High Canopy Cover of Invasive Acer Negundo L. Affects Ground Vegetation Diversity


 We assessed the link between canopy cover degree and ground vegetation diversity under alien the ash-leaved maple (Acer negundo) and other (native or alien) tree species. We investigated urban and suburban forests in the large city of Yekaterinburg, Russia. Forests were evaluated on two spatial scales. Through an inter-habitat comparison completed over three years, we recorded canopy cover and plant diversity among 13 sample plots of 20 × 20 m where A . negundo dominated and 13 plots where other tree species dominated. In an intra-habitat comparison, we recorded canopy cover and ground vegetation diversity among 800 sample plots measuring 1 m2 in the extended urbanised forest, which featured abundant alien (308 plots) and native trees (492 plots). We observed decreased diversity among vascular ground plant species by 40% (inter-habitat) and 20% (intra-habitat) in areas dominated by the A. negundo compared to areas dominated by native tree and shrub species. An abundance of A . negundo was accompanied by increased canopy cover. We found a negative relationship between canopy cover and the number of understory herbaceous species. Thus, the interception of light and the restriction of its amount for other species is a main factor supporting the negative influence of A . negundo on native plant communities.


Introduction
Ash-leaved maple (Acer negundo L.) is an invasive tree in the territory of Northern Eurasia that is currently colonising disturbed and semi-natural territories ( The ash-leaved maple is not only an alien and invasive species but also a transformer species. Transforming species signi cantly change the conditions in the invaded ecosystems (Richardson et al. 2000). The impact of transformer species is realised by in uences on the light regime of communities (Gorchov and  In this study, we focused on testing the hypothesis of whether the associated effects of the canopy cover of A. negundo can explain its effect on living ground vegetation, namely shrubs and herbs. In general, light availability determines the productivity of the ground vegetation ( (Canham 1994;Knight et al. 2008; Barbier et al. 2008). Competition for light or specialisation in its use is the leading determinant in the organisation of plant communities. Therefore, we assumed that the main direction of the relationship between canopy cover and ground cover diversity is negative; that is, with an increase in canopy cover, the number of ground cover species decreases.
The reduced richness of ground cover under A. negundo due to its in uence on illumination levels can be explained in several ways. Firstly, it can be assumed that the ash-leaved maple shades the soil surface more completely than trees of other species (Fig. 1 a), implying that the canopy cover in thickets of A. negundo is denser than in thickets of other trees. However, at the same time, similarly dense canopy cover decreases the richness of ground cover in the same way regardless of the tree species forming the canopy. This assumption seems easy to understand but has no de nitive experimental con rmation, while there is much evidence that invasive plants create a denser canopy than native plants (Reinhart et  In addition to the hypothesis about the strong shading of A. negundo, it is possible that even with the same canopy cover, the ash-leaved maple has a stronger effect on ground cover than other trees ( Fig. 1 b -c). This suggestion implies that the reduced richness of the ground cover under A. negundo is associated not only with shading but also with certain other mechanisms. The reasons for this are potentially very diverse and may include increased litter formation (Nilsson et al. 2008 1 Hypothetical mechanisms of the generation of reduced diversity of the ground cover throughout shading in communities with Acer negundo (dark-grey circle) in comparison with other trees (white circle); rhombus-average rates We emphasise the need for further research, including in new geographic regions, to identify the mechanisms of A. negundo invasion and the impact of this invasion on native plant communities. In this study, we conducted two eld investigations to estimate the impacts of ash-leaved maple invasion on native plant communities. This paper analyses whether the ability of A. negundo to create a dense canopy explains its in uence on ground cover.
We tested two hypotheses. First, we assumed that A. negundo trees create a denser (closed) leaf canopy than other tree species of the southern taiga.
Second, we assumed that the angles of linear regression describing the relationship between the density of the canopy and the species richness of the ground cover are the same in both A. negundo thickets and those of other tree species.

