Methodology of the habitat classification of anthropogenic urban areas in Budapest (Hungary)

The vast majority of floristical and coenological studies focus on the natural vegetation and its remnants in urban environments; however, there is an increasing interest in the systematic exploration of built environments. This survey of the inhabited areas of Budapest launched in 2018 applied a new and detailed methodology. The territorial units of the survey were street sections, where the presence of vascular plants was recorded in ecologically different habitats separately. We laid down the foundations to establish new habitat categories for the habitat classification and mapping of urban ecosystems using a different approach than the Hungarian habitat classification system (ÁNÉR). The method enables us to study the flora of urban ecosystems in detail. We established 18 urban habitat categories, surveyed more than 27,000 survey units and registered 647 spontaneously growing plant species in Budapest. Our survey evinced the spontaneous occurrence of 193 alien species in Budapest, 81 of them are data deficient, while 10 of them are new introductions in Hungary. 53.9% of the aliens are cultivated frequently in Hungary.


Introduction
The invasion of plants is a major problem of recent biology, having serious negative impacts on natural habitats and human health (Early et al. 2016;Mazza et al. 2014;Rumlerová et al. 2016). The invasion of plants is preceded by a period which lasts from their introduction to their naturalisation in appropriate habitats (Blackburn et al. 2011). Urban ecosystems host a higher number of non-native species and can be the starting points of their expansion (Padayachee et al. 2017;von der Lippe et al. 2005). The dispersal and spreading of newly introduced species mostly depend on human activities (Ricotta et al. 2010). The extent of urbanisation nowadays is larger than ever in our history, more than half of the human population lives in cities now and the ratio of urban population is expected to grow even more in the next decades (United Nations 2019). Because of these facts, the role of biologists and ecologists in studying the urban ecology and invasion biology has an increasing significance.
Urban ecosystems differ from natural and semi-natural habitats in many ways. The temperatures are higher (Mohajerani et al. 2017), air (Jorquera et al. 2019;Pascal et al. 2013) and soil pollution (Horváth et al. 2014;Li et al. 2018) are stronger than in the surrounding areas. Extremities and quick changes in the weather of urban areas are frequent, and the microclimate of urban areas is greatly influenced by city planning principles (Bourbia and Boucheriba 2010). Due to varying moisture, light and soil conditions, the surrounding natural environment, the different substrates, the orientation of the streets, the structure of the city, the level of human disturbance and the maintenance methods of the local government the occurring habitats in a city can be diverse (Rebele 1994), but the main determinant of the species composition of a city is the geographical location (Celesti Grapow and Blasi 1998). The habitats of an urban ecosystem show high heterogeneity both vertically and horizontally and change dynamically (Pickett et al. 2017). On account of the attributes mentioned above, supplemented by the cultural, commercial and touristic embeddedness of the given city, the biodiversity and ecosystem services of urban ecosystems can cover a very large scale (McPhearson et al. 2018). The study of the way of entry of alien plants to an urban ecosystem is difficult in the present, as historic data on the early occurrences of a species is necessary when investigating the origin of alien plant populations (Kowarik et al. 2013).
The number of systematic floristic and ecological studies of urban ecosystems increased at the end of the last century and the turn of the millennium, and an enormous amount of data gathered worldwide from urban habitats and numerous publications reached completion since then (e.g. Maslo 2014Maslo , 2015Pyšek 1998).
In Hungary, there are examples for urban floristic research from the nineteenth century, e.g. Simonkai (1877) provides scattered data from the inner city of Budapest, but these researches were not systematic. In the twentieth century, Polgár (e.g. 1914) focused on the adventive flora of Győr (W Hungary), discovering the first occurrences of a high number of alien species in Hungary. The research of urban ecosystems is evolving in Hungary, an increasing number of papers, like the studies on the adventive woody flora of the Buda Castle (Czúcz 2005) and on the fern flora appearing on old built walls (Tamás et al. 2017), have been published in recent decades. There are many protected areas and (semi-) natural habitats in Budapest, and their floristic research has garnered a lot of attention recently (e.g. Nagy and Malatinszky 2019; Tóth and Papp 2012). The flora of Pécs (S Hungary) has been studied for a long period and is still being studied (Wirth et al. 2013(Wirth et al. , 2018).
An easily applicable habitat classification system (ÁNÉR) has been in use in Hungary for a long time (Fekete et al. 1997;Bölöni et al. 2007Bölöni et al. , 2011; however, its application in urban ecosytsems is very limited, as it treats all these areas under one or a few habitat classes. The purpose of ÁNÉR is to provide a habitat classification system for vegetation mapping in Hungary that is up-to-date, nationwide, coherent and covers the whole landscape, but ÁNÉR does not deal with micro-habitats below the usual vegetation mapping scale (1: 5000).
The majority of urban floristic studies focus on the species composition of urban areas and the comparison of urban floras (e.g. Pyšek 1998;Lososová et al. 2011), without considering ecological character, the distribution of habitats and high-resolution mapping. The floristic survey of Mostar (Bosnia-Herzegovina) was carried out in five different regions of the city and the main objective of this study was to enumerate the taxa of the urban ecosystem of Mostar (Maslo 2014(Maslo , 2015. Lundholm and Marlin (2006) analysed the microhabitat preference and habitat origin of spontaneously occurring plant species in a university campus in Halifax (E Canada) in 11 different appointed habitat types. The work of Lososová et al. (2011) focused on the comparison of the urban floras of several Central European cities (including Budapest). These studies had integrated methodology using complex habitat categories and appointed study areas. Salinitro et al. (2018) assess the urban flora of Bologna (Italy) according to the various ecological niches distributed horizontally and vertically. The research of the urban and (semi-)natural flora of Podgorica (Montenegro) (Stešević et al. 2014) was very detailed and resulted in grid cell-based distribution maps of the found taxa and the very detailed typification and analysis of the urban habitats of Podgorica.
This paper presents a study on the urban flora of Budapest. The aim of our work was to (1) gain detailed and reliable reference datasets of the anthropogenic urban flora of Budapest (2) elaborate a method to comparably study the floristic composition of anthropogenic urban (micro-)habitats; (3) determine the habitat preference of the plant species in Budapest; (4) evaluate the taxa according to the latest Hungarian monographs; (5) investigate the way of entry of the species to the urban ecosystem of Budapest.
The hypotheses of our research were: (1) anthropogenic urban (micro-)habitats of Budapest are more diverse than can be described by the known habitat classification systems in Hungary; (2) the anthropogenic urban habitats of Budapest cannot be treated uniformly from an ecological point of view and the species have preferences among them; (3) the studied anthropogenic urban habitats of Budapest host a number of vascular plats not revealed by recent floristic studies in Hungary; (4) the rate of spontaneous occurrences of cultivated alien species (ornamentals, herbs, spices, fruits, vegetables and crops) is high in Budapest.

