Effect of aphid abundance and urbanization on the abundance of Harmonia axyridis ( Coleoptera : Coccinellidae )

The factors that affect the local distribution of the invasive Harmonia axyridis are not yet completely resolved. Hypotheses predicting positive and independent effects of prey abundance and degree of urbanization on the adult abundance of this species in Central Europe were tested. Populations of H. axyridis were sampled in a period when it was most abundant, by sweeping lime trees (Tilia spp.) at 28 sites along a 20 km transect across urban (western Prague) and surrounding rural areas. The sites differed in aphid abundance (number of Eucallipterus tiliae per 100 sweeps) and degree of urbanization (percentage of the surrounding area within a 500 m radius covered by impervious human constructions). Multiple linear regression analysis of log-transformed data revealed that abundance of H. axyridis (number of adults per 100 sweeps) increased signifi cantly with both aphid abundance (P = 0.015) and urbanization (P = 0.045). The positive relationship between degree of urbanization and abundance of H. axyridis was thus not a side effect of variation in aphid abundance, which was also greater in urban than rural areas. The effect of urbanization might constrict the habitat available to H. axyridis and force this species to aggregate in urban green “refugia”. These results point to a plurality of factors that determine coccinellid abundance at natural sites.

areas because in these areas its life cycle is not synchronized with that of its potential prey (Honek et al., 2017).
The pattern of the local distribution of H. axyridis in invaded areas is also interesting.It is determined by several factors, the most important of which is its preference for particular plants.In contrast to in its native and some of the areas it has invaded, where H. axyridis is abundant in low-growing vegetation (Grez et al., 2014;Lu et al., 2015), in the western Palearctic it prefers trees and rarely occurs in stands of wild herbaceous plants and crops unless they are heavily infested with aphids (Adriaens et al., 2008).
Harmonia axyridis aggregates at sites where aphids are abundant (Jovicic et al., 2016;Soares et al., 2017;Viglasova et al., 2017) as an abundance of food is a prerequisite for reproduction (Ovchinnikova et al., 2016) and the adults are capable of locating and populating such sites (Osawa, 2000;With et al., 2002).

INTRODUCTION
Harmonia axyridis (Pallas), an aphidophagous lady beetle species native to the eastern Palearctic, Japan, China, Mongolia and Russian Siberia east of the Urals (Kovar, 2007;Bidinger et al., 2012), is worthy of attention because of its recent spread across the rest of world except Australia and New Zealand (Brown et al., 2011).The history, causes and consequences of this invasion are well investigated and authoritatively reviewed by Roy et al. (2016).
Great attention is paid to the factors affecting the distribution of H. axyridis in invaded areas (Bidinger et al., 2012).In Europe, this species gradually spread east and west from France, Belgium and Great Britain, where it fi rst became established in the early 2000s (Adriaens et al., 2008).The spread was rapid in a longitudinal direction resulting in H. axyridis arriving in central and southern Russia in the 2010s (Korotayev, 2013;Ukrainsky & Orlova-Bienkowskaja, 2014;Zakharov, 2015) but less rapid in a latitudinal direction (Brown et al., 2008;Honek et al., 2017;Soares et al., 2017) resulting in this species arriving later in Greece (Ceryngier & Romanowski, 2017) and Bulgaria (Tomov et al., 2009).It seems that H. axyridis, despite optimistic predictions (Poutsma et al., 2008), is possibly kept from colonizing dry subtropical Mediterranean at particular sites consisted of ≥ 5 lime trees (Tilia cordata Mill., Tilia platyphyllos Scop.) infested with the aphid Eucallipterus tiliae (L.) (Sternorrhyncha: Aphididae).Continuous monitoring of coccinellid abundance on trees (Honek et al., 2018) revealed that Tilia spp.were the most infested trees at that time.There were very few ladybirds on the other species of trees growing in the urban and rural areas (Acer spp., Betula spp., Quercus spp.).

Sampling period
The lime trees were sampled in one week 18 to 24 July 2018, which was when H. axyridis was most abundant on the lime trees.This is confi rmed by the results of a season-long monitoring of H. axyridis abundance at sites 19, 20, 22, 23, 24 and 28 (Fig. 1), during which the abundance of H. axyridis was monitored at weekly intervals.This revealed that sampling occurred when the abundance of adult H. axyridis was at a maximum.One-way ANOVA using abundance of H. axyridis at particular sites from 15-31 July (abundance at the date included in this study and one week before and one week after, i.e., 3 records for each site, total n = 18) as the response variable and site as factor revealed that in this period, differences between sites accounted for 95.1% of the variance in H. axyridis abundance.

Sampling method
Beetles were swept from the canopy up to a height of cca 3 m.Sampling was standardised by using a standard sweep net (35 cm diameter, 140 cm handle) operated by the same person (A.H.).All the days on which samples were collected were sunny and calm, and other sucking Hemiptera (Hanks & Denno, 1993) is also higher in urban areas.The positive correlation between the above two factors, prey abundance and urbanization, complicates and confounds an analysis of the factors determining the abundance of H. axyridis.
In order to resolve this problem, the relationships between the abundance of H. axyridis, abundance of aphids and degree of urbanization, was studied using a standardized protocol that minimizes the confounding effects of temporal (Honek et al., 2015) and regional (Viglasova et al., 2017) differences in the abundance of H. axyridis.This study focused on adults that were sampled over a short period and on a single host-plant/aphid system.We tested three hypotheses: The abundance of H. axyridis increases (i) with aphid abundance and (ii) with the degree of urbanization (iii), and that these increases are independent of one another.The aim was to elucidate the relationship between both of the factors that are associated with the abundance of this coccinellid.