Materials And Methods
Study area. The data was collected in urbanised habitats in the southern taiga subzone of the boreal zone of the Middle Urals, located in the territory of the Yekaterinburg urban agglomeration. The average annual temperature is +3°C, which has shown an upward trend (1°C increase in the last 25 years). The average annual precipitation is 542 mm. The average height of snow cover in February is 20-25 cm, and the average duration of snow cover is 160-180 days. Yekaterinburg is a large city with a population of General design. To test our working hypotheses, we executed two studies: the rst was an inter-habitat comparison, while the second was carried out as an intra-habitat comparison.
Sample plots. The inter-habitat comparison was carried out using 13 sites of intra-urban vegetation located in the city of Yekaterinburg and its suburbs in the Middle Urals, Russia. The area of the site allowed the placement of at least two plots measuring 20 × 20 m. At each site, we selected two plots, the rst in the thickets of ash-leaved maple and the second in thickets of other tree species. Two plots in one area formed a linked pair of one plot with the studied impact of A. negundo; (An+) and one without its impact, (An-; factor 'plot type'). Paired plots were located as close together as possible and no further than 0.4 km from each other. Moreover, they were similar by a relievo and location relative to human housing and infrastructure, and they had close values in terms of canopy cover. The level of moisture was normal (10 sites) or increased (three sites). According to the European nature information system (EUNIS) habitat classi cation, 10 sites were classi ed as small green spaces completely or almost surrounded by buildings (X22) or roads (X23), three sites as large parks (X11) and one site as low forest or shrubs in wetlands (F9) (Hill et al. 2004). In total, we laid 26 plots. An-plots were dominated by Ulmus laevis Huds. Estimation of canopy cover. Every year in mid-July, we took 10 colour photos of the canopy at each plot. We took the photos in randomly selected places, pointing the camera straight up to a height of 0.8-1.2 m. We used a Lumix DMC-FP2 digital camera (CCD sensor: 1/2.5"/10.3 million pixels/primary colour lter; photo resolution: 3648 × 2736 pixels). In total, we took 720 photos. To prepare images for analysis, we used Adobe Photoshop 11.0 (Adobe System Inc., 2008). Each photo was converted into binary so that crowns, tree trunks and other obstacles to natural sunlight were rendered as black pixels. The open sky was displayed as white pixels. The analysis of canopy cover was performed in Matlab R2018b Intra-habitat comparison.
Site. The second study was performed in June 2018 in the Yugo-Zapadnyi Forest Park in the city of Yekaterinburg. According to the EUNIS classi cation, a forest park is an X11 habitat type (large parks) (Hill et al. 2004). We laid the site measuring 795 × 20 m on the 7°-northerly slope from the top of a small ridge to the middle of the slope. The site was generally populated by nettle pine forests with an undergrowth of Rubus idaeus L., forb-grass and small-grass. The pine stand was strongly disrupted by selective felling, and derivative communities with Populus balsamifera L. and A. negundo in the tree layer were formed at certain places. On a 795 × 20 m site, we laid ve parallel transects 795 m in length on each of which 160 plots were marked at 5 m intervals. As a result, we formed a square grid with a step of 5 m and 800 plots in its nodes. Situation of the study site and generalised data of its characteristics are given in the supplementary material.
Characteristics of plant communities and the estimation of canopy cover.We took two photos with the Lumix DMC-FP2 digital camera at each of the 800 plots. The rst photo was taken straight up from a height of 0.8-1.2 m, while the second was taken straight down from a height of 1.5-2 m. At the same time, a 1 × 1 m frame was laid on the ground.
In each photo of the canopies, we visually estimated the total canopy cover as a percentage of the eld of view with an accuracy of 5%. We also identi ed the species or at minimum the genus of trees and shrubs which fell into the frame of each photo. The proportion of coverage of each taxon was estimated by eye.
In each photo of the ground cover, we identi ed the species, the genus or at minimum the family of low shrubs, herbs and individual trees less than 1.5 m in height. We counted only the number of taxa that were located within a 1 × 1 m frame.
During plant identi cation, we used a list including 131 species based on 27 preliminary geobotanical descriptions of this section of the forest park ). We were able to identify some of the plants in the photos only to the family or genus level. At the same time, plants we were unable to identify to the species level were considered as different conditional species if we could distinguish their morphological features. This made it possible to su ciently fully calculate the total number of vascular plant species on each plot. We used the term 'species richness' when evaluating alpha diversity.
At each site, we identi ed a dominant taxon of woody plants without dividing their life forms (shrubs, undergrowth trees or trees of the second and rst layers). The taxon which occupied the largest share of the total canopy cover in the image was considered dominant.
All woody plant species encountered were assigned to one of two groups: native () or alien (). Data analysis. We used different versions of GLM (general linear models) to carry out analysis of effects that may in uence on ground cover diversity. The values of the canopy cover and species richness of the ground cover were analysed for study 1 in a two-way ANOVA (factors: 'plot type' and 'year of observation') and for study 2 in a one-way ANOVA (factor: 'woody dominant'). The relationship between canopy cover and the species number of vascular plants for ground cover were assessed using three-way (study 1) and two-way (study 2) GLMs with the discrete factors 'plot type' and 'year' and the continual factor of 'canopy

