Urbanisation and its effects on species: the ants of Lyon as a case in point.
Bernard Kaufmann from the Artificial and Natural Hydrosystems Ecology Lab at the University of Lyon presented his take on how cities and their attendant processes and structures affect wildlife, using his studies of ant ecology in Lyon and surrounds to make his case, the main ideas and findings of which we summarise here.
Defining urban areas and measuring urbanisation.
When studying the effects of urbanisation, one must first begin with a working definition of an urban area; of course, no definition is or ever will be universally applicable.
Briefly, urban areas are often defined by administrative status, or by a threshold population density of inhabitants with differing threshold densities for a site to qualify as urban.
In general, urban area landuse patterns present a gradient of density of buildings, and of densities and sizes of agglomerations of buildings, outwards from a central high density location. More simply, a city has fewer building and artificial structures towards its outskirts.
On the other hand, by a definition from a landuse point of view, an urban area has 60% or greater of its land surface covered by artificial or impermeable material, such as asphalt on roads, or concrete for infrastructure.
A novel way of quantifying the level of human presence is to examine urban lighting: night photos of the land can accurately measure city sizes. This approach is easier, but might not always be applicable, since some very low population density sites are very brightly lit (such as oil fields due to natural gas burning) and some high density cities don’t light facilities such as roads.
This makes urbanisation a fairly easily measured process, especially when combined with historical data and modern satellite imagery. As with Lyon, many parts of the world have seen massive increases in urban areas in the second half of the 20th century. A gradient usually exists, both of the level of urbanisation as well as the size of structures: there’s usually a decreasing percentage of large buildings when starting from the centre and moving outwards.
This pattern differs depending on the region, with the US model of city centre < suburbs < rural areas not particularly applicable to Europe or Asia. Cities, especially older ones, have been designed, and later grow, depending on geography, and might be modelled as concentric (also nuclear, or US model), polycentric (also multipolar, as with twin cities), or as constrained entities (also linear, often following natural features such as rivers or relief). Cities may have two or more models applicable to them at once, or none at all. Lyon, in the south-centre of France, was both constrained by low hills to the west and the Rhône and Saône rivers to the east. Then again, modern Lyon has managed to cross the river and resembles more a nuclear settlement, with the old town at its centre.
Scale is another factor that must be taken into account when studying urban ecology, since heterogeneity in the landscape only becomes evident when the area is studied at a finer resolution, so Lyon might appear at one scale to be a homogeneous urban block, while a finer examination might reveal considerable differences in the type of landuse.
The effects of urbanisation.
The presence of settled humans in high densities quite predictably disrupts local ecology, beginning with the fragmentation of existing landscapes, and continuing with their increased artificialisation. Since cities typically draw on their hinterlands for food and water, logistical infrastructure, such as roads, airports and surrounding agriculture can transform by fragmentation or destruction pre-existing landscapes quite far from cities where definitions of the urban area don’t usually apply.
Within cities proper, the high concentrations of artificial ground surfaces and above-ground structures cause increased temperatures, something known as the urban heat-island effect, which is most pronounced in summer. Pollutants too are more heavily concentrated in cities and in the water tables of urban and peri-urban areas. Cities provide novel stressors for both humans and animals in the form of light, air and noise pollution. For wildlife, merely the constant presence of humans might easily constitute a major disturbance. Soil structure is also often different: cities tend to be built on impermeable ground up to 4-5m deep resulting from either the selection of a site with underlying bedrock or from the use of rubble as filling material. Water sources tend to be concentrated and natural aquifers often provide most of a city’s drinking water since nearby surface water-bodies are often too polluted by discharge from the city itself; keeping water sources clean is a major challenge for urban areas today.
Underground water sources may also change in temperature which might result in novel pollutant dynamics; however, some areas in and around cities also clean water and return it to the water table.
Urbanisation and wildlife.
