Loading...
Please wait, while we are loading the content...
Similar Documents
Trait-based modelling in ecology: lessons from two decades of research
| Content Provider | Semantic Scholar |
|---|---|
| Author | Zakharova, Liubov Meyer, Katrin M. Seifan, Merav |
| Copyright Year | 2019 |
| Abstract | 12 Trait-based approaches are an alternative to species-based approaches for functionally linking individual 13 organisms with community structure and dynamics. In the trait-based approach, the focus is on the traits, 14 the physiological, morphological, or life-history characteristics, of organisms rather than their species. 15 Although used in ecological research for several decades, this approach only emerged in ecological 16 modelling about twenty years ago. We review this rise of trait-based models and trace the occasional 17 transfer of trait-based modelling concepts between terrestrial plant ecology, animal and microbial ecology, 18 and aquatic ecology. Trait-based models have a variety of purposes, such as predicting changes in species 19 distribution patterns under climate and land-use change, planning and assessing conservation 20 management, or studying invasion processes. In modelling, trait-based approaches can reduce technical 21 challenges such as computational limitations, scaling problems, and data scarcity. However, we note 22 inconsistencies in the current usage of terms in trait-based approaches and these inconsistencies must be 23 resolved if trait-based concepts are to be easily exchanged between disciplines. Specifically, future trait24 based models may further benefit from incorporating intraspecific trait variability and addressing more 25 complex species interactions. We also recommend expanding the combination of trait-based approaches 26 with individual-based modelling to simplify the parameterization of models, to capture plant-plant 27 interactions at the individual level, and to explain community dynamics under global change. 28 29 PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.27484v1 | CC BY 4.0 Open Access | rec: 14 Jan 2019, publ: 14 Jan 2019 Introduction 30 Understanding community structure and dynamics is a key element of modern ecology, 31 especially in the light of global change (Harte and Shaw, 1995; Knapp, 2002). This 32 understanding was traditionally mediated by species-based approaches. More recently, such 33 approaches were complemented by approaches based on traits. Trait-based approaches are 34 popular, because they allow the direct connection of organism performance to its functions and to 35 the functions of higher levels of organization such as populations, communities and ecosystems. 36 While trait-based approaches are now firmly established in empirical research (Violle et al., 37 2007; Suding and Goldstein, 2008), they were only introduced to modelling about twenty years 38 ago. Given that modelling is important for understanding community structure and dynamics, 39 trait-based modelling can reduce some of the challenges faced by species-based modelling. For 40 example, species-based models are usually complex, difficult to parameterize and often produce 41 outcomes that cannot be generalized to other species. Trait-based models often require less 42 parameterization effort than species-based models, facilitate scaling-up, and produce more 43 generalizable results. Here, we review the rise of trait-based models over the past twenty years, 44 highlight their main fields of application and point out avenues for future trait-based modelling. 45 46 Traits arose from the concept of plant functional groups and these groups were the first published 47 classification of organisms according to function instead of taxonomy (Raunkiaer, 1934; Grime, 48 1974). The next wave of interest into functional groups was led by the desire to predict 49 community and ecosystem responses to environmental change (Diaz and Cabido, 1997; Lavorel 50 et al., 1997; Chapin et al., 2000). The focus then shifted from functional groups to functional 51 traits and thus from species grouped because they use similar strategies to the similar 52 characteristics underlying those strategies (Yanzheng Yang et al., 2015). Distinct aspects of 53 strategies were reflected in sets of correlated traits that were defined as trait dimensions (Westoby 54 et al., 2002). This shift from a species-based approach to a trait-based approach is described as 55 the ‘Holy Grail of Ecology’ (Lavorel and Garnier, 2002). This approach involves the use of plant 56 PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.27484v1 | CC BY 4.0 Open Access | rec: 14 Jan 2019, publ: 14 Jan 2019 functional traits, rather than species identities, to generalize complex community dynamics and to 57 predict the effects of environmental changes (Suding and Goldstein, 2008). 58 59 Functional traits not only help derive individual strategies (Wright et al., 2004), but also to 60 connect them to functions at organizational levels higher than those of the species such as the 61 community or ecosystem level. There are four requirements for a trait (Lavorel et al. 2007). It 62 should be connected with a function. It should be relatively easy to observe and quantify. It 63 should be possible to measure it in a standardized way across a wide range of species and 64 environmental settings. And it should have a consistent ranking. Trait-based ecology is further 65 based on the assumption that trade-offs and constraints have shaped phenotypic variation in 66 different trait dimensions (Messier et al., 2017). 