly, as forest habitats are commonly reported as natural refugia for D. suzukii, we expected a strong increase of fruit infestations across sampling sites with an increase in neighbouring forest cover. Using elderberry fruits collected from 215 sites along a latitudinal gradient in France, we estimated the variation in D. suzukii infestation rates and tested for an effect of macroclimatic (temperature, rain, latitude), landscape (land cover, hedgerow and watercourse length around sites) and local abiotic and biotic variables (slope, aspect, altitude, habitat, fleshy-fruited plant species presence), as well as plant traits (number of fruits, maturity, diameter, tree height, leaf and corymb sizes). We assessed the variation in resource quality among sites using fruit maturity (colour and associated sugar content) and the variation in resource quantity by counting the number of fruits per corymb and corymbs per shrub. Complementary laboratory experiments were thereafter carried out to document the effect of fruit maturity on D. suzukii oviposition preference and development success. Indeed, sugar content in host fruits is a major attractant of D. suzukii (Travaillard, 2020) and a primary compound of its diet (Biolchini et al., 2017).

Collection and measures of Sambucus nigra fruits and plant traits
Elderberry corymbs were sampled between July 17th and September 1st 2020 in 13 different regions in France (from North to South, see Figure 1 and Table 1). Depending on their availability in the field (the shrubs were scattered and rare in some regions due to the dominance of open crop fields in the agricultural landscape) and after a maximum of one week of search effort to locate elderberry trees bearing fruits, between 15 and 30 shrubs were randomly sampled per region, resulting in a total of 215 sampled shrubs (Table 1). The shrubs were at least 500 m apart from each other.
One corymb, bearing between 26 and 535 fruits, was collected per shrub to avoid spatial dependency between samples. The following reproductive and vegetative traits were measured for each shrub: the size of the sampled corymb, the number of fruits of each stage of maturity (green, red, reddish-black, black, overripe or dry), the fruit diameter (measured from five black fruits randomly chosen within each corymb), the number of corymbs on the shrub, the height of the shrub, the size and number of leaflets of the largest leaf. For each elder shrub, we also estimated the mean volume of a fruit (mean radius 3 Â π Â 4/3), the volume of fruits on the collected corymb and the total volume of fruits on the shrub.

Environmental variables
To examine the effects of regional and local environmental conditions on the number of fruits and infestation rates, local, landscape and global environmental variables (Table S1) Table S1. Climatic conditions were characterized for each sampling site. The daily meteorological data between 01/01/2020 and the day of sampling of elderberry fruits were retrieved from the nearest meteorological station of each site (https://www.historique-meteo.net/ F I G U R E 1 Location of the 13 French cities which determined the studied regions france/). Daily minimum, mean and maximum temperatures were extracted for each day from January 1st to the day of sampling, allowing us to calculate the mean daily difference in temperature (daily maximum -daily minimum) as well as the number of frost days since the beginning of the year. Degree-days were calculated using a lower threshold of 0 C between 01/01/2020 and the day of sampling (Baskerville & Emin, 1969). The baseline value of 0 C is a standard threshold commonly used in insect and plant studies (McNeil et al., 2020;White et al., 2012). This threshold is particularly adapted to study the temporal synchrony between flies and plant resources (Iler et al., 2013). This threshold was also chosen as we observed and

Emergences of drosophila species
After collection, the elderberry corymbs were individually placed on paper towels in mesh-covered plastic containers. They were kept at room temperature and humidified to keep the paper towels wet. Adult flies were picked up, as they emerged from the corymbs, and placed in ethanol. Flies were then identified to the species level and the D.
suzukii individuals were sexed using a Leica M205C stereomicroscope equipped with a Leica MC170 HD camera and the software Leica Application Suite.
Laboratory study of the effect of fruit maturity on D. suzukii oviposition preference and egg-to-adult viability Fruit collection and D. suzukii rearing conditions To better understand the role of the maturity of the fruits in D. suzukii preferences and infestation success, we carried out experimental assays in controlled conditions at the site of Amiens. Indeed, results from our observational field data (see GLM results, hereafter) strongly suggested that the number of D. suzukii flies emerging from S. nigra fruits may depend on their ripening level. Colour changes in S. nigra fruits are well pronounced during maturation (Künsch & Temperli, 1978). Therefore, laboratory experiments were developed to assess the attractiveness and suitability of three ripening stages of S. nigra fruits, according to their colour (Künsch & Temperli, 1978;Mitsui et al., 2006;Poyet et al., 2014): entirely green (

