Digging deeper in urban ecology: the urban ecosystem convergence hypothesis

By Dietrich Epp Schmidt, Graduate Reserach Assistant, University of Maryland

“Even the mightiest of us return to dust, they say. Nothing remains but these shattered fragments of their kingdom… But that's not really the point, is it? These shattered fragments remain- that's the point. We look upon the magnificent temples and stelae and ball courts of Caracol in awe. There's no despair here. The Maya built something astounding and permanent. Look on our works, ye mighty, and revere. The ancient Maya speak to the twenty-first century through those temples and say: We did something amazing here.

What will our descendants think when they come upon Chalillo [dam]? When they scrape away the deep layer of dirt covering its stepping-stone facade, what will they make of the dogleg design, the Chinese gauges, the long-stopped turbines? What will they make of the skeletons and fossils of birds long gone? Will they connect the two?"


-Bruce Barcott, The last flight of the Scarlet Macaw


For a very long time, humans have experienced the world as having two fundamental domains: that which we understand as being under human control, and that which is not. Canonically, we call the latter “natural” and the former “unnatural;” and while the broader society may perceive there to be a distinction between the two, we know that our existence has an impact on ecological process, and vice versa. Our social processes, which are as basic to us as defining our identity and competing for status within our community, are intricately linked to our individual and global consumption of resources. Our economic trade is itself an ecological process that transports resources and organisms across the globe [see telecoupling]; our civilization alters community composition and function wherever we can make a living, all the while affecting global biogeochemical cycles. Our framework for understanding the role of social process on ecosystem function is just in its infancy, and there is no vernacular language for describing the built environment as an ecosystem. This is not how our society understands the environments we inhabit. One goal of the Global Urban Soil Ecology and Education Network (GLUSEEN) is to increase the exposure of urban citizens to the important role that soils play in maintaining ecosystem health.

Within the discipline of ecology, there exist frameworks to describe the outcomes of human behavior in terms of ecosystem process. For instance, biotic homogenization (BH) describes a process of convergence among biotic communities; generally communities become more similar (converge) when endemic specialist species are extirpated and generalist species come to dominate. Convergence is a process that is often applied to understanding the effect of both urbanization and agriculture, where the implied (or assumed) mechanisms are generally anthropogenic disturbance and/or facilitated dispersal. In this instance, BH describes how land-use conversion (habitat loss) drives local extinctions, while the cultivation of exotic species facilitates the dispersal of a common set of organisms. In this context, BH helps to explain the paradox of high urban and peri-urban biodiversity concomitant with significant global biodiversity loss. Endemic species go extinct, while opportunists thrive in the human-disturbed landscape (see, for example, this). To bring the mechanisms into focus, BH has been reformulated somewhat as the Urban Ecosystem Convergence Hypothesis, which relates structural changes in the built environment to changes in community process. It predicts that if urban landscapes are constructed and maintained in a similar manner (causing a convergence of habitat characteristics across biomes), then their biotic community and ecosystem processes should converge as well. For example, in their paper entitled “Ecological homogenization of urban USA,” Groffman et al. show that the practice of maintaining irrigated lawns causes a convergence of biophysical conditions across biomes within urban areas in the United States; in temperate forest systems, land-use conversion to lawns increased surface temperatures by reducing shade and evaporative cooling that normally occurs in the canopy. Whereas irrigating arid land for lawns increased evaporative cooling at the ground level, causing the two environments to converge with respect to temperature as well as humidity.

Image recreated from Pouyat et al., 2017
Journal of Urban Ecology.

The Global Urban Soil Ecology and Education Network (GLUSEEN) applied the Urban Ecosystem Convergence Hypothesis to urban soils. We sampled from soils that occurred within four different land-uses, which were categorized to represent land-use histories that are typical of urbanized landscapes (published here). These four categories were reference, remnant, turf and ruderal. Reference sites served as our control; they were sites located outside of the urban matrix, which are representative of the historic state of the ecosystem and are being managed to mitigate human impact. Many reference sites were areas set aside for habitat conservation. Remnant sites are similar to reference sites in community structure but occur within the urban matrix, and thus exposed to urban environmental factors. Turf sites were defined as sites under management to maintain a turf-grass system, which include municipal, residential, or park lawns. Ruderal sites were defined as sites that have experienced recent and substantial disruption of the soil profile, and typically were areas with a history of demolition or construction activity. Using these land-use and cover categorizations, we asked whether specific types of land-use and cover (both largely an outcome of cultural processes) caused physicochemical properties and microbial communities of soils to converge; and whether these changes result in a convergence of function among these soils.

Within-group variance of edaphic factors, among land-use;
2a shows the convergence of soil pH, OC, and N
under turf and ruderal land-use relative to the reference;
2b shows the divergence of P and K under turf
and ruderal land-use relative to the reference.
Recreated from: Pouyat et al., 2015.

