Challenge 1: Methodological approaches
General Application — Studies of adaptation have historically relied on a mix of observational and experimental methods. Adaptation research often focuses on divergent habitats, although clines across environmental transitions have also been instrumental in studying local adaptation (Endler 1986, Kawecki and Ebert 2004, Hereford 2009). Yet it can be difficult to define the boundaries of habitats and populations in heterogeneous landscapes (see Challenge 4) and environmental variation may present as mosaics rather than gradients. Certain organisms may be more tractable for the quantification of natural selection because of their reproductive cycle, demography, generation time, and geography, which may bias the organisms we choose to study. In particular, approaches requiring movement of organisms between habitats, such as reciprocal transplantation, are not feasible for all organisms, can be prohibitively expensive and time consuming, may require unattainably large numbers of replicates to obtain sufficient statistical power, can facilitate the spread of diseases and parasites, and may be impossible for ethical or legal reasons (Cunningham 1996, Kawecki and Ebert 2004, Blanquart et al. 2013, Johnson et al. 2021). Common approaches for adaptation research, such as mark recapture and long-term monitoring, which have been crucial in disentangling the temporal dynamics of adaptive evolution (e.g., Grant and Grant 2014), may be compromised by external factors such as natural disasters and logistics of carrying out such projects (e.g., funding and researcher continuity). Genomic approaches to identifying local adaptation are becoming increasingly common and may be valuable complements to field research methods, yet genomic approaches come with their own methodological limitations as well (Hoban et al. 2016, Perrier et al. 2020). Lastly, interpersonal interactions between researchers and local community members in any environment can be friendly and educational — offering opportunities for broader impacts of research activities — but can also pose safety risks for researchers (Demery and Pipkin 2021).
Human Element — Some methods that may be relatively easy to employ in non-urban settings may be untenable in urban environments (or vice versa). Urban adaptation can be influenced by factors related to greater human activity that are difficult to control using traditional manipulative experiments or may be difficult to predict. Direct and indirect human interactions with wildlife can shape behavioral responses and adaptations (e.g., pedestrian behavior, Bateman and Fleming 2014) and human activities can drastically transform urban environments even on short timescales (see Challenge 3). Rapid or unanticipated anthropogenic modifications in cities limit the establishment and success of studies that involve repeated sampling and long-term monitoring (McPhearson et al. 2016). The mosaic of private and public lands in urban environments intersecting with human and wildlife activity adds additional complexity to the methods that can be employed to conduct urban adaptation research. For example, mark-recapture methods to estimate selection gradients can be challenging because marked individuals can move into inaccessible anthropogenic spaces that dominate urban landscapes, such as restricted-access private property (e.g., backyards or inside homes). Similarly, a random sample of the environment for population genomic analyses could be hampered by private property access in non-random ways across the urban landscape. Some methods may unintentionally facilitate human-wildlife conflict (Treves et al. 2006, Kansky et al. 2016, Schell et al. 2020), disease transmission between urban wildlife and domesticated animals and humans (Bradley and Altizer 2007, Brearley et al. 2013), and biological invasion (Hufbauer et al. 2012, Borden and Flory 2021). Community members tend to be more concerned and vocal about these potential threats when they occur near their homes (Dickman et al. 2014, Drake et al. 2020). Additionally, urban areas are characterized by a higher density of human presence, which increases interactions between researchers and the public and law enforcement, both positive and negative, and can be problematic when urban sites are repeatedly accessed (Dyson et al. 2019, Des Roches et al. 2021).
Misconceptions — A misconception perpetuated by methodological challenges to urban adaptation research is that only specific approaches, such as reciprocal transplants, provide strong support for local adaptation (e.g., Donihue and Lambert 2015, Lambert et al. 2021, Diamond et al. 2022). Although common garden and reciprocal transplant studies are informative for evaluating evidence of local adaptation in some taxa, such as invertebrates or plants (Yakub & Tiffin 2017, Gorton et al. 2018, Chick et al. 2020, Tüzün and Stoks 2020, Yilmaz et al. 2020, Diamond et al. 2022), they are not informative or feasible for all taxa. Advocating broadly for “gold standard” methods might lead to an overrepresentation in urban adaptation research of organisms, microhabitats, or geographic regions most amenable to these approaches. Extrapolating findings based on a restricted set of methods or taxa could lead to incorrect conclusions regarding the generalizability and prevalence of urban adaptive responses.
Moving Forward — To address the methodological challenges associated with human presence and activity in urban landscapes, research efforts that employ complementary and innovative methods will provide different pieces of the adaptation puzzle (Figure 1). As in non-urban environments, multifaceted approaches will be most robust for detecting and characterizing local adaptation (Kawecki and Ebert 2004, Barrett and Hoekstra 2011). As a result of increased human interactions in urban areas, collaborations among diverse disciplines can become more commonplace and bring new technology and methodology into urban adaptation research (McPhearson et al. 2016). Interdisciplinary approaches may be particularly valuable in urban ecosystems, where both empirical and applied science involve human activities and have the potential to promote human well-being (McPhearson et al. 2016). Inclusion of local communities in urban and non-urban systems alike can improve the success of methodological approaches via the incorporation of local knowledge (Uprety et al. 2012, Camacho et al. 2021) and will help improve researcher outcomes in terms of safety, access, and study continuity (e.g., continued or repeated access, reduced vandalism). There are several examples where integrated approaches have been used to build a more comprehensive picture of urban adaptation: research onAnolis lizards has incorporated behavioral, phenotypic, experimental, and genomic analyses to understand adaptation to thermal and structural habitats (Winchell et al. 2016, Aviles-Rodriguez et al. 2019, Campbell-Staton et al. 2020); work on white clover (Trifolium repens ) has involved the global community in sampling efforts complemented with experimental, phenotypic, and whole genome sequencing analyses to test for parallelism (Thompson et al. 2016, Santangelo et al. 2020b, 2022); research on Galapagos finches (Geospiza spp.) has employed morphometrics and behavioral approaches to understand how access to human foods alter historical patterns of diet-based selection on beak shape (De León et al. 2011, 2018); and a combination of reciprocal transplants, phenotypic variation, and mate choice experiments in Tungara frogs (Engystomops pustulosus ) has revealed adaptive sexual selection (Halfwerk et al. 2019).