Statistical analyses
We examined the relationships between explanatory environmental variables and each of three fruit infestation variables that where     (Table 2a). Within each region, the number of D. suzukii emerging per corymb increased with forest cover in a 500 m radius around the sampling site (Table 2a). The rate of fruit infestation (INFEST) decreased with the mean daily maximum temperature, both across and within regions (Table 2b). The rate of fruit infestation T A B L E 2 Effect of environmental variables and elderberry traits on (a) the number of emerging Drosophila suzukii per corymb (SUZUt), (b) the fruit infestation rate per corymb (INFEST = 100 × SUZUt/number of fruits in a corymb) and (c) the frequency of corymb infestation per region (Finfest = number of infested corymbs in a region/number of sampled corymbs in the same region) at the scale of the sampling site (n = 215) and/ or the region (n = 13).   Table 2b). Forest cover was also negatively correlated with the cover of buildings (r = À0.402, p < 0.001), but the latter variable was not retained in models. The infestation frequency of corymbs (Finfest, region scale, Table 2c)  infestation rate and local environmental variables is given in Table S2).

Laboratory experiments
We found a significant effect of colour/maturity of fruits collected in the field near Amiens on the total number of eggs laid (χ 2 = 18.99, p < 0.001; Figure 3a), the number of infested fruits (χ 2 = 19.35, p < 0.001; Figure 3b) and the number of eggs laid per infested fruit (χ 2 = 9.45, p = 0.009; Figure 3c). The preference increasing order was always green, red and black. The preferences of D. suzukii for a fruit maturity/colour were however not significant in the laboratory choice tests, although a trend was observed toward more eggs being laid on black fruits than on red fruits, black than on green and red than on green ( Figure 4).
We also found a significant effect of fruit colour/maturity on the number of D. suzukii adults emerging from fruits collected in the field, with the same order as for colour preference measured previously (i.e. green < red < black; χ 2 = 28.32, p < 0.001; Figure 3d). Since there was a significant difference (Wilcoxon, W = 547.5, p = 0.011) between the mean number of adults emerging from the 50 fruits on which we counted the eggs (4.92 AE 1.18) and the 50 fruits that were left unhandled (13.95 AE 2.17), we assumed that the manipulation of fruits impacted greatly the egg-to-adult survival rate. Knowing both the mean infestation rate found in the wild (Figure 3a) and the mean number of flies that emerged (Figure 3d), an estimated egg-to-adult survival rate could be calculated for each maturity stage. The egg-to-adult development success (number of adults Â 100/number of eggs) was estimated at 8.24%, 63.78% and 70.2% in green, red and black fruits, respectively.
Chemical analyses of sugar in S. nigra fruit revealed that for each type of sugar tested (glucose, fructose and sucrose), differences between the maturity stages were significant and sugar contents followed the order green<red<black ( Figure 5). This indicated that sugar content increased during fruit maturation, with a total concentration 8.10 and 3.47 folds higher in black fruits than in green and red fruits, respectively.

DISCUSSION
Our study showed that three main categories of environmental factors systematically influenced fruit infestation: (i) latitude and associated climatic factors, (ii) landscape composition and (iii) local environmental factors. We also demonstrated that a plant generally considered as a "good host" for D. suzukii such as S. nigra shows a high variability of fruit infestation rates both across and within regions.
Fruit infestation rates varied greatly among geographic regions (min = Nîmes 0.83%, max = Boulogne-Sur-Mer 23.5%), which translated into a significant inter-regional effect of environmental factors (climatic gradient). Fruit infestation rates also varied among sampling sites within each region (e.g., from 0.37% to 46.44% in the region of Amiens) due to an intra-regional heterogeneity of the environment (e.g., contrasting landscapes).