Materials and methods
The survey of urban habitats was carried out from September to November 2018. During this period, the data collecting lasted 65 days altogether. The study area covered all districts of Budapest ( Fig. 1). Budapest is the capital of Hungary and is located in Central Hungary, between northern latitude 47.613437° and 47.349269°, and eastern longitude 18.926381° and 19.334538°. The altitude of the city varies from about 95 m to about 520 m. Budapest is divided by the river Danube and surrounded by the Buda Mountains and the Pest plains.
The survey was limited to anthropogenic urban areas, natural and semi-natural habitats have been excluded from the research. The habitats of the following ÁNÉR categories (Bölöni et al. 2011) were studied: U1-Cities, areas with blocks of flats; U2-Suburbs and recreation areas; U3-Villages (and village-like suburbs); U4-Yards, wastelands, dumping grounds.
The smallest possible, clearly identifiable units were investigated during the survey. The territorial units were street sections, public and domestic spaces. The applied street sections mean 50-150 m long sections of roads 1 3 between crossing streets, blocks and squares (Fig. 2). Open public places (e.g. squares) are delimited by surrounding streets and blocks, while closed domestic areas (e.g. yards) are delimited by fences or walls (Fig. 2) and both cover an area of 100-1000 m 2 . The survey units (survey plots) were the distinguished habitats in the territorial units.
The survey attended to vascular plants, the nomenclature follows Király (2009), except for the ones that are missing from it, as in those cases we listed the authors. In the case of problematic taxonomic groups, aggregated names were used (e.g. Taraxacum officinale agg., Polygonum aviculare agg., etc.). A complete list of spontaneously occurring species has been recorded in every distinguished habitat within all surveyed units. The seedlings of cultivated plants were considered spontaneous occurrences. Planted individuals were recorded additionally to explore the local origin of spontaneously occurring specimens. In the case of descendants of cultivated species, the nearest occurrence of the planted (supposedly parent) individual was also recorded. During the preparation process, we aimed to distinguish as many (micro-) habitats as possible, after their revision the ones that were undistinguishable ecologically and in terms of use were merged.
The borders of the units, tracks and the occurrences of further noteworthy species (rare, protected or newly introduced species) were recorded with GPS devices. The collected specimens are stored in the herbarium of the Hungarian National History Museum (BP).
The terminology related to alien plants (introduced, casual, naturalised, and invasive) follows the categorisation of Richardson et al. (2000). The information on the status of alien plants in Hungary is gathered from Balogh et al. (2004). Our species list has been compared with the latest Hungarian monographs (Balogh et al. 2004;Király 2009;Bartha et al. 2015; University of Sopron 2018-) to test their frequency and novelty. Data managing, filtering and counting the relative frequency of species in habitats and habitat groups were done using MS Excel.