As an assessment of the soil habitat characteristics, and to test the first question, we measured edaphic features such as phosphorus (P), nitrogen (N), and potassium (K) availability, as well as other characteristics such as organic carbon (OC) and pH. To quantify and identify the archaeal, bacterial, and fungal community, and to test the second question, we used quantitative PCR and amplicon sequencing. And finally, as a test of soil function, we conducted a decomposition experiment using tea bags in each of the study sites (see here). First, we found that in fact some physicochemical properties of soils converged under turf and ruderal land-use and cover types. Soil pH, OC, and N in particular converged. However, not all characteristics converged as K and P actually diverged under urban land-use (Figure 1).  We believe that it’s likely that cultural differences in how fertilizers are formulated (N vs N:P vs N:P:K fertilizers) and the variability in their rates of application may explain the increased variability among P and K nutrients; while N is also enriched systematically by fossil fuel combustion that leads to consistent atmospheric deposition of N in urban areas (and thus convergence). The convergence of soil pH is likely related to the widespread use of concrete, and the resulting concrete dust in urban areas; the calcium oxides and carbonates found in concrete effectively act as a liming agent as they dissolve, buffering soil pH towards a more alkaline condition. And finally, while specific land-use and cover types might have differing effects on the soil OC concentration, the effects within each land-use are consistent; disturbances often result in lower OC, while irrigation and fertilization in the absence of disturbance may actually increase carbon storage in soils. Thus, cultural factors may drive convergence of some habitat characteristics while causing other habitat characteristics to diverge.

Within-group variance of the fungal, bacterial, and archaeal
communities; the fungi converge in ruderal sites
relative to reference,the bacteria do not converge,
and the archaea converge in the turf and ruderal sites.
Recreated from Epp Schmidt et al., 2017.

Our next question was whether communities of organisms living in the soil converge under similar land-use and cover. We found that of the three phylogenetic domains making up the soil microbial community, the archaea and fungi exhibited a marked convergence, while bacteria did not (published here). We also showed that this convergence may be driven by different ecological factors. For example, we found that convergence in the fungal community was largely due to the loss of ectomycorrhizal fungi (ECM), while the convergence of archaeal communities was due to the increased abundance of ammonia oxidizing archaea. ECM function as symbionts with woody plants, and thus are highly reliant on the abundance of their host species. When land is converted from forest to any non-forested urban land-use, it appears that there is no longer viable habitat for most of these species. Archaea, on the other hand, actually increased in overall abundance under lawn use, and their community became dominated by organisms that derive energy from the oxidation of reduced nitrogen species. This means that, unlike for fungi, the N enrichment of urban environments is favorable to these members of the archaeal community. Moreover, since absolute abundance increased and richness also increased, it appears to be the case that the metabolic differentiation among archaea allowed convergence to occur without competitive exclusion. Therefore our dataset demonstrates the two mechanisms by which convergence might happen; a loss of unique species, or an increased dominance of just a few species that can be found in all sites. It is of course possible that both mechanisms operate in tandem to cause convergence.

The urban landscape represents the zenith of human development; it is the cultural hub of our civilization and the control center for our social process. It is also the area with the highest land-use intensity, and therefore is the most significantly disrupted ecosystem. With our dataset, we are able to show some effects that human culture has on ecosystems; the cultural process that drives the globalization of our economy and the homogenization of our global culture also has global impacts on ecosystem process via the local decisions that land managers make. Our data demonstrates that human culture may cause either a convergence (soil pH, N and OC) or divergence (P and K) of soil habitat characteristics, and that microbial communities may (fungi, and archaeal) or may not (bacteria) converge as a result. We were also able to determine multiple mechanisms that drive convergence. The fungi likely converged due to our impacts on cover (reducing the abundance of host species), while the archaeal likely converged because of our N enrichment of the urban landscape (enhancing the fitness of organisms that rely on certain forms of N metabolism). Thus there are specific interactions between human alteration of the landscape and biotic community response. The impacts of urbanization on the function and makeup of the biotic community may be long lasting, and indeed, might even outlive civilization itself. In 2008, when he published his book The last flight of the scarlet macaw, Bruce Barcott could scarcely have known that within a decade scientists studying the Mesoamerican landscape would be using the color of tree leaves (by reflecting laser light off of them from high altitudes) to discover the location of lost ancient Mayan structures. It is remarkable that urban centers, constructed and abandoned nearly a millennium ago, can still be discovered using their legacy impact on the biotic community.


Read the full manuscript here:

Epp Schmidt, D. et al. 2017. Urbanization erodes ectomycorrhizal fungal diversity and may cause microbial communities to converge. Nature Ecology & Evolution doi:10.1038/s41559-017-0123