Latitude and associated climatic factors
We found that latitude, which is negatively and positively related to temperature and cumulative rainfall respectively, was a major variable . Indeed, we found that leaf length, corymb diameter and area varied with latitude but none of these variables were related to infestation rates over the whole study including 215 sites. We did not find any correlation between latitude and total fruit number per corymb or fruit diameter. Other local factors such as competition for light with neighbouring shrubs, soil characteristics, shrub management history and age, local genetic diversity or the abundance of local pollinators are likely to modulate plant performances (e.g., growth, nutritional quality, etc.), thereby explaining the observed phenotypic variations (Llorens et al., 2018;Patten & Wang, 1994).
The total number of corymbs and total volume of fruits produced per shrub were also negatively correlated with temperature and positively with rainfall showing that the temperate climate conditions of northern France may be more favourable to S. nigra reproduction output. The number of D. suzukii emerging per corymb increased with this overall volume of fruits hanging on the whole shrub. These results confirm our hypothesis that the northern temperate climate, more suitable to S. nigra, promotes the overall production of fruits, which has a positive cascading effect on infestation rates. Nevertheless, this hypothesis is only partially validated as most of the other elderberry traits did not respond to latitude. Olfactory or visual cues such as the size of the fruit resource are known to be used by other pests such as Tephritidae fruits flies to forage for food (Prokopy et al., 1973;Prokopy & Roitberg, 1984). We thus can hypothesize that shrubs bearing larger amounts of fruits in temperate regions could be more easily detected by the flies when moving across the landscape, and consequently more attractive.
The rate of infestation of S. nigra was strongly correlated with latitude (SUZUt and INFEST variables). Studies on other model species, particularly agricultural pests (Yadav et al., 2018), have shown positive relationships between insect abundance and latitude. This could be the case here as the rate of infestation is likely to increase with D. suzukii abundance. Moreover, the proportion of dipteran species among flower-visiting insects was shown to increase with the latitude, at least in Europe (Elberling & Olesen, 1999). However, the impact of latitude on insect distribution is complex and cannot be reduced to a pure geographical (i.e. spatial) effect. Many other factors such as macroclimatic ones are probably involved in this latitude-insect infestation success relationship. For example, in northern Europe during cold years, abundance of birch feeding leafminers significantly decreased with latitude, while during warm years their abundance increased with latitude (Kozlov et al., 2013). In the meantime, other data suggest that insects living at lower latitude (i.e. tropics) are less tolerant to climate warming than insects living at higher latitude (Deutsch et al., 2008), although this relationship may be modulated by microclimatic patterns in herbivores (Pincebourde & Casas, 2019).
D. suzukii, living mostly in a temperate climate, has developed specific adaptations to survive cold winters (Enriquez & Colinet, 2017 (Kimura, 2004;Kinjo et al., 2014;Tochen et al., 2014). The dry environment may also negatively affect D. suzukii populations, as this species shows a better development and reproduction at high humidity levels (Guédot et al., 2018;Hamby et al., 2016;Tochen et al., 2016). In agreement with this, a low relative humidity seems to reduce its resistance to heat (Enriquez & Colinet, 2017). Moreover, in the dry climatic conditions of southern produces fruits (Poyet et al., 2015). This delayed phenology and lateseason build-up of D. suzukii populations also found in colder locations were already shown by Gutierrez et al. (2016) in their analysis of trapping records by Dalton et al. (2011) in the United States.