Results
Canopy cover. In both the inter-habitat and intra-habitat comparisons, higher canopy cover was observed for A. negundo in comparison with other species of woody plants in urban pine forests.
Inter-habitat comparison. We observed a slightly higher canopy cover in communities dominated by the ash-leaved maple compared to communities with other woody dominants. In a two-way ANOVA, the differences between plots An+ and An-in canopy cover were signi cant (Fplot type (1; 66) = 6.06; P = 0.0165), while the effect of the year (Fyear (2; 66) = 1.91; P = 0.1555) and the interaction between option and year were insigni cant (Fplot type × year (2; 66) = 0.18; P = 0.8344). The absolute differences in the mean values of canopy cover between the plots were small (Fig. 2): 90 ± 1% in the An+ plots and 86 ± 1% in the An-plots. Fig. 2 Average rates (±SE, ±95CI) of canopy cover in communities dominated by Acer negundo (An+; darkgrey plots) and other tree species (An-; white plots) Intra-habitat comparison. We found a higher canopy cover in areas dominated by A. negundo compared to areas with other tree dominants. In the one-way ANOVA, the differences in canopy cover between the plots dominated by native and alien tree species versus A. negundo were signi cant (Fplot type (2; 797) = 4.21; P = 0.0151). The absolute differences in the mean values of canopy cover between the options were small (Fig. 3): areas with native dominants showed 66 ± 1%, with A. negundo 70 ± 1% and with other alien dominants 70 ± 2%. Fig. 3 Average rates (±SE, ±95CI) of canopy cover in areas dominated by native (white plots), alien, excluding Acer negundo (white plots) species and individually -Acer negundo (dark-grey plots) We found that native and alien woody species resulted in heterogeneous canopy cover (Fig. 4). Some alien species, such as Populus balsamifera, had thin canopies, while some native species, such as Prunus padus and Sorbus aucuparia, had dense canopies. Moreover, among the alien species, the canopies of A. negundo were not the densest. Some alien woody plants (namely Ulmus laevis, Acer tataricum, and Malus baccata) had higher average canopy cover than A. negundo. Richness of ground cover. In both inter-habitat and intra-habitat comparisons, we observed reduced species richness of the ground cover on plots with a tree layer formed by the ash-leaved maple.
Inter-habitat comparison. We observed reduced species richness of the ground cover in communities dominated by A. negundo compared to communities with other dominant trees. In a two-way ANOVA, the differences in the number of species per 400 m 2 were signi cant between plots An+ and An-(  Intra-habitat comparison. We found that on the plots with alien woody dominants, including A. negundo, the number of ground cover species was lower in comparison to sites with native dominants (Fplot type (2; 797) = 28.58; P < 0.0001). We observed that absolute differences in average values of species richness between the plots with native woody dominants were 6.6 ± 0.1 species per m 2 ; with alien woody dominants, excluding A. negundo, -5.2 ± 0.2 species per m 2 ; and with A. negundo -5.3 ± 0.2 species per m 2 (Fig. 6).