Sites sampled
Abundance of H. axyridis was recorded at 28 sites scattered along an ~ 20 km long longitudinal transect that extended from the west Prague into the surrounding area (Fig. 1).The tree stands and maximum temperatures (experienced in the afternoon) were between 27.7 and 32.1°C.Sampling occurred between 08.00 and 18.00 h and consisted of 100 sweeps at each site.In each sampling session, we recorded (i) abundance of H. axyridis (number of adults caught per 100 sweeps), (ii) abundance of aphids (approximate number of aphids caught per 100 sweeps, pilot observations revealed that aphid numbers were determined with ~ 10% precision), and urbanization quantifi ed as (iii) the percentage of the surrounding area (a circle of 500 m radius centred at the sampling site) covered by impervious surfaces (buildings, streets, parking places, etc.).The latter value was called "degree of urbanization".It was determined using maps available at (https:// en.mapy.cz).

Data analysis
The relationships between the abundance of H. axyridis, abundance of aphids and degree of urbanization were obtained using raw data and linear regression, y = a + bx, in which y is the abundance of H. axyridis and x either aphid abundance or degree of urbanization.The relationship between the abundance of aphids and degree of urbanization was tested using the same formula where y is the abundance of aphids and x the degree of urbanization.To linearize the relationship between data variables (abundance of aphids, degree of urbanization and abundance of H. axyridis), the data were log transformed.Using log-transformed data, we used multiple linear regression, log(y + 1) = a + b 1 log (x 1 + 1) + b 2 log (x 2 + 1), in which y is the abundance of H. axyridis, x 1 abundance of aphids and x 2 degree of urbanization.We then tested the signifi cance of the regression coeffi cients b 1 and b 2 in order to determine the effect of abundance of aphids and degree of urbanization on H. axyridis abundance.Calculations were performed using SigmaStat 3.5 (Systat Software Inc., 2006).

DISCUSSION
The results revealed a positive effect of the abundance of aphids on the abundance of adults of H. axyridis, which is consistent with the prediction of hypothesis (i).This result is expected and consistent with earlier evidence (Hodek & Evans, 2012;Honek, 2012).While there is no doubt that the abundance of aphids has an important role in determin-ing the abundance of many coccinellids, identifi cation of the relative importance of the population processes that determine the maximum abundance of H. axyridis at a particular site needs further study.Adult populations consist of individuals arriving from the surrounding area as well as individuals that developed at the site.Population size is thus determined by a combined effect of the settling of adult immigrants at a site (Osawa, 2000;With et al., 2002), intensity of oviposition (Osawa, 2005;Seagraves, 2009) and success of pre-imaginal development (Hironori & Katsuhiro, 1997), which are all positively affected by aphid abundance.
The results further reveal a weaker but signifi cant positive effect of urbanization, consistent with hypothesis (ii).The positive effect of the urban environment on the abundance of H. axyridis (Adriaens et al., 2008;Mukwevho et al., 2017;Wagner et al., 2017) and its role in facilitating the rapid spread of this species (Veran et al., 2016) are well documented, but contrary effects are also reported (Rocha & Fellowes, 2018;Rocha et al., 2018).The mechanism behind the effect of urbanization is not yet resolved (Honek et al., 2017;Sloggett, 2017).We hypothesize that in rural areas, coccinellids may be distributed evenly, while in cities, where the suitable habitat is reduced to small patches of green (parks, green courtyards, tree lines bordering streets), coccinellids are forced to aggregate at these sites.Urban greens thus serve as small refugia in an otherwise hostile urban environment.
The positive effect of urbanization on H. axyridis abundance may be confounded by its parallel effect on aphid abundance.A positive relation between aphid abundance and urban environment has been established several times (Fluckiger & Braun, 1999;Mackos-Iwaszko et al., 2015), although contrary cases describing lower aphid abundance in urban areas compared to rural area are also reported (Sienkiewicz-Paderewska et al., 2017).In this study, a signifi cant positive relationship was found between degree of urbanization and abundance of E. tiliae.Mechanisms underlying the effect of urbanization on aphid abundance include biotic (natural enemies) and abiotic (salinity and drought) factors (Sienkiewicz-Paderewska et al., 2017;Andrade et al., 2017;Rocha & Fellowes, 2018).These effects, while outside the scope of this study, deserve more investigation.
Our study thus established and quantifi ed the separate action of two factors on the abundance of H. axyridis, abundance of aphids and urbanization, both of which affected the abundance of H. axyridis independently of one another.This result is consistent with the prediction of hypothesis (iii).Both factors contribute to the local diversifi cation of the abundance of H. axyridis and point to a plurality of factors that determine coccinellid abundance at natural sites.

Fig. 1 .
Fig. 1.Map showing the sites where the populations of H. axyridis were sampled.The size of the pictured area is 23 × 16 km and centred at 50.134°N, 14. 207°E.Insert: Standardized abundances (percentage of maximum at each site) of H. axyridis recorded at sites 19, 20, 22, 23, 24 and 28 in the vegetative season in 2018.Shaded area -period when the sampling occurred.

Fig. 2 .
Fig. 2. The relationship between the abundance of Harmonia axyridis (n individuals per 100 sweeps) and A -the abundance of aphids (n Eucallipterus tiliae individuals per 100 sweeps), B -the degree of urbanization, and C the interrelationship between abundance of aphids and degree of urbanization.