Habitats
Within the study area, 18 habitats were distinguished ( Table 1). The most frequent habitats were: cracks, road verges, amenity grasslands, planting pits, urban hedges, flower boxes, courtyards and front gardens (Table 2). Based on the availability status and on the prevailing ecological factors and impacts, the habitats were grouped in 3 main groups: habitats of (1) road attached areas; (2) public areas and (3) domestic areas ( Table 1). The habitats of road attached and public areas were usually more disturbed (trampled, mowed etc.) and polluted than the habitats of domestic areas. Walls, urban hedges, cracks and flower boxes occurred in all groups, so these habitats are treated as overlapping subcategories (Table 1).
Within the territory of Budapest, 2241 survey plots were studied, 90% of which were road attached and public areas, while domestic areas were less covered (Table 2). During the study, a total of 27105 plant occurrence records were collected which referred to 647 vascular plant taxa (Table 2). Altogether, the highest number of taxa were found in road verges, cracks and amenity grasslands, but regarding the single survey units, the most species-rich unit was an amenity grassland at Laczkovich street (Budapest District IX), with 96 species, while the average species number was highest in road slopes, ruderals and along railways and lowest in urban hedges (Fig. 3). The abundance of alien plants was highest in urban hedges and lowest in roadslopes, parks, amenity grasslands and road verges (Fig. 4). The number of occurrence records of rare (and partly protected) plants was highest on walls and in road slopes.

Species pool
According to our study, the top ten most frequent species are four alien and six native species. The most frequent species in the urban areas of Budapest are Polygonum aviculare agg., Taraxacum officinale agg. and Setaria viridis (Table 2). Polygonum aviculare agg. colonises all habitats, but is less common in domestic areas (gardens, courtyards and plots) and urban hedges. Taraxacum officinale agg. and Setaria viridis show less habitat preference, therefore they are frequent in almost each habitats (except urban hedges). In general, most weeds prefer permanently disturbed places, like public areas, road attached areas and flower boxes, while trees, shrubs and woody climbers prefer urban hedges and domestic areas, where there is only moderate disturbance. Ferns bear an outstanding preference for walls and courtyards. One of our most remarkable observations was that Thelypteris palustris, a protected fern species in Hungary, occurred relatively abundantly on built walls in Budapest ( Fig. 1) and it was found to be the 7th most frequent plant species on walls with 11 records.
We found 117 species exclusively in road attached habitats, all of them with a single or few known localities there (e.g. Astragalus glycyphyllos, Peucedanum alsaticum). Most of the widespread taxa are characteristic in road attached habitats and cracks, as Eleusine indica (with 689 records and 82.9% of them in road attached habitats and in cracks), Eragrostis minor (with 611 records and 81.9% of them in road attached habitats and in cracks), and Polygonum aviculare agg. (with 1172 records and 78.0% of them along roads and in cracks). On the other hand, 136 taxa avoid road attached habitats and have no known localities there (e.g. Ranunculus repens, Asplenium ruta-muraria).
Taxa preferring public areas have usually less than 20 known localities altogether (e.g. Hieracium pilosella, Thrincia nudicaulis), while some taxa found in public areas have only one single known locality (e.g. Leontodon autumnalis, Mentha longifolia). Taxa of natural habitats-as Asparagus officinalis, Campanula persicifolia, Carpinus betulus, Hieracium sabaudanum, Fragaria vesca, Vicia sepium, Astragalus onobrychis-occur exclusively in or near domestic areas (especially gardens). There are some exclusive subspontaneous occurrences of remnants or descendants of planted ornamentals-as Taxus baccata, Vinca major, Melissa officinalis-in or near gardens, in front gardens, courtyards, parks and flower boxes.
The survey resulted in the spontaneous occurrence of 81 taxa which are missing in Bartha et al. (2015) (Table 3). It means that we know nothing or our knowledge is incomplete on the distribution of these species in Hungary. Twenty-one taxa are missing even in the extended and updated online version of this work (University of Sopron 2018-), as a consequence of the lack of any records of them in the last 15 years in Hungary (Table 3). Twenty-one Mostly dry, grassy slopes on banks, ditches and railway embankments Safety islands (Si) Paved islands in the middle of roads Planting pits (Pp) Pits for planted trees, often fenced off, covered with grid or plastic. These are located in roadside lawns or in any paved urban areas. Island beds (Ib) Maintained, usually fenced off and/or elevated beds located on the street Shrubby roadsides, alleys (Sh) Stands of spontaneously growing shrubs and trees along roads and in urban green areas Public areas Front gardens (Fg) Council owned public areas in front of private houses, mostly maintained by the residents. Amenity grasslands (Ag) Lawns that are not in the immediate vicinity of roads, mostly between the pavement and residential buildings. Parks (Pa) Maintained and cultivated, public, green areas with mowed lawns and ornamental plantings Ruderals (Ru) Areas continuously exposed to disturbance, paved with gravel or any other material, mostly in enclosed areas Domestic areas Gardens (   taxa are completely missing even in the latest Hungarian flora monograph (Király 2009) and further 28 are mentioned there only as a cultivated or ornamental plant (Table 3). Twenty-two taxa are missing from the latest list of neophytes in Hungary (Balogh et al. 2004), and ten taxa are missing in all the latest Hungarian monographs (Table 3); therefore, these are considered here as new introductions in Hungary.