For wildlife, all resources in cities are different from those in natural areas, with either a lot of or very little food around. Environmental seasonality is highly reduced, since humans produce a lot of waste that’s edible. Pollinators, for example, have access to more, and more diverse resources: as an example, linden trees, Tilia sp. impose non-natural constraints on bees, whose population dynamics in cities are driven by their flowering.
The modified climates of cities allow tropical and subtropical species to thrive in cities because temperatures more closely match their native ones, and the urban heat island effect often allows more vulnerable species to survive in cities in winter. The excess of sounds so recognisable in cities particularly affects birds and amphibians which communicate primarily through vocalisation.
To generalise then, urban landscapes impose restrictions or filters on which species can and cannot inhabit cities. If at the regional scale a certain species pool exists, urban areas might select for certain life-history traits. Forest or cliff birds, like pigeons and some falcons manage city life quite well, while birds of open fields, like bustards or galliformes, cannot.
Birds, especially, are the best studied taxa in cities, because many birdwatchers live in cities worldwide, creating a large and competent cadre of citizen-scientists who regularly and accurately gather data. However, in decreasing order, plants, beetles, and mammals are also well studied. With birds, the pattern is that synanthropic, or human commensal species like blackbirds and pigeons are more abundant in urban areas, and while species diversity decreases with urbanisation, the abundance of individuals of those species that do exist is much higher, since resources are more abundant too. Since natural predators are often filtered out, there is rarely any control on these city adapted birds. Human commensal mammals such as cats, dogs and rats also have some effect on species composition in cities, but not a great one, with disturbance by commensals disproportionately affecting ground nesters when compared to tree nesters. While non-volant mammals show uniform drops in diversity with increased urbanisation, bats are more like birds in how cities affect them.
Studying urbanisation and its effects on spatial distribution.
How then, can we better explain species occurrences in cities: which species tolerate, avoid and exploit cities, and which of these are invasive? By studying an assemblage of Lasius and Tetramorium ants, including the invasive L. neglectus, Kaufmann and others aimed to do just that. They chose ants to study the effects of urbanisation because ants are ubiquitous, especially in open areas, and respond quickly to environmental perturbations. Sampling in Lyon, at 1248 points, they searched for every formicary of the two genera. Sampling density was affected by accessibility, with each sampled site never more than 50m from a road, and always within 500m of another site in the outskirts, and within 200m in the city of Lyon itself.
Data were gathered for various factors including land cover and urban heat-island existence, among others. The resolution of the data varied depending on the source used, from satellite images of 1pixel = 1m2, to 30m pixels. Urban history was calculated as change in NDVI (Normalized Difference Vegetation Index) between 2013 and 1986. Buffer distances around each point varied: landscape fragmentation was calculated in a 500m buffer, for example. Human activity was measured as the distance to the nearest road, the density of secondary roads and the distance to the nearest embankment. Climate and altitude were also measured. A PCA (principal component analysis) resulted in fewer factors.
In Lyon, the ant assemblage comprised 7 species of Lasius, with L. niger the most common being present in 74% of samples, and 4 Tetramorium species, with T. sp E the most common, present in 49% of samples. T. sp U2 was found to avoid urban, built up Lyon. T. sp E inhabits urban areas, but not U2. L. neglectus proved to be a generalist, inhabiting all areas. L. niger was also ubiquitous, as were some arboreal Lasius species which profited from the presence of trees. An Outlying Mean Index method PCA with the origin as the mean environmental condition and with Axis 1 representing fragmentation and first land cover and Axis 2 the urban history reinforced the cosmopolitan nature of L. niger, which was close to the origin, but T. sp E was found to prefer warm urban areas.
Finally, urbanisation was found to be neither a uniform process, nor were invasives found to be favoured by all urban processes. L. niger and L. neglectus for example, are capable of both competition or coexistence in different niches.