The chemical composition of plants plays an important role in modulating herbivory. For example, phosphorus is a nutrient present in plants that play an important role in the reproduction of female Drosophila (Markow & O'Grady, 2008) and its content seems to be an important cue in D. suzukii oviposition preferences (Olazcuaga et al., 2019). This primary compound is present in leaves, fruits and seeds of Sambucus species (Pandia et al., 2018;Short & Epps, 1976 Tait et al., 2020). Moreover, D. suzukii abundance is known to increase from the forest edge toward the forest interior (Poyet et al., 2014), indicating the close association between this fly species and wooded habitats. The affinity of D. suzukii for forest habitats is also reported in its native range . Forest patches act as refuge areas for D. suzukii for two reasons. First, forest ecosystems provide a multitude of shelters including branches, leaves and litter (Briem et al., 2018) and is characterized by a buffered microclimate protecting the pest from extreme temperatures (Wallingford et al., 2018). Second, forest environments offer varied food resources for both the adults and the larvae such as fleshy-fruited species and mushrooms (Kenis et al., 2016;Poyet et al., 2015;Stockton et al., 2019a). In our study, this positive effect of wood cover was mainly observed in a radius inferior or equal to 500 m (i.e. in a 500 m or a 100 m radius and in the immediate vicinity). A previous study showed that D. suzukii has a relatively low dispersal capacity, with a daily flight distance below 100 m (Vacas et al., 2019), which could explain this positive forest effect only at small distance ranges.
Local trophic resources D. suzukii infestation rates were positively associated with the number of fruits per corymb in the field (Table 2). This could be caused by bigger populations where S. nigra bear higher numbers of fruits. D. suzukii could also be able to estimate the local resource quantity on the shrub and target the optimal infructescence to oviposit. In agreement with this, D. suzukii is more attracted by large shapes (Poyet et al., 2015;Rice et al., 2016)  Our laboratory experiments suggest that ripe (black) fruits are subjected to higher infestations in the wild and allow a higher egg-to-adult survival rate than unripe fruits on cut corymbs in laboratory conditions.
Even if the oviposition preference was not clearly demonstrated in controlled conditions, likely due to a small number of repetitions, observation on wild fruits showed a consistent trend of infestation following the maturity of the fruit in the order green<red<black. Red and black are the two most visually attractive colours for D. suzukii (Rice et al., 2016), which could explain the higher number of eggs found on fruits of these colours. As shown in our chemical analyses, these fruits also have higher sugar contents, which are essential cues in the fly's oviposition preferences (Biolchini et al., 2017;Travaillard, 2020). Thus, both skin colour and sugar content could act in synergy in fruit attractiveness to the fly.
Other chemical compounds that were not measured in this study could also interfere in fruit attractiveness. For instance, S. nigra fruits are known to be rich in anthocyanins (Veberic et al., 2009), pigments that are usually red but can appear purple, blue or even black depending on various chemical factors (Lev-Yadun & Gould, 2009). In many species, anthocyanin content is positively correlated with latitude (Åkerström et al., 2010;Jaakola & Hohtola, 2010;Lätti et al., 2008). Altitude, which has an effect similar to latitude, seems to be positively correlated to the anthocyanin content in S. nigra fruits (Senica et al., 2017). Anthocyanins being a source of antioxidants, their consumption has been proven beneficial for Drosophila melanogaster (Valenza et al., 2018;Wang et al., 2016). A higher number of reddish-black fruits as well as higher contents of anthocyanins could therefore mean resources of better quality for D. suzukii and explain the higher infestation of elderberry at high latitudes. Although it is likely that other factors than the sole fruit maturity play a role in D. suzukii's preferences, black fruits were more infested than red ones in the wild even if D. suzukii is generally considered to prefer ripening fruits to fully ripe ones (Poyet et al., 2014). However, this could be explained, at least partly, by a delay between the egg-laying and the sampling, leaving time for the fruit to mature from red to black.

CONCLUSION AND PERSPECTIVES
Our study shows the importance of taking multiple scales and multiple factors into consideration when studying the interaction between predict changes in the pest invasion process. The present study predicts that global warming, by increasing the phenological mismatch between the fly and its host plant, should reduce its infestation, especially in southern Europe, but the risks of S. nigra infestation should be displaced toward northern Europe, generating a new ecosystem dis-service to agriculture in higher latitudes.

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.