Alien taxa
Altogether 70.2% of the recorded taxa are native and 29.8% (193 species) are aliens. 61.3% of the alien taxa are casuals, 11.1% are naturalised, 21.6% are considered invasive in the natural and semi-natural habitats of Hungary and 6% of them are new introductions in Hungary (e.g. Artemisia verlotiorum) (Fig. 5). One hundred four (53.9%) of the alien species are cultivated or their fruits or seeds are sold and eaten in Budapest. Eighty-six of them are planted frequently as ornamentals (e.g. Brunnera macrophylla, Euonymus japonicus), while the rest are herbs (e.g. Salvia officinalis, Calendula officinalis), fruits (e.g. Armeniaca vulgaris, Persica vulgaris), vegetables (e.g. Solanum lycopersicon, Cucumis sativus) and crops (e.g. Zea mays, Triticum aestivum) that escaped from cultivation or derived from households. In the case of 38 species, the potential parent plants were found within the same survey units 0.5-60 m away from the escapees (Table 4), while in other cases no planted parent plants could be identified there, only much further away. The range of the dispersal of cultivated plants is usually not more than 10 m.
Considering the ten newly introduced taxa (Table 3), eight of them are garden escapees, which are as follows: Begonia cf. cucullata Willd, Celosia argentea L. [mentioned in Király (2009)