Some interacting factors exist, like higher temperature and embankments, which favour L. neglectus, but cold embankments do not L. paralienus also prefers cold banks, while T. sp U2’s distribution follows the urban heat island effect, while land cover and landuse history are not as important. Invasives are not really favoured by cities, but classification based on the concept of urban filters is possible. T. sp E thrives in cities, while T. sp U2 doesn’t, which goes to show that disentangling the various factors that together make up the process of urbanisation is necessary to understand how ants and other taxa are affected.
Studying urbanisation’s effects on community composition.
In order to understand in depth how ant communities were affected by cities, Kaufmann and others actively searched for formicaries, using survey methods that resulted in high detection probabilities and made density measures possible. Landuse was measured as the only abiotic factor, with the others being biological traits of each species. Four major types of areas – urban, peri-urban, rural and agricultural – were sampled. The type of site was always a permanently vegetated, mostly grassy meadow.
The Shannon-Weaver diversity index was calculated for each type of site, with urban and rural areas very different in diversity. Nest density was actually higher in urban, periurban and rural areas, with ants much less abundant in agrosystems. The Lyon ant community was biased towards L. myops, Solenopsis fugax and L. niger. This result was obtained because S. fugax and other highly abundant species are deep nesters and were not detected in the previous study, which managed to find mostly epigeous nests and species. In agricultural areas, though, the dominance of deep-nesters was broken by L. niger. A multivariate plot showed that there really is a species gradient which follows the level of urbanisation, with the hyperabundant S. fugax and L. myops preferring urban areas.
Interestingly, species with smaller workers were found to be more abundant in forests, but queen body size was not important in determining distribution, though species preferring warm climates were more common in cities, while those preferring dry habitats were found farther from forests and wooded areas. Finally, it was shown that the suburban area doesn’t restrict ant diversity, but the agricultural area does not favour high colony densities. Since the type of agriculture is based on frequent deep-tillage of large fields, many subterranean ants are likely filtered out of the area by this.
Urbanisation and gene flow.
Kaufmann and others then asked whether Lyon’s urban centre was affecting gene flow in the landscape, to wit, whether Lyon’s highway was a barrier to gene flow, and whether hybridisation occurred between some of the more common species. They also asked if intraspecifically, a distance, barrier or filter/resistance effect could be seen on gene flow.
Using the gene for Cytochrome oxidase 1, and comparing between T. sp E and T. sp U2 from 778 samples of which 453 were E and 325 U2, they attempted to answer their question using 17 microsatellite markers.
They found that nearly 16% of the ants sampled were hybrids and in the peri-urban zones, hybridisation is of the introgressive type, which means that hybrids are backcrossing with the two species, indicating some level of hybrid fertility. This might mean that hybridisation is advantageous, and hybrid colonies might spread. Since the queens of both species mate with multiple males, they could be mating with males from both species to produce a colony that’s adapted to nearly any condition in terms of urbanisation: a hybrid species could also produce males of either of the two original species, so it wouldn’t be selected against in any kind of environment. Intraspecifically, there was significant isolation by distance, but not by the highway. Resistance was shown to be important, with open areas oddly resistant, especially for T. sp E, which found it easier to move in urban areas, this species having been noted previously as an urban specialist, T. sp E was found differentiated into a northern and a southern group, but it was not clear whether this was due to genetic isolation by distance or whether Lyon was on the boundary of two hitherto separate populations.
T. sp U2 was found to treat the landscape as a fairly homogeneous one, and it appears that its nuptial flights are long and fast; while how long or how fast are not known, the dispersal ability and distance is fairly high.
To conclude, Kaufmann and others determined that cities offer novel habitats that might promote the existence and persistence of hybrids, to the extent that speciation becomes possible. They now propose to test whether urban areas actually promote the existence of hybrids. The idea is to study 20 cities for the same effect, across a 400km gradient in France, and with a 4℃ temperature change between the most geographically distant points, and to subsequently survey nests to check the percentage of hybrids in each nest along this distance.
Pratik Gupte and Kevin Moull, IMAE 2015-2017