Choices of sampling method bias functional components estimation and ability to discriminate assembly mechanisms

Abstract

The understanding and prediction of species distributions have been advanced by the development of community assembly theories and functional trait‐based approaches. Coupled with null models, trait dispersion patterns are commonly used to gauge the relative importance of niche versus neutral‐based processes on shaping local communities. However, asymmetrical trait dispersion patterns of identical communities may artificially arise as a result of sampling methods choice. This non‐random filtering on traits is overlooked, and it may bias the estimation of functional components as well as the detection of underlying assembly mechanisms—especially for species‐rich arthropod communities. We sampled subtropical ground‐dwelling ant communities using a paired treatment design comprising pitfall traps and Winkler extractors (Winklers), two commonly used sampling techniques for terrestrial arthropods, in 17 shrublands and 33 secondary forest 400 m2 plots in Hong Kong to determine the effect of sampling filters on functional composition (community means of six morphological traits) and functional diversity (range and variance of individual traits; size and filling properties of multidimensional trait space) using a null model approach for each habitat. We found, as hypothesized, that sampling filters from different methods affected the trait composition and diversity of ant communities asymmetrically. In shrublands and forests, the trait compositions of communities sampled by pitfall traps were dominated by larger sized, slender‐shaped and long‐legged ants as compared to those sampled by Winklers. The latter method also underestimated the diversity of individual traits related to body size as well as the size and filling properties of multidimensional trait space. Our results highlight that the usage of particular sampling methods without prior knowledge on their potential filtering effects on traits can affect the detection of assembly processes, and interpretation of functional proxies, such as body size. Additional care should be taken when comparing trait components obtained from studies using different sampling methods, so as to distinguish the sampling artefacts from actual ecological phenomena. Hence, it is important to consider the potential biases that different sampling methods may introduce to functional trait‐based research.

Trait‐modulated decline of carabid beetle occurrence along elevational gradients across the European Alps

“…The choice of survey method can bias against particular traits, for example, typically pitfall traps over‐represent larger species of arthropod as they tend to be more active and hence more likely to be trapped (Hancock & Legg ; Lee & Guénard, ). Similarly, temperature can affect activity and hence pitfall trap efficiency, leading to a positive correlation between ambient temperature and carabid abundance (Saska et al, ).…”

Finding missing links in interaction networks

“…Ecological networks are sparse and hence any unobserved interactions could be either "missing links" (interactions that would be observed with complete sampling), or true negatives. Differences or deficiencies in sampling can lead to divergent conclusions about underlying mechanisms (Lee and Gu enard 2019). Furthermore, under-sampling will skew the apparent prevalence of interactions toward frequent interactions and may miss infrequent interactions entirely (Dormann et al 2017).…”

On the Effectiveness of Hand Collection to Complement Baits When Studying Ant Vertical Stratification in Tropical Rainforests

“…Myrmecologists have developed many ingenious sampling methods to collect ants (García-Martínez et al, deserve special attention in the literature, and these works provide valuable information to be known before starting sampling (Gotelli et al, 2011;Lee et al, 2019;Souza et al, 2016). Specifically, comparisons between distinct sampling methods should help researchers to evaluate the trade-offs between sampling completeness and the costs and time required (Souza et al, 2012).…”

“…Specifically, comparisons between distinct sampling methods should help researchers to evaluate the trade-offs between sampling completeness and the costs and time required (Souza et al, 2012). Also, the choice of sampling method should be adapted to the scientific question under consideration to avoid bias in the interpretation of results and conclusions (Lee et al, 2019). Therefore, understanding the shortcomings and advantages of each sampling method is important for myrmecological investigations (Gotelli et al, 2011).…”