67 68 Sets of plant traits that reliably represent the processes of growth, survival, and reproduction 69 (Violle et al., 2007) make it possible to facilitate and generalize empirical and modelling studies. 70 Therefore, researchers attempted to define a universal set of traits. Pachepsky et al. (2001) 71 identified twelve critical traits that affected resource uptake, the area over which resource is 72 captured, the internal allocation of resources between structure, storage and reproduction, time of 73 reproduction, number of progeny produced, dispersal of progeny, and survival. Other researchers 74 used smaller numbers of traits. The leaf economics spectrum, for example contains only six 75 (Wright et al., 2004). Díaz et al. (2015) also used six traits but not those of the leaf economics 76 spectrum, and several researchers even used a set with as few as three traits (Westoby, 1998; 77 Westoby et al., 2002; Wright et al., 2004; Chave et al., 2009; Garnier and Navas, 2012). Thus, 78 rather than applying a universal trait set, modern use of the concept implies a selection of a small 79 set of critical functional traits specific to the needs of a specific study and dependent on the 80 specific organisms for which strategies are being described. 81 82 Using trait-based approaches overcomes some of the well-known problems of species-based 83 approaches. In trait-based approaches, for example, it is possible to directly connect community 84 functions such as production to environmental changes via functional traits. Moreover, the trait85 based approach is an intuitive approach for addressing evolutionary processes because evolution 86 selects organisms in a community according to their function and not their taxonomy. Trait-based 87 PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.27484v1 | CC BY 4.0 Open Access | rec: 14 Jan 2019, publ: 14 Jan 2019 approaches are, furthermore, more suitable than species-based approaches for generalizations 88 across species as they are not tied to taxonomy. In addition, trait data are, in most cases, more 89 readily available than species data due to the rapid expansion of trait databases. Trait databases 90 are especially well developed for plants (Kleyer et al., 2008; Kattge et al., 2011). 91 92 Although current trait-based approaches have several benefits, they also have some shortcomings 93 not present in species-based approaches. One of these is the choice of appropriate functional traits 94 and their trade-offs with other traits given that a great diversity of traits are available (Funk et al., 95 2017). Furthermore, traits differ intraspecifically but these differences are often neglected (Violle 96 et al., 2012; Bolnick et al., 2011). Existing trait databases are usually of limited use when it 97 comes to species interactions, intraspecific trait variation and variable environmental settings 98 (Funk et al., 2017). In addition, the theoretical assumptions of trait-based studies are not always 99 supported by experimental data (Suding and Goldstein, 2008). These shortcomings can be 100 overcome by closer cooperation between empirical and theoretical researchers and by the 101 development of standards for trait data collection (e.g. Garnier and Shipley, 2001; Pérez102 Harguindeguy et al., 2013). 103 104 In the most recent 20 years trait-based approaches have entered ecological modelling. The main 105 advantage of modelling over empirical approaches is that it allows the comparison of several 106 scenarios with different sets of assumptions, so conducting virtual experiments. This makes 107 possible the systematic exploration of the outcomes under each set of assumptions and the 108 elucidation of the mechanisms underlying the patterns observed. Using models therefore avoids 109 the costs and risks of real-world experiments. In principal, trait-based models consist of 110 combinations of functional traits that respond to environmental changes (response traits) and 111 affect community and ecosystem properties (effect traits) (Fig. 1). Implementing trait-based 112 approaches for modelling may also help overcome the high data demand of species-based models 113 (Garrard et al., 2013; Weiss et al., 2014), simply due to the fact that traits usually represent more 114 than one species. For the same reason, trait-based modelling may also reduce computing times. 115 Moreover, using traits in modelling can facilitate scaling of physiological processes to global 116 PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.27484v1 | CC BY 4.0 Open Access | rec: 14 Jan 2019, publ: 14 Jan 2019 scales (Shipley, Vile, & Garnier, 2006; Lamarque et al., 2014) because traits can function as a 117 |
| File Format | PDF HTM / HTML |
| DOI | 10.7287/peerj.preprints.27484v1 |
| Alternate Webpage(s) | https://peerj.com/preprints/27484.pdf |
| Alternate Webpage(s) | https://doi.org/10.7287/peerj.preprints.27484v1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
