Multidimensional scaling for animal traits in the context of dynamic energy budget theory

Sebastiaan A L M Kooijman; Konstadia Lika; Starrlight Augustine; Nina Marn

doi:10.1093/conphys/coab086

Multidimensional scaling for animal traits in the context of dynamic energy budget theory

Conserv Physiol. 2021 Dec 15;9(1):coab086. doi: 10.1093/conphys/coab086. eCollection 2021.

Authors

Sebastiaan A L M Kooijman¹, Konstadia Lika², Starrlight Augustine³, Nina Marn⁴

Affiliations

¹ Department of Theoretical Biology, VU University Amsterdam, de Boelelaan 1087, 1081 HV Amsterdam, The Netherlands.
² Department of Biology, University of Crete, 70013, Heraklion, Greece.
³ Akvaplan-niva AS, Fram High North Research Centre for Climate and the Environment, 9296 Tromsø, Norway.
⁴ Division for Environmental and Marine Research, Rudjer Bošković Institute, 10000 Zagreb, Croatia.

Abstract

The method of multidimensional scaling (MDS) has long existed, but could only recently be applied to animal traits in the context of dynamic energy budget (DEB) theory. The application became possible because of the following: (i) the Add-my-Pet (AmP) collection of DEB parameters and traits (approximately 280) recently reached 3000 animal species with 45000 data sets of measurements; (ii) we found a natural distance measure for species based on their traits as a side result of our research on parameter estimation in DEB context; and (iii) we developed plotting code for visualization that allows labelling of taxonomic relationships. Traits, here defined as DEB parameters or any function of these parameters, have different dimensions, which hamper application of many popular distance measures since they (implicitly) assume that all traits have the same dimensions. The AmP collection follows the workflow that measured data determine parameters and parameters determine trait values. In this way we could fill up the species traits table completely, which we could not do by using measured values only, as data availability varies considerably between species and is typically poor. The goodness of fit of predictions for all data sets is generally excellent. This paper discusses links between the MDS method and parameter estimation and illustrates the application of MDS for the AmP collection to five taxa, three ectothermic and two endothermic, which we consider to be 'complete', in the sense that we expect that it will be difficult to find more species with data in the open literature. This application of MDS shows links between traits and taxonomy that supplements our efforts to find patterns in the co-variation of parameter values. Knowledge about metabolic performance is key to conservation biology, sustainable management and environmental risk assessment, which are seen as interlinked fields.

Keywords: Add-my-Pet collection; Dynamic Energy Budgets; animals; classical multidimensional scaling; conservation biology; environmental risk assessment; life history; loss function; metabolism; parameter estimation; principal coordinate analysis; sustainable management; traits.