Discussion
Using small sampling units and ecologically well-distinguished habitats, it was possible to collect over 27,000 habitat-specific occurrence data in a short period of time. The majority of the found species occur at different frequencies in the studied habitats. Some recent studies also suggested that the use of small, ecologically well-defined habitat categories is important to monitor the distribution and habitat preference of plants and the selectiveness of habitats (Lundholm and Marlin 2006;Salinitro et al. 2018;Steševic et al. 2014). The purpose of the applied habitat classification, which is based on the ecological differences of the (micro-)habitats studied, is remarkably different from that of the only system used in the region (ÁNÉR, Bölöni et al. 2011), which is used primarily for site mapping, while our method is developed for a more detailed classification on ecological basis. This study managed to include domestic areas unevenly, which, according to Wirth et al. (2018) can significantly influence the species list in a grid-based, fullcoverage study. In this case, however, the focus is on the habitats and the habitat preference of species, so it has no significant importance. The low abundance of records from domestic areas is the clear consequence of their restricted availability. The majority of the species found in Budapest are connected to exact habitats or habitat groups, while The numbers in the 'Rank' columns indicate that if the plants are sorted by frequency, what would be the given plant in the row; '-' indicates that the plant is absent from that habitat one sixth of them occur sporadically and have one or few records, which support the connections of hosting habitats with species-rich ecosystems (Lundholm and Marlin 2006;Tikka et al. 2001).
The high average number of species in road attached habitats in Budapest confirms the role of roads as corridors, which support the large-scale introduction of alien species (Ascensão and Capinha 2007) and the introduction of natural grassland species at the same time (Tikka et al. 2001). The relative species richness found in unpaved road attached habitats (amenity grasslands and road verges) coincides with the results of Salinitro et al. (2018), who found that habitats with (semi-)natural soils are the most species-rich habitats in the city of Bologna.
The taxa that avoid road attached habitats, occurring mainly in domestic areas, parks and flower boxes in    Lundholm and Marlin (2006), who found that species of wet natural habitats are less abundant in paved roadsides and more abundant in planted areas.
The high rate of species of natural habitats, especially ferns (Tamás et al. 2017) on built walls is a consequence of the constancy of the ecological conditions experienced, and the appearance of species of wet habitats is a consequence of the specific moisture conditions of the walls (Segal 1969). The frequent presence of the rare and protected marsh fern (Thelypteris palustris) is owing to the presence of abundant natural populations in the adjacent areas along the river Danube (Reszler 1997).
The ten species we found for the first time in Hungary and the 81 we confirmed shows the importance of recent urban floristic research and monitoring. Eight of the new plants are cultivation escapees, which, according to our results are in the early stages of their naturalisation and have a very small distribution area in Hungary so far. The high number of cultivated species occurring spontaneously in Budapest is because of the overall presence of gardening in the city, and more than half of the alien taxa are also cultivated in the whole country. Further monitoring of these taxa is very important in the future, as tracking the spread of these species, can yield an enormous amount of information. Though the introduction of alien species is studied worldwide (Hulme et al. 2008;Kowarik et al. 2013), the related terminology is developed for national and larger levels where the way of entry of aliens is relatively well definable and traceable. For cities, it is more difficult to determine whether individuals of a particular species have escaped directly from cultivation or from an already established population, so categories and direct origins are not always traceable (Salinitro et al. 2018). Furthermore, some native species are locally introduced in Budapest as well (e.g. Sedum hispanicum, Amygdalus nana). The newly recognised Artemisia verlotiorum is naturalised in most European countries (Mosyakin et al. 2019), but was missing from the Hungarian monographs. Due to its resemblance to Artemisia vulgaris, further field investigations and herbarium revisions are needed to determine the distribution of the plant in Hungary. Eragrostis virescens is also missing from Hungarian monographs, but Király et al. (2019) found that the plant has been sporadically present in Hungary ever since 1929 and was misidentified as Eragrostis parviflora (R. Br.) Trin.
Comparing the total flora, much higher values were found for other Central European capitals (Wien: 1476 species, 51% of them are aliens, Berlin: 1418 species, 41% of them are aliens and Warszawa: 1109 species, 31% of them are aliens) (Pyšek 1998). The number of the species (647) and the ratio of aliens (29.8%) in Budapest are small compared to these numbers. Pyšek (1998) also found that the number of species correlates with city size and the density of human population. The reasons for the low number of species are that our research was carried out only in autumn, so the plants of spring and summer aspects were not recorded and our research is restricted to anthropogenic habitats of Budapest. Because of this, the 647 plant species found in Budapest are not to be considered small.

Conclusion for Future Biology
Biological invasions are one of the greatest threats to natural habitats, biodiversity and human health. Halting or mitigating the expansion of invasive plants is the common responsibility of researchers (especially in the field of biology), decision makers, horticultural, economic and industrial actors. On the part of biologists and ecologists this means that future researchers should pay more attention to the understanding of invasion biology. One of the hotspots of biological invasions are urban areas, so one of the main objectives of future biology is to concentrate on the systematic research of urban (and other anthropogenic) habitats to gain a better understanding of invasions from the starting point to the unstoppable expansion of invasive species. Our research serves this need by starting the systematic research of a Central European capital, Budapest, with integrated methods. The development of habitat categories of secondary habitats (such as urban areas) in habitat classification systems is essential if researchers wish to understand the species selection traits of secondary habitats and the preferences of species of secondary habitats. We hope our research contributes to the understanding of the habitat preferences of invasive plants, so the scientific community can provide solutions to one of the greatest challenges of humanity in the near future. We would also like to point out that this research does not only focus on adventive or invasive plants, as it also raises our knowledge on the preferences of many members of the native flora, even protected plant species, so that we can evaluate the role of urban habitats in nature conservation. Finally, we would like to emphasize that we need a better understanding on secondary habitats and the species occurring in them, the surveys should be done on the smaller scale as possible. The secondary habitats are and will be very important in future biology, because of the decrease of (semi-)natural habitats.
Baráth for providing us with valuable advice by revising the manuscript. We would also like to thank Tamás Rigó for the linguistic revision of the manuscript.
Authors' contributions ZB designed the protocol of the research. AR and ZB gathered the data on field, analysed the data, drafted the article and created figures and tables. AR compiled the literature cited in the article.

Data accessibility
The data of our research is still broadening, and therefore not available for the public. Not every data is included in this paper.

Conflict of interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.