Fig. 6 Average number (±SE, ±95CI) of observed ground cover species per m 2 on plots dominated by native (white plots), alien, excluding Acer negundo (light-grey plots) species, and individually -on plots dominated by Acer negundo (dark-grey plots)
Plots were relatively homogeneous in terms of richness of the ground layer under the canopy regardless of whether the dominant species was native or alien (Fig. 7). Areas dominated by different species of alien plants did not differ in terms of ground richness. Among the native plants, two groups of species were distinguished according to the richness of ground cover communities: 1) underbrush shrubs Prunus padus and Sorbus aucuparia and 2) trees of the rst and second tiers of Pinus sylvestris, Salix sp., Betula sp., Populus tremula.
Relationship between canopy cover and species richness of ground cover. In both the inter-habitat and intra-habitat comparisons, a negative correlation was found between canopy cover and the number of vascular plant species comprising ground cover. Fig. 7 Average number (±SE, ±95CI) of observed ground cover species per m 2 on plots under the canopies of different native (white plots) and alien (light-grey and dark-grey plots) tree species; rare species -the indication for a group of plots dominated by species rarely found in the studied site (Amelanchier spicata, Caragana arborescens, Cotoneaster lucidus, Crataegus sanguinea, Larix sibirica, Lonicera xylostella, Tilia cordata) Inter-habitat comparison. In the three-way GLM using the factors 'plot type', 'canopy cover' and 'year', the species richness of the ground cover signi cantly depended on the main effects 'plot type' (Fplot type (1; 60) = 13.61; P = 0.0005) and 'canopy cover' (Fcover (1; 60) = 6.02; P = 0.0170). No other effects, including interaction effects, were signi cant. This means that the angle of inclination of the lines (coe cients b in the equation y = a + bx) describing the relationship between canopy cover and the number of species of ground cover on plots dominated by the ash-leaved maple and other tree species did not differ (Fig. 8, b -d). An increase of 10% in canopy cover induces a decrease in the number of ground cover species by 5.82 ± 3.30 species per 400 m 2 (P = 0.0866) in the An+ plots and by 4.77 ± 2.66 species per 400 m 2 (P = 0.0820) in the An-plots. The estimates for species richness of the ground cover in the absence of the shading effect of trees (coe cient a in the equation y = a + bx) in the plots An-and An+ were similar: in the An-plots a = 70.05 ± 23.09 species (P = 0.0046) and in the An+ plots a = 69.29 ± 29.60 species (P = 0.0252). Intra-habitat comparison. In the second study's two-way GLM with the factors 'plot type' and 'canopy cover', the number of ground cover species on the site signi cantly depended only on the main effects Fplot type (2; 794) = 25.73 (P < 0.0001) and Fcover (1; 794) = 12.43 (P = 0.0004). The 'plot type × canopy cover' interaction was not signi cant: Fplot type × cover (2; 794) = 1.69; P = 0.1855; this indicates that we did not establish a difference in the angle of inclination of the lines describing the relationship between the canopy cover and the number of ground cover species per m 2 on plots with different woody dominants (Fig. 8, a). However, the species richness of the ground cover changed with the growth of the canopy cover depending on the dominant plants. A signi cant decrease in species richness with an increase of canopy cover was observed in plots with a dominance of both native (with a 10% increase in canopy cover by 0.27 ± 0.07 species per m 2 ; P < 0.0001) and alien plants (with a 10% increase in canopy cover by 0.38 ± 0.13 species per m 2 ; P = 0.0035). On plots dominated by A. negundo, the number of ground cover species did not change signi cantly with an increase in canopy cover (with an increase in canopy cover of 10%, the number of ground cover species per m 2 decreased only by 0.04 ± 0.11; P = 0.7465).
In the plots dominated by A. negundo, the species richness of the ground cover was reduced even at the lowest shading levels: in the equation y = a + bx, a = 5.51 ± 0.79 species per m 2 (P < 0.0001). On plots dominated by native or alien plants-excluding A. negundo-initial levels of ground cover richness were higher: under native plants, a = 8.40 ± 0.46 species per m 2 (P < 0.0001); under alien plants, a = 7.92 ± 0.93 species per m 2 (P < 0.0001).

Discussion
Our results show the general consistency of the appearance at the two spatial scales we observed. In both inter-habitat and intra-habitat comparisons, we registered a reduced ground cover species richness under canopies of ash-leaved maple. This result con rms the generally accepted idea that alien plants produce a negative impact on the diversity of native communities. Our earlier inter-habitat comparison showed a decrease in alpha diversity in uenced by A. Additionally, we found that the results regarding the transformation of the light regime correspond between inter-habitat and intra-habitat comparisons. In both studies, we con rmed that the canopy cover of A. negundo is higher than the canopy covers of other tree species. Additionally, we observed a decrease in ground layer species richness linked with an increase in canopy cover. In the inter-habitat comparison, this effect was detected both in thickets of ash-leaved maple and in thickets of other tree species. In our intra-habitat comparison, this effect was traced under the canopies of native and alien trees but not under the canopies of A. negundo.
Our rst working hypothesis was that A. negundo can create strong shading. The ash-leaved maple seems to produce a denser canopy than many other woody species. This thesis cannot be strictly proven by the inter-habitat comparison because we purposefully selected plots in pairs of communities An+ and An-with high and equal canopy cover. However, the intra-habitat comparison also clearly shows that the average cover of A. negundo is greater. In intra-habitat comparison, estimates of canopy cover under different types of trees and shrubs were gathered randomly. However, the ash-leaved maple does not form the highest canopy cover of leaves present in this study. There are several native and alien plants that form denser canopies. Therefore, it is more correct to summarise our results as the following: A. negundo creates a canopy of leaves which is equally as dense as the canopy of other alien trees but denser than that of native trees. Our second hypothesis assumed that with an increase in shading level, the number of grass species equally decreases both under A. negundo and under other tree species. In our inter-habitat comparison, an increase in canopy cover was linked with a decrease in plant diversity both under ash-leaved maple shading and that of other trees. However, the intra-habitat comparison illustrated an increase in the canopy cover of A. negundo that was not accompanied by a decrease in plant diversity.
However, on both spatial scales, the number of grass species under crowns of A. negundo was less than under crowns of native trees over the whole range of canopy cover. Consequently, the transformation of the light regime of communities does not explain the decrease in plant richness in A. negundo thickets (Fig. 9). We cannot con rm that our results describe all possible effects of light regime transformation under an ash-leaved maple canopy. Notably, in the Middle Urals, the maximum sun height above the horizon during the summer season is 43-56°. Although the photos shot vertically upward show the canopy cover at the point the camera was located, they do not fully describe the lighting conditions for each plot because they ignore the intensity of side light ux. The difference between the estimates of canopy cover and illumination may be especially large for the intra-habitat comparison, but this difference is smaller for the inter-habitat comparison. One more possible consequence of the transformation of light conditions under the canopy of A. negundo is a speci c change in the spectral composition of light. Fig. 9 A probable mechanism of the generation of reduced diversity of the ground cover throughout canopy shading in communities dominated by Acer negundo (dark-grey circle) in comparison with other trees (white circle); rhombus -average rates Along with a decrease in illumination, the mechanisms of the suppression of ground cover under the in uence of the ash-leaved maple may be different, depending on spatial canopy structure. For example, the fact that A. negundo has a dense and low crown may not only decrease illumination in A. negundo thickets but also blocks the ow of seeds or other plant diaspores. It is known that the diversity and pool of seed banks may decrease in native communities invaded by alien plants (Gioria et al. 2012; Gioria and Osborne 2014), including ash-leaved maple ).
In addition, the ability of alien species to reduce diversity via allelopathic effects is often discussed (Call and Nilsen 2005; Gruntman et al. 2017), and allelopathic activity has been con rmed for the water extracts from A. negundo leaves (Csiszár 2009;Csiszár et al. 2013) and for the soil from its thickets (Yeryomenko 2014). However, in another experiment, the allelopathic activity of the soil from the thickets of A. negundo was not con rmed (Veselkin et al. 2019b). Accompanying the direct allelopathic effects on plants of substances secreted by A. negundo, indirect allelopathy is possible and is a probable effect of A. negundo on native plants through its primary effect on soil microorganisms. In this experiment, though we did not con rm the effect of soil from ash-leaved maple thickets on seed germination, we observed suppression of mycorrhiza in model grasses (Veselkin et al. 2019b). Consequently, direct and indirect allelopathy are possible mechanisms for the in uence of A. negundo on native plants, although their ecological impact requires further evaluation.

Conclusion
We observed decreased ground cover plant species in large (inter-habitat comparison) and small (intrahabitat comparison) vegetation areas dominated by the alien maple Acer negundo compared to areas dominated by native tree and shrub species. At the same time, the dominance of the ash-leaved maple was accompanied by a recorded higher canopy cover of a height of 1-1.2 m. Additionally, in both studies, we established a negative correlation between the canopy cover of trees and the number of vascular plant species of ground cover. Thus, the capture of light and restriction of its amount for other species is a central mechanism that causes several of the negative effects of A. negundo on native communities. Similar to other studies (Lanta et al. 2013), we found that high canopy cover, which produces high shading, is not the only mechanism by which A. negundo realises its potential as a transformer species. Most likely, the ash-leaved maple affects native plants and the structure of communities via additional factors. We believe that further studies are needed to understand which of these mechanisms (in uence on the seed bank; direct allelopathy; in uence on soil microorganisms; transformation of the physicochemical properties of soil) can explain the impact of Acer negundo on the diversity and composition of plant communities.

DATA AVAILABILITY
The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

CODE AVAILABILITY
The script for estimating the canopy cover is given in supplementary material. DVV conceived the ideas, designed the research, and analyzed data. DID collected the eld data. DID and LAP performed processing and prepared data for analysis. DVV wrote the text of the paper. DDV, DID, and LAP corrected